ES2365813T3 - DETECTION OF LISOPHOSFATIDILCOLINA FOR PROGNOSIS OR DIAGNOSIS OF A SYSTEMIC INFLAMMATORY DISORDER. - Google Patents

Abstract

A method for the diagnosis or prognosis of sepsis in a patient comprising the stage of monitoring or evaluating the amount of total lysophosphatidylcholine in a sample of fluid or tissue of the patient, where the diagnosis or prognosis is made prior to the appearance of sepsis.

Description

The present invention relates to methods for diagnosis or prognosis of sepsis in a patient.

Early detection of a disease condition typically allows a more effective therapeutic treatment with the correspondingly more favorable clinical outcome. In many cases, however, early detection of disease symptoms is problematic due to the complexity of the disease; therefore, a disease can become relatively advanced before diagnosis is possible. Systemic inflammatory conditions represent one of these kinds of diseases. These conditions, particularly sepsis, typically, but not always, result from an interaction between a pathogenic microorganism and the host defense system that causes an excessive and deregulated inflammatory response in the host. The complexity of the host response during the systemic inflammatory response has complicated efforts to understand the pathogenesis of the disease (reviewed in Healy, 2002, Ann. Pharmacother, 36: 648-54). An incomplete understanding of the pathogenesis of the disease, in turn, contributes to the difficulty of finding useful diagnostic biomarkers. Early and reliable diagnosis is imperative, however, due to the markedly rapid advance of sepsis becoming a life-threatening condition.

The development of sepsis in a subject follows a well-described course, which progresses from a negative systemic inflammatory response syndrome ("SIRS"), to positive SIRS, and then to sepsis, which can then progress to severe sepsis, septic shock, dysfunction multiple organs ("MOD"), and finally death.

Documenting the presence of pathogenic microorganisms that are clinically significant for sepsis has proven difficult. The causative microorganisms are typically detected by culturing the subject's blood, sputum, urine, wound secretion, permanent line catheter surfaces, etc. The causative microorganisms, however, may reside only in certain environments of the body such that the particular material that is grown may not contain the contaminating microorganisms. Detection can also be complicated by the low numbers of microorganisms at the site of infection. The low numbers of blood-borne pathogens present a particular problem in diagnosing sepsis through blood culture. In one study, for example, positive culture results were obtained in only 17% of the subjects presenting clinical manifestations of sepsis (Rangel-Frausto et al., 1995, JAMA 273: 117-123). The diagnosis can also be complicated by contamination of the samples by non-pathogenic microorganisms. For example, only 12.4% of the microorganisms detected were clinically significant in a study of 707 subjects with septicemia (Weinstein et al., 1997, Clinical Infectious Deseases 24: 584-602).

The difficulty in the early diagnosis of sepsis is reflected by the high morbidity and mortality associated with the disease. Sepsis is currently the tenth cause of death in the United States and is especially prevalent among patients hospitalized in non-coronary intensive care units (ICUs), where it is the most common cause of death. The total mortality rate is as high as 35%, with 750,000 estimated cases per year occurring in the United States alone. The annual cost to treat sepsis in the United States is only in the order of one billion dollars.

Most existing sepsis classification systems or prediction models predict only the risk of late stage complications, which includes death, in patients who are already considered septic and do not predict the development of the same sepsis. Often, the diagnosis of sepsis is based on clinical suspicion with an empirical classification system such as APACHE II (Knaus et al., 1985, Crit. Care Med. 13: 818-829), followed by blood culture. It may take 48 hours or more to confirm any systemic infection. By then, it is often too late to save some patients. If the diagnosis or prognosis of sepsis can be made early, treatment can be made available to prevent or decrease the speed of sepsis progression in severe sepsis or septic shock.

Therefore, there is a need for methods to diagnose systemic inflammatory conditions, including sepsis, by using techniques that have satisfactory specificity and sensitivity, early enough to allow effective intervention and prevention. The present invention relates to a method for diagnosis

or prognosis of sepsis in a patient that comprises the stage of monitoring or evaluating the amount of total lysophosphatidylcholine in a sample of fluid or tissue of the patient, where the diagnosis or prognosis is made prior to the appearance of sepsis.

The present invention also relates to a method for the diagnosis or prognosis of sepsis in a patient comprising the step of detecting lysophosphatidylcholine in the patient by contacting a sample of fluid or tissue of the patient with: an enzyme or reagent capable of react lysophosphatidylcholine to form glycerophosphatidylcholine; an enzyme or reagent capable of reacting glycerophosphatidylcholine to form choline; an enzyme or reagent capable of reacting choline, water and oxygen to form peroxide; a peroxidase; and a fluorogenic substrate of said peroxidase; under appropriate conditions for the formation of a fluorescent product where the fluorescent product indicates lysophosphatidylcholine; where the diagnosis or prognosis is made prior to the appearance of sepsis.

5

10

fifteen

twenty

25

30

35

40

In one aspect, the present invention is based, in part, on the discovery that the amount of total lysophosphatidylcholine in a sample of a patient can be used for the rapid, sensitive and accurate diagnosis or prognosis of a systemic inflammatory condition in a subject. In aspects of the invention, total lysophosphatidylcholine in a sample of a patient is used to assess the presence of or risk of systemic inflammatory condition. As shown in the examples below, the methods of the invention can be used for the diagnosis or prognosis of the systemic inflammatory condition. While the systemic inflammatory condition in the methods of the invention is sepsis, any systemic inflammatory condition known to those skilled in the art including systemic inflammatory response syndrome ("SIRS"), sepsis, severe sepsis, is described herein. septic shock, multiple organ dysfunction ("MOD") and mortality.

In certain embodiments, total lysophosphatidylcholine is evaluated in a sample of the subject to assess the presence or risk of a systemic inflammatory condition. As described herein, it has been found that the amount of total lysophosphatidylcholine in a sample of a patient may be correlated with the appearance of a systemic inflammatory condition and may still indicate the condition in anticipation of its occurrence. As a result, in certain embodiments, total lysophosphatidylcholine in a patient sample can be used as a prognosis for a systemic inflammatory condition. The evaluation can proceed according to any technique to evaluate the total lysophosphatidylcholine known to those skilled in the art.

Exemplary techniques are described herein. However, the present invention provides methods based on any technique for evaluating the total lysophosphatidylcholine evident to those skilled in the art.

In another aspect, the present invention is based, in part, on the discovery of a class of lysophosphatidylcholine biomarkers that are useful for the rapid, sensitive and accurate diagnosis or prognosis of a systemic inflammatory condition in a subject. In aspects of the invention, the evaluation of a biomarker of the invention in the subject is used to assess the presence of or risk of systemic inflammatory condition. As shown in the examples below, the methods of the invention can be used for the diagnosis or prognosis of the systemic inflammatory condition with accuracy up to 80% or more in such a short time as, for example, eight or even six hours.

In certain embodiments, the biomarker is a 1-0-acyl-2-smooth-sn-glycerol-3-phosphocholine. The acyl group can be any acyl group known to those skilled in the art. In certain embodiments, the acyl group is C14-C22 acyl. In other embodiments, the acyl group is C16-C18 acyl. In particular embodiments, the acyl group is C16 acyl. In a preferable embodiment, the acyl group is palmitoyl. In other particular embodiments, the acyl group is C18 acyl. In a preferable embodiment, the acyl group is stearoyl. The biomarker can be any form of the subject's biomarker, for example any salt or solvate of the biomarker that can be identified by those skilled in the art.

In certain embodiments, the biomarker is a compound in accordance with formula (I):

image 1

or a salt or solvate thereof. In formula (I), R may be any acyl group known to those skilled in the art. In certain embodiments, R is C14-C22 acyl. In other embodiments, R is C16-C18 acyl. In particular embodiments, R is C16 acyl. In a preferable embodiment, R is palmitoyl. In other particular embodiments, R is C18 acyl. In a preferable embodiment, R is stearoyl.

Exemplary salts of formula (I) are provided by formula (Ia): wherein said salt can be coordinated with any physiological inorganic or organic anion, or any physiological inorganic or organic cation, or both, known to those skilled in the art. Exemplary physiological anions include chloride, bromide, phosphate, acetate, carbonate, bicarbonate and sulfate. Exemplary physiological cations include sodium, potassium, calcium, magnesium and ammonium.

image 1

In certain aspects, the biomarker in the subject is evaluated to evaluate the presence or risk of the systemic inflammatory condition. The evaluation can proceed according to any method to evaluate a biomarker known to those skilled in the art. In certain embodiments, the amount of the biomarker is measured in a fluid of a subject. However, the present invention provides methods based on any technique for evaluating a biomarker of the invention apparent to those skilled in the art.

In the following description, the term lysophosphatidylcholine may refer to total lysophosphatidylcholine or a biomarker of lysophosphatidylcholine, unless otherwise specified. Of course, the present invention provides prognosis or diagnosis based on total lysophosphatidylcholine, prognosis or diagnosis based on one or more biomarkers of lysophosphatidylcholine of the invention, and prognosis or diagnosis based on total lysophosphatidylcholine together with one or more biomarkers of lysophosphatidylcholine of the invention.

In some embodiments, a plurality of measurements of lysophosphatidylcholine are performed on the subject over time. The time intervals may be, for example, 3 hours, 4 hours, 6 hours, 12 hours, 24 hours or other intervals according to the criteria of the practitioner in the art. In these embodiments, the relative amounts of lysophosphatidylcholine are evaluated for the diagnosis or prognosis of the systemic inflammatory condition by a practitioner with experience in the art. In particular embodiments, decreasing amounts of lysophosphatidylcholine indicate the increased risk of systemic inflammatory condition, increasing amounts of total lysophosphatidylcholine indicate the decreasing risk of systemic inflammatory condition.

In advantageous embodiments, a single sample of the subject may be sufficient for the diagnosis or prognosis of the systemic inflammatory condition. The amount of lysophosphatidylcholine can be compared to one or more biomarkers present in the sample that are known to those skilled in the art for maintaining a relatively constant amount in the sample in individuals similar to the subject. The amount of lysophosphatidylcholine of the invention can be evaluated against this internal standard for the diagnosis or prognosis of the systemic inflammatory condition by the experienced practitioner. In particular embodiments, low amounts of lysophosphatidylcholine indicate increased risk of systemic inflammatory condition, and high amounts of lysophosphatidylcholine indicate reduced risk of systemic inflammatory condition.

In other embodiments, the evaluation is based on a comparison of the amount of lysophosphatidylcholine with respect to a reference amount of lysophosphatidylcholine. The reference amount may be, for example, the amount of lysophosphatidylcholine in a reference individual that manifests, or manifests within a defined period of time, one or more symptoms of the known systemic inflammatory condition. The quantity may be, for example, an absolute value or an absolute value with a margin of error or a range of values, as determined by those skilled in the art. In certain embodiments, the reference individual exhibits, or will exhibit, symptoms of SIRS, sepsis, severe sepsis, septic shock or MOD or no symptoms of a systemic inflammatory condition. In particular embodiments, low amounts of lysophosphatidylcholine (for example with respect to a reference amount) indicate increased risk of systemic inflammatory condition, and high amounts of lysophosphatidylcholine indicate reduced risk of systemic inflammatory condition.

Advantageously, the reference amount does not need to be determined by not carrying out a method of the invention. Instead, the reference amount of lysophosphatidylcholine can be identified by consultation data available to those skilled in the art. Such data can be obtained from any source available to those skilled in the art. In certain embodiments, the sources may be developed with reference amounts of lysophosphatidylcholine collected by those skilled in the art in accordance with methods described herein.

In certain embodiments, the reference amount is from a reference individual who exhibits symptoms of the systemic inflammatory condition. The reference individual may present with symptoms of SIRS, sepsis, severe sepsis, septic shock, MOD or mortality or no symptoms of a systemic inflammatory condition. In certain embodiments, the reference amount may be evaluated at a time prior to or after the presentation of symptoms. For example, in an advantageous embodiment, a reference amount may be the amount measured in an individual with positive SIRS 12, 24, 36 or 48 hours prior to the onset of sepsis. The measurement of said reference quantities is within the experience of those of the art.

In other embodiments, the reference amounts are from a plurality of individuals exhibiting symptoms of one or more systemic inflammatory conditions. Reference quantities may be calculated according to any appropriate statistical method known to those skilled in the art. For example, the reference amounts may be based on the statistical average of the reference amounts of the reference individuals presenting with a systemic inflammatory condition. In advantageous embodiments, the comparison is made with a value or range of values in terms of the amount of lysophosphatidylcholine. The value or range of values may be obtained according to what is described herein and may be available to a practitioner of the methods of the invention. In particular embodiments, low amounts of lysophosphatidylcholine (for example with respect to a reference amount) indicate increased risk of systemic inflammatory condition, and high amounts of lysophosphatidylcholine indicate reduced risk of systemic inflammatory condition.

The comparison can be according to any technique to compare the amounts of biomarkers known to those skilled in the art. In one embodiment, the diagnosis or prognosis of a systemic inflammatory condition is based on the difference between the amount of lysophosphatidylcholine in the subject and the reference amount. In certain embodiments, the difference between the amount of lysophosphatidylcholine in the subject and the reference amount is inversely correlated with the risk of systemic inflammatory condition. In other embodiments, the reference amount is a limit - in other words, the subject can be evaluated for having or having a risk of systemic inflammatory condition if the amount of lysophosphatidylcholine in the subject is less than a reference limit amount. Said reference limit amounts may be calculated according to methods described herein.

The amount of total lysophosphatidylcholine in the subject can be determined according to any technique known to those skilled in the art without limit. In certain embodiments, one with experience can measure an amount that correlates with the amount of total lysophosphatidylcholine in a sample. For example, in particular embodiments, one with experience can measure total free lysophosphatidylcholine, total bound lysophosphatidylcholine or total bound and free Iisophosphatidylcholine in the sample to indicate the amount of total lysophosphatidylcholine in the sample. In other words, in certain embodiments, measurement of free or bound lysophosphatidylcholine, or both free and bound, can be correlated with the amount of total lysophosphatidylcholine. In certain embodiments, the technique for evaluating total lysophosphatidylcholine is not critical to the invention and needs to be carried out by one who practices the methods herein. For example, in particular embodiments, the methods of the invention may comprise the only step of comparing the amount of total lysophosphatidylcholine in a subject with an amount of total reference lysophosphatidylcholine to assess the risk of systemic inflammatory condition regardless of how any quantity. In other embodiments, the total lysophosphatidylcholine of the subject is evaluated by a technique described herein followed by comparison with a total reference lysophosphatidylcholine to assess the risk of systemic inflammatory condition. In certain embodiments, total lysophosphatidylcholine is evaluated by spectrometry, chromatography, immunoassay, electrophoresis or enzymatic assay as described in detail below.

In one aspect, the present invention provides fluorescent methods for testing total lysophosphatidylcholine in a sample of a subject. In certain embodiments, the methods comprise contacting the subject's sample with one or more reagents capable of generating a fluorescent product indicative of total lysophosphatidylcholine in the sample. The methods can be used to detect the presence of total lysophosphatidylcholine or to detect the amount of total lysophosphatidylcholine, or both, in the sample. In particular embodiments, the sample is contacted with a fluorogenic substrate of one or more of the reagents. This fluorogenic substrate can be converted into the fluorescent product indicating total lysophosphatidylcholine. In advantageous embodiments, the reagents comprise peroxidase, choline oxidase, glycerophosphatidylcholine diesterase and lysophospholipase. A useful fluorogenic substrate is 10-acetyl-3,7-dihydroxyphenoxazine, a compound that can be converted into the 7-hydroxy3H-phenoxazin-3-one fluorescent product. Advantageously, said methods may provide the sensitivity necessary to detect low amounts of total lysophosphatidylcholine in subjects who have, or are at risk of, a systemic inflammatory condition.

The amount of the biomarker in the subject can be determined according to any technique known to those skilled in the art without limit. In certain embodiments, the technique for evaluating the biomarker is not critical to the invention and does not need to be carried out by one who practices the methods herein. For example, in particular embodiments, the methods of the invention may comprise the only step of comparing the biomarker in a subject with a reference biomarker to assess the risk of systemic inflammatory condition regardless of how any biomarker is measured. In other embodiments, the subject biomarker is evaluated by a technique described herein followed by comparing the biomarker in a subject with a reference biomarker to assess the risk of systemic inflammatory condition. In certain embodiments, the biomarker is evaluated by spectrometry, chromatography, immunoassay or electrophoresis as described in detail below.

The amount of lysophosphatidylcholine can be measured in fluids or tissues of the subject as provided herein. The processes for preparing the fluid or tissue, for example, processes for extracting or purifying lysophosphatidylcholine are described herein. In addition, techniques for measuring lysophosphatidylcholine are provided herein.

In another aspect, the present invention provides methods for monitoring a systemic inflammatory condition in a subject. In such methods, the evaluation of lysophosphatidylcholine is used to monitor a systemic inflammatory condition in the subject. In such methods, changes in the amount of lysophosphatidylcholine indicate changes in the systemic inflammatory condition. For example, in certain embodiments, increasing amounts of lysophosphatidylcholine indicate decreased severity or less risk of systemic inflammatory condition, and decreasing amounts of lysophosphatidylcholine indicate increased severity or increased risk of systemic inflammatory condition. In some embodiments, the evaluation of lysophosphatidylcholine may indicate the conversion of one systemic inflammatory condition to another such as SIRS, sepsis, severe sepsis, septic shock, MOD or mortality or no systemic inflammatory condition.

In another aspect, the present invention provides methods for monitoring the treatment of a systemic inflammatory condition in a subject. In such methods, the evaluation of lysophosphatidylcholine is used to monitor the systemic inflammatory condition in the subject. In such methods, changes in the amount of lysophosphatidylcholine indicate changes in the systemic inflammatory condition. For example, in certain embodiments, increasing amounts of lysophosphatidylcholine indicate decreased severity or less risk of systemic inflammatory condition, and decreasing amounts of total lysophosphatidylcholine indicate increased severity or increased risk of systemic inflammatory condition. Advantageously, the treatment of the systemic inflammatory condition can be adjusted based on monitoring. In some embodiments, the evaluation of lysophosphatidylcholine may indicate the conversion of one systemic inflammatory condition to another such as SIRS, sepsis, severe sepsis, septic shock, MOD or mortality or no systemic inflammatory condition. In preferable embodiments, a subject that is SIRS negative can be monitored for conversion to SIRS or sepsis or other inflammatory condition. In other preferable embodiments, a subject that is SIRS positive can be monitored for conversion to sepsis or other systemic inflammatory condition, or for conversion to negative SIRS.

Kits for the diagnosis or prognosis of a systemic inflammatory condition are also described herein. In some embodiments, the kits comprise a composition suitable for the evaluation of lysophosphatidylcholine. The kits may also comprise a label or label with instructions for using the total lysophosphatidylcholine evaluation for the diagnosis or prognosis of one or more systemic inflammatory conditions. In certain embodiments, the kit may comprise a label or label with reference amounts, or citations of said reference amounts, of lysophosphatidylcholine to facilitate the prognosis or diagnosis of a systemic inflammatory condition with the kit composition.

FIG. 1 provides an exemplary system of the invention;

FIG. 2 provides analysis of the time course of biomarker 496.3;

FIG. 3 provides analysis of the time course of biomarker 518.3;

FIG. 4-1 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIG. 4-2 illustrates the evolution of feature inclusion parameters during the optimization of the methods of the invention;

FIG. 4-3 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIGS. 4-4 and 4-5 illustrate the evolution of feature inclusion parameters during the optimization of methods of the invention;

FIG. 4-6 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIG. 4-7 to 4-9 illustrate the evolution of feature inclusion parameters during the optimization of methods of the invention;

FIG. 5-1 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIG. 5-2 illustrates the evolution of feature inclusion parameters during the optimization of methods of the invention;

FIG. 5-3 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIGS. 5-4 and 5-5 illustrate the evolution of feature inclusion parameters during the optimization of methods of the invention;

FIG. 5-6 illustrates the sensitivity and specificity versus threshold in a performance model according to the invention;

FIG. 5-7 to 5-9 illustrate the evolution of feature inclusion parameters during the optimization of methods of the invention;

FIG. 6 provides a scheme for the detection of lysophosphatidylcholine according to the embodiments of the invention; Y

FIG. 7 provides lysophosphatidylcholine fluorescently detected in samples from patients who showed symptoms of SIRS or sepsis;

Definitions

As used herein, the following terms will have the following meanings:

The term "subject" refers to animals such as mammals, which include, but are not limited to, primates (eg, humans), cows, sheep, goats, horses, dogs, cats, rabbits, rats, mice and the like. In preferable embodiments, the subject is a human being. The term "patient" is interchangeable with a human subject, unless otherwise indicated.

"Systemic inflammatory response syndrome," or "SIRS," refers to a clinical response to a variety of severe clinical attacks, as manifested by two or more of the following conditions within a 24-hour period:

- body temperature greater than 38 ° C (100.4 ° F) or less than 36 ° C (96.8 ° F);

- heart rate (HR) greater than 90 beats / minute;

-? respiratory rate (RR) greater than 20 breaths / minute, or PCO2 less than 32 mmHg, or requiring mechanical ventilation; Y

- white blood cell count (WBC) greater than 12.0 x 10 9 / L or less than 4.0 x 109 / L or having more than 10% of immature band shapes.

These symptoms of SIRS represent consensus definitions of SIRS that can be modified or supplanted by other definitions in the future. This definition is used to clarify current clinical practice and does not represent a critical aspect of the invention (see, for example, American College of Chest Physicians / Society of Critical Care Medicine Consensus Conference: Definitions for Sepsis and Organ Failure and Guidelines for the Use of Innovative Therapies in Sepsis, 1992, Crit. Care. Med 20, 864-874).

A subject with SIRS has a cynical presentation that is classified as SIRS, as defined above, but not considered clinically to be septic. Methods to determine which subjects are at risk of developing sepsis are well known to those in the art. Such subjects include, for example, those in an intensive care unit ("ICU") and those who have otherwise suffered a physiological trauma, such as a burn, surgery or other attack. A distinctive feature of SIRS is the creation of a pro-inflammatory state that may be marked by tachycardia, tachypnea or hyperpnea, hypotension, hypoperfusion, oliguria, leukocytosis or leukopenia, pyrexia or hypothermia and the need for volumetric infusion. SIRS characteristically does not include a documented source of infection (for example, bacteremia).

"Sepsis" refers to a systemic response of the host to SIRS infection plus a documented infection (for example, a subsequent laboratory confirmation of a clinically significant infection such as a positive culture of an organism). Thus, sepsis refers to the systemic inflammatory response to a documented infection (see, for example, American College of Chest Physicians Society of Critical Care Medicine, Chest, 1997, 101: 1644-1655). As used herein, "sepsis" includes all stages of sepsis that include, but are not limited to, the occurrence of sepsis, severe sepsis, septic shock and multiple organ dysfunction ("MOD") associated with the final stages. of sepsis

The "appearance of sepsis" refers to an early stage of sepsis, for example, prior to a stage when conventional clinical manifestations are sufficient to support a clinical suspicion of sepsis. Because the methods of the present invention can be used to detect sepsis prior to a time when sepsis would be suspected by using conventional techniques, in certain embodiments, the disease status of the subject in early sepsis is retrospectively confirmed, when The manifestation of sepsis is more clinically obvious. The exact mechanism by which a subject becomes septic is not a critical aspect of the invention. The methods of the present invention can detect the occurrence of sepsis regardless of the origin of the infectious process.

"Severe sepsis" refers to sepsis associated with organ dysfunction, hypoperfusion abnormalities, or sepsis-induced hypotension. Hypoperfusion abnormalities include, but are not limited to, lactic acidosis, oliguria, or an acute alteration in mental state.

"Septic shock" refers to sepsis-induced hypotension that is not sensitive to adequate intravenous fluid overload and with manifestations of peripheral hypoperfusion.

A "converter" or "converting subject" refers to a positive SIRS subject that advances to clinical suspicion of sepsis during the period in which the subject is monitored, typically during an ICU stay.

A "non-converting" or "non-converting subject" refers to a positive SIRS subject that does not advance until clinical suspicion of sepsis during the period in which the subject is monitored, typically during a stay in ICU.

A "biomarker" is a compound that is present in or obtained from a biological sample. "Obtained from" as used in this context refers to a compound that, when detected, is indicative of a particular molecule that is present in the biological sample. For example, the detection of a particular fragment of a compound may be indicative of the presence of the same compound in the biological sample. A biomarker, for example, can be isolated from the biological sample, directly measured in the biological sample, or detected in or determined to be in the biological sample. A biomarker, for example, can be functional, partially functional, or non-functional.

As used herein, "conventional techniques" in the context of the diagnosis or prognosis of a systemic inflammatory condition are those techniques that classify a subject based on phenotypic changes without evaluating a biomarker according to the present invention.

"Predicting the development of sepsis" is a determination as to whether the subject develops sepsis. Such prediction is limited by the accuracy of the means used to make this determination. The present invention provides a method, for example, by using a decision rule / s, to make this determination with an accuracy that is 60% or greater. As used herein, the terms "predict the development of sepsis" and "predict sepsis" are interchangeable. In some embodiments, the act of predicting the development of sepsis (predicting sepsis) is achieved by evaluating one or more biomarker profiles of a subject by using a decision rule that is indicative of the development of sepsis and, as a result of this evaluation, by receiving a result of the decision rule that indicates that the subject will become septic. Said evaluation of one or more biomarker profiles of a test subject through the use of a decision rule uses some or all quantities in one or more biomarker profiles to obtain said result.

As used herein, the term "specifically," and analogous terms, in the context of an antibody, refers to peptides, polypeptides, and antibodies or fragments thereof that specifically bind to an antigen or a class of antigens, or fragments thereof, and do not specifically bind to other antigens or other fragments. A peptide or polypeptide that specifically binds to an antigen can bind to other peptides or polypeptides with lower affinity, as determined by standard experimental techniques, for example, by any immunoassay well known to those skilled in the art. Such immunoassays include, but are not limited to, radioimmunoassays (RIAs) and enzyme-linked immunoenzymatic assay (ELISAs). Antibodies or fragments that specifically bind to an antigen may have cross reactivity with related antigens. Preferably, the antibodies or fragments thereof that specifically bind to an antigen do not cross-react with other antigens. See, for example, Paul, ed., 2003, Fundamental Immunology, 5th ed., Raven Press, New York on pages 69-105, for a discussion regarding antigen-antibody interactions, specificity and cross-reactivity, and methods to determine all of the above.

As used herein, a "reference population" is a population of subjects that can be used to construct a decision rule for the evaluation of a biomarker of subjects at risk of developing a systemic inflammatory condition.

A "reference subject" is a subject that has been diagnosed, or will be diagnosed within a defined period of time, with a systemic inflammatory condition according to standards recognized by those skilled in the art. A reference subject is useful for establishing a biomarker reference amount that can be used to evaluate a biomarker amount in a test subject for the diagnosis or prognosis of a systemic inflammatory condition.

A "biomarker profile" comprises a plurality of one or more types of biomarkers (for example, a polypeptide, peptide, lipid, nucleic acid, metabolite, mRNA, cDNA, and / or a carbohydrate, etc.), or an indication of they, together with a characteristic, such as a measurable aspect (for example, abundance, level of expression) of biomarkers. A biomarker profile comprises at least two of said biomarkers or indications thereof, where the biomarkers can be in the same or different classes, such as, for example, a nucleic acid and a carbohydrate. A biomarker profile may also comprise at least three, four, five, 10, 20, 30 or more biomarkers or indications thereof. In one embodiment, a biomarker profile comprises hundreds, or even thousands, of biomarkers or indications thereof. A biomarker profile can also comprise one or more internal controls or standards. In one embodiment, the biomarker profile comprises at least one biomarker, or indication thereof, that serves as an internal standard. In another embodiment, a biomarker profile comprises an indication of one or more types of biomarkers. The term "indication" as used herein in this context simply refers to a situation where the biomarker profile contains symbols, data, abbreviations or other similar indications for a biomarker, rather than the same molecular entity of the biomarker.

Each biomarker in a biomarker profile includes a corresponding "feature". A "characteristic", as used herein, refers to a measurable aspect of a biomarker. A characteristic may include, for example, the presence or absence of the biomarker in the biological sample, the abundance or quantity of the biomarker in the sample, the ratio of amounts of molecules in the sample, etc. A characteristic may also be the difference between a measurable aspect of the corresponding biomarker that is taken from two samples, where the two samples are collected from a subject at two different time points. Those skilled in the art will appreciate that other methods of computing a feature can be invented and all such methods are within the scope of the present invention. For example, a characteristic may represent the average of an abundance of a biomarker in all biological samples collected from a subject at two or more time points. In addition, a characteristic may be the difference or ratio of the abundance of two or more biomarkers of a biological sample obtained from a subject at a single time point. A biomarker profile can also comprise at least three, four, five, 10, 20, 30 or more features. In one embodiment, a biomarker profile comprises hundreds, or even thousands, of features.

A "phenotypic change" is a detectable change in a parameter associated with a given state of the subject. For example, a phenotypic change may include an increase or decrease of a biomarker in a body fluid, where the change is associated with SIRS, sepsis, the appearance of sepsis or a particular stage in the progression of sepsis. A phenotypic change can also include a change in a detectable aspect of a given state of the subject that is not a change in a measurable aspect of a biomarker. For example, a change in the phenotype may include a detectable change in body temperature, breathing rate, pulse, blood pressure, or other physiological parameter. Such changes can be determined through clinical observation and measurement by using conventional techniques that are well known to the skilled technician.

A "decision rule" is a method used to evaluate biomarker profiles. Such decision rules may take one or more forms that are known in the art, as exemplified in Hastie et al., 2001, The Elements of Statistical Learning, Springe Verlag, New York. A decision rule can be used to act on a set of characteristic data to, inter alia, predict the occurrence of sepsis, to determine the progress of sepsis, or to diagnose sepsis. Exemplary decision rules that can be used in some embodiments of the present invention are described in more detail in Section 5.5, below.

"Predicting the development of sepsis" is a determination as to whether the subject develops sepsis. Such prediction is limited by the accuracy of the means used to make this determination. The present invention provides a method, for example, by using a decision rule / s, to make this determination with an accuracy that is 60% or greater. As used herein, the terms "predict the development of sepsis" and "predict sepsis" are interchangeable. In some embodiments, the act of predicting the development of sepsis (predicting sepsis) is achieved by evaluating one or more biomarker profiles of a subject by using a decision rule that is indicative of the development of sepsis and, as a result of this evaluation, by receiving a result of the decision rule that indicates that the subject will become septic. Said evaluation of one or more biomarker profiles of a test subject through the use of a decision rule uses one or all of the characteristics in one or more biomarker profiles to obtain said result.

As used herein, a "training population" is a set of samples from a population of subjects used to construct a decision rule, through the use of a data analysis algorithm, for the evaluation of the biomarker profile. of subjects at risk of developing sepsis. In a preferable embodiment, a training population includes samples of subjects that are converters and subjects that are non-converters.

As used herein, a "data analysis algorithm" is an algorithm used to construct a decision rule by using biomarker profiles of subjects in a training population. The representative data analysis algorithms are described in Section 5.5. A "decision rule" is the end product of a data analysis algorithm, and is characterized by one or more sets of values, where each of these sets of values is indicative of an aspect of SIRS, the appearance of sepsis, sepsis, or a prediction that a subject will acquire sepsis. In a specific example, a set of values represents a prediction that a subject will develop sepsis. In another example, a set of values represents a prediction that a subject will not develop sepsis.

As used herein, a "set of values" is a combination of values, or ranges of values for the characteristics in a biomarker profile. The nature of this set of values and the values in it depends on the type of characteristics present in the biomarker profile and the data analysis algorithm used to construct the decision rule imposed by the set of values. For example, a biomarker profile can be obtained from each member of a training population. Each of said biomarker profile includes a characteristic measured for each biomarker. These characteristic values can be used by a data analysis algorithm to construct a decision rule. The data analysis algorithm can be a decision tree, described below. A decision rule defines sets of values. One of said set of values predicts the appearance of sepsis. A subject whose biomarker characteristic values satisfy this set of values will possibly become septic. For example, a set of values may comprise the amount of biomarker A being less than a first value and the amount of biomarker B being less than a second value. Another set of values predicts a septic free state. A subject whose biomarker characteristic values satisfy this set of values is not possible to become septic. An exemplary set of values of this could comprise biomarker A being greater than the first value and biomarker B being greater than a third value.

When the data analysis algorithm is a neural network analysis and the end product of this neural network analysis is a properly weighted neural network, a set of values is those ranges of biomarker profile characteristic values that will make the network Weighted neuronal indicate that the occurrence of sepsis is likely. Another set of values is those ranges of biomarker profile characteristic values that will cause the weighted neural network to indicate that the occurrence of sepsis is not likely.

"Preventing" or "prevention" refers to a reduction in the risk of acquiring a disease or disorder (that is, making at least one of the clinical symptoms the disease does not develop in a subject that may be exposed or predisposed to disease but not yet experience or show symptoms of the disease). Preferably, prevention refers to the use of a compound or composition in a subject not yet affected by the disease or disorder or that does not yet exhibit a symptom of the disease or disorder, for example a subject not yet infected or not yet exhibiting the Symptoms of the infection

"Treat" or "treatment" of any disease or disorder refers to, in one embodiment, alleviating the disease or disorder (ie, stopping or reducing the development of the disease or at least one of the clinical symptoms thereof) that It exists in a subject. In another embodiment, "treating" or "treatment" refers to alleviating at least one physical parameter, which may be imperceptible by the subject. In yet another embodiment, "treating" or "treatment" refers to modulating the disease or disorder, physically (for example, stabilization of a noticeable symptom) or physiologically (for example, stabilization of a physical parameter) or both. In yet another embodiment, "treating" or "treatment" refers to delaying the onset of the disease or disorder.

The term "label" refers to a sample of written, printed or graphic material on the immediate packaging of an article, for example the written material displayed on a vial containing a pharmaceutically active agent.

The term "labeling" refers to all labels and other written, printed or graphic material in any article.

or any of its packages or wrappings or that accompanies said article, for example, a leaflet or audio of instructions or videos, for example videotapes or DVDs, that accompany or are associated with a container of a pharmaceutically active agent.

"Acyl" refers to a radical -C (O) R, where R is alkyl.

"Alkyl," by itself or as part of another substituent, means, unless otherwise stated, a branched or linear chain hydrocarbon radical that can be completely saturated, mono or polyunsaturated, which has the number of atoms of designated carbon (i.e. C1-C22 means one to twenty one carbons). Examples of saturated hydrocarbon radicals include groups such as ethyl, ethyl, n-propyl, isopropyl, n-butyl, t-butyl, isobutyl, sec-butyl, cyclohexyl, (cyclohexyl) methyl, cyclopropylmethyl, homologs and isomers of, for example, n-pentyl, n-hexyl, n-heptyl, n-octyl, and the like. An unsaturated alkyl group is one that has one or more double bonds or triple bonds. Examples of unsaturated alkyl groups include vinyl, 2-propenyl, crotyl, 2-isopentenyl, 2- (butadienyl), 2,4-pentadienyl, 3- (1,4-pentadienyl), ethynyl, 1-and 3-propynyl, 3-butynyl, and homologs and isomers.

"Physiologically acceptable salt" refers to a salt of a compound of the invention that is pharmaceutically acceptable and that possesses the desired pharmacological activity of the parent compound. Such salts include: (1) acid addition salts formed with organic or inorganic acids such as hydrochloric, hydrobromic, sulfuric, nitric, phosphoric, acetic, trifluoroacetic, trichloroacetic, propionic, hexanoic, cyclopentylpropionic, glycolic, glutaric, pyruvic, lactic acid, Malonic, succinic, sorbic, ascorbic, malic, maleic, fumaric, tartaric, citric, benzoic, 3- (4-hydroxybenzoyl) benzoic, picrico, cinnamic, mandelic, phthalic, lauric, methanesulfonic, ethanesulfonic, 1,2-ethane-disulfonic , 2-hydroxyethanesulfonic acid, benzenesulfonic acid, 4-chlorobenzenesulfonic acid, 2naphthalenesulfonic acid, 4-toluenesulfonic acid, camphoric acid, canophosulfonic acid, 4-methylbicyclo [2.2.2] -oct-2-ene-1-carboxylic acid, glucoheptonic, 3-phenylpropionic, trimethyl butylacetic, lauryl sulfuric, gluconic, benzoic, glutamic, hydroxynaphthoic, salicylic, stearic, muconic and similar acids; or (2) salts formed when an acidic proton present in the parent compound (a) is replaced by a metal ion, for example, an alkali metal ion, an alkaline earth ion or an aluminum ion, or alkali metal hydroxides or alkaline earth, such as sodium hydroxide, potassium, calcium, magnesium, and barium, ammonia or (b) is coordinated with an organic base, such as aromatic, aliphatic or alicyclic organic amines, such as methylamine, dimethylamine, diethylamine, picolina, ethanolamine, danolamine, triethanolamine, N-methylglucamine and the like.

The salts also include, by way of example only, sodium, potassium, calcium, magnesium, ammonium, tetraalkylammonium and the like, and when the compound contains a basic functionality, salts of non-toxic organic or inorganic acids, such as hydrochloride, hydrobromide, tartrate , mesylate, acetate, maleate, oxalate and the like. The term "physiologically acceptable cation" refers to a physiologically acceptable, non-toxic cationic counterion of an acid functional group. Said cations are exemplified by sodium, potassium, calcium, magnesium, ammonium and tetraalkylammonium cations and the like.

"Solvate" refers to a compound of the present invention or a salt thereof, which also includes a stoichiometric or non-stoichiometric amount of solvent bound by non-covalent intermolecular forces. Where the solvent is water, the solvate is a hydrate.

It should be understood that compounds that possess the same molecular formula but differ in the nature or sequence of union of their atoms or in the arrangement of their atoms in space are called "isomers." Isomers that differ in the arrangement of their atoms in space are called "stereoisomers."

Stereoisomers that are not mirror images of each other are called "diastereomers" and those that are mirror images that are not superimposable on each other are called "enantiomers." When a compound has an asymmetric center, for example, when it is attached to four different groups, a pair of enantiomers is possible. An enantiomer can be characterized by the absolute configuration of its asymmetric center and is designated (R) or (S) according to the rules of Cahn and Prelog (Cahn et al., 1966, Angew. Chem. 78: 413-447, Angew . Chem., Int. Ed. Engl. 5: 385-414 (errata: Angew. Chem., Int. Ed. Eng /. 5: 511); Prelog and Helmchen, 1982, Angew, Chem. 94: 614631, Angew. Chem. Internal. Ed. Eng. 21: 567-583; Mata and Lobo, 1993, Tetrahedron: Asymmetry 4: 657-668) or can be characterized by the way in which the molecule rotates the plane of polarized light and is designated dextrorotatory or levorotatory (that is, as isomers (+) - or (-) -, respectively). A chiral compound can exist as an individual enantiomer or as a mixture thereof. A mixture that contains equal proportions of enantiomers is called a "racemic mixture."

In certain embodiments, the compounds of the present invention may possess one or more asymmetric centers; said compounds can therefore be produced as (R) -o (S) -enantiomers individually or as a mixture thereof. Unless otherwise indicated, for example by designation of stereochemistry at any position of a formula, the description or name of a particular compound in the specification and claims is intended to include both individual enantiomers and mixtures, racemic or not, of the same. Methods for the determination of stereochemistry and separation of stereoisomers are well known in the art. In particular embodiments, the present invention provides the stereoisomers of the compounds represented herein in the base treatment.

Embodiments of the Invention

The present invention allows rapid and accurate diagnosis or prognosis of a systemic inflammatory condition by evaluating lysophosphatidylcholine in a subject. The amounts of lysophosphatidylcholine can be constructed from one or more biological samples of subjects at a single point of time ("instantaneous"), or multiple time points, during the time period in which the subject is at risk of developing a systemic inflammatory condition. Advantageously, the systemic inflammatory condition can be diagnosed or anticipated prior to the appearance of conventional clinical symptoms, thereby allowing the most effective therapeutic intervention.

Subjects

In certain embodiments of the invention, the subject is an animal, preferably a mammal, more preferably a non-human primate. In much more preferable embodiments, the subject is a human being.

Although the methods of the invention can be used for the diagnosis or prognosis of a systemic inflammatory condition in any subject, particularly useful subjects include those who are at risk of systemic inflammatory condition. The subject may be at risk of systemic inflammatory condition according to any criteria known to the practitioner with experience in the art.

In certain embodiments, the subject is negative SIRS. In the context of the present invention, negative SIRS subjects include healthy subjects who, for any reason according to the judgment of the practitioner of the technique, need the diagnosis or prognosis of sepsis. Such subjects include, but are not limited to, negative SIRS patients in hospital intensive care units and similarly situated subjects. The methods of the invention can be used for the diagnosis or prognosis of SIRS, sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality. Other methods of the invention can be used to monitor a treatment or prevention for the increased or decreased risk of SIRS, sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality. In particular embodiments, the subject is a patient who could be at risk of a systemic inflammatory condition, such as a patient in an intensive care unit.

In other embodiments, the subject is SIRS positive. The methods of the invention can be used for diagnosis.

or prognosis of sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality, or may be used for the diagnosis or prognosis of conversion into negative SIRS. Other methods of the invention can be used to monitor a treatment or prevention for the increased or decreased risk of sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality, or to monitor the possible conversion to negative SIRS.

In other embodiments, the subject has sepsis. The methods of the invention can be used for the diagnosis or prognosis of severe sepsis, septic shock, multiple organ dysfunction or mortality, or they can be used for the diagnosis or prognosis of conversion into positive SIRS (and negative sepsis) or conversion into negative SIRS. Other methods of the invention can be used to monitor a treatment or prevention for increased risk.

or decreased from severe sepsis, septic shock, multiple organ dysfunction or mortality, or to monitor the possible conversion into positive SIRS (and negative sepsis) or conversion into negative SIRS.

In other embodiments, the subject has severe sepsis. The methods of the invention can be used for the diagnosis or prognosis of septic shock, multiple organ dysfunction or mortality, or they can be used for the diagnosis or prognosis of conversion in sepsis or in positive SIRS (and negative sepsis) or in negative SIRS. Other methods of the invention can be used to monitor a treatment or prevention for increased risk.

or decreased septic shock, multiple organ dysfunction or mortality, or to monitor the possible conversion into sepsis or positive SIRS (and negative sepsis) into negative SIRS.

In other embodiments, the subject has septic shock. The methods of the invention can be used for the diagnosis or prognosis of multiple organ dysfunction or mortality, or they can be used for the diagnosis or prognosis of conversion into severe sepsis, sepsis, positive SIRS (and negative sepsis) or negative SIRS. Other methods of the invention can be used to monitor a treatment or prevention for the increased or decreased risk of multiple organ dysfunction or mortality, or to monitor the possible conversion to severe sepsis, sepsis, positive SIRS (and negative sepsis) or SIRS negative.

In other embodiments, the subject has multiple organ dysfunction. The methods of the invention can be used for the diagnosis or prognosis of mortality, or they can be used for the diagnosis or prognosis of conversion in septic shock, severe sepsis, sepsis, positive SIRS (and negative sepsis) or negative SIRS. Other methods of the invention can be used to monitor a treatment or prevention for the increased or decreased risk of mortality, or to monitor the possible conversion into septic shock, severe sepsis, sepsis, positive SIRS (and negative sepsis) or negative SIRS .

In preferable embodiments, the subject is negative SIRS (ie the subject may be healthy) but needs the diagnosis or prognosis of a systemic inflammatory condition according to the criteria of a practitioner with experience in the technique. The subject could be, for example, a patient in an intensive care unit. Similarly, in preferable embodiments, the subjects are SIRS negative, and the methods of the invention are used to monitor the prevention of a systemic inflammatory condition. For example, in a negative SIRS subject considered at risk of a systemic inflammatory condition, for example according to a method of the invention, a course of intervention could be administered to the subject to prevent a systemic inflammatory condition. Said prevention of a systemic inflammatory condition can be monitored with a method of the invention.

In other preferable embodiments, the subjects are SIRS positive, and the methods of the invention are used for the diagnosis or prognosis of another systemic inflammatory condition. Similarly, in preferred embodiments, the subjects are SIRS positive, and the methods of the invention are used to monitor the treatment of SIRS or prevention of another systemic inflammatory condition. For example, in a positive SIRS subject considered at risk of another systemic inflammatory condition, for example according to a method of the invention, a course of intervention could be administered to the subject to prevent systemic inflammatory condition. Said prevention of the systemic inflammatory condition can be monitored with a method of the invention.

In specific embodiments of the invention, subjects at risk of developing sepsis or SIRS are selected by using the methods of the invention. In accordance with these embodiments, the methods of the present invention can be used to select, for example, subjects admitted to an intensive care unit and / or those who have experienced some kind of trauma (such as, for example, surgery, vehicular accident , shot wound, etc.).

In specific embodiments, a subject is selected by utilizing the methods and compositions of the

5

10

fifteen

twenty

25

30

35

40

Four. Five

invention as frequently as necessary (for example, during your stay in an intensive care unit) for the diagnosis or prognosis of a systemic inflammatory condition. In a preferable embodiment, the subject is selected soon after arrival in an intensive care unit. In some embodiments, the subject is selected daily after they arrive at an intensive care unit. In some embodiments, the subject is selected every 1 to 8 hours, 8 to 12 hours, 12 to 16 hours, or 16 to 24 hours after they arrive at an intensive care unit.

Lysophosphatidylcholine - total

In one aspect, the present invention provides prognosis or diagnosis of a systemic inflammatory condition based on total lysophosphatidylcholine.

In certain embodiments, total lysophosphatidylcholine in a sample of a subject is used for the prognosis or diagnosis of systemic inflammatory condition. Total lysophosphatidylcholine refers to an amount that corresponds to all lysophosphatidylcholine (free or bound or both) in the sample. For example, total lysophosphatidylcholine may refer to those molecules in the sample that are in accordance with formula (I):

image 1

or any salt or solvate thereof, where R is any acyl group. The acyl group can be any acyl group known to those skilled in the art. Exemplary acyl groups include caproyl, lauroyl, myristoyl, palmitoyl, stearoyl, palmitoleyl, oleyl, arachidonyl and linoleyl. Preferably, the total lysophosphatidylcholine includes at least 1-0-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine and 1-0-stearoyl-2-lyso-sn-glycerol-3-phosphocholine. In typical embodiments, total lysophosphatidylcholine is measured regardless of the identity of the acyl group. Useful techniques are described herein.

In certain embodiments, total lysophosphatidylcholine alone is used for the prognosis or diagnosis of systemic inflammatory condition. In other embodiments, one or more additional biomarkers are used for the prognosis or diagnosis of systemic inflammatory condition. In other embodiments, one or more clinical measurements are additionally used for the prognosis or diagnosis of systemic inflammatory condition. In certain embodiments, one or more additional biomarkers and one or more clinical measurements are additionally used for the prognosis or diagnosis of systemic inflammatory condition.

The experienced practitioner can use any technique to measure or indicate total lysophosphatidylcholine in a sample. In certain embodiments, the experienced practitioner can measure an amount or value from a sample that correlates with total lysophosphatidylcholine. For example, in certain samples of subjects, a fraction of total lysophosphatidylcholine may be free of other molecules while another fraction of total lysophosphatidylcholine may be bound by other molecules. For example, a fraction of total lysophosphatidylcholine may be bound by albumin. Sample preparation and measurement techniques used by the experienced practitioner can affect the amount of lysophosphatidylcholine actually measured. For example, precipitation and / or purification techniques can separate free and bound lysophosphatidylcholine. Detection techniques may be more sensitive to free lysophosphatidylcholine or bound lysophosphatidylcholine. This measured amount can be correlated with the amount of total lysophosphatidylcholine in the sample according to methods available to the experienced practitioner. In certain embodiments, a measurement of free lysophosphatidylcholine is used to indicate the amount of total lysophosphatidylcholine in the sample. In certain embodiments, a measurement of bound lysophosphatidylcholine is used to indicate the amount of total lysophosphatidylcholine in the sample. In certain embodiments, a measurement of free and bound lysophosphatidylcholine is used to indicate the amount of total lysophosphatidylcholine in the sample. In certain embodiments, free lysophosphatidylcholine can be used for prognosis or diagnosis in the methods of the invention. In certain embodiments, bound lysophosphatidylcholine can be used for prognosis or diagnosis in the methods of the invention. In certain embodiments, free and bound lysophosphatidylcholine can be used for prognosis or diagnosis in the methods of the invention.

Each additional biomarker can be any type of biomarker for a systemic inflammatory condition known to those skilled in the art that includes biomarkers of protein, peptide, nucleic acid, lipid, phospholipid and metabolite (e.g., protein, peptide, nucleic acid, nucleoside, lipid or phospholipid, metabolite).

Other exemplary biomarkers for the prognosis or diagnosis of a systemic inflammatory condition, and the methods for their evaluation, are described in US Patent Application Publication No. 20030194752, 20040096917, 20040097460, 20040106142, 20040157242, and US Provisional Application No. 60 / 671,620, filed on April 15, 2005, 60 / 671,941, filed on April 15, 2005, and 60 / 674,046, filed on April 22, 2005. Other exemplary biomarkers for sepsis include endotoxin; Bacterial DNA; acute phase proteins such as protein C, procalcitonin, LBP-LPS binding protein; coagulation factors such as fibrin degradation products, antithrombin III, dimer D; membrane cell markers such as HLA-DR, CD-64, E-selectin; hormones such as cortisol, ACTH; soluble receptors such as CD-14, sTNFRI. sTNF-RII; and cytokines such as TNF, IL-6, IL-8 and IL-10; and others such as D-dimer, prothrombin time, activated partial thromboplastin time, plasminogen activator inhibitor-1, soluble thrombomodulin, IL-6, IL-10, IL-8, protein C, thrombin activatable fibrinolysis inhibitor , protein S, antithrombin, TNF-a. See, for example, Kinasewitz et al., 2004, Critical Care 8: R82-R90, Bozza et al., 2005, Mem Inst. Oswaldo Cruz 100 (s) 1: 217-221, the contents of which are incorporated by this by reference in its entirety. Preferable biomarkers include C-reactive protein, procalcitonin and IL-6.

In certain embodiments, the present invention provides methods for evaluating a panel of biomarkers of a subject for the diagnosis or prognosis of a systemic inflammatory condition. The panel may comprise any number of biomarkers sufficient to carry out a diagnosis or prognosis of the systemic inflammatory condition according to the criteria of one skilled in the art. The panel must comprise total lysophosphatidylcholine. The panel may additionally comprise other biomarkers for the systemic inflammatory condition, including those described in the previous paragraph. In some embodiments, the panel comprises 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20 or 25 or more biomarkers for the systemic inflammatory condition. Each biomarker must be evaluated using appropriate techniques for the biomarker class. Exemplary techniques are described herein, and other techniques should be apparent to those skilled in the art. In convenient embodiments, biomarkers of the same class, or biomarkers that can be evaluated by the same technique, can be evaluated together in the panel. For example, in a particular embodiment, the metabolite and / or protein biomarkers can be evaluated by immunoassay techniques in an array or chip format, as described below.

The clinical measurement for the systemic inflammatory condition may be any clinical measurement for the systemic inflammatory condition known to those skilled in the art. In certain embodiments, the clinical measurement is according to a clinical severity model for sepsis. Such models include, but are not limited to, the Chronic Health Assessment and Acute Physiology score (APACHE, and its APACHE II and III improvements) (Knaus et al., 1985, Crit Care Med 13: 818-829; Knaus et al., 1991, Chest 100: 1619-1636), the Mortality Prediction Model (MPM) (Lemeshow et al., 1993, JAMA 270: 2957-2963), the Simplified Acute Physiology Score (SAPS) (Le Gall et al., 1984, Crit Care Med 12: 975-977), Multiple Organ Dysfunction score (MODS) (Marshall et al., 1995, Crit Care Med 23: 1638-1652), Failure Assessment score of Sequential Organs (SOFA) (Ferreira et al., 2002, JAMA 286: 1754-1758), the Logistic Organ Dysfunction Score (LODS) (Le Gall et al., 1996, JAMA 276: 802-810) and the concept of predisposition, infection, response, and organ dysfunction (PIRO) (Levy et al., 2003, Intensive Care Med 29: 530-538). In certain embodiments, the clinical measurement comprises one or more measurements used by those skilled in the art to assist in the prognosis or diagnosis of sepsis. Such measurements include, but are not limited to, temperature, heart rate, white blood cell count, differential white blood cell count (monocytes, lymphocytes, granulocytes and / or neutrophils), immature neutrophil ratio and total neutrophil count. of platelets, serum creatinine, urea, lactate, excess bases, pO2 and HCO3-.

Biomarker

In another aspect, the present invention provides the prognosis or diagnosis of a systemic inflammatory condition with a lysophosphatidylcholine biomarker of the invention.

In certain embodiments, the biomarker is a 1-0-acyl-2-smooth-sn-glycerol-3-phosphocholine. For example, in certain embodiments the biomarker is a compound in accordance with formula (I):

image 1

5

10

fifteen

twenty

25

30

35

40

45 or any salt or solvate thereof, where R is an acyl group.

image2

wherein said salt may be coordinated with any ion or ions known to those skilled in the art. The ion or ions can be physiological but do not need to be physiological. For example, the ion or ion may result from contact with the salt during the preparation of the subject's sample, as described below. In some embodiments, the salt is coordinated with an anion, for example, a physiological anion known to those skilled in the art. Exemplary anions include chloride, bromide, phosphate, acetate, carbonate, bicarbonate and sulfate. In some embodiments, the salt is coordinated with a cation, for example, a physiological cation, known to those skilled in the art. Exemplary cations include sodium, potassium, calcium, magnesium and ammonium. In some embodiments, as will be recognized by those skilled in the art, the salt is coordinated with one

or more anions and with one or more cations.

The acyl group can be any acyl group known to those skilled in the art. In certain embodiments the acyl group is saturated. Saturated acyl groups include caproyl, lauroyl, myristoyl, palmitoyl and stearoyl. In other embodiments, the acyl group is monounsaturated. The monounsaturated acyl groups include palmitoleyl and oleyl. In other embodiments, the acyl group is polyunsaturated. Polyunsaturated acyl groups include arachidonyl and linoleyl.

More systematically, in certain embodiments, the acyl group is C10-C22 acyl. In certain embodiments, the acyl group is C14-C22 acyl. Exemplary acyl groups include 16: 0, 18: 0, 18: 1, 18: 2, 20: 4 (n-6) and 22: 6 (n-3), according to family nomenclature for those experts in the technique. In said nomenclature, the first number indicates the number of carbon atoms in the acyl group, and the second number indicates the number of double bonds in the group. For example, "18: 1" indicates an acyl group with 18 carbon atoms and a double bond. Numbers in parentheses, if any, indicate the location of the double bond, and the note "(n-x)" indicates a double bond x positions away from the terminal methyl of the longest fatty acid chain. See Biochem. J., 1978, 171, 21-35; Chem. Phys. Lipids, 1978,21, 159-173; Eur. J. Biochem., 1977, 79, 11-21; Hoppe-Seyler's Z. Physiol. Chem., 1977, 358, 617-631;

J. Lipid Res., 1978, 19, 114-128; Lipids, 1977, 12, 455-468; Mol. Cell Biochem., 1977, 17, 157-171; Biochemical Nomenclature and Related Documents, 2nd edition, Portland Press, 1992, pages 180-190.

In certain embodiments, the acyl group is C14-C22 acyl. In certain embodiments, the acyl group is C16-C20 acyl. In other embodiments, the acyl group is C16-C18 acyl. In certain embodiments, the acyl group is hexadecanoyl or octadecanoyl. In particular embodiments, the acyl group is C16 acyl. In a preferable embodiment, the acyl group is hexadecanoyl. In other particular embodiments, the acyl group is C18 acyl. In a preferable embodiment, the acyl group is octadecanoyl.

The biomarker can be any form of the subject's biomarker, for example any salt or solvate of the biomarker that can be identified by those skilled in the art. In preferable embodiments, the biomarker is in the form of a sodium salt.

In certain embodiments, the biomarker is a metabolite of a compound in accordance with formula (I). For example, the biomarker can be a precursor of a compound in accordance with formula (I) known to those skilled in the art. The precursor may be one or two or three, or in some other embodiments, steps prior to the compound in accordance with formula (I) in a biosynthetic pathway known to those skilled in the art. In other embodiments, the biomarker can be a metabolite in the 3 'direction of a compound in accordance with formula (I) in a biosynthetic pathway known to those skilled in the art. The metabolite in the 3 'direction may be one or two or three, or in some other embodiments, steps after the compound in accordance with formula (I) in a biosynthetic pathway known to those skilled in the art. In certain embodiments, the biosynthetic pathway is a de novo pathway for the synthesis of the platelet activating factor known to those skilled in the art. In other embodiments, the biosynthetic pathway is a remodeling pathway for the synthesis of the platelet activating factor known to those skilled in the art.

In particular embodiments, the metabolite is a 1-0-acyl-2-0-acyl-sn-glycerol-3-phosphocholine. In preferable embodiments, the 2-0-acyl group is any acyl group described above or acetyl. In particular embodiments, the metabolite is a 1-0-acyl-2-0-alkyl-sn-glycerol-3-phosphocholine. In preferable embodiments, the 2-0-alkyl group is any group known to those skilled in the art to modify a glycerol-3-phosphocholine, for example any C14-C22 alkyl.

In certain embodiments, a single biomarker is used for the prognosis or diagnosis of systemic inflammatory condition. In other embodiments, a plurality of biomarkers is used for the prognosis or diagnosis of systemic inflammatory condition. The biomarkers in the plurality may be according to the invention according to what is described above, or the plurality may comprise biomarkers according to the invention together with other biomarkers for the diagnosis or prognosis of a systemic inflammatory condition known to those skilled in the art. The technique. The biomarker can be any type of biomarker for a systemic inflammatory condition known to those skilled in the art that includes the biomarkers protein, peptide, nucleic acid, lipid, phospholipid and metabolite (e.g., protein, peptide, nucleic acid, nucleoside, lipid or phospholipid metabolite).

Other exemplary biomarkers for the prognosis or diagnosis of a systemic inflammatory condition, and methods for their evaluation, are described in US Patent Application Publication No. 20030194752, 20040096917, 20040097460, 20040106142, 20040157242, and US Provisional Application No. 60 / 671,620, filed on April 15, 2005, 60 / 671,941, filed on April 15, 2005, and 60 / 67,046, filed on April 22, 2005. Other exemplary biomarkers for sepsis include endotoxin; Bacterial DNA; acute phase proteins such as protein C, procalcitonin, LBP-LPS binding protein; coagulation factors such as fibrin degradation products, antithrombin III, dimer D; membrane cell markers such as HLA-DR, CD-64, E-selectin; hormones such as cortisol, ACTH; soluble receptors such as CD-14, sTNFRI, sTNF-RII; and cytokines such as TNF, IL-6, IL-8 and IL-10; and others such as D-dimer, prothrombin time, activated partial thromboplastin time, plasminogen activator inhibitor-1, soluble thrombomodulin, IL-6, IL10, IL-8, protein C, thrombin activatable fibrinolysis inhibitor, protein S, antithrombin, TNF-a, See, for example, Kinasewitz et al., 2004, Critical Care 8: R82-R90, Bozza et al., 2005, Mem. Inst. Oswaldo Cruz 100 (s) 1: 217

221. Preferable biomarkers include C-reactive protein, procalcitonin and IL-6.

In certain embodiments, the present invention provides methods for evaluating a panel of biomarkers of a subject for the diagnosis or prognosis of a systemic inflammatory condition. The panel may comprise any number of biomarkers sufficient to carry out a diagnosis or prognosis of the systemic inflammatory condition according to the criteria of one skilled in the art. The panel must comprise one or more biomarkers of the invention, for example, a biomarker in accordance with formula I or formula Ia. The panel may additionally comprise other biomarkers for the systemic inflammatory condition, including those described in the previous paragraph. In some embodiments, the panel comprises 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20 or 25 or more biomarkers for the systemic inflammatory condition. Each biomarker must be evaluated using appropriate techniques for the biomarker class. Exemplary techniques are described herein, and other techniques should be apparent to those skilled in the art. In convenient embodiments, biomarkers of the same class, or biomarkers that can be evaluated by the same technique, can be evaluated together in the panel. For example, in a particular embodiment, the metabolite and / or protein biomarkers can be evaluated by immunoassay techniques in an array or chip format, as described below.

Lysophosphatidylcholine measurement

In this section and the sections that follow, unless otherwise specified, the term "lysophosphatidylcholine" refers to the total lysophosphatidylcholine or a biomarker of lysophosphatidylcholine of the invention.

In certain embodiments of the invention, the method of measuring lysophosphatidylcholine is not critical. Accordingly, the present invention provides methods for the diagnosis or prognosis of a systemic inflammatory condition comprising the only step of assessing the risk of systemic inflammatory condition from lysophosphatidylcholine.

Total lysophosphatidylcholine may be measured by one who practices a method of the invention in any manner whatsoever. Exemplary techniques are described herein. As described above, any technique indicating lysophosphatidylcholine in the sample can be used in the methods of the invention. In certain embodiments, the methods are based on free lysophosphatidylcholine in the sample. In certain embodiments, the methods are based on the lysophosphatidylcholine bound in the sample. In certain embodiments, the methods are based on free and bound lysophosphatidylcholine in the sample.

The amount of a lysophosphatidylcholine biomarker can be measured by one who practices a method of the invention in any manner whatsoever. Exemplary techniques are described herein.

When a plurality or panel of biomarkers must be evaluated, each individual biomarker must be evaluated according to an appropriate technique for that biomarker. In advantageous embodiments, biomarkers that can be evaluated by the same technique or by compatible techniques can be evaluated together. For example, the protein, peptide, lipid, phospholipid and metabolite biomarkers that can be evaluated by immunoassays can be evaluated together or in groups according to techniques known to those skilled in the art.

In one embodiment, only a single biological sample taken at a single point in time of the subject is used to make a prognosis or diagnosis of sepsis. In another embodiment, a plurality of biological samples taken at different points in the subject's time are used to make a prognosis or diagnosis of sepsis.

In a specific embodiment, the amount of lysophosphatidylcholine is obtained by using samples collected from the subject at a time point. In another specific embodiment, the amount of lysophosphatidylcholine is obtained by using samples obtained from the subject at separate time points. In one example, these samples are obtained from the subject once or, alternatively, on a daily basis, or more frequently, for example, every 2, 3, 4, 6, 8 or 12 hours.

Lysophosphatidylcholine can be obtained from any biological sample, which can be, by way of example and not as a limitation, blood, plasma, serum, saliva, sputum, urine, cerebral spinal fluid, cells, a cell extract, a tissue sample, a biopsy tissue, a stool sample or any sample that can be obtained from a subject by using techniques well known to those skilled in the art. The precise biological sample that is taken from the subject may vary, but the sampling is preferably minimally invasive and easily performed by conventional techniques.

The biological sample can be processed or purified according to the criteria of those skilled in the art based on, for example, the type of biomarker and the measurement technique. For example, when the biomarker is a lipid or phospholipid metabolite, the sample can be processed by extraction and / or chromatography. When the biomarker is a protein or peptide, for example when a panel of biomarkers must be evaluated, the sample can be processed by precipitation, centrifugation, filtration and / or chromatography. When the biomarker is a nucleic acid, for example when a panel of biomarkers must be evaluated, the sample can be processed to isolate the nucleic acids by extraction, precipitation and / or chromatography.

Some part of the mixture of peptides, proteins, nucleic acids, phospholipids, and metabolites (for example, metabolites or peptides, proteins, phospholipids or nucleosides) and / or other molecules in the sample can then be resolved as a biomarker profile. This can be achieved by measuring the quantities of the biomarkers in the biomarker profile. A biomarker profile comprises a plurality of one or more types of biomarkers (e.g., a phospholipid, an mRNA, a cDNA, a protein and / or a carbohydrate, etc.), or an indication thereof, together with amounts of biomarkers A biomarker profile may comprise at least one of said biomarker or indication thereof. Multiple biomarkers may be in the same or different classes, such as, for example, phospholipid and a polypeptide.

These quantities can be determined through the use of any reproducible measurement technique or combination of measurement techniques. Such techniques include those that are well known in the art that include any technique described herein. Typically, such techniques are used to measure amounts by utilizing a biological sample taken from a subject at a single point in time or multiple samples taken at multiple points in time. In a preferred embodiment, an exemplary technique for obtaining a biomarker profile of a sample taken from a subject is the immunoassay. The biomarker profile can be generated by using a kit, such as a kit described below.

In certain embodiments, the biomarker detection methods include their detection through interaction with a biomarker specific antibody. For example, antibodies directed to the biomarker of the invention. Antibodies can be generated by using standard techniques well known to those skilled in the art. In specific embodiments, the antibodies may be polyclonal, or more preferably, monoclonal. For example, an intact antibody, or an antibody fragment (for example, scFv, Fab or F (ab ') 2) can be used.

For example, antibodies, or fragments of antibodies, specific to a biomarker can be used to quantitatively or qualitatively detect the presence of a biomarker. This can be achieved, for example, by immunofluorescence techniques. The antibodies (or fragments thereof), additionally, can be used histologically as in immunoelectronic microscopy or immunofluorescence, for the on-site detection of a biomarker. In situ detection can be achieved by extracting a biological sample (for example, a biopsy specimen) from a patient, and applying it to a labeled antibody that is directed to a biomarker. The antibody (or fragment) is preferably applied by covering the antibody (or fragment) on a biological sample. Through the use of such a procedure, it is possible to determine not only the presence of the biomarker, but also its distribution, in a particular sample. A wide variety of well-known histological methods (such as staining procedures) can be used to achieve such in situ detection.

Immunoassays for a biomarker typically comprise incubating a biological sample of a detectably labeled antibody capable of identifying a biomarker, and detecting the bound antibody by any of a number of techniques well known in the art. As discussed in greater detail below, the term "labeling" may refer to the direct labeling of the antibody through, for example, the coupling of (i.e., physical connection) a detectable substance to the antibody, and may also be referred to as labeling. indirect antibody by reactivity with another reagent that is directly labeled. Examples of indirect labeling include detection of a primary antibody by the use of a fluorescently labeled secondary antibody.

The biological sample may be contacted and immobilized in a solid phase carrier or vehicle such as nitrocellulose, or other solid support that is capable of immobilizing soluble cells, cell particles or proteins.

The support can then be washed with appropriate buffers followed by treatment with the specific antibody of the detectably labeled typographic identity gene. The solid phase support can then be washed with the buffer a second time to remove unbound antibody. The amount of label attached on solid support can then be detected by conventional methods.

By "solid phase carrier or vehicle" is meant any support capable of binding an antigen or an antibody. Well-known supports or vehicles include glass, polystyrene, polypropylene, polyethylene, dextran, nylon, amylases, natural and modified celluloses, polyacrylamides and magnetite. The nature of the vehicle may be soluble to some extent or insoluble for the purposes of the present invention. The support material can have virtually any possible structural configuration as long as the coupled molecule is capable of binding to an antigen or antibody. Thus, the configuration of the support can be spherical, as in an account, or cylindrical, as in the inner surface of a test tube, or the outer surface of a rod. Alternatively, the surface can be flat such as a sheet, test strip, etc. Preferable supports include polystyrene beads. Those skilled in the art will know many other appropriate vehicles for binding the antibody or antigen, or will be able to find out the same by using routine experimentation.

One of the ways in which a biomarker-specific antibody can be detectably labeled is by connecting it to an enzyme and using it in an enzyme immunoassay (EIA) (Voller, 1978, "The Enzyme Linked Immunosorbent Assay ( ELISA) ", Diagnostic Horizons 2: 1-7, Microbiological Associates Quarterly Publication, Walkersville, MD; Voller et al., 1978, J. Clin. Pathol. 31: 507-520; Butler, JE, 1981, Meth. Enzymol. 73: 482-523; Maggio, E. (ed.), 1980, Enzyme Immunoassay, CRC Press, Boca Raton, FL; Ishikawa, E. et al. (Eds.), 1981, Enzyme Immunoassay, Kgaku Shoin, Tokyo The enzyme that is bound to the antibody will react with an appropriate substrate, preferably a chromogenic substrate, in such a way as to produce a chemical moiety that can be detected, for example, by spectrophotometric, fluorimetric means or by visual means. to detectably label the antibody include, but n or are limited to, malate dehydrogenase, staphylococcal nuclease, delta-5-steroid isomerase, alcohol dehydrogenase in yeast, alphaglycerophosphate, dehydrogenase, triosaphosphate isomerase, radish peroxidase, alkaline phosphatase, asparaginase, glucose oxidase, beta-galactosidase urease , catalase, glucose-6-phosphate dehydrogenase, glucoamylase and acetylcholinesterase. Detection can be achieved by colorimetric methods that employ a chromogenic substrate for the enzyme. Detection can also be achieved by visual comparison of the degree of enzymatic reaction of a substrate compared to standards prepared in a manner.

Detection can also be achieved by using any variety of other immunoassays. For example, by radioactively labeling antibodies or antibody fragments, it is possible to detect a biomarker through the use of a radioimmunoassay (RIA) (see, for example, Weintraub, B., Principles of Radioimmunoassays, Seventh Training Course on Radioligand Assay Techniques , The Endocrine Society, March 1986. The radioactive isotope (for example, 125I, 131I, 35S or 3H) can be detected by such means as the use of a gamma counter or scintillation counter or by autoradiography.

It is also possible to label the antibody with a fluorescent compound. When the fluorescently labeled antibody is exposed to light of appropriate wavelength, its presence can then be detected due to fluorescence. Among the most commonly used fluorescent labeled compounds are fluorescein isothiocyanate, rhodamine, phycoerythrin, phycocyanin, allophycocyanin, o-phthaldehyde and fluorescamine.

The antibody can also be detectably labeled by the use of fluorescence emitting metals such as 152Eu, or others of the lanthanide series. These metals can be bound to the antibody by using said metal chelating groups such as diethylenetriaminepentaacetic acid (DTPA) or ethylenediaminetetraacetic acid (EDTA).

The antibody can also be detectably labeled by coupling it to a chemiluminescent compound. The presence of the chemiluminescent labeled antibody is then determined by detecting the presence of luminescence that arises during the course of a chemical reaction. Examples of particularly useful chemiluminescent labeled compounds are luminol, isoluminol, thermic acridinium ester, imidazole, acridinium salt and oxalate ester.

Similarly, a bioluminescent compound can be used to label the antibody of the present invention. Bioluminescence is a type of chemiluminescence found in biological systems in which a catalytic protein increases the efficiency of the chemiluminescent reaction. The presence of a bioluminescent protein is determined by detecting the presence of luminescence. The bioluminescent compounds important for labeling purposes are luciferin, luciferase and aequorin.

In advantageous embodiments, the antibodies may be in the form of an arrangement. Said arrangement can be used for the measurement or detection of a plurality or panel of biomarkers simultaneously. The arrangement can be any arrangement for antibodies known to those skilled in the art. In certain embodiments, a plurality of antibodies may be in the form of an antibody chip for the detection of a plurality or panel of biomarkers. Exemplary antibody arrays and antibody chips are described in U.S. Patent Application Publication No. 20050048566, 20050054015, 20050037343, 20050095591, 20050100947, 20040161748, 20040097460, and 20040096917. Commercial antibody arrays may be used or adapted for this invention, and include, but not limited to, those available from Whatman Schliecher & Schuell, Eurogentec, Stigma Aldrich, Novagen and Chemicon International. The binding of the antigen to the antibody can proceed according to the appropriate techniques for the antibody or antibody chip arrangement, and include, but is not limited to, fluorescence, surface plasmon resonance and mass spectrometry.

In another embodiment, specific binding molecules other than antibodies, such as aptamers, can be used to bind to biomarkers. In yet another embodiment, the biomarker profile may comprise a measurable aspect of an infectious agent (for example, lipopolysaccharides or viral proteins) or a component thereof.

The amounts of biomarkers in a biomarker profile can also, for example, be generated by using one or more of the following methods described below. For example, the methods may include nuclear magnetic resonance spectroscopy (NMR), a mass spectrometry method, such as mass spectrometry - electrospray ionization (ESI-MS), ESI-MS / MS, ESI-MS / (MS ) "(n is an integer greater than zero), matrix-assisted laser desorption ionization -flight time-mass spectrometry (MALDI-TOF-MS), surface-enhanced laser desorption ionization -flight time -speed spectrometry mass (SELDI-TOFMS), silicon desorption / ionization (GOD), secondary ion mass spectrometry (SIMS), quad time flight (QTOF), chemical ionization at atmospheric pressure - mass spectrometry (APCIMS), APCI- MS / MS, APCI- (MS) n, atmospheric pressure photoionization-mass spectrometry (APPI-MS), APPIMS / MS, and APPI- (MS) n. Other mass spectrometry methods may include, inter alia, quadrupole , Fourier transform mass spectrometry (FTMS) and trap i Other appropriate methods may include chemical extraction division, column chromatography, ion exchange chromatography, hydrophobic liquid chromatography (reverse phase), isoelectric focusing, polyacrylamide gel electrophoresis in one dimension (PAGE), polyacrylamide gel electrophoresis in two dimensions (2D-PAGE) or other chromatography, such as liquid, gas or thin layer chromatography, or any can be fractionated prior to the application of the separation method.

In one embodiment, flight time laser desorption / ionization mass spectrometry is used to determine the amount of a biomarker where the biomarker is a molecule that has been ionized and evaporated from an immobilization support by incident laser radiation. A variety of laser desorption / ionization techniques are known in the art (see, for example, Guttman et al., 2001, Anal. Chem. 73: 1252-62 and Wei et al., 1999, Nature 399: 243-246 .

Desorption / ionization laser flight time mass spectrometry allows the generation of large amounts of information in a relatively short period of time. A biological sample is applied to one of several varieties of a support that binds to all biomarkers, or a subset of them, in the sample. Cell lysates or samples are applied directly to these surfaces in volumes as small as 0.5 PL, with or without prior purification or fractionation. The lysates or sample can be concentrated or diluted prior to application on the surface of the support. Then laser desorption / ionization is used to generate mass spectra of the sample, or samples, in as little as three hours.

The analysis by liquid chromatography - mass spectrometry produces a spectrum of mass intensity, the peaks of which represent various components of the sample, where each component has a characteristic mass and load ratio (m / z) and characteristic retention time (rt) . The presence of a peak with the ratio m / z and retention time of a biomarker indicates that the marker is present. The peak representing a marker can be compared with a corresponding peak of another spectrum (for example, from a control sample) to obtain an internal standard) can be used when a quantitative measurement is desired. In addition, deconvolution software is available to separate overlapping peaks. The retention time depends to some extent on the conditions employed in performing the separation by liquid chromatography.

In MALDI mass spectrometry (MALDI-MS), various mass analyzers can be used, for example, magnetic deflection / magnetic sector instruments in single or triple quadrupole mode (MS / MS), Fourier transform and flight time (TOF) ), which include orthogonal flight time (O-TOF), configurations as is known in the mass spectrometry technique. For the desorption / ionization process, numerous matrix / laser combinations can be used. Ion trap and reflectron configurations can also be used.

Mass-electrospray ionization spectrometry (ESI-MS) is widely applicable for the analysis of macromolecules, which include proteins, nucleic acids, and carbohydrates (Fenn et al., 1989, Science 24: 64-71; Crain et al. , 1998, Curr. Opin. Biotechnol. 9: 25-34; Smith et al., 1990, Anal Chem. 62: 882-99; Han & Gross, 1994, Proc Natl Acad Sci USA 91: 10635-10639). Electrospray techniques have been used to separate and measure biomarkers such as those of formula I and formula Ia (see Petkovic et al., 2001, Anal Biochem. 289 (2): 202-16; Pulfer & Murphy, 2003, Mass Spec Rev 22: 332-364; Han & Gross, 1995, J. Amer. Soc. Mass Spec. 6: 1202-1210).

For the following classes of metabolites, the following sources provide additional guidance in the mass spectral analysis of said (1) lipid molecules (see, for example, Fenselau, C., ACS Symp. Ser., 541: 1-7 (1994 )); (2) volatile metabolites (see, for example, Lauritsen and Lloyd, D., ACS Symp Ser. 541: 91-106 (1994)); (3) carbohydrates (see, for example, Fox and Black, ACS Symp. Ser. 541: 107-131 (1994); (4) nucleic acids (see, for example, Edmonds et al., ACS Symp. Ser., 54: 14.7-158 (1994); and (5) proteins (see, e.g., Vorm, O. et al, Anal Chem 66: 3281-3287 (1994); and Vorm and Mann, J. Am.... Soc. Mass. Spectrom. 5: 955-958 (1994)).

In certain embodiments, the species in the biological sample can be separated by liquid chromatography in conjunction with any of the mass spectrometry techniques described above. Such techniques of liquid / mass spectrometry chromatography have proven useful for the separation of biomarkers such as proteins, peptides, nucleic acids, lipids, phospholipids and metabolites. For example, because it has achieved the separation of species sensitive phospholipids LC / MS (see Kim et al, 1994, Anal Chem 66 (22): W>} A77-82; Ma & Kim, 1995,.. Anal Biochem. 226 (2): 293-301), said techniques can be used to separate and measure a biomarker in accordance with formula I or formula Ia.

In specific embodiments of the invention, the biomarkers in a plurality or panel of biomarkers are nucleic acids. Such biomarkers and corresponding amounts of the biomarker profile can be generated, for example, by detecting the expression product (for example, a polynucleotide or polypeptide) of one or more genes described herein. In a specific embodiment, the biomarkers and corresponding amounts in a biomarker profile are obtained by detecting and / or analyzing one or more nucleic acids by using any method well known to those skilled in the art that includes, but is not limited to. in no way a, hybridization, microarray analysis, RT-PCR, nuclease protection assays and Northern blot analysis. As will be recognized by those skilled in the art, in convenient embodiments, the biological sample can be divided, with a portion evaluated for nucleic acid biomarkers, and another part evaluated for other biomarkers such as proteins, peptides, lipids, phospholipids and metabolites. . In fact, the biological sample can be divided as many times as desired by the experienced practitioner to facilitate the evaluation or measurement of each biomarker in a plurality or panel of biomarkers.

In certain embodiments of the invention, nucleic acid arrays are employed to generate amounts of biomarkers in a biomarker profile by detecting the expression of any or more of the genes described herein. In one embodiment of the invention, a microarray, such as a cDNA microarray, is used to determine the amounts of biomarkers in a biomarker profile. The diagnostic use of cDNA arrays is well known in the art. (See, for example, Zou et. al., 2002, Oncogene 21: 4855-4862; as well as Draghici, 2003, Data Analysis Tools for DNA Microarrays, Chapman & Hall / CRC. Exemplary methods for the analysis of cDNA microarrays are described below, and in the examples in Section 6, index.

In certain embodiments, amounts for biomarkers in a biomarker profile are obtained by hybridization to detectably labeled nucleic acids of the array that represent or correspond to nucleic acid sequences in mRNA transcripts present in a biological sample (e.g. , CDNA labeled in fluorescent form synthesized from the sample) to a microarray comprising one or more probe stains.

Nucleic acid arrays, for example, microarrays, can be produced in a number of ways, of which several are described herein below. Preferably, the arrangements are reproducible, which allows multiple copies of a given arrangement to be produced and result from such microarrays compared to each other. Preferably, the arrangements are made from materials that are stable under binding conditions (eg, nucleic acid hybridization). Those skilled in the art will know supports, substrates or carriers for hybridizing appropriate test probes with probe spots on an array, or be able to ascertain the same for the use of routine experimentation.

Several chromatographic techniques can be used to separate biomarkers. For example, amplification products can be separated by agarose, agarose acrylamide or polyacrylamide gel electrophoresis by using conventional methods. See Sambrook et al., 2001. Several techniques for detecting biomarkers quantitatively without electrophoresis can also be used according to the invention (see, for example, PCR Protocols, A Guide to Methods and Applications, Innis et al., 1990, Academic Press, Inc. NY). For example, chromatographic techniques can be used to effect separation. There are many kinds of chromatography that can be used in the present invention: adsorption; division, ion exchange and molecular sieve, HPLC, and many specialized techniques to use them including column chromatography, paper, thin film and gas (Freifelder, Physical Biochemistry Applications to Biochemistry and Molecular Biology, 2nd ed., Wm. Freeman and Co. , New York, NY, 1982).

In certain embodiments, one or more of the biomarkers is a protein. In a specific embodiment, a biomarker profile is generated by detecting and / or analyzing one or more proteins and / or discriminating fragments thereof by using any method known to those skilled in the art to detect proteins that includes , but not limited to microarray protein analysis, immunohistochemistry and mass spectrometry.

Standard techniques can be used to determine the amount of protein or proteins of interest present in a sample. For example, standard techniques can be employed by using, for example, immunoassays such as, for example Western blotting, immunoprecipitation followed by polyacrylamide gel electrophoresis with sodium dodecyl sulfate, (SDS-PAGE), immunocytochemistry, and the like to determine the amount of protein or proteins of interest present in a sample. An exemplary agent for detecting a protein of interest is an antibody capable of specifically binding a protein of interest, preferably a detectably labeled antibody, directly or indirectly.

For such detection methods, if it is desired that a protein in the sample be analyzed, it can be easily isolated by using techniques that are well known to those skilled in the art. Protein isolation methods may, for example, be such as those described in Harlow and Lane, 1988, Antibodys: A Laboratory Manual. Cold Spring Harbor Laboratory Press (Cold Spring Harbor, New York).

In certain embodiments, the methods of detection of the protein or proteins of interest include their detection through interaction with a specific protein antibody. For example, antibodies directed to a protein of interest. Antibodies can be generated by using standard techniques well known to those skilled in the art. In specific embodiments, the antibodies may be polyclonal, or more preferably, monoclonal. For example, an intact antibody, or an antibody fragment (for example, scFv, Fab or F (ab ') 2) can be used. Exemplary immunoassays are described above.

In some embodiments, a protein chip assay (see, for example, Zhu & Snyder, 2003, Curr. Opin. Chem. Biol. 7: 55-63; Mitchell, 2002, Nature Biotechnology 20: 225-229) is used to measure quantities for biomarkers in the biomarker profile. See also, for example, Lin, 2004, Modern Pathology, 1-9; Li, 2004, Journal of Urology 171, 1782-1787; Wadsworth, 2004, Clinical Cancer Research, 10, 1625-1632; Prieto, 2003, Journal of Liquid Cromatography & Related Technologies 26, 2315-2328; Coombes, 2003, Clinical Chemistry 49, 1615-1623; Mian, 2003, Proteomics 3, 1725-1737; Lehre et al., 2003, BJU International 92, 223-225; and Diamond, 2003, Journal of the American Society for Mass Spectromtry 14, 760-765. Particularly useful in certain embodiments of the invention are antibody chips that facilitate detection by MALDI or SELDI (see, for example, Wang, et al., 2001, Int'l. J. of Cancer 92: 8 71-876; Figeys , 2002, Proteomics 2: 373-382: Sonksen et al. 1998, Anal. Chem. 70: 2731-6; Gl6kler, & Angenendt, 2003, J. Cromatography B, 797: 229-240.

"Selective" antibodies to "myomarkers" of the "invention"

Antibodies against the biomarkers of the invention can be produced or obtained according to any technique apparent to those skilled in the art. For example, antibodies against phospholipids and metabolites of the invention can be prepared or isolated according to the techniques described in Mourdjeva et al., 2005 Apoptosis 10 (1): 209-17, von Landenberg et al., 1999, J Autoimmun. 13: 215-23, or Menon et al. J Autoimmun. 10:43

57.

Antibodies against polypeptides, or metabolites thereof, can be prepared according to standard techniques.

An isolated phospholipid of the invention, or a metabolite or fragment thereof, can be used as an immunogen to generate antibodies by using standard techniques for the preparation of polyclonal and monoclonal antibody. An immunogen is typically used to prepare antibodies by immunization of an appropriate subject, (eg, rabbit, goat, mouse or other mammal). An appropriate immunogenic preparation may comprise, for example, chemically synthesized or recombinantly expressed polypeptide. The preparation may also include an adjuvant, such as Freud's incomplete or complete adjuvant, or similar immunostimulatory agent. In certain embodiments, to facilitate antibody production, the phospholipid of the invention, or a metabolite or fragment thereof, can be coupled to a vehicle, for example a protein such as hemocyanin extracted from mollusk called 'lapa californian', bovine serum albumin , thyroglobulin, and ovalbumin.

Polyclonal antibodies can be prepared by immunization of an appropriate subject with a phospholipid of the invention, or a metabolite or fragment thereof, of the invention as an immunogen. The antibody titer in the immunized subject can be monitored over time by standard techniques, such as with an enzyme linked assay inmonoenzimático (ELISA) using immobilized polypeptide techniques. If desired, the antibody molecules can be isolated from the subject (for example, from the blood) and further purified by well known techniques, such as protein A chromatography to obtain the IgG fraction. Alternatively, antibodies specific to a phospholipid of the invention, or a metabolite or fragment thereof, of the invention may be selected for (for example, partially purified) or purified by, for example, affinity chromatography. For example, a recombinant and purified (or partially purified) protein of the invention is produced according to what is described herein, and is covalently or non-covalently coupled to a solid support such as, for example, a column of chromatography The column can then be used to affinity purify the antibodies specific for the proteins of the invention from a sample comprising antibodies directed against a large number of different epitopes, thereby generating a substantially purified antibody composition, that is, a that is substantially free of contaminating antibodies. By a composition of substantially purified antibodies it is meant, in this context, that the antibody sample comprises a maximum of only 30% (dry weight) of contaminating antibodies directed against epitopes other than those in the desired protein or polypeptide of the invention. , and preferably at most 20%, even more preferably at most 10%, and much more preferably at most 5% (dry weight) of the sample are contaminating antibodies. A composition of purified antibodies means that at least 99% of the antibodies in the composition are directed against the desired protein or polypeptide of the invention.

At an appropriate time after immunization, for example, when the specific antibody titres are higher, antibody producing cells can be obtained from the subject and used to prepare monoclonal antibodies by standard techniques, such as the hybridoma technique originally described by Kohler and Milstein (1975, Nature 256: 495-497), the B-cell hybridoma technique (Kozbor et al., 1983, Immunol Today 4:72), the EBV-hybridoma technique (Cole et al., 1985 , in Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, Inc., pp. 77-96) or triomas techniques. The technology for producing hybridomas is well known (see for example, Current Protocols in Immunology, Coligan et al. (Eds.) John Wiley & Sons, Inc., New York, N.Y. (1994)). The hybridoma cells that produce a monoclonal antibody of the invention are detected by the selection of hybridoma culture supernatants for antibodies that bind to the polypeptide of interest, for example, by the use of a standard ELISA assay.

Alternatively to the preparation of the hybridoma secreting monoclonal antibody, a monoclonal antibody directed against a polypeptide of the invention can be identified and isolated by the selection of a recombinant combinatorial immunoglobulin library (eg, an antibody phage expression library ) with the polypeptide of interest. Kits for generating and selecting phage expression libraries are commercially available (e.g., Pharmacy Recombinant Antibody System, Catalog No. 279400-01; and Stratagene SurfZAP Phage Expression Kit, Catalog No. 240612). Additionally, examples of particularly receptive methods and reagents for use in the generation and selection of an antibody expression library can be found in, for example, US Patent No. 5,223,409; PCT Publication No. WO 92/18619; PCT Publication No. WO 91/17271; PCT Publication No. WO 92/20791; PCT Publication No. WO 92/15679; PCT Publication No. WO 93/01288; PCT Publication No. WO 92/01047; PCT Publication No. WO 92/09690; PCT Publication No. WO 90/02809; Fuchs et al., 1991, BioTechnology 9: 1370-1372; Hay et al., 1992, Hum Antibod Hybridomas 3: 81-85; Huse et al., 1989, Science 246: 1275-1281; Griffiths et al., 1993, EMBO J. 12: 725-7 4.

Additionally, recombinant, such as chimeric and humanized monoclonal antibodies, comprising both human and non-human portions, which can be made using standard recombinant DNA techniques, antibodies are within the scope of the invention. A chimeric antibody is a molecule in which different parts are obtained from different animal species, such as those that possess a variable region obtained from a murine mAb and a constant region of human immunoglobulin (See, for example, Cabilly et al., Patent U.S. No. 4,816,567; and Boss et al., U.S. Patent No. 4,816,397). Humanized antibodies are antibodies from molecules of non-human species that possess one or more regions that determine complementarity (CDRs) of a non-human species and a framework region of a human immunoglobulin molecule (see, for example, Queen, US Patent No. 5,585,089).

Such chimeric and humanized monoclonal antibodies can be produced by recombinant DNA techniques known in the art, for example using methods described in PCT Publication WO 87/02671 No.; European Patent Application 184,187; European Patent Application 171,496; European Patent Application 173,494; PCT Publication No. WO 86/01533; U.S. Patent No. 4,816,567; European Patent Application 125,023; Better et al., 1988, Science 240: 1041-1043; Liu et al., 1987, Proc Natl Acad Sci. 84: 34393443; Liu et al., 1987, J. Immunol. 139: 3521-3526; Sun et al., 1987, Proc Natl Acad Sci. 84: 214-218; Nishimura et al., 1987, Cancer Res. 47: 999-1005; Wood et al., 1985, Nature 314: 446-449; and Shaw et al., 1988, J. Natl. Cancer Inst. 80: 1553-1559; Morrison, 1985, Science 229: 1202-1207; OI et al., 1986, BioTechniques 4: 214; US Patent No. 5,225,539, Jones et al., 1986. Nature 321: 552-525; Verhoeyan et al., 1988, Science 239: 1534; and Beidler et al., 1988, J. Immunol. 141: 4053-4060.

Completely human antibodies can be produced, for example, by the use of transgenic mice that are unable to express endogenous immunoglobulin light and heavy chain genes, but which can express human heavy and light chain genes. Transgenic mice are normally immunized with a selected antigen, for example, all or a part of a polypeptide of the invention. Monoclonal antibodies directed against the antigen can be obtained by using conventional hybridoma technology. Human immunoglobulin transgenes housed by transgenic mice are redisposed during B cell differentiation, and subsequently undergo class change and somatic mutation. Thus, by using said technique, it is possible to produce therapeutically useful IgG, IgA and IgE antibodies. For an overview of this technology to produce human antibodies, see Lonberg and Huszar (1995, Int. Rev Immunol. 13: 65-93). For a detailed discussion of this technology for producing human antibodies and human monoclonal antibodies and protocols for producing such antibodies, see, for example, US Patent No. 5,625,126; U.S. Patent No. 5,633,425; U.S. Patent No. 5,569,825; U.S. Patent No. 5,661,016; and U.S. Patent No. 5,545,806. In addition, companies such as Abgenix, Inc. (Fremont, Calif.), May undertake to provide human antibodies directed against a selected antigen by using technology similar to that described above.

Fully human antibodies that recognize a selected epitope can be generated by using a technique called "guided selection." In this approach a selected non-human monoclonal antibody, for example, a mouse antibody, is used to guide the selection of a completely human antibody that recognizes the same epitope. (Jespers et al., 1994, BioTechnology 12: 899-903).

Enzymatic assay for myomarkers of the invention

In advantageous embodiments, the total lysophosphatidylcholine can be detected, measured or monitored by one or more enzymatic assays. Enzymatic assays may be any enzymatic assay known to those skilled in the art that is useful for detecting, measuring or monitoring one or more of the biomarkers of the invention.

In certain embodiments, enzyme assays may be according to published application JP 2002-17938 (Kishimoto et al., 2002, Method of Measuring Phospholipid), or according to Kishimoto et al., 2002, Clinical Biochem. 35: 411-416.

In certain embodiments, total lysophosphatidylcholine can be measured by contacting a sample of the invention with an enzyme capable of hydrolyzing lysophosphatidylcholine to produce glycerophosphorylcholine. The enzyme can be any enzyme known to those skilled in the art. Exemplary enzymes include lysophospholipases such as EC 3.1.1.5 (commercially available from, for example, Asahi Chemical Co.). In certain embodiments, lysophospholipase preferentially hydrolyzes lysophospholipids relative to other phospholipids. In certain embodiments, the lysophospholipase is from Bacillus. In certain embodiments, lysophospholipase is according to JP 200217938.

The resulting glycerophosphorylcholine can be detected, measured or monitored according to any technique evident to those skilled in the art. For example, in certain embodiments, glycerophosphorylcholine may be contacted with a diesterous glycerophosphorylcholine known to those skilled in the art (eg EC 3.1.4.2) under appropriate conditions to produce choline. The resulting choline can be contacted with a choline oxidase known to those skilled in the art (eg EC 1.1.3.17) under appropriate conditions to produce peroxide. The use of choline oxidase allows the method to detect lysophosphatidylcholine instead of other lysophospholipids such as lysophospholipids comprising serine or ethanolamine. The resulting peroxide can be detected by any technique evident to those skilled in the art that includes, for example, colorimetric techniques.

The detection of hydrogen peroxide can be achieved by any obvious technique for one with experience in the art. Exemplary techniques include chemiluminescence (Kiba et al., 2003, Analytical Science 19 (6): 823-827), fluorescence (Zhang et al., 199, Talanta 48 (5): 1031-1038; Chen et al., 2001 , Analytica Chimica 434 (1): 51-58), and spectrophotometry (Pappas et al., 2002, Analytica Chimica 455 (2): 305-313). Other exemplary techniques include metal complexes (Paleologos, 2002, Analytical Chemistry 74 (1): 100-106) as well as redox-mediated electrochemical detection (eg, commercially available glucose meters).

In advantageous embodiments, peroxidase activity can be detected with a fluorogenic substrate. Such embodiments provide rapid and sensitive techniques for the detection of total lysophosphatidylcholine in the sample. These techniques thereby provide rapid and sensitive tests for the diagnosis or prognosis of a systemic inflammatory condition as described herein. The fluorogenic substrate may be any fluorogenic substrate known to those skilled in the art that is capable of conversion into a fluorescent product by a peroxidase in the presence of peroxide under appropriate conditions, for example with water and oxygen. In particular embodiments, the fluorogenic substrate is 10-acetyl-3,7-dihydroxyphenoxazine. This substrate can be obtained from commercial suppliers (for example Amplex Red, Invitrogen). Those skilled in the art will recognize that this fluorogenic substrate can be converted into the 7-hydroxy-3-phenoxazin-3-one (resorufin) fluorescent product, detectable by techniques apparent to those skilled in the art. Useful detection techniques include, of course, fluorescence detection. Preferably, the detection methods are carried out under conditions in which the product can be formed and detected. Useful conditions and results are described in the working examples below.

In certain embodiments, the glycerophosphorylcholine can contact glycerophosphorylcholine phosphodiesterase known to those skilled in the art in appropriate glycerol-3-phosphate to produce conditions. The resulting glycerol-3-phosphate can be contacted with a glycerol-3-phosphate oxidase known to those skilled in the art under appropriate conditions to produce peroxide. Useful glycerol-3-phosphate oxidases include those obtained from Streptococcus, Aerococcus, and Pediococcus, and those described in JP 200217938. The resulting peroxide can be detected by any technique evident to those skilled in the art that includes, for example, colorimetric techniques.

In certain embodiments, the glycerophosphorylcholine can contact glycerophosphorylcholine phosphodiesterase known to those skilled in the art in appropriate glycerol-3-phosphate to produce conditions. The resulting glycerol-3-phosphate may be contacted with a glycerol-3-phosphate dehydrogenase known to those skilled in the art under appropriate conditions to produce a detectable product. For example, the contact may be in the presence of NAD + to produce detectable NADH. The contact can also be in the presence of NADP + to produce detectable NADPH.

Techniques for detecting, measuring or monitoring detectable products such as peroxide, NADH and NADPH are well known to those skilled in the art. Useful techniques are described in JP 2002-17938, Misaki, 1999, Modern Medical Laboratory 27 (8): 973-980, (1999), Japanese Patent No. 1594750, Japanese Patent in Process No. 05-229993, and Aoyama, 1997, Journal of Medical Technology 14: 1014-1019.

Diagnosis or prognosis of the systemic inflammatory condition

In certain methods of the invention, total lysophosphatidylcholine in the subject is used for the diagnosis or prognosis of systemic inflammatory condition. As described above, total lysophosphatidylcholine can be measured directly, or a measurement can be made that correlates with the amount of total lysophosphatidylcholine. In certain embodiments, free lysophosphatidylcholine is measured. In certain embodiments, bound lysophosphatidylcholine is measured. In certain embodiments, free lysophosphatidylcholine and bound lysophosphatidylcholine are measured.

In certain methods of the invention, the amount of one or more biomarkers of lysophosphatidylcholine in the subject is used for the diagnosis or prognosis of systemic inflammatory condition.

In certain methods of the invention, the amount of total lysophosphatidylcholine and one or more biomarkers of lysophosphatidylcholine in the subject are used for the diagnosis or prognosis of systemic inflammatory condition.

In some embodiments, a single sample of the subject is sufficient for the diagnosis or prognosis of the systemic inflammatory condition. In such embodiments, the amount of lysophosphatidylcholine can be compared with an internal reference in the biological sample that is present in a relatively constant amount in individuals similar to the subject. The internal reference can be any reference considered appropriate for one with experience in the art and preferably does not relate to the biomarker or systemic inflammatory conditions. In certain embodiments, the internal reference is phosphatidylcholine, phosphatidylethanolamine or phosphatidylserine.

In some embodiments, a plurality of biological samples of the subject are evaluated for the diagnosis or prognosis of the systemic inflammatory condition. In such embodiments, the change in the amount of lysophosphatidylcholine is indicative of the risk of systemic inflammatory condition.

In certain embodiments, increasing amounts of lysophosphatidylcholine indicate a decreasing risk of a systemic inflammatory condition. For example, in certain embodiments, a second amount that is at least 110%, 125%, 150%, 175%, 200%, 300%, 400% or 500% of a previous amount indicates decreased risk of systemic inflammatory condition.

In certain embodiments, decreasing amounts of lysophosphatidylcholine indicate an increased risk of a systemic inflammatory condition. For example, in certain embodiments, a second amount that is less than 95%, 90%, 80%, 75%, 50%, 33%, 25%, 20% or 10% of a previous amount indicates decreased risk of the condition systemic inflammatory

In certain embodiments, the diagnosis or prognosis of the systemic inflammatory condition may be based on a comparison of the amount of lysophosphatidylcholine in a sample of the subject with a reference amount of lysophosphatidylcholine. Reference quantities are described in the section below. Significantly, the amount of the reference amount need not be obtained or measured by a practitioner of a method of the invention. Instead, the reference quantity can be identified by consulting the reference quantities in the reference populations available to those skilled in the art. Such quantities may be published, for example, in the scientific literature in electronic databases.

In preferable embodiments, the reference amount is measured by the same technique or a comparable technique used to measure the amount in the sample. For example, preferably, if free lysophosphatidylcholine is measured in the sample, the reference amount can be based on free lysophosphatidylcholine in a subject reference or a reference population. For example, preferably, if the bound lysophosphatidylcholine is measured in the sample, the reference amount may be based on the bound lysophosphatidylcholine. Of course, if the total lysophosphatidylcholine and the reference amount are measured by different techniques, the correlation of the two quantities should be within the capacity of the experienced practitioner.

If a lysophosphatidylcholine biomarker is measured in the sample, the reference amount can be based on the lysophosphatidylcholine biomarker in a subject reference or reference population. Of course, if the lysophosphatidicoline biomarker and the reference amount are measured by different techniques, the correlation of the two quantities should be within the capacity of the experienced practitioner

When a reference quantity is used, the difference between the reference amount and the amount in the test subject is used by a practitioner skilled in the art to make a diagnosis or prognosis of the systemic inflammatory condition. In certain embodiments, if the amount in the test subject is between 10,190%, 20-180%, 30-170%, 40-160%, 50-150%, 75-125%, 80-120%, 90- 110% or 95-105% of the reference amount, the diagnosis or prognosis of the systemic inflammatory condition is indicated.

If a reference limit amount is used, the difference between the limit and the amount in the test subject is used by a practitioner with experience in the art to make a diagnosis or prognosis of the systemic inflammatory condition. In certain embodiments, if the amount in the test subject is below, or substantially below, the reference limit amount, the diagnosis or prognosis of the systemic inflammatory condition is indicated, and if the amount in the test subject is above or substantially above the reference limit amount, no diagnosis or prognosis of the systemic inflammatory condition would be indicated.

In certain embodiments, the difference between the amount of lysophosphatidylcholine in the test subject and the reference amount is inversely correlated with the risk to the systemic inflammatory condition. Such correlation can be determined by those skilled in the art.

When reference quantities of a plurality of reference subjects are used, the evaluation may be based on any statistical technique known to those skilled in the art. Similarly, when a plurality of biomarkers are used, the diagnosis or prognosis may be based on the plurality of quantities according to techniques known to those skilled in the art, such as those described in US Patent Application Publication No. No. 20030194752, 20040096917, 20040097460, 20040106142, 20040157242, and US Provisional Application No. 60 / 671,620, filed on April 15, 2005, 60 / 671,941, filed on April 15, 2005, and 60 / 674,046, filed on December 22 April 2005

Biomarker of reference

In certain methods of the invention, the amount of total lysophosphatidylcholine of the subject is compared with a corresponding reference amount of total lysophosphatidylcholine. The reference amount is typically the amount of total lysophosphatidylcholine in a reference subject (not the subject of the method) that has, or will have, within a defined period of time, a known systemic inflammatory condition. While we do not attempt to adhere to any particular theory of operation, the present invention is based, in part, on the discovery of a correlation between total lysophosphatidylcholine and a systemic inflammatory condition in a subject. Accordingly, one who practices a method of the invention can compare the amount of total lysophosphatidylcholine in a subject with a reference amount of total lysophosphatidylcholine to make a prognosis or diagnosis of the systemic inflammatory condition.

In certain methods of the invention, the amount of the subject's lysophosphatidylcholine biomarker is compared with a corresponding reference amount of the biomarker. The reference amount is typically the amount of the same biomarker, or derivative thereof, in a reference subject (not the subject of the method) that has, or will have within a defined period of time, a known systemic inflammatory condition. While we do not attempt to adhere to any particular theory of operation, the present invention is based, in part, on the discovery of a correlation between a lysophosphatidylcholine biomarker of the invention and a systemic inflammatory condition in a subject. Accordingly, one who practices a method of the invention can compare the amount of a lysophosphatidylcholine biomarker in a subject with a reference amount of the lysophosphatidylcholine biomarker to make a prognosis or diagnosis of systemic inflammatory condition.

Advantageously, in order to practice the methods of the invention, one does not need to collect amounts of reference lysophosphatidylcholine in the reference populations. Said reference quantities may be identified in sources available to those skilled in the art, such as public or private databases, or by reference to the data provided herein. As such, in methods that use a reference amount of lysophosphatidylcholine, one only needs to make the comparison described in the method.

A reference amount can be measured according to techniques known to those skilled in the art that include those described herein. Advantageously, in certain embodiments, the amount of lysophosphatidylcholine in the reference subject and the amount of lysophosphatidylcholine in the test subject are obtained by the same technique.

The reference subject can be any subject that presents, or will present within a defined period of time, symptoms of the systemic inflammatory condition according to one with experience in the art. In certain embodiments, the reference amount is obtained at a time when the reference subject exhibits the symptoms. In certain embodiments, the reference amount may be obtained at a time before or at a time after the reference subject exhibits symptoms of, or is diagnosed, the systemic inflammatory condition. For example, in certain embodiments, the reference amounts are obtained from the reference subjects 48, 36, 24 or 12 hours prior to the onset of sepsis. Those skilled in the art will recognize that such amounts can be obtained by measuring quantities in a reference population at risk of sepsis and monitoring the diagnosis of the reference subject over time.

The reference subject may have any systemic inflammatory condition or may be free of a systemic inflammatory condition. In certain embodiments, the reference subject may be negative SIRS or have symptoms of SIRS, sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality. These reference amounts can be used for the diagnosis or prognosis of the condition.

Methods for the diagnosis of systemic inflammatory condition should be carried out according to the knowledge of those skilled in the art. Such methods are routine and will not be described herein.

In certain embodiments, the diagnosis or prognosis of a systemic inflammatory condition is based on a reference limit amount. A reference limit amount is an absolute value for the amount that indicates a risk of systemic inflammatory condition. For example, a reference limit amount of 100 for a biomarker of the invention may indicate a risk of a systemic inflammatory condition when the test subject possesses a quantity of the biomarker that is less than 100 (or greater than 100 in alternative embodiments). The reference limit amounts can be determined by using statistical techniques known to those skilled in the art based on reference amounts obtained from reference subjects. For example, a limit for a particular systemic inflammatory condition can be determined so that a new reference subject can have a diagnosis or prognosis within an appropriate confidence interval for those skilled in the art, for example with 60%, 70%, 80%, 85%, 90%, 95% or 99% confidence.

Methods to monitor the treatment or prevention of systemic inflammatory conditions

In certain aspects, the present invention provides methods for monitoring a systemic inflammatory condition,

or the risk of systemic inflammatory condition, in a subject who needs it. In certain embodiments, the amount of lysophosphatidylcholine is obtained from the subject and is used to assess the condition or the risk of the condition.

In particular embodiments, the present invention provides methods for monitoring a treatment of a systemic inflammatory condition. The methods of the invention can be used to evaluate the effectiveness of the treatment and alter the treatment depending on the results of the method. Such methods generally comprise the step of measuring the amount of lysophosphatidylcholine to assess the risk of a systemic inflammatory condition in a subject to which it has been administered, or will be administered, a treatment or prevention of a systemic inflammatory condition.

In certain embodiments, the present invention provides methods for monitoring a treatment or prevention of SIRS, sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality. The treatment or prevention can be any treatment or prevention of SIRS, sepsis, severe sepsis, septic shock, multiple organ dysfunction or mortality known to those with experience.

In certain embodiments, the treatment or prevention comprises the administration of antibiotics according to known techniques for those with experience. The antibiotic may be any appropriate antibiotic for the treatment or prevention of the systemic inflammatory condition known to those skilled in the art. In some embodiments, the antibiotic may be effective against gram positive bacteria such as gentamicin, ceftriaxone, tobramyna or ceftazidine. In some embodiments, the antibiotic may be effective against gram-negative bacteria such as vancomycin. In some embodiments, the antibiotic may be effective against anaerobic bacteria, such as metrionidazole.

In other embodiments, the treatment or prevention comprises steroid administration according to techniques known to those with experience. The steroid can be any appropriate steroid for the treatment or prevention of the systemic inflammatory condition known to those skilled in the art. In certain embodiments, the steroid is hydrocortisone or dexamethasone.

In other embodiments, the treatment or prevention comprises the administration of a vasoconstrictor or inotropic therapy according to known techniques for those with experience. The vasoconstrictor or inotropic therapy may be any vasoconstrictor or inotropic therapy appropriate for the treatment or prevention of the systemic inflammatory condition known to those skilled in the art. In certain embodiments, the vasoconstrictor or inotropic therapy is norepinephrine, dopamine or dobutaima.

In other embodiments, the treatment or prevention comprises the administration of XIGRIS® (drotrecogin alfa (activated), Eli Lilly and Company). XIGRIS® is a recombinant form of human activated protein C approved by the United States Food and Drug Administration for the reduction of mortality in adult patients with severe sepsis.

In particular embodiments, the present invention provides methods for treating or preventing a systemic inflammatory condition. The methods comprise the steps of administering a treatment or prevention of the systemic inflammatory condition and monitoring the risk or severity of the systemic inflammatory condition according to a method of the invention. The risk or severity of the systemic inflammatory condition can be assessed according to the methods described herein. In certain embodiments, another treatment or prevention administration may be adjusted based on the result of the monitoring according to the criteria of one skilled in the art.

In one embodiment, the present invention provides a method for treating or preventing a systemic inflammatory condition in a subject in need of the same, which comprises the steps of administering to the subject an effective amount of XIGRIS® and monitoring the systemic inflammatory condition according to a method. of the invention. In certain embodiments, the systemic inflammatory condition is sepsis or severe sepsis.

In some embodiments, a plurality of amounts of lysophosphatidylcholine are obtained over time to monitor the systemic inflammatory condition. In certain embodiments, increasing amounts indicate the decreased risk of systemic inflammatory condition or decreased severity of the condition. For example, in certain embodiments, a second amount that is at least 110%, 125%, 150%, 175%, 200%, 300%, 400% or 500% of a previous amount indicates the decreased risk of systemic inflammatory condition. or decreased severity of the condition. In other embodiments, decreasing amounts indicate the increased risk of systemic inflammatory condition or increased severity of the condition. For example, in certain embodiments, a second amount that is less than 95%, 90%, 80%, 75%, 50%, 33%, 25%, 20% or 10% of a previous amount indicates increased risk of the condition. systemic inflammatory or increased severity of the condition.

In some embodiments, one or more amounts are obtained over time. In such embodiments, the difference between the amount of lysophosphatidylcholine and a reference amount of total lysophosphatidylcholine is used for a systemic inflammatory condition for the diagnosis or prognosis of systemic inflammatory condition. The comparison with the reference amount can be carried out as described above.

Kits for diagnosis or prognosis, or monitoring, of a systemic inflammatory condition

Also described herein are kits that are useful for the diagnosis or prognosis, or monitoring, of a systemic inflammatory condition in a subject. In some embodiments, the kits comprise a reagent that specifically binds to total lysophosphatidylcholine. The reagent can be part of an arrangement, or the reagent can be packaged separately and / or individually. The kit may also comprise at least one internal standard that should be used in the evaluation of total lysophosphatidylcholine.

In particular embodiments, the kits comprise an antibody array or antibody chip with specificity for one or more biomarkers in a plurality or panel of biomarkers of the invention.

The kits may contain reagents that can be used to detect the biomarkers contained in the biological samples from which the biomarker profiles are generated. In a specific embodiment, a kit for predicting the development of sepsis in a test subject comprises an antibody that specifically binds to total lysophosphatidylcholine. In accordance with this embodiment, the kit may comprise an antibody or functional fragment or derivative thereof (eg, Fab, F (ab ') 2, Fv, or scFv fragments) that preferentially bind to total lysophosphatidylcholine. In certain embodiments, the antibodies may be detectably labeled.

In certain embodiments the kits comprise reagents useful for the detection of total lysophosphatidylcholine in a sample. In certain embodiments, the reagents comprise one or more enzymes and one or more substrates useful for the detection of lysophosphatidylcholine. In particular embodiments, the kits may comprise a fluorogenic substrate useful for the detection of total lysophosphatidylcholine. Certain kits comprise an enzyme or reagent capable of reacting lysophosphatidylcholine to form glycerophosphatidylcholine under appropriate conditions, an enzyme or reagent capable of reacting glycerophosphatidylcholine to form choline under appropriate conditions, an enzyme or reagent capable of reacting choline to form peroxide under appropriate conditions, a peroxidase and a fluorogenic substrate of said peroxidase. Certain kits comprise a lysophospholipase, a glycerophosphatidylcholine diesterase, a choline oxidase, a peroxidase and 10-acetyl-3,7-dihydroxyphenoxazine. Certain kits comprise EC 3.1.1.5, EC 3.1.4.2, EC 1.1.3.17, horseradish peroxidase and 10-acetyl-3,7-dihydroxyphenoxazine. The kits may also comprise one or more reference standards for evaluating total lysophosphatidylcholine according to methods of the invention.

Also described herein are kits that are useful for the diagnosis or prognosis, or monitoring, of a systemic inflammatory condition in a subject. In some embodiments, the kits comprise a reagent that specifically binds to a biomarker of the present invention. The reagent can be part of an arrangement, or the reagent can be packaged separately and / or individually. The kit may also comprise at least one internal standard for use in the evaluation of a biomarker of the present invention. In particular embodiments, the kits comprise an antibody array or antibody chip with specificity for one or more biomarkers in a plurality or panel of biomarkers of the invention.

The kits may contain reagents that can be used to detect biomarkers contained in the biological samples from which the biomarker profiles are generated. In a specific embodiment, a kit for predicting the development of sepsis in a test subject comprises an antibody that specifically binds to a biomarker of the invention. In accordance with this embodiment, the kit may comprise an antibody or functional fragment or derivative thereof (eg, Fab, F (ab ') 2, Fv, or scFv fragments) that specifically bind to one or more of the biomarkers of the invention. In certain embodiments, the antibodies may be detectably labeled.

In other embodiments, a kit may comprise a specific biomarker binding reagent, such as an aptamer. If the biomarkers comprise a nucleic acid, the kit may provide an oligonucleotide probe that is capable of forming a duplex with the biomarker or with a complementary strand of a biomarker. The oligonucleotide probe can be labeled detectably.

The kits may also include reagents such as buffers, or other reagents that can be used in obtaining the amount of a biomarker. The prevention of the action of microorganisms can be ensured by the inclusion of various antibacterial and antifungal agents, for example, paraben, chlorobutanol, sorbic phenol acid, and the like. It may also be desirable to include isotonic agents such as sugars, sodium chloride, and the like.

In embodiments that use biomarkers in addition to lysophosphatidylcholine, such as nucleic acid biomarkers, the kits may advantageously comprise a microarray. In one embodiment this microarray comprises a plurality of probe stains for biomarkers that must be evaluated with the kit. In some embodiments, the microarray consists of between about three and about one hundred probe stains on a substrate. In some embodiments, the microarray consists of between about three and about one hundred probe stains on a substrate. As used in this context, the term "approximately" means within the interest rate of the established value, within ten percent of the established value, or within twenty-five percent of the established value.

In certain embodiments, the kits further comprise a label or label with instructions for carrying out a method of the invention. For example, the tag or label may provide a reference amount or reference amounts of lysophosphatidylcholine corresponding to one or more systemic inflammatory conditions. Labeling or labeling may provide one or more reference amounts of lysophosphatidylcholine limits corresponding to one or more systemic inflammatory conditions. In addition, the label or label may provide citations or connections to sources of said reference quantities.

Some kits may also comprise a computer program product for use in conjunction with a computer system, where the computer program product comprises a computer readable storage medium and a computer program mechanism included therein. In such kits, the computer program mechanism comprises instructions for evaluating whether one or more amounts of lysophosphatidylcholine of a test subject at risk of developing a systemic inflammatory condition satisfy a first set of values. The satisfaction of the first set of values predicts that the test subject may develop the systemic inflammatory condition. In some kits, the computer program product further comprises instructions for evaluating whether one or more amounts of lysophosphatidylcholine of the test subject satisfy a second set of values. The satisfaction of a second set of values predicts that the test subject is not possible to develop the systemic inflammatory condition.

Algorithms for the diagnosis or prognosis, or monitoring, of a systemic inflammatory condition

The present invention provides total lysophosphatidylcholine and biomarkers of lysophosphatidylcholine useful for the diagnosis, prognosis and monitoring of a systemic inflammatory condition. As described above, lysophosphatidylcholine can be used alone or as part of a plurality or panel of biomarkers. A plurality or panel of biomarkers may comprise a lysophosphatidylcholine of the invention, biomarkers known to those skilled in the art that are useful, or both. In advantageous embodiments, these biomarkers are capable of discriminating between converters and non-converters.

The identity of these biomarkers and their corresponding characteristics (for example, quantity or expression values) can be used to develop a decision rule, or plurality of decision rules, that discriminate between converters and non-converters. The data analysis algorithms can be used to construct a number of said decision rules. Data analysis algorithm uses characteristics (for example, quantity or expression values) of a subset of the biomarkers of the present invention in a whole training population that includes converters and non-converters. Typically, a SIRS subject is considered a non-converter when the subject does not develop sepsis in a defined period of time (for example, observation period). This defined period of time may be, for example, twelve hours, twenty-four hours, forty-eight hours, one day, one week, one month, or more. The specific data analysis algorithms for constructing a decision rule, or plurality of decision rules, that discriminate between subjects who develop sepsis and subjects who do not develop sepsis during a defined period will be described in the subsections below. Once a decision rule has been constructed by using these exemplary data analysis algorithms or other techniques known in the art, the decision rule can be used to classify a test subject into one of two or more phenotypic classes ( for example, a converter or a non-converter). These are achieved by applying the decision rule to a biomarker profile obtained from a test subject. These decision rules, therefore, have value as diagnostic indicators.

The present invention provides, in one aspect, the evaluation of a biomarker profile of a test subject with respect to the biomarker profile obtained from a training population. In some embodiments, each biomarker profile obtained from subjects in the training population, as well as the test subject, comprises a characteristic for each of a plurality of different biomarkers. In some embodiments, this comparison is achieved by (i) the development of a decision rule by using the biomarker profiles of the training population and (ii) the application of the decision rule to the biomarker profile of the subject of test. As such, the decision rules applied in some embodiments of the present invention are used to determine whether a test subject possessing SIRS will possibly acquire or not sepsis.

In some embodiments of the present invention, when the results of the application of a decision rule indicate that the subject may acquire sepsis, the subject is diagnosed as a "sepsis" subject. If the results of an application of a decision rule indicate that the subject will not acquire sepsis, the subject is diagnosed as a "SIRS" subject. Thus, in some embodiments, the result in the binary decision situation described above has four possible outcomes:

•

Truly septic, where the decision rule indicates that the subject will acquire sepsis and the subject actually acquires sepsis during the defined period of time (true positive, TP);

•

(ii) falsely septic, where the decision rule indicates that the subject will acquire sepsis and the subject does not actually acquire sepsis during the defined period of time (false positive, FP);

•

(iii) truly SIRS, where the decision rule indicates that the subject will not acquire sepsis and the subject does not actually acquire sepsis during the defined period of time (true negative, TN); or

•

(iv) falsely SIRS, where the decision rule indicates that the subject will not acquire sepsis and the subject actually acquires sepsis during the defined period of time (false negative, FN).

It will be appreciated that other definitions can be made for TP, FP, TN, FN. For example, TP could have been defined as cases where the decision rule indicates that the subject will not acquire sepsis and the subject does not actually acquire sepsis during the defined period of time. While all such alternative definitions are within the scope of the present invention, for ease of understanding of the present invention, the definitions for TP, FP, TN, and FN given by the definitions (i) to (iv) above will be used herein, unless otherwise stated.

As will be appreciated by those skilled in the art, a number of quantitative criteria can be used to communicate the performance of the comparisons made between a test biomarker profile and the reference biomarker profiles (for example, the application of a rule of decision to the biomarker profile of a test subject). These include positive predictive value (PPV), negative predictive value (NPV), specificity, sensitivity, accuracy, and certainty. In addition, other constructions such as receiver operator curves (ROCs) can be used to evaluate the performance of the decision rule. As used herein:

•

PPV = TP / (TP + FP)

•

NPV = TN / (TN + FN)

•

specificity = TN / (TN + FP)

•

sensitivity = TP / (TP + FN)

•

accuracy = certainty = (TN + TP) / N

Here, N is the number of samples compared (for example, the number of test samples for which a sepsis or SIRS determination is requested). For example, consider the case in which there are ten subjects for whom the SIRS / sepsis classification is requested. Biomarker profiles are constructed for each of the ten test subjects. Then, each of the biomarker profiles is evaluated by applying a decision rule, where the decision rule was developed based on the biomarker profiles obtained from a training population. In this example, N, from the equations above, is equal to 10. Typically, N is a number of samples, where each sample was collected from a different member of a population. This population can, in fact, be of two different types. In one type, the population comprises subjects whose phenotypic samples and data (for example, biomarker characteristic values and an indication of whether the subject acquired sepsis or not) was used to construct or debug a decision rule. Such population is referred to herein as a training population. In the other type, the population includes subjects that were not used to construct the decision rule. Such population is referred to herein as a validation population. Unless stated otherwise, the population represented by N is exclusively a training population or exclusively a validation population, as opposed to a mixture of the two types of population. It will be appreciated that scores such as accuracy will be higher (closer to the unit) when they are based on a training population as opposed to a validation population. However, unless explicitly stated otherwise herein, all the criteria used to evaluate the performance of a decision rule (or other forms of evaluation of a biomarker profile of a test subject) that includes certainty (Accuracy) refers to the criteria that were measured by applying the decision rule corresponding to the criteria regarding a training population or a validation population. In addition, the definitions for PPV, NPV, specificity, sensitivity, and accuracy defined above can also be found in Draghici, Data Analysis Tools for DNA Microanalysis, 2003, CRC Press LLC, Boca Raton, Florida, pages 342-343.

In some embodiments the training population comprises non converters and converters. In some embodiments, biomarker profiles are constructed from this population through the use of biological samples collected from the training population at some time prior to the appearance of sepsis by population converters. As such, for converters in the training population, a biological sample can be collected two weeks before, one week before, four days before, three days before, one day before, or any other period of time before the converters are Come back septic. In practice, such collections are obtained by collecting a biological sample at regular time intervals after admission to the hospital with a diagnosis of SIRS. For example, in one approach, subjects who have been diagnosed with SIRS in a hospital are used as a training population. Once admitted to the hospital with SIRS, the biological samples are collected from the subjects at selected times (for example, every hour, every eight hours, every twelve hours, daily, etc.). A part of the subjects acquires sepsis and a part of the subjects does not acquire sepsis. For subjects who acquire sepsis, the biological sample taken from the subjects just before the onset of sepsis is called the T-12 biological sample. All other biological samples of the subjects are indexed retroactively with respect to these biological samples. For example, when a biological sample has been taken from a subject on a daily basis, the biological sample taken the day before the T-12 sample is referred to as the T-36 biological sample. The time points for the biological samples for a non-converter in the training population are identified by the "time comparison" of the non-converting subject with a converting subject. In order to illustrate, consider the case in which a subject in the training population becomes clinically defined as septic on his sixth day of recruitment. For the purpose of illustration, for that subject, T-36 is day four of the study, and the biological sample T-36 is the biological sample that was obtained on day four of the study. Similarly, T-36 for the subject that does not become compared is considered to be the fourth day of the study in this subject that does not become pairs.

In some embodiments, N is more than one, more than five, more than ten, more than twenty, between ten and 100, more than 100, or less than 1000 subjects. A decision rule (or other forms of comparison) can be at least about 99% certain, even more, in some embodiments, against a training population or a validation population. In other embodiments, the certainty is at least about 97%, at least about 95%, at least about 90%, at least about 85%, at least about 80%, at least about 75%, or at least about 70% against a training population or a validation population. The degree of useful certainty may vary, depending on the particular method of the present invention. As used herein, "certainty" means "accuracy." In one embodiment, the sensitivity and / or specificity is at least about 97%, at least about 95%, at least about 90%, at least about 85%, at least about 80%, at least about 75%, or at least approximately 70% against a training population or a validation population. The number of characteristics that can be used by a decision rule to classify a test subject with adequate certainty is typically approximately four. Depending on the degree of certainty requested, however, the number of features used in a decision rule may be more less, but in all cases it is at least two. In one embodiment, the number of features that can be used by a decision rule to classify a test subject is optimized to allow a classification of a test subject with high certainty.

In the examples below, the metabolite abundance data was collected for a plurality of biomarkers in each subject. That is, for each biomarker in a biomarker profile, a characteristic, metabolite abundance data for the biomarker was measured. The decision rules are developed from said biomarker profiles of a training population through the use of data analysis algorithms to predict sample phenotypes based on observed gene expression patterns. While new classification tools are constantly being developed, the existing pattern recognition body and prediction algorithms provide effective data analysis algorithms to build decision rules. See, for example, National Research Council; Panel on Discriminant Analysis Classification and Clustering, Discriminant Analysis and Clustering, Washington, D.C .: National Academy Press.

In addition, the techniques described in Dudoit et al., 2002, "Comparison of discrimination methods for the classification of tumors using gene expression data." JASA 97; 77-87, can be used to develop these decision rules.

The relevant data analysis algorithms for developing a decision rule include, but are not limited to, discriminant analyzes that include linear, logistic, and more flexible discrimination techniques (see, for example, Gnanadesikan, 1977, Methods for Statistical Data Analysis of Multivariate Observations, New York: Wiley 1977); Tree-based algorithms such as classification and regression trees (CART) and variants (see, for example, Breiman, 1984, Regression and Classification Tree, Belmont, California: Wadsworth International Group, as well as Section 5.1.3, below ); generalized additive models (see, for example, Tibshirani, 1990, Generalized Additive Models, London: Chapman and Hall); and neural networks (see, for example, Neal, 1996, Bayesian Learning for Neural Networks, New York: Springer-Verlag; and Insua, 1998, Feedforward neural networks for nonparametric regression In: Practical Nonparametric and Semiparametric Bayesian Statistics, pages 181-194 , New York: Springer; as well as Section 1.6, below).

In one embodiment, comparing the biomarker profile of a test subject with the biomarker profiles obtained from a training population is carried out, and comprises applying a decision rule. The decision rule is constructed by using a data analysis algorithm, such as a computer pattern recognition algorithm. Other appropriate data analysis algorithms for constructing decision rules include, but are not limited to, logistic regression algorithm (see Section 1.10, below) or a non-parametric algorithm that detects differences in the distribution of characteristic values ( for example, a Ranking Test with Wilcoxon Signs (adjusted and not adjusted). The decision rule can be based on two, three, four, five, 10, 20 or more characteristics, corresponding to observables measured from one, two, three, four, five, 10, 20 or more biomarkers. In one embodiment, the decision rule is based on hundreds of features or more. Decision rules can also be constructed by using a classification tree algorithm. For example, each biomarker profile of a training population may comprise at least three characteristics, where the characteristics are predictors in a classification tree algorithm (see Section 1.1, below). The decision rule predicts the grouping within a population (or class) with an accuracy of at least about at least about 70%, of at least about 75%, of at least about 80%, of at least about 85%, of at least about 90%, at least about 95%, at least about 97%, at least about 98%, at least about 99%, or about 100%.

Appropriate data analysis algorithms are known in the art, some of which are reviewed in Hastie et al., Supra. In a specific embodiment, a data analysis algorithm of the invention comprises Regression and Classification Tree (CART; Section 1.1, below), Multiple Additive Regression Tree (MART; Section 1.4, below), Prediction Analysis for Microarrays (PAM; Section 1.2, below) or Random Tree Analysis (Section 1.1, below). These algorithms classify complex spectra of biological materials, such as a blood sample, to distinguish between subjects as normal or as having biomarker expression levels characteristic of a particular disease state. In other embodiments, a data analysis algorithm of the invention comprises ANOVA and non-parametric equivalents, linear discriminant analysis (Section 1.10, below), logistic regression analysis (Section 1.10, below), nearest neighbor classifier analysis ( Section 1.9, below), neural networks (Section 1.6, below), component analysis (Section 1.8, below), discriminant quadratic analysis (Section 1.11, below), regression classifiers (Section 1.5, below) and support vector machines (Section 1.12, below). While these algorithms can be used to construct a decision rule and / or increase the speed and efficiency of the application of the decision rule and avoid the diversion of the researcher, one with common experience in the art will realize that the algorithms based Computers are not required to carry out the methods of the present invention.

Decision rules can be used to evaluate biomarker profiles, regardless of the method that was used to generate the biomarker profile. For example, the appropriate decision rules that can be used to evaluate the profiles of biomarkers generated by using gas chromatography, as discussed in Harper, "Pyrolysis and GC in Polymer Analysis," Dekker, New York (1985). In addition, Wagner et al., 2002, Anal. Chem. 74: 1824-1835 discloses a decision rule that improves the ability to classify subjects based on spectra obtained by static flight time - secondary ionic mass spectrometry (TOF-SIMS). Additionally, Bright et al., 2002, J. Microbiol. Methods 48: 127-38, discloses a method to distinguish between bacterial strains with high certainty (79-89% of correct classification rates) by analyzing the MALDI-TOF-MS spectra. Dalluge, 2000, Fresenius J. Anal. Chem. 366: 701-711, discusses the use of MALDI-TOF-MS and liquid chromatography-electrospray mass spectrometry (LC / ESI-MS) to classify biomarker profiles in complex biological samples.

Decision guns

One type of decision rule that can be constructed by utilizing the characteristic values of the biomarkers identified in the present invention is a decision tree. Here, the "data analysis algorithm" is any technique that can build the decision tree, while the final "decision tree" is the decision rule. A decision tree is built by utilizing a training population and

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

Specific data analysis algorithms. Decision trees in general are described by Duda, 2001, Pattern Classification, John Wiley & Sons, Inc., New York. pages 395-396. Tree-based methods divide the characteristic space into a set of rectangles, and then fit a model (as a constant) in each.

The training population data includes the characteristics (eg, expression values, or some other observable) for the biomarkers of the present invention in a whole training set population. A specific algorithm that can be used to build a decision tree is a classification and regression tree (CART). Other specific decision tree algorithms include, but are not limited to, ID3, C4.5, MART and Random Trees. CART, ID3, and C4.5 are described in Duda, 2001, Pattern Classification, John Wiley & Sons, Inc., New York, pages 396-408 and pages 411-412. CART, MART, and C4.5 are described in Hastie et al., 2001, The Elements of Statistical Learning, Springer-Verlag, New York, Chapter 9. Random Trees are described in Breiman, 1999, "Random Forests -Random Features , "Technical Report 567, Statistics Department, UCBerkeley, September 1999.

In some embodiments of the present invention, decision trees are used to classify subjects by utilizing features for combinations of biomarkers of the present invention. Decision tree algorithms belong to the class of supervised knowledge algorithms. The purpose of a decision tree is to induce a classifier (a tree) of exemplary real-world data. This tree can be used to classify unseen examples that have not been used to obtain the decision tree. As such, a decision tree is obtained from training data. Exemplary training data contains data for a plurality of subjects (the training population). For each respective subject there is a plurality of characteristics the class of the respective subject (for example, sepsis / SIRS). In one embodiment of the present invention, training data is expression data for a combination of biomarkers in the entire training population.

The following algorithm describes a decision tree derivation: Tree (Examples, Class, Characteristics)

Create a root node

If all Examples have the same Class value, give the root this label

Otherwise, if the Feature is empty, label the root according to the most common value.

Otherwise start

Calculate the gain of information for each Feature

Select Feature A with the highest information gain and convert it to Feature

root for each possible value, v, of this characteristic

Add a new branch below the root, corresponding to A = v

Let Examples (v) be those examples with A = v

If Examples (v) are empty, convert the new branch into a leaf node labeled with the value

most common among the Examples

Otherwise let the new branch be the tree created by

Tree (Examples (v), Class, Characteristics - {A})

Finalize

A more detailed description of the information gain calculation is shown below. If the possible classes vi of the examples have probabilities P (vi) then the information content I of the real answer is given by:

image 1

The value I shows how much information we need to be able to describe the result of a classification for the specific data set used. Assuming that the data set contains examples (for example subjects) p positive (for example it will develop sepsis) and n negative (for example it will not develop sepsis), the information contained in a correct response is:

image 1

where log2 is the logarithm that base two uses. By testing the unique characteristics the amount of information needed to perform a correct classification can be reduced. The rest of a specific characteristic A (for example, representing a specific biomarker) shows how much information that is necessary can be reduced.

 Rest

"v" is the number of unique attribute values for characteristic A in a certain data set, "i" is a certain attribute value, "pi" is the number of examples for characteristic A where the classification is positive ( for example it will develop sepsis), "ni" is the number of examples for characteristic A where the classification is negative (for example it will not develop sepsis). The information gain of a specific characteristic A is calculated as the difference between the information content for the classes and the rest of the characteristic A:

image3

Information gain is used to assess how important the different characteristics are for classification (how well the examples divide), and the characteristic with the highest information.

10 In general there are a number of different decision tree algorithms, many of which are described in Doubt, Pattern Classification, Second Edition, 2001, John Wiley & Sons, Inc. Decision tree algorithms often require consideration of processing of characteristics, impurity measurement, interruption criteria, and pruning. Specific decision tree algorithms include, but are not limited to classification and regression trees (CART), multivariate decision trees, ID3, and C4.5.

In one approach, when a decision tree is used, gene expression data for a select combination of genes described in the present invention in a whole training population is standardized to have mean zero and unit variance. The members of the training population are randomly divided into a training set and an essay set. For example, in one embodiment, two thirds of the members of the training population are placed in the training set and one third of the members of the training

20 training population are placed in the trial set. Expression values for a select combination of biomarkers described in the present invention are used to construct the decision tree. Next, the ability of the decision tree to correctly classify the members in the test set is determined. In some embodiments, this computation is carried out several times for a given combination of biomarkers. In each computational iteration, members of the training population are assigned

25 randomly to the training set and the test set. Then, the quality of the biomarker combination is taken as the average of each iteration of the decision tree computation.

In addition to the univariate decision trees in which each division is based on a characteristic value for a corresponding biomarker, among the set of biomarkers of the present invention, or the relative characteristic values of two of said biomarkers, the decision trees multivariates can

30 be implemented as a decision rule. In said multivariate decision trees, some or all decisions actually comprise a linear combination of characteristic values for a plurality of biomarkers of the present invention. Said linear combination can be trained by using known techniques such as descending gradient in a classification by using a squared error sum criterion. To illustrate this decision tree, consider the expression:

35 0.04? X1 +? 0.16? X2 <? 500

Here, x1 and x2 refer to two different characteristics for two different biomarkers from among the biomarkers of the present invention. To prune the decision rule, characteristic values x1 and x2 are obtained from measurements obtained from the unclassified subject. These values are then inserted into the equation. If a value less than 500 is computed, then a first branch is taken in the decision tree. Otherwise, it

40 takes a second branch in the decision tree. Multivariate decision trees are described in Duda, 2001, Pattern Classification, John Wiley & Sons, Inc., New York, pages 408-409.

Another approach that can be used in the present invention is multivariate adaptive regression through the use of slots (MARS). MARS is an adaptive regression procedure, and it fits well for the large-scale problems addressed by the present invention. MARS can be seen as a generalization of the

45 regression line in stages or a modification of the CART method to improve the performance of CART in the regression adjustment. MARS is described in Hastie et al., 2001, The Elements of Statistical Learning, Springer-Verlag, New York, pages 283-295.

Predictive microarray analysis (PAM)

An approach to developing a decision rule by utilizing characteristic values of 50 biomarkers of the present invention is the closest centroid classifier. These techniques compute, for each

5

10

fifteen

twenty

25

30

35

40

Four. Five

class (sepsis and SIRS), a centroid given by the average characteristic levels of the biomarkers in the class, and then assign new samples to the class whose centroid is closest. This approach is similar to the k-mean grouping except that the groups are replaced by known classes. This algorithm can be sensitive to noise when a large number of biomarkers are used. An improvement in the technique uses depreciation: for each biomarker, differences between class centroids are set to zero if they are considered to be due to chance. This approach is implemented in the Microarray Prediction Analysis, or PAM. See, for example, Tibshirani et al., 2002, Proceedings of the National Academy of Science USA 99; 65676572.

Depreciation is controlled by a lower threshold whose differences are considered noise. Biomarkers that show no difference above the noise level are eliminated. A threshold can be chosen by cross validation. As the threshold decreases, more biomarkers are included and the estimated classification errors decrease, until they reach a floor and begin to climb again as a result of noise biomarkers - a phenomenon known as overfitting.

Bagging, amplification, and random sumespace method

Bagging, amplification, the random subspace method, and additive trees are data analysis algorithms known as combination techniques that can be used to improve weak decision rules. These techniques are designed to, and usually apply to, decision trees, such as the decision trees described in Section 1.1, above. In addition, such techniques can also be useful in the decision rules developed by using other types of data analysis algorithms such as linear discriminant analysis.

In bagging, one obtains a sample of the training set, which generates random independent starting replicas, constructs the decision rule in each of them, and adds them by a simple majority vote in the final decision rule. See, for example, Breiman, 1996, Machine Learning 24, 123-140; and Efron & Tibshirani, An Introduction to Boostrap, Chapman & Hall, New York, 1993.

In amplification, the decision rules are constructed in weighted versions of the training set, which are dependent on the results of previous classification. Initially, all the characteristics under consideration have equal weights, and the first decision rule is built on this data set. Then, the weights are changed according to the performance of the decision rule. The misclassified characteristics obtain larger weights, and the next decision rule is amplified in the reweighted training set. In this way, a sequence of training sets and decision rules is obtained, which is then combined by a simple majority vote or by a weighted majority vote in the final decision rule. See, for example, Freund & Schapire, "Experiments with a new boosting algorithm," Proceedings 13th International Conference on Machine Learning, 1996, 148-156.

To illustrate the amplification, consider the case where there are two phenotypes exhibited by the study population, phenotype 1 (for example, which acquires sepsis for a defined period of time), and phenotype 2 (for example, SIRS only, which means that the subject acquires sepsis within a defined period of time). Given a vector of indicator biomarkers (for example, a feature vector representing said biomarkers) of the training data set, a decision rule G (X) produces a prediction that takes one of the type values in the two value sets : {phenotype 1, phenotype 2}. The error rate on the training sample is

image 1

where N is the number of subjects in the training set (the total sum of the subjects that has phenotype 1 or phenotype 2). For example, there are 49 organisms that acquire sepsis and 72 organisms that remain in the SIRS state, N is 121. A weak decision rule is one whose error rate is only slightly better than the random assumption. In the amplification algorithm, the weak decision rule is repeatedly applied to modified versions of the data, thereby producing a sequence of weak decision rules Gm (x), m, = 1, 2, ..., M. The predictions of all the decision rules in this sequence are then combined through a weighted majority vote to produce the final decision rule:

image 1

5

10

fifteen

twenty

25

30

35

40

Here a1, 2, ..., M are computed using the amplification algorithm and its purpose is to weigh the contribution of each respective decision rule Gm (x). Its effect is to give more influence to the most exact decision rules in the sequence.

The data modifications at each amplification stage consist of applying the weights w1, w2, ..., wn to each of the training observations (xi, yi), i = 1, 2, ..., N. Initially All weights are set to wi = 1 / N, so that the first stage simply teaches the decision rule on the data in the usual way. For each successive iteration m = 2, 3, ..., M the observation weights are modified individually and the decision rule is reapplied to the weighted observations. In stage m, those observations that were misclassified by the decision rule Gm-1 (x) induced in the previous stage have their weights increased, while the weights are reduced for those that were correctly classified. Thus, as the iterations proceed, observations that are difficult to classify correctly receive ever increasing influence. Each successive decision rule is thereby forced to concentrate on those training observations that have been lost by previous observations in the sequence.

The amplification algorithm is summarized as follows:

one.

Initialize the observation weights wi = 1 / N, i = 1, 2, ..., N.

2.

For m = 1 to M:

(to)

Adjust a decision rule Gm (x) to the training set by using weights wi.

(b)

Compute

(C)

Compute am = log ((1-errm) / errm).

(d)

3.

 Outcome

image 1

image4

image 1

In an embodiment in accordance with this algorithm, each object is, in reality, a factor. In addition, in the algorithm, the current decision rule Gm (x) is induced in the observations weighted on it is computed on line 2b. Line 2c calculates the weighting am given to Gm (x) in the production of the final classifier G (x) (line 3). The individual weights of each of the observations are updated for the next iteration on line 2d. Observations misclassified by Gm (x) have their weights scaled by an exp factor (am), which increases their relative influence to induce the following classifier Gm + 1 (x) in the sequence. In some embodiments, the modifications of Freund and Schapire, 1997, Journal of Computer and System Sciences 55, pages. 119-139, amplification methods are used. See, for example, Hasti et al., The Elements of Statistical Learning, 2001, Springer, New York, Chapter 10. For example, in some embodiments, feature preselection is carried out by using a technique such as non-parametric scoring methods of Park et al., 2002, Pac. Symp Biocomput 6, 52-63,

Feature preselection is a form of dimensionality reduction in which the genes that discriminate between classifications are best selected for use in the classifier. Then, the LogitBoost procedure introduced by Friedman et al., 2000, Ann Stat 28, 337-407 is used instead of the Freund and Schapire amplification procedure. In some embodiments, amplification and other classification methods of Ben-Dor et al., 2000, Journal of Computational Biology 7, 559-583 are used in the present invention. In some embodiments, amplification and other classification methods of Freund and Schapire, 1997, Journal of Computer and System Sciences 55, 119-139 are used.

5

10

twenty

25

30

35

In the random subspace method, the decision rules are constructed in random subspaces of the data characteristics space. These decision rules are usually combined by a simple majority vote in the final decision rule. See, for example, Ho, "The Random subspace method for constructing decision forests," IEEE Trans Pattern Analysis and Machine Intelligence, 1998; 20 (8): 832-844.

Weapons of regression of multiple additives

Multiple Additive Regression Trees (MART) represents another way to construct a decision rule that can be used in the present invention. A generic algorithm for MART is:

one.

Initialize

2.

 For m = 1 to M:

(to)

 For 1 = 1.2, ..., N compute

(b)

Fit a regression tree to the rim whites that give terminal regions Rjm, j = 1,2, ..., Jm.

(C)

For j = 1, 2, ..., Jm compute

(d)

 To update

image 1

image 1

image5

image6

Specific algorithms are obtained by inserting different loss criteria L (y, f (x)). The first line of the algorithm initializes the optimal constant model, which is a single terminal node tree. The components of the negative gradient computed on line 2 (a) are called pseudo generalized residues, r. The gradients for the commonly used loss functions are summarized in Table 10.2, from HHastie et al., 2001, The Elements of Statistícal Learníng, Springer-Verlag, New York, page 321. The algorithm for classification is similar and is described in Hastie et al., Chapter 10. The tuning parameters associated with the MART procedure are the number of iterations M and the sizes of each of the constituent trees Jm, m = 1, 2, ..., M.

Rules of "decision" omitted from "regression"

In some embodiments, a decision rule is used to classify subjects by using regression. In such embodiments, the decision rule can be characterized as a regression classifier, preferably a logistic regression classifier. Said regression classifier includes a coefficient for each of the biomarkers (for example, a characteristic for each of said biomarkers) used to construct the classifier. In such embodiments, the coefficients for the regression classifier are computed by using, for example, a maximum probability approach. In said computation, the characteristics are used for biomarkers (for example, RT-PCR, microarray data). In particular embodiments, the molecular marker data of only two subgroups of traits are used (eg, subgroup of traits a: will acquire sepsis in a defined period of time and subgroup of traits b: will not acquire sepsis in a defined period of time) and the dependent variable is absent or in the presence of a particular feature in the subjects for which biomarker data is available.

In another specific embodiment, the training population comprises a plurality of feature subgroups (for example, three or more feature subgroups, four or more specific feature subgroups, etc.). These multiple subgroups of traits may correspond to discrete stages in the phenotypic progress of healthy, to SIRS, to sepsis, to more advanced stages of sepsis in a training population. In this specific embodiment, a generalization of the logistic regression model that manipulates responses from multiple categories can be used to develop a decision that discriminates between the various subgroups of traits found in the training population. For example, the measured data for selected markers can be applied to any of the mu category logit models described in Agresti, An Introduction to Categorical Data Analysis, 1996, John Wiley & Sons, Inc., New York, Chapter 8, to develop a classifier capable of discriminating between any of a plurality of subgroups of traits represented in a training population.

Neural Networks

In some embodiments, the characteristic data measured for select biomarkers of the present invention (eg, RT-PCR data, mass spectrometry data, microarray data) can be used to train a neural network. A neural network is a decision rule for classification or regression of two stages. A neural network has a layered structure that includes a layer of input units (and deviation) connected by each one of weights to a layer of output units. For regression, the output unit layer typically includes only one output unit. However, neural networks can manipulate multiple quantitative responses in a perfect way.

In multi-layered neural networks, there are input units (input layer), hidden units (hidden layer), and unit is output (output layer). There is also a single deviation unit that is connected to each unit other than the input units. Neural networks are described in Duda et al., 2001, Pattern Classification, Second Edition, John Wiley & Sons, Inc., New York; and Hastie et al., 2001, The Elements of Statistical Learning, Springer-Verlag, New York.

Neural networks are also described in Draghici, 2003, Data Analysis Tools for DNA Microarrays, Chapman & Hall / CRC; and Mount, 2001, Bioinformatics: sequence and genome analysis, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. What is disclosed below is some exemplary forms of neural networks.

The basic approach to the use of neural networks is to start with an untrained network, present a training pattern to the input layer, and pass signals through the network and determine the output in the output layer. These outputs are then compared with the target values; Any difference corresponds to an error. This error or criterion function is some scalar function of the weights and is minimized when the network outputs coincide with the desired outputs. In this way, the weights are adjusted to reduce this error measurement. For regression, this error can be sum of squared errors. For classification, this error may be squared error or cross entropy (deviation). See, for example, Hastie et al., 2001, The Elements of Statistical Learning, Springer-Verlag, New York.

Three commonly used training protocols are stochastic, batch, and online. In stochastic training, the patterns are chosen randomly from the training set and the net weights are updated for each pattern presentation. Non-linear multilayer networks trained by descending gradient methods such as stochastic backward propagation perform a maximum probability estimate of the weighting values in the classifier defined by the network topology. In batch training, all patterns submitted to the network before learning takes place. Typically, in batch training, several passes are made through training data. In online training, each employer is presented once and only once to the network.

In some embodiments, starting values for weights are considered. If the weights are near zero, then the sigmoid operating part commonly used in the hidden layer of a neural network (see, for example, Hastie et al., 2001, The Elements of Statistical Learning, Springer Verlag, New York) is approximately linear, and therefore the neural network collapses into an approximately linear classifier. In some embodiments, the start values for the weights are chosen to be random values near zero. Therefore the classifier begins almost linear, and becomes nonlinear as the weights increase. Individual units are located at addresses and introduce nonlinearities where necessary. The use of exact zero weights leads to zero derivatives and perfect symmetry, and the algorithm never moves. Alternatively, starting with large weights often leads to poor solutions.

Because the escalation of inputs determines the effective escalation of the weights in the lower layer, it can have a great effect on the quality of the final solution. Thus, in some embodiments, at first all expression values are standardized to have zero mean and a standard deviation of one. This ensures that all entries are treated equally in the regularization process, and allows one

5

10

fifteen

twenty

25

30

35

40

Four. Five

Choose a significant range for random start weights. With standardization entries, it is typical to take random uniform weights in the interval [-0.7, +0.7].

A recurring problem in the use of three-layer networks is the optimal number of hidden units to use in the network. The number of inputs and outputs of a three layer network is determined by the problem to be solved. In the present invention, the number of inputs for a given neural network will be equal to the number of biomarkers selected from the training population. The output number for the neural network will typically be one. However, in some embodiments, more than one output is used so that more than just two states can be defined by the res. For example, a multi-exit neural network can be used to discriminate between healthy phenotypes, various stages of SIRS, and / or various stages of sepsis. If too many hidden units are used in a neural network, the network will have too many degrees of freedom and is trained too much, there is a danger that the network will overfit the data. If there are very few hidden units, the training set cannot be learned. Generally speaking, however, it is better to have too many hidden units than too few. With very few hidden units, the classifier may not have enough flexibility to capture nonlinearities on the date; With too many hidden units, the extra weighting may shrink to zero if proper regularization or pruning is used, as described below. In typical embodiments, the number of hidden units is somewhat in the range of 5 to 100, with the number growing with the number of entries and number of training cases.

A general approach to determine the number of hidden units to use is to apply a regularization approach. In the regularization approach, a new criterion function is built that depends not only on the classic training error, but also on the complexity of the classifier. Specifically, the new criterion function penalizes highly complex classifiers; to seek the minimum in this criterion is to balance the error in the training set with the error in the training set plus a regularization term, which expresses restrictions or desirable properties of solutions: J = Jpal + AJreg. Parameter A is adjusted to impose regularization more or less strongly. In other words, higher values for A will tend to shrink the weights towards zero: cross-validation with a validation set is typically used to estimate A. This validation set can be obtained by setting aside a random subset of the training population. . Other forms of penalty have been proposed, for example the weighting elimination penalty (see, for example, Hastie et al., 2001, The Elements of Statistical Learning, Springer Verlag, New York).

Another methodology to determine the number of hidden units to use is to eliminate -weigh - the weights that are least needed. In one methodology, weights with the smallest magnitude are eliminated (set to zero). Such pruning based on magnitude may work, but it is not optimal; Sometimes weights with small quantities are important for training and learning data. In some embodiments, instead of using a pruning approach based on magnitude, Wald statistics are computed. The fundamental idea in Wald statistics is that they can be used to estimate the importance of a hidden unit (weighting) in a classifier. Then, hidden units that have the least importance are eliminated (by setting their input and output weights at zero). Two algorithms in this regard are the Optimal Brain Injury (OBD) and the Optimal Brain Surgeon (OBS) algorithms that use the second-order approximation to predict how training error depends on a weighting, and eliminates the weighting that leads to the most increase. Small in training error.

The Optimal Brain Injury and the Optimal Brain Surgeon share the same basic approach to train a network to minimum local error in the weighting w, and then prune a weighting that leads to the smallest increase in training error. The expected functional increase in the error for a change in the full weighting vector or w is:

image 1

where: is the matrix of Hessian. The first term disappears at a local minimum in error; Third order terms and higher orders are ignored. The general solution to minimize this function given the restriction of eliminating a weighting is:

image 1

image 1

Here, uq is the unit vector along the direction qth in the weighting space and Lq is the approximate point approximation of the weighting q - the increase in training error if the weighting q is pruned and the other weights are updated Sw. These equations require the inverse of H. One method to calculate the inverse matrix is to start with a small value,

image 1

10 where a is a small parameter - effectively a weighting constant. Then the matrix is updated with each pattern according to:

image 1

Where subscriptions correspond to the pattern being presented and am decreases with m. After the training set has been presented, Hessian's inverse matrix is given by H -1 = Hn-1 In algorithmic form, the Optimal Brain Surgeon method is:

Start initializing nH, w, 8

Train a reasonably large network to the minimum error computes H-1 by equation 1

image 1

up to J (w)> 8

20 return w

Finalize

The Optimal Brain Injury method is simpler at the computational level because the calculation of the Hessian inverse matrix on line 3 is particularly simple for a diagonal matrix. The algorithm above ends when the error is greater than an initialized criterion to be 8. Another methodology is to change line 6

25 to finish when the change in J (w) due to the elimination of a weighting is greater than some criterion value. In some embodiments, the backward neural network. See, for example, Abdi, 1994, "A neural network primer," J. Biol System. 2, 247-283.

Grouping

In some embodiments, the characteristics for the selected biomarkers of the present invention are used to group a training set. For example, consider the case in which ten characteristics (corresponding to ten biomarkers) described in the present invention are used. Each member m of the training population will have characteristic values (eg expression values) for each of the ten biomarkers. These values of a member m in the training population define the vector:

X1m, X2m, X3m, X4m, X5m, X6m, X7m, X8m, X9m, X10m

where Xim is the level of expression of the biomarker ith in the organism m. If there are m organisms in the training set, the selection of i biomarkers will define m vectors. Note that the methods of the present invention do not require that each expression value of every single biomarker used in the vectors be represented in each single vector m. In other words, the data of a subject in which one of the ith biomarkers cannot be found for clustering can still be used. In such cases, the lost expression value is assigned a "zero" or some other normalized value. In some embodiments, prior to grouping, characteristic values are normalized to have an average value of zero unit variance. Those members of the training population that exhibit similar expression patterns throughout the training group will tend to group together. A particular combination of genes of the present invention is considered to be a good classifier in this aspect of the invention when vectors are grouped into groups of traits found in the training population. For example, if the training population includes class a: subjects who do not develop sepsis, and class r: subjects who develop sepsis, an ideal grouping classifier will group the population into two groups, with one group representing only class a and the other group representing only class b.

Clustering is described on pages 211-256 of Duda and Hart, Pattern Classification and Scene Analysis, 1973, John Wiley & Sons, Inc., New York, (hereinafter "Doubt 1973"). According to what is described in Section 6.7 of Duda 1973, the problem of grouping is described as one of finding natural groupings in a data set. To identify natural groupings, two topics are discussed. First, a way of measuring the similarity (or dissimilarity) between two samples is determined. This metric (similarity measurement) is used to ensure that the samples in a cluster are more similar to each other than they are to samples in other groups. Second, a mechanism is determined to divide the data into groups by using similarity measurement.

Similarity measurements are discussed in Section 6.7 of Doubt 1973, which states that one way to begin a cluster investigation is to define a distance function and compute the distance matrix between all pairs of samples in a data set. If the distance is a good measure of similarity, then the distance between the samples in the same group will be significantly less than the distance between the samples in different groups. However, according to what is established on page 215 of Duda 1973, grouping does not require the use of a distance metric. For example, a non-metric similarity function s (x, x ') can be used to compare two vectors x and x'. Conventionally, s (x, x ') is a symmetric function whose value is large when x and x' are something "similar." An example of a non-metric similarity function s (x, x ') is provided on page 216 of Doubt 1973.

Once a method to measure "similarity" or "dissimilarity" between points in a data set has been selected, the grouping requires a criterion function that measures the grouping quality of any data division. The divisions of the data set that ends the criterion function are used to group the data. See page 217 of Duda 1973. The criteria functions are discussed in Section 6.8 of Duda 1973.

More recently, Duda et al., Pattern Classification, 2nd edition, John Wiley & Sons, Inc. New York has been published. Pages 537-563 describe the grouping in detail. More information on clustering techniques can be found in Kaufman and Rousseeuw, 1990, Finding Groups in Data: An Introduction to Cluster Analysis, Wiley, New York, NY; Everitt, 1993, Cluster analysis (3d ed.), Wiley, New York, NY; and Backer, 1995, Computer-Assisted Reasoning in Cluster Analysis, Prentice Hall, Upper Saddle River, New Jersey. Particular exemplary clustering techniques that may be used in the present invention include, but are not limited to, hierarchical clustering (agglomerative clustering through the use of the nearest neighbor algorithm, furthest neighbor algorithm, the average connection algorithm, the centroid algorithm, or the sum of squares algorithm), k-media grouping, fuzzy k-means clustering algorithm, and Jarvis-Patrick grouping.

Principal component analysis

Principal component analysis (PCA) has been proposed to analyze gene expression data. More generally, PCA can be used to analyze the characteristic value data of the biomarkers of the present invention to construct a decision rule that discriminates non-converter converters. Principal component analysis is a classic technique to reduce the dimensionality of a set of data by transforming the data into a new set of variable (main components) that summarize the characteristics of the data. See, for example, Jolliffe, 1986, principle component analysis, Springer, New York. The analysis of

5

fifteen

25

35

Four. Five

55

Main components are also described in Draghici, 2003, Data Analysis Tools for DNA Microarrays, Chapman & Hall / CRC. The following are the non-restrictive examples of principal component analysis.

The main components (PCs) are not correlated and are arranged in such a way that the kth PC has the largest variance kth between the PCs. The kth PC can be interpreted as the direction that maximizes the variation of the projections of the data points of so that it is orthogonal to the first k -1 PCs. The first few PCs capture most of the variation in the data set. In opposition, the last few PCs are often supposed to capture only the residual 'noise' in the data.

PCA can also be used to create a classifier in accordance with the present invention. In such an approach, the vectors for the selected biomarkers of the present invention can be constructed in the same manner described for clustering above. In reality, the set of vectors, where each vector represents characteristic values (for example, abundance values) for the elected genes of a particular member of the training population, can be seen as a matrix. In some embodiments, this matrix is represented in a Free-Wilson method of qualitative binary description of monomers (Kubinyi, 1990, 3D QSAR in drug design theory methods and applications, Pergamon Press, Oxford, pages 589-638), or is distributed in a maximally compressed space by using PCA so that the first principal component (PC) captures as much variance information as possible, The second principal component (PC) captures the second largest amount of all variance information, etc. until all variance information in the matrix has been considered.

Then, each of the vectors is plotted (where each vector represents a member of the training population). Many different types of planes are possible. In some embodiments, a one-dimensional plane is made. In this one-dimensional plane, the value for the first main component of each of the members of the training population is plotted. In this form, the expectation is that the members of a first subgroup (for example those subjects who do not develop sepsis in a given period of time) are grouped into a range of values of the first main component and the members of a second subgroup (for example, those subjects who develop sepsis in a given period of time) will be grouped into a second range of values of the first principal component.

In an ideal example, the training population comprises two subgroups: "sepsis" and "SIRS." The first major component is computed by utilizing the expression values of the molecular marker for the selected biomarkers of the present invention in the entire data set of the entire training population. Then, each member of the training set is plotted as a function of value for the first main component. In this ideal example, those members of the training population in which the first principal component is positive are the "responders" and those members of the training population in which the first principal component is negative are "subjects with sepsis."

In some embodiments, members of the training population are plotted against more than a major component. For example, in some embodiments, the members of the training population are drawn on a two-dimensional plane in which the first dimension is the first principal component and the second dimension is the second principal component. In this two-dimensional plane, the expectation is that the members of each subgroup represented in the training population will be grouped into discrete groups. For example, a first group of members in the two-dimensional plane will represent subjects that develop sepsis in a given period of time and a second group of members in the two-dimensional plane will represent subjects that do not develop sepsis in a given period of time.

Nearest neighbor analysis

The classifiers of the nearest neighbor are based on memory and do not require that any classifier be adjusted. Given a question point xO, k training points x (r), r, ..., k are closest in distance to x0 and then point x0 is classified by using k closest neighbors. Links can be broken randomly. In some embodiments, the Euclidean distance in the feature space is used to determine the distance as:

image 1

Typically, when the closest neighbor algorithm is used, the expression data used to compute the linear discriminant is standardized to have zero mean and variance 1. In the present invention, the members of the training population are randomly divided in a training set and an essay set. For example, in one embodiment, two thirds of the members of the training population are placed in the training set and one third of the members of the training population are placed in the test set. A select combination of biomarkers of the present invention represents the feature space in which the members of the test set are plotted. Next, the ability of the training set to correctly characterize the members of the test set is computed. In some embodiments, the nearest neighbor count is carried out several times for a given combination of biomarkers of the present invention. At each iteration of the computation, the members of the training population are randomly assigned to the training set and the test set. Next, the quality of the biomarker combination is taken as the average of each iteration of the nearest neighbor count.

The nearest neighbor rule can be debugged to address issues of prior unequal class, costs of poor differential classification, and feature selection. Many of these purifications include some form of weighting vote for neighbors. For more information about the closest neighbor analysis, see Duda, Pattern Classification, Second Edition, 2001, John Wiley & Sons, Inc; and Hastie, 2001, The Elements of Statistical Learning, Springer, New York.

Linear discriminant analysis

Linear discriminant analysis (LDA) attempts to classify a subject into one of two categories based on certain object properties. In other words, LDA tests whether object attributes measured in an experiment predict the categorization of objects. LDA typically requires continuous independent variables and a dichotomous categorical dependent variable. In the present invention, the characteristic values for the selected combinations of the biomarkers of the present invention throughout a subset of the training population serve as the required continuous independent variables. The subgroup of classification of traits of each of the members of the training population serves as the dichotomous categorical dependent variable.

LDA requests the linear combination of variables that maximizes the relationship between the group variance and within the group variance through the use of grouping information. Implicitly, the linear weights used by LDA depend on how the characteristic values of a molecular marker throughout the training set are separated into two groups (for example, a group that develops sepsis over a defined period of time and a group b which does not develop sepsis for a defined period of time) and how these characteristic values correlate with the characteristic values of other biomarkers. In some embodiments, LDA is applied to the data matrix of N members in the training sample using K biomarkers in a combination of biomarkers described in the present invention. Then, the linear discriminant of each member of the training population is drawn. Ideally, those members of the training population that represent a first subgroup (for example those subjects who develop sepsis in a defined period of time) will be grouped into a range of linear discriminant values (for example, negative) and those members of the population Training that represent a second subgroup (for example those subjects who will not develop sepsis in a defined period of time) will be grouped into a second range of linear discriminant values (for example, positive). The LDA is considered more successful when the separation between discriminant value groups is greater. For more information on linear discriminant analysis, see Doubt, Pattern Classification, Second Edition, 2001, John Wiley & Sons, Inc; and Hastie, 2001, The Elements of Statistical Learning, Springer, New York; and Venables & Ripley, 1997, Modern Applied Statistics with s-plus, Springer, New York.

Quadratic discriminant analysis

Quadratic discriminant analysis (QDA) takes the same input parameters and returns the same results as the LDA. QDA uses quadratic equations, instead of linear equations, to produce the results. LDA and QDA are interchangeable, and which one to use is a matter of preference and / or availability of software to support the analysis. Logistic regression takes the same input parameters and returns the same results as LDA and QDA.

Support vector machine

In some embodiments of the present invention, support vector machines (SVMs) are used to classify subjects by utilizing characteristic values of the genes described in the present invention. SVMs on a relatively new type of learning algorithm. See, for example, Cristianini and Shawe-Taylor, 2000, An Introduction to Support Vector Machines, Cambridge University Press, Cambridge; Boser et al., 1992, "A training algorithm for optimal margin classifiers," in Proceedings of the 5th Annual ACM Workshop on Computational Learning Theory, ACM Press, Pittsburgh, PA, pages 142-152; Vapnik, 1998, Statistical Learning Theory, Wiley, New York; Mount, 2001, Bioinformatics: sequence and genome analysis, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, Doubt, Pattern Classification, Second Edition, 2001, John Wiley & Sons, Inc .; and Hastie, 2001, The Elements of Statistical Learning, Springer, New York; and Furey et al., 2000, Bioinformatics 16, 906-914. When used for classification, SVMs separate a given set of binary tagging data training data with a hyperplane that separates maximally from them. For cases where no linear separation is possible, SVMs can work in combination with the 'kernels' technique, which automatically performs a non-linear mapping in a feature space. The hyperplane found by the SVM in the feature space corresponds to the non-linear decision limit in the input space.

In one methodology, when an SVM is used, the characteristic data is standardized to have zero mean and unit variance and the members of a training population are randomly divided into a training set and a test set. For example, in one embodiment, two thirds of the members of the training population are placed in the training set and one third of the members of the training population are placed in the test set. Expression values for a gene count described in the present invention are used to train the SVM. Then the ability of a trained SVM to correctly classify the members in the test set is determined. In some embodiments, this computation is carried out several times for a given combination of molecular markers. At each iteration of the computation, the members of the training population are randomly assigned to the training set and the test set. Next, the quality of the biomarker combination is taken as the average of each iteration of the SVM computation.

Evolutionary methods

Inspired by the process of biological evolution, the evolutionary methods of the decision rule design employ stochastic search for a decision rule. In a broad perspective, such methods create several decision rules - a population - from a combination of biomarkers described in the present invention. Each decision rule varies somewhat from the other. Then, the decision rules obtain a score in relation to the characteristic data in the entire training population. In analogy with biological evolution, the resulting (scalar) score is sometimes called fitness. The decision rules are marked according to your score and the best decision rules are retained (some part of the total population of decision rules). Again, maintaining the biological terminology, this is called survival as fit. The decision rules are stochastically altered in the next generation - children or progeny. Some progeny decision rules will have higher scores than their parents in the previous generation, some will have lower scores. The general process is then repeated for the next generation: the decision rules get a score and the best ones are retained, randomly altered to give yet another generation, and so on. In part, due to the classification, each generation has, on average, a slightly higher score than the previous one. The process is interrupted when the best single decision rule in a generation has a score that exceeds a desired criterion value. More information about evolutionary methods can be found in, for example, Duda, Pattern Classification, Second Edition, 2001, John Wiley & Sons, Inc.

Other data analysis algorithms

The data analysis algorithms described above are simply examples of the types of methods that can be used to construct a decision rule to discriminate converters from non-converters. In addition, combinations of the techniques described above may be used. Some combinations have been described, such as the use of the combination of decision trees and amplification. However, many other combinations are possible. In addition, other techniques in the art such as Projection Search and Weighted Voting can be used to construct decision rules.

Devices for diagnosis or prognosis, or monitoring, of a systemic inflammatory condition

Devices useful for the diagnosis, prognosis or monitoring of a systemic inflammatory condition are also described herein. Some devices comprise a computer that has a central processing unit and a memory coupled to the central processing unit. The memory stores the instructions to assess whether one or more quantities of one or more biomarkers of a test subject at risk of developing a systemic inflammatory condition satisfy a first set of values. Satisfying the first set of values predicts that the test subject may develop the systemic inflammatory condition.

FIG. 1 details an exemplary system that supports the functionality described above. The system is preferably a computer system 10 that has:

•

a central processing unit 22;

•

a main non-volatile storage unit 14, for example, a hard disk drive, for storing software and data, the storage unit 14 controlled by the storage controller 12;

•

a system memory 36, preferably high speed random access memory (RAM), for storing the system control programs, data, and application programs, comprising programs and data loaded from the non-volatile storage unit 14; system memory 36 may also include read-only memory (ROM);

•

a user interface 32, comprising one or more input devices (eg, keyboard 28) and a display 26 or other output device;

•

a network interface card 20 for connecting to any wired or wireless communication network 34 (for example, a wide area network such as the Internet);

•

an internal bus 30 to interconnect the aforementioned elements of the system; Y

•

a power source 24 to drive the aforementioned elements.

The operation of the computer 10 is basically controlled by the operating system 40, which is executed by the central processing unit 22. The operating system 40 can be stored in the system memory 36. In addition to the operating system 40, a system Typical implementation memory 36 includes:

•

file system 42 for controlling access to various files and data structures used by the present invention;

•

a set of training data 44 for use in the construction of one or more decision rules in accordance with the present invention;

•

a data analysis algorithm module 54 for processing training data and constructing decision rules;

•

one or more decision rules 56;

•

a biomarker profile evaluation module 60 for determining whether a plurality of quantities in a biomarker profile of a test subject satisfies a first set of values;

•

a biomarker profile of test subject 62 comprising biomarkers 64 and, for each of said biomarkers, amounts 66; Y

•

a database 68 of selected biomarkers of the present invention.

According to what is illustrated in FIG. 1, computer 10 comprises software program modules and data structures. Data structures stored in computer 10 may include training data set 44, decision rules 56, biomarker profile of test subject 62, and / or biomarker database 68. Each of these data structures may comprise any form of data storage system that includes, but is not limited to, a flat ASCII or binary file, an Excel spreadsheet, a database of relational data (SQL), or an online analytical processing database (OLAP) (MDX and / or variants thereof). In some specific embodiments, said data structures are each in the form of one or more databases that include hierarchical structure (eg, a star scheme). In some embodiments, said data structures are each in the form of databases that do not have an explicit hierarchy (for example, dimension tables that are not arranged hierarchically).

In some embodiments, each of the data structures stored or accessible to system 10 are unique data structures. In other embodiments, said data structures actually comprise a plurality of data structures (for example, databases, files) that may or may not be housed by the same computer 10. For example, in some embodiments, the data set Training 44 comprises a plurality of Excel spreadsheets that are stored in computer 10 and / or in computers that are addressable by computer 10 in the entire wide area network 34. In another example, the training data set 44 It comprises a database that is stored in computer 10 or distributed to one or more computers that are addressable by computer 10 throughout the wide area network 34.

It will be appreciated that many of the modules and data structures illustrated in FIG. 1 can be located on one or more remote computers. For example, some embodiments of the present application are network service type implementations. In said embodiments, the biomarker profile evaluation module 60 and / or other modules may reside on a client computer that is in communication with the computer 10 through the network 34. In some embodiments, for example, the evaluation module Biomarker profile 60 can be an interactive web page.

In some embodiments, the training data set 44, decision rules 56, and / or biomarker database 68 illustrated in Figure 1 are on a single computer (computer 10) and in other embodiments one or more of said structures Data and module are housed by one or more remote computers (not shown). Any arrangement of the data structures and software modules in one or more computers is within the scope of the present invention provided that these data structures and software modules are addressable with respect to each other in an entire network or by electronic means. Thus, the present invention fully encompasses a broad array of computer systems.

EXAMPLES

6.1 Example 1: Analysis of serum metamolites of patients with Sepsis and SiRS

Examples 1 and 2 demonstrate the usefulness of the biomarkers of the invention for the diagnosis and prognosis of sepsis and SIRS. Reference biomarker profiles were established for two populations of volunteer patients.

Patient pomulations

The patients were divided into two populations. The first population ("the group with SIRS") represents the patients who developed SIRS and who entered the present study on "Day 1" but did not progress to sepsis during their hospital stay. The second population ("the group with sepsis") represents patients who similarly developed SIRS and entered the present study on Day 1 but who progressed to sepsis typically at least several days after entering the study.

Sample collection

Blood samples were taken approximately every 24 hours from each study group. The clinical suspicion of sepsis in the group with sepsis occurred at "hour 0." The samples were taken in the "hour -12 hours", "hour -36 hours" and "hour -60 hours" before the day of clinical suspicion of the occurrence of sepsis in the group with sepsis. That is, the samples from the group with sepsis included those taken on the day of admission to the study (Day 1), 60 hours prior to the clinical suspicion of sepsis (hour -60 hours), 36 hours prior to the clinical suspicion of sepsis (time -36 hours), and on the day of clinical suspicion of the appearance of sepsis (time -12 hours).

Sample Preparation

At 200 PL of ice-cold methanol (stored at -20 C) 20 PL of human plasma was added. The mixture was then vortexed for 60 seconds before it was incubated at 4 ° C for 20 minutes to aid the precipitation of palsmatic proteins. The mixture was then centrifuged at 13k rpm for 10 minutes at 4 ° C until pellet proteins. The supernatant (100 PL) containing metabolites was then transferred to an Eppendorf tube and evacuated to dry by Speed-Vac. For mass spectrometry (MS) analysis, the metabolites were prepared with a final dilution of 1:50 with respect to the original plasma in 50% methanol and 0.5% formic acid in deionized water with 200 nM reserpine as a internal control.

 Mass spectrometry of the metamolites

The MS analysis of serum metabolite extracts was carried out in a Qstar XL MS / MS system (Applied Biosystems). The sample was poured into MS by the Nanomate (Advion Bioscience), an automated nano-spraying system. The electrospray was carried out at a spray voltage of 1.5 kV and a spray pressure of 0.15 psi. The MS spectrum accumulated over 3 minutes with mass on load (M / Z) ranging from 100 to 1200 Dalton.

6.2 Example 2: Diagnosis and prognosis of Sepsis and SiRS with the miomarker 496.3 or myomarker 518.3 of the invention

This Example provides studies over time of the biomarkers 496.3 and 518.3 and demonstrates their usefulness for the diagnosis, prognosis and monitoring of sepsis and SIRS. Biomarker 496.3 is 11-0-palmitoyl-2-lisosn-glycerol-3-phosphocholine, and biomarker 518.3 is the hate salt of 1-0-palmitoyl-2-smooth-sn-glycerol-3-phosphocholine.

Blood was taken daily from the patients recruited in this study up to 13 days from the day of admission to ICU as described in Example 1. The relative concentrations of the metabolite biomarkers disclosed in the patent were measured with ESI-MS after of extraction with cold methanol, also according to what is described in Example 1.

FIGS. 2 and 3 provide relative concentration changes of biomarkers 496.3 and 518.3 over a period of 13 days for the patient with M113 sepsis and the patient with M245 SIRS.

As can be seen in FIG. 2, the concentration of biomarker 496.3 of the M245 patient with SIRS gradually increases from day 1 to day 13, indicating a gradual recovery of the patient with SIRS from an initial shock trauma.

On the other hand, the concentration of the same biomarker of the patient with sepsis gradually decreases up to two days after the clinical diagnosis of sepsis on day 12. It is also shown in FIG. 2, there is a sudden drop in the concentration of biomarker 518.3 between day 5 and day 7 for the patient with sepsis M113, which indicates a beginning or worsening of systemic infection.

As can be seen in FIG. 3, the concentration of biomarker 518.3 of the M245 patient with SIRS gradually increases from day 1 to day 13, indicating a gradual recovery of the patient with SIRS from an initial shock trauma.

On the other hand, the concentration of the same biomarker of the patient with sepsis gradually decreases until two days after the clinical diagnosis of sepsis on day 12. It is also shown in FIG. 3, there is a fall

5

10

fifteen

twenty

25

30

Sudden concentration of biomarker 518.3 between day 6 and day 7 for the patient with sepsis M113, indicating a beginning or worsening of systemic infection.

Thus, it is possible to diagnose the occurrence of sepsis for the M113 patient between day 6 and day 7, which is approximately four days before clinical symptoms occur. The patient with sepsis M113 suffered a septic shock on day 12 before eventual recovery on day 13, as indicated by the recovery of biomarkers 496.3 and 518.3 (see FIGS. 2 and 3).

An early diagnosis of sepsis between day 6 and day 7 accompanied by an early prophylactic or therapeutic intervention could have saved the patient from septic shock, along with all the costs associated with septic shock treatment.

As shown above, the number of biomarkers 496.3 and 518.3 were useful for the diagnosis and monitoring of SIRS in patients M245 (see days 1-13) and M113 (see days 1-6), In addition, biomarkers 496.3 and 518.3 were also useful for the diagnosis or prognosis of conversion to sepsis and septic shock in patient M113 (see days 6-12) and monitoring the recovery thereof (see days 12-13).

6.3? Example? 3:? Identification? Of? Mymarkers? By? MSSMS

Biomarkers in general were identified by carrying out tandem MS in the peak MS signals of the plasma samples. The peak MS / MS signals were then assigned to the corresponding fragments of the proposed biomarker. The biomarkers were finally confirmed by MS / MS spectrum of commercially purchased compounds.

Table 1. MS / MS of 496.3 (1-0-palmitoyl -2-lyso-sn-glycerol-3-phosphocholine)

"Ionic" peak (Dalton)

Fragments? Assigned

496.34

Ion of progenitor molecule (MH +)

478.33

MH + H2O

313.27

MH +-Phosphorylcholine

184.07

Phosphorylcholine

104.11

Hill

Table 2. MS / MS of 524.3 (1- (1-0-stearoyl -2-lyso-sn-glycerol-3-phosphocholine)

Ionic Peak (Dalton)

Fragments? Assigned

524.37

Ion of progenitor molecule (MH +)

506.35

MH + H2O

341.02

MH +-Phosphorylcholine

184.07

Phosphorylcholine

104.11

Hill

6.4 Example 4: Performance of minimum squares of 1-0-palmitoyl-2-smooth-sn-glycerol-3-phosphocholine as a diagnosis for Sepsis by using minimum square analysis

This example demonstrates how well the assays perform based on the 1-0-palmyl-2-lyso-glycerol-3-phosphocholine marker for the diagnosis of sepsis. In these exemplary trials, the minimum square curve adjustments are used to model the data and diagnose the imminence of sepsis. The performance of the models demonstrates the usefulness of the methods of the invention for the diagnosis, prognosis and monitoring of systemic inflammatory conditions in subjects.

The tests evaluated two characteristics in the subject, 1-0-palmitoyl-2-smooth-sn-glycerol-3-phosphocholine with an apparent molecular weight of 496.3 and the corresponding sodium salt with an apparent molecular weight of 518.3 .

There were 55 total patients, of which 23 were 23 patients with SIRS and 32 were patients with sepsis. Tables 41 and 4-2 below provide distributions of race, sex, age, and septic status for these samples.

5

10

fifteen

twenty

25

30

35

40

? amla? 4-1

State

Sex Afro-American Caucasian Other

SIRS Sepsis

Woman Man Woman Man 0 6 0 5 5 10 11 16 0 2 0 0

? amla? 4-2

? rupo SIRS Sepsis

Minimum 18 19 Mean 52.5 40.7 Medium 57 40 Dev.?st. 19.4 18.4 Maximum 84 79

The patients obtained a score in time points based on the imminence of sepsis. Patients with SIRS obtained a score of 0. On the date of admission, patients with sepsis obtained a score of 1. Two days prior to the onset of sepsis, patients obtained a score of 2. One day before the onset of Sepsis patients obtained a score of 3. Immediately prior to the appearance of sepsis, patients obtained a score of 4.

The "delayed" characteristics were constructed from the characteristics for each patient over time. The data with Delay 0 for a given patient in a given time is the data of that patient in that time only. The data with Delay 1 for that patient at that time is the data with Delay 0, as well as the data of the previous time point for that patient. Similarly, the data with Delay 2 for that patient at that time is the data available for that patient at that time as well as the previous two time points.

A consequence of constructing delays greater than 0 is that data for patients at early time points is incomplete, since delayed data is not available. These cases are discarded from the analysis. Therefore, the data with delay 0 is complete, the data with delay 1 does not have any case of the first available time point, and the data with delay 2 has no case of the first and second time point available. Higher delays produce more information per patient in a given time, but fewer cases to train. Of course, the longer the delay, the longer the time before a prediction can be made, although that prediction may be more accurate finally.

Multiple linear regression models were adjusted, by varying the included characteristics as well as the model complexity parameters, using the SPSA optimization routine (stochastic approximation by simultaneous disturbance). For multiple regression models, no complexity parameters are available. The models were selected to optimize the loss of mean square error.

A model plan with a small expected loss in general will give better patient results (more exact calls, earlier calls, etc.) than a model plan with a high expected loss. However, the same loss does not tell us which patients will be called, or when. To evaluate this, cross validation is carried out with the model plan identified. Patients are divided into K groups (typically 10), each of similar size and with a similar number of septic patients and with SIRS.

(All data of a given patient is assigned to the same group.) One model is adjusted according to the plan, excluding each group at a time. The adjusted model is then applied to the excluded patient data at each time point, and the expected SI values are calculated. Then, a sequence of thresholds is applied to each expected SI value. For each threshold, we can then determine whether the expected Si value of a patient ever exceeded the threshold, and if so, when. The expected results of all patients are then assembled to form estimates of sensitivity and joint specificity for all thresholds.

6.4.1. Delay Results 0

The characteristics selected by the best-estimated model include characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-0-palmyltoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt.

It is estimated that iteration 1000 has the best performance. The model features and complexity selected in this iteration are estimated to produce the loss 1,4482. This model plan was evaluated with 10 times cross validation. FIG. 4-1 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 4-3 presents the data of the selected points. The

5

10

fifteen

twenty

25

Table 4-4 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 4-3 provides the estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

? amla? 4-3

? mmral

Sensibility Specificity

1,329

96.88 13.04

1,7107

75 52.17

2,1145

6.25 91.3

Table 4-4 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is the day of admission, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 4-4

Study Day:

one 2 3 4 i

Sensitivity

Sepsis 28 2 0 one one

SIRS

19 one 0 0 3

Agreement

Sepsis 10 4 5 5 8

SIRS

7 one 0 3 12

Specificity

Sepsis 0 0 0 2 30

SIRS

0 0 one one twenty-one

FIG. 4-2 illustrates the evolution of feature inclusion parameters during optimization. While the last iteration is typically the best, the process is random and the best iteration is estimated; In this case, iteration 1000 is estimated to be the best.

6.4.2. Delay Results 1

The characteristics selected by the best estimated model include characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-O-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt.

Iteration 1000 is estimated to have the best performance. The characteristics and model complexity selected in this iteration are estimated to produce the loss 1.4638. This model plan was evaluated with 10 times cross validation. FIG. 4-3 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 4-5 presents the data of the selected points. Table 4-6 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 4-5 provides estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

? amla? 4-5

? mmral

Sensibility Specificity

1,7107

90.62 30.43

2,1145

75 65.22

2,2587

62.5 91.3

5

10

fifteen

twenty

25

Table 4-6 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is the day of admission, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 4-6

Study Day:

one 2 3 4 i

Sensitivity

Sepsis 0 twenty-one 4 4 3

SIRS

0 5 3 8 7

Agreement

Sepsis 0 8 10 6 8

SIRS

0 one 0 7 fifteen

Specificity

Sepsis 0 5 8 7 12

SIRS

0 0 0 2 twenty-one

FIGS. 4-4 and 4-5 illustrate the evolution of feature inclusion parameters during optimization. While the last iteration is typically the best, the process is random and thus we estimate the best iteration; In this case, iteration 1000 is estimated to be the best.

6.4.3. Delay Results 2

The characteristics selected by the best estimated model include characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-0-palmyltoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt.

Iteration 1000 is estimated to have the best performance. The characteristics and model complexity selected in this iteration are estimated to produce the loss of 1.6208. This model plan was evaluated with 10 times cross validation. FIG. 4-6 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 4-7 presents the data of the selected points. Table 4-8 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 4-7 provides estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

? amla? 4-?

? mmral

Sensibility Specificity

1,7107

90.62 26.09

2.3465

68.75 69.57

2,5329

59.38 91.3

Table 4-8 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is the day of admission, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 4-8

Study Day:

one 2 3 4 i

Sensitivity Concordance

Sepsis SIRS Sepsis 0 0 0 0 0 0 24 8 19 5 9 3 3 6 10

Specificity

SIRS Sepsis SIRS 0 0 0 0 0 0 0 11 0 7 8 2 16 13 21

5

10

fifteen

twenty

25

30

35

FIGS. 4-7 to 4-9 illustrate the evolution of feature inclusion parameters during optimization. While the last iteration is typically the best, the process is random and thus we estimate the best iteration; In this case, iteration 1000 is estimated to be the best.

6.5 Example 5: Performance of 1-0-palmitoyl-2-liso-sn-glycerol-3-phosphocholine as a diagnosis of Sepsis through? The? Utilization? Of? Neural network? Analysis?

This example demonstrates how well the tests perform based on the 1-0-palmitoyl-2-lyso-glycerol-3-phosphocholine marker for the diagnosis of sepsis. In these exemplary trials, adjustments of the neural network curve are used to model data diagnosing sepsis imminence. The performance of the models demonstrates the usefulness of the methods of the invention for the diagnosis, prognosis and monitoring of systemic inflammatory conditions in subjects.

These trials evaluated two characteristics in the subjects, 1-0-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine with an apparent molecular weight of 496.3 and the corresponding sodium salt with an apparent molecular weight of 518.3 .

There were 55 patients in total, of which 23 were 23 patients with SIRS and 32 were patients with sepsis. Tables 5-1 and 5-2 below provide the distributions of race, sex, age, and septic status for these samples.

? amla? 5-1

African American Sex Status

Caucasian Other

SIRS Woman 0 Man 6

5 10 0 2

Sepsis Woman 0 Man 5

11 16 0 0

? amla? 5-2

? rupo Minimum Medium Medium SIRS 18 52.5 57 Sepsis 19 40.7 40

Dev.?st. 19.4 18.4 Maximum 84 79

The patients obtained a score in the time points based on the imminence of sepsis ("SI"). Patients with SIRS obtained a score of 0. On the date of admission, patients with sepsis obtained a score of 1. Two days prior to the onset of sepsis, patients obtained a score of 2. One day before the onset of Sepsis patients obtained a score of 3. Immediately prior to the appearance of sepsis, patients obtained a score of 4.

The "Delayed" characteristics were constructed from the characteristics for each patient over time. The data with Delay 0 for a given patient in a given time is the data of that patient in that time only. The data with Delay 1 for that patient at that time is the delay 0 data, as well as the previous time point data for that patient. Similarly, the data with Delay 2 for that patient at that time is the data available for that patient at that time as well as the previous two time points.

A consequence of constructing delays greater than 0 is that the data for patients at early time points is incomplete, since delayed data is not available. These cases are discarded from the analysis. Therefore, the data with delay 0 is complete, the data with delay 1 does not have any case of the first available time point, and the data with delay 2 has no case of the first and second time point available. Higher delays produce more information per patient in a given time, but fewer cases to train. Of course, the longer the delay, the longer the time before a prediction can be made, although that prediction may be more accurate finally

Many neural network models were adjusted by varying the included characteristics as well as the model complexity parameters, using the SPSA optimization routine (stochastic approximation by simultaneous disturbance). For neural network models, complexity is governed by the number of hidden nodes and decreased weighting. The models were selected to optimize the loss by mean square error.

5

10

fifteen

twenty

25

30

35

A model plan with an expected loss in general will give better patient results (more exact calls, early calls, etc.) than a model plan with a high expected loss. However, the same loss does not tell us which patients will be called, or when. To evaluate this, cross validation is carried out with the model plan identified. Patients are divided into K groups (typically 10), each of similar size and with a similar number of septic patients and with SIRS. (All data of a given patient is assigned to the same group.) One model is adjusted according to the plan, excluding each group at a time. The adjusted model is then applied to the excluded patient data at each time point, and the expected SI values are calculated. Then, a sequence of thresholds is applied to each expected SI value. For each threshold, we can then determine whether the expected Si value of a patient ever exceeded the threshold, and if so, when. The expected results of all patients are then assembled to form joint sensitivity and specificity estimates for all thresholds.

6.5.1. Delay Results 0

The characteristics selected by the best estimated model include the characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-0-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt. The complexity parameters of the model identified are 5 hidden nodes and the weighting decrease of 0.3003.

Iteration 1000 is estimated to have the best performance. The model features and complexity selected in this iteration are estimated to produce the loss of 1.3537. This model plan was evaluated with 10 times cross validation. FIG. 5-1 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 5-3 presents the data of the selected points. Table 5-4 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 5-3 provides estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

? amla? 5-3

? mmral

Sensibility Specificity

1.1336

96.88 4.35

2,0036

68.75 69.57

2,5551

37.5 91.3

Table 5-4 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is the day of admission, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 5-4

Study Day:

one 2 3 4 i

Sensitivity

Sepsis twenty-one 6 3 one one

SIRS

17 2 0 3 one

Agreement

Sepsis 5 2 9 6 10

SIRS

3 one 0 3 16

Specificity

Sepsis one 0 4 7 twenty

SIRS

0 one 0 one twenty-one

FIG. 5-2 illustrates the evolution of feature inclusion parameters during optimization. While the last iteration is typically the best, the process is random and the best iteration is estimated; In this case, iteration 1000 is estimated to be the best.

5

10

fifteen

twenty

25

30

35

6.5.2. Results of Delay 1

The characteristics selected by the best estimated model include the characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-O-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt. The complexity parameters of the model identified are 4 hidden nodes and the weighting decrease of 1,0563:

Iteration 847 is estimated to have the best performance. The model features and complexity selected in this iteration are estimated to produce the loss of 1,3603. This model plan was evaluated with 10 times cross validation. FIG. 5-3 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 5-5 presents the data of the selected points. Table 5-6 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 5-5 provides estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

? amla? 5-5

? mmral

Sensibility Specificity

1.3215

90.62 26.09

2,1481

68.75 69.57

2.5311

59.38 91.3

Table 5-6 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is admission day, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 5-6

Study Day:

one 2 3 4 i

Sensitivity

Sepsis 0 twenty-one 3 5 3

SIRS

0 6 2 9 6

Agreement

Sepsis 0 10 9 3 10

SIRS

0 one 0 6 16

Specificity

Sepsis 0 7 5 7 13

SIRS

0 0 0 2 twenty-one

FIGS. 5-4 to 5-5 illustrate the evolution of feature inclusion parameters during optimization. While the last iteration is typically the best, the process is random and thus we estimate the best iteration; In this case, iteration 847 is estimated to be the best.

6.5.3. Delay Results 2

The characteristics selected by the best estimated model include characteristics with apparent molecular weights of 496.3 and 518.3. According to what is shown in Example 3 above, these characteristics correspond to the 1-0-palmitoyl-2-lyso-sn-glycerol-3-phosphocholine marker and the corresponding sodium salt. The complexity parameters of the model identified are 3 hidden nodes and the weighting decrease of 1.6457.

Iteration 1000 is estimated to have the best performance. The characteristics and model complexity selected in this iteration are estimated to produce the loss of 1,5287. This model plan was evaluated with 10 times cross validation. FIG. 5-6 illustrates the sensitivity and specificity that would be obtained with respect to patients in their full time course. Table 5-7 presents the data of the selected points. Table 5-8 summarizes the estimated time indices of the first call of impending sepsis, for the thresholds selected for high sensitivity, high total concordance, and high specificity, respectively.

Table 5-7 provides estimated sensitivity and specificity for the thresholds selected for high sensitivity (lowest threshold), high specificity (highest threshold), and high total concordance.

5

10

fifteen

twenty

25

30

? amla? 5-?

? mmral

Sensibility Specificity

1.0748

90.62 13.04

2,1886

68.75 65.22

2,6667

62.5 91.3

Table 5-8 provides the estimated distribution of the times of the first sepsis imminence call for the selected thresholds that give high sensitivity, high concordance, and high specificity, respectively. Day 1 is the day of admission, Day 2 is the next day rehearsed, etc. "I" indicates no call of sepsis imminence during the course of the study.

? amla? 5-8

Study Day:

one 2 3 4 i

Sensitivity

Sepsis 0 0 27 2 3

SIRS

0 0 eleven 9 3

? amla? 5-8 (continued)

Study Day:

one 2 3 4 i

Agreement

Sepsis 0 0 19 3 10

SIRS

0 0 2 6 fifteen

Specificity

Sepsis 0 0 13 7 12

SIRS

0 0 0 2 twenty-one

FIGS. 5-7 to 5-9 illustrate the evolution of feature inclusion parameters during optimization. While the last iteration is typically better, the process is random and thus we estimate the best iteration; In this case, iteration 1000 is estimated to be the best.

6.6 Example 6: Colorimeter enzyme assay for the determination of total lysophosphatidylcholine

Patient pomulations

The patients were divided into two populations. The first population ("the group with SIRS") represents the patients who developed SIRS and who entered the present study on "Day 1" but did not progress to sepsis during their hospital stay. The second population ("the group with sepsis") represents patients who similarly developed SIRS and entered the present study on Day 1 but who progressed to sepsis typically at least several days after entering the study. ? 0362? Sample collection

Blood samples were taken approximately every 24 hours from home study group. The clinical suspicion of sepsis in the group with sepsis occurred at "time 0." The samples were taken at "time of 12 hours", "time of 36 hours" and "time of 60 hours" before the day of clinical suspicion of the appearance of sepsis in the group with sepsis. That is, samples from the group with sepsis included those taken on the day of admission to the study (Day 1), 60 hours prior to the clinical suspicion of sepsis (time 60 hours), 36 hours prior to the clinical suspicion of sepsis (time 36 hours), and the day of clinical suspicion of the occurrence of sepsis (time 12 hours).

Reagent Sources

Lysophospholipase (EC 3.1.1.5) was obtained from Asachi Chemical Co. (Tokyo, Japan). Glycerophosphorylcholine phosphodiesterase (GPCP; EC 3.1.4.2), choline oxidase (COD; EC 1.1.3.17), peroxidase (EC 1.11.1.7), N-ethyl-N- (2-hydroxy-3-sulfopropyl) -3-methylaniline dehydrate Sodium (TOOS) and 4-aminoantipyrine were purchased from Aldrich (Milwaukee, WI). 1-palmitoyl-2-hydroxy-phosphatidylcholine was obtained from Avanti Polar-Lipids Inc. (Alabaster, AL). 10-acetyl3,7-dihydroxyphenoxazine was obtained from Invitrogen (Carlsbad, CA).

Mechanism

In the test mechanism shown in Scheme 1 (FIG. 6), a lysophosphatidylcholine (LPC) molecule

5

10

fifteen

twenty

25

30

35

40

It can pass through a series of enzymatic reactions to produce two molecules of hydrogen peroxide. The resulting hydrogen peroxide can oxidize 4-aminoantipyrine in the presence of sodium N-ethyl-N- (2-hydroxy-3-sulfopropyl) -3-methylaniline dehydrate (TOOS) salt to produce a tincture of quinoneimine. The absorbance intensity of quinoneimine can be measured at the wavelength of 590 nm.r

Reagents

Reagent A comprises 100 mM Tris-HCl (pH 8.0), 0.01% Triton X-100, 1 mM calcium chloride, 3 mM TOOS (NFthylN- (2-hydroxy-3-sulfopropyl) -3-methylaniline, sodium salt, dihydrate), 10 K Units / L of peroxidase, 0.1 K Units / L GPCP of (glycerophosphorylcholine phosphodiesterase), and 10 K Units / L of COD (choline oxidase). Reagent B comprises 100 mM Tris-HCl (pH 8.0), 0.01% Triton X-100, 5 mM 4-aminoantipyrine and 30 K Units / L of lysophospholipase.

LPC Colorimetric Enzyme Assay

An 8 µl plasma sample was pre-incubated with 240 µl of reagent A for 5 minutes at 37 ° C, and the absorbance between 570 nm (primary wavelength) and 700 nm (secondary wavelength) was measured in a reader of plate (Perkin Elmer Victor3). The reaction was started by adding 80 µl of reagent B. After 5 minutes, the absorbance was measured between 570 nm and 700 nm. The total LPC concentration was determined with the aid of a calibration curve of the sequential dilution of 500 umol / L of 1-palmitoyl-2-hydroxy-phosphatidylcholine.

Results

Sample

Rep1 (absorbance) Rep2 (absorbance) Medium (absorbance)

STD1: 600 µM

0.449 0.445 0.447

STD2: 400 µM

0.336 0.334 0.335

STD3: 200 µM

0.169 0.233 0.201

STD6: 0? M

0.071 0.116 0.094

Healthy subject

0.231 0.223 0.228

Patient 113 D10 (sepsis)

0.081 0.078 0.080

Patient 237 D09 (SIRS)

0.165 0.193 0.179

As shown in the table above, a patient with sepsis and a patient with SIRS have lower concentrations of LPC compared to a normal healthy subject. The concentration of LPC in the patient with sepsis is also decreased compared to the patient with SIRS. Both differences were observed in a larger sample set.

6.? Example?: Fluorescent enzyme assay for the determination of LPC? Total

Porlacíones of patients

The patients were divided into two populations. The first population ("the group with SIRS") represents patients who developed SIRS and who entered the present study on "Day 1" but did not progress to sepsis during their hospital stay. The second population ("the group with sepsis") represents patients who similarly developed SIRS and entered the present study on Day 1 but who progressed to sepsis typically at least several days after entering the study.

Sample collection

Blood samples were taken approximately every 24 hours from each study groups. The clinical suspicion of sepsis in the group with sepsis occurred in "time 0." The samples were taken in "12-hour time", "36-hour time" and "60-hour time" prior to the day of clinical suspicion of the occurrence of sepsis in the sepsis group. That is, samples from the group with sepsis included those taken on the day of admission to the study (Day 1), 60 hours prior to the clinical suspicion of sepsis (time of 60 hours), 36 hours prior to the clinical suspicion of sepsis ( 36 hours time), and the day of clinical suspicion of the appearance of sepsis (12 hours time).

Reaction sources

Lysophospholipase (EC 3.1.1.5) was obtained from Asachi Chemical Co. (Tokyo, Japan). Glycerophosphorylcholine phosphodiesterase (GPCP; EC 3.1.4.2), choline oxidase (COD; EC 1.1.3.17), peroxidase (EC 1.11.1.7), N-ethyl-N- (2-hydroxy-3-sulfopropyl) -3-methylaniline dehydrate Sodium (TOOS) and 4-aminoantipyrine were purchased from Aldrich (Milwaukee, WI). 1-Palmitoyl-2-hydroxy-phosphatidylcholine was obtained from Avanti Polar-Lipids Inc. (Alabaster, AL). 10-Acetyl-3,7-dihydroxyphenoxazine was obtained from Invitrogen (Carlsbad, CA).

Mechanism

Similar to the colorimetric enzyme assay, one molecule of LPC can produce two molecules of hydrogen peroxide through a series of enzymatic reactions (Scheme 1; see FIG. 6). The resultant can then oxidize 10-acetyl-3,7-dihydroxyphenoxazine and form a fluorescent product, 7-hydroxy-3H-phenoxazine-3-one. The fluorescent intensity of 7-hydroxy-3H-phenoxazine-3-one can be measured at 590 nm with an excitation wavelength of approximately 530 nm.

Reacts

Reagent A comprises 100 mM Tris-HCl (pH 8.0), 0.01% Triton X-100, 10 mM magnesium chloride, 0.1 K units / L GPCP (glycerophosphorylcholine phosphodiesterase), 10 K units / L peroxidase, 10 K units / L of COD (choline oxidase) and 10 m-acetyl-3,7-dihydroxyphenoxazine. Reagent B comprises 100 mM Tris-HCl (pH 8.0), 0.01% Triton X-100 and 30 K units / L of lysophospholipase.

240 PL of reagent A was added to each well of a 96 black microplate for fluorescence. 8 PL of each sample and or standards are added in each well containing reagent A. 80 PL of reagent B are added in each well. The resulting mixture is incubated for 60 minutes or more at 37 ° C, protected from light. The fluorescence is read on the Perkin Elmer Victor 3 plate reader with an excitation wavelength of 530 nm, an emission wavelength of 590 nm, lamp energy of '10000', and "Small" emission aperture. The total LPC is determined with the aid of a calibration curve of 500 umol / L of 1-palmitoyl-2-hydroxy-phosphatidylcholine sequentially diluted in a buffer solution comprising 100 mM Tris-HCl (pH 8.0), Triton X -100 to 0.01% and 10 mM magnesium chloride.

Results

The total LPC concentrations for subjects with sepsis 29 and 22 subjects with SIRS on the day of admission, intervals T-60, T-36 and T-12 (the day of clinical suspicion of the occurrence of sepsis) are shown in the FIG. 7. As can be seen in FIG. 7, the concentration of total LPC for most patients with SIRS gradually increases from the day of admission, indicating a gradual recovery of each patient with SIRS from an initial shock trauma. On the other hand, the total LPC concentration for most patients with sepsis begins to fall or is leveled two days before the day of clinical suspicion of the occurrence of sepsis (T-12).

In this data set, the clinical prognosis of sepsis can be performed according to the methods of the invention one to two days before the appearance of sepsis with 90% specificity and 60% sensitivity, by using the following criteria:

-concentration of LPC (Dn)? 60 PM on the day of the call (n day); Y

-Dn -Dn-1? 0 PM (LPC concentration dropped or leveled).

An early diagnosis of sepsis accompanied by an early prophylactic or therapeutic intervention could have saved the patient from septic shock along with all the costs associated with septic shock treatment.

Although the foregoing invention has been described in some detail by way of illustration and example for the purposes of clarity of understanding, it will be readily apparent to those with common experience in the art in light of the training of the present invention that certain changes and modifications they can be made thereto without departing from or scope of the appended claims.

Claims (60)

 REI? INDICATIONS

one.

A method for the diagnosis or prognosis of sepsis in a patient comprising the stage of monitoring or evaluating the amount of total lysophosphatidylcholine in a sample of fluid or tissue of the patient, where the diagnosis or prognosis is made prior to the appearance of sepsis.

2.

The method of Claim 1, wherein the lysophosphatidylcholine is a compound according to formula (I)

image 1 or a salt or solvate thereof, where R is C10-C22 acyl. 3. The method of Claim 2, wherein R is C14-C22 acyl, preferably C16-C20 acyl, more preferably C16-C18 acyl.

Four.

The method of Claim 1 comprising the step of measuring over time a plurality of amounts of total lysophosphatidylcholine in the patient fluid or tissue sample.

5.

The method of Claim 4 wherein said lysophosphatidylcholine comprises 1-0-palmitoyl-2-lyso-sn-glycer 3-phosphocholine and 1-0-stearoyl-2-lyso-sn-glycerol-3-phosphocholine.

The method of Claim 4 wherein the total free lysophosphatidylcholine and / or total bound lysophosphatidylcholine indicates the amount of total lysophosphatidylcholine. 7. The method of Claim 4, wherein the lysophosphatidylcholine is a compound in accordance with formula (I) image 1 20 or a salt or solvate thereof, where R is C10-C22 acyl.

8.

The method of Claim 7 wherein R is C14-C22 acyl, preferably C16-C20 acyl, more preferably C16-C18 acyl.

9.

The method of Claim 7 wherein R is saturated acyl, preferably C16 saturated acyl, or C18 saturated acyl.

The method of Claim 7 wherein R is unbranched C16-C18 acyl, preferably palmitoyl, or stearoyl. 5 10 fifteen twenty 25 30 35

eleven.

The method of Claim 4 or 7 wherein a decreasing amount indicates increasing risk of sepsis or an increasing amount indicates decreasing risk of sepsis, or where a decreasing amount indicates diagnosis of sepsis or an increasing amount indicates against diagnosis of sepsis.

12.

The method of Claim 4 or 7 wherein the patient is SIRS positive.

13.

The method of Claim 4 or 7 which additionally comprises measuring one or more biomarkers, wherein said biomarker is distinct from lysophosphatidylcholine.

14.

The method of Claim 4 or 7 wherein the amount is measured by spectrometry, chromatography, immunoassay and, electrophoresis or enzymatic assay.

fifteen.

The method of Claim 4 or 7 wherein the fluid or tissue is blood, plasma, saliva, serum, sputum, urine, cells, cell extract or tissue biopsy.

16.

The method of Claim 4 or 7 wherein the fluid or tissue is blood.

17.

The method of Claim 1 or 2 comprising the step of monitoring or evaluating one or more clinical measurements.

18.

The method of Claim 17 wherein the clinical measurement is according to a clinical severity model for sepsis, wherein the clinical model is selected from the group consisting of the Chronic Health Assessment and Acute Physiology, the Chronic Health Assessment and Acute Physiology II, Chronic Health Assessment and Acute Physiology III, Mortality Prediction Model, Simplified Acute Physiology score, Multiple Organ Dysfunction score, Sequential Organ Failure Assessment score, Logistic Dysfunction Score Organs, and predisposition, infection, response, and organ dysfunction concept.

19.

The method of Claim 1 comprising the step of comparing the amount of total lysophosphatidylcholine in the patient's fluid or tissue sample against a reference amount indicative of the amount of total lysophosphatidylcholine in a fluid or tissue sample of an individual who have sepsis

twenty.

The method of Claim 19, wherein the lysophosphatidylcholine is a compound according to formula (I)

image 1 or a salt or solvate thereof, where R is C10-C22 acyl.

twenty-one.

The method of Claim 20 wherein R is C14-C22 acyl, preferably C16-C20 acyl, more preferably C16-C18 acyl.

22

The method of Claim 19 or 20 wherein the difference between the amounts is inversely correlated with the risk of sepsis.

2. 3.

The method of Claim 19 or 20 wherein the reference amount is a limit reference amount.

24.

The method of Claim 23, wherein an amount less than the limit reference amount indicates the diagnosis of sepsis.

25.

The method of Claim 19 or 20 wherein the amount is measured 12, 24, 36 or 48 hours prior to the onset of sepsis.

26.

The method of Claim 19 or 20 wherein the patient is SIRS positive.

27.

The method of Claim 26 wherein the individual is SIRS positive.

28.

The method of Claim 26 wherein the individual is positive sepsis.

29.

A method for the diagnosis or prognosis of sepsis in a patient comprising the step of detecting lysophosphatidylcholine in the patient by contacting a sample of fluid or tissue from the patient with:

an enzyme or reagent capable of reacting lysophosphatidylcholine to form glycerophosphatidylcholine; an enzyme or reagent capable of reacting glycerophosphatidylcholine to form choline; an enzyme or reagent capable of reacting choline, water and oxygen to form peroxide; a peroxidase; Y a fluorogenic substrate of said peroxidase; under appropriate conditions for the formation of a fluorescent product where the fluorescent product indicates lysophosphatidylcholine; where the diagnosis or prognosis is made prior to the appearance of sepsis.

30

The method of Claim 29 wherein said enzyme or reagent capable of reacting lysophosphatidylcholine to form glycerophosphatidylcholine is a lysophospholipase.

31.

The method of Claim 29 wherein said enzyme or reagent capable of converting lysophosphatidylcholine into glycerophosphatidylcholine is EC 3.1.1.5.

32

The method of Claim 29 wherein said enzyme or reagent capable of reacting the glycerophosphatidylcholine to form choline is a diesterase glycerophosphatidylcholine.

33.

The method of Claim 29 wherein said enzyme or reagent capable of reacting glycerophosphatidylcholine to form choline is EC 3.1.4.2.

3. 4.

The method of Claim 29 wherein said enzyme or reagent capable of reacting the choline to form peroxide is a choline oxidase.

35

The method of Claim 29 wherein said enzyme or reagent capable of reacting choline with peroxide is EC 1.1.3.17.

36.

The method of Claim 29 wherein said peroxidase is horseradish peroxidase.

37.

The method of Claim 29 wherein said fluorogenic substrate is 10-acetyl-3,7-dihydroxyphenoxazine and said fluorescent product is 7-hydroxy-3H-phenoxazin-3-one.

38.

The method of claim 29 wherein said fluorescent product is detected by fluorescent detection.

39.

The method of Claim 29 wherein the sample is contacted with a lysophospholipase, a glycerophosphatidylcholine diesterase, a choline oxidase, a peroxidase and 10-acetyl-3,7-dihydroxyphenoxazine under conditions suitable for the formation of 7-hydroxy-3H phenoxazin-3-one where 7-hydroxy-3H-phenoxazin-3-one indicates lysophosphatidylcholine.

40

The method of Claim 29 wherein the sample is contacted with EC 3.1.1.5, EC 3.1.4.2, EC 1.1.3.17, horseradish peroxidase and 10-acetyl-3,7-dihydroxyphenoxazine under conditions suitable for the formation of 7 -hydroxy-3H-phenoxazin-3-one where 7-hydroxy-3H-phenoxazin-3-one indicates lysophosphatidylcholine.

41.

The method of Claim 29 wherein the amount of fluorescent product indicates the amount of lysophosphatidylcholine.

42

The method of Claim 19, wherein the amount of total lysophosphatidylcholine in the patient's fluid or tissue sample is compared with a reference amount indicative of the amounts of total lysophosphatidylcholine in fluids or tissues of a plurality of individuals having sepsis.

43

The method of Claim 1 comprising the step of comparing the amount of total lysophosphatidylcholine in the patient's fluid or tissue sample with an internal standard in the fluid or tissue.

44.

The method of Claim 17 wherein the clinical measurement / s are selected from the group consisting of respiratory rate, temperature, heart rate, white blood cell count, monocyte count, lymphocyte count, granulocyte count, blood count neutrophils, ratio of immature neutrophil and total neutrophil, platelet count, serum creatinine concentration, urea concentration, lactate concentration, excess bases, pO2 concentration and HCO3-.

Four. Five.

The method of Claim 17 wherein a clinical measurement is monitored or evaluated.

5 10 fifteen twenty 25 30 35

46.

The method of Claim 45 wherein the respiratory rate is monitored or evaluated.

47

The method of Claim 45 wherein the temperature is monitored or evaluated.

48.

The method of Claim 17 wherein more than a clinical measurement is monitored or evaluated.

49.

The method of Claim 48 wherein the respiratory rate and temperature are monitored or evaluated.

fifty.

The method of Claim 17 which further comprises monitoring or evaluating one or more biomarkers in the patient's fluid or tissue sample, wherein the biomarkers are selected from the group consisting of endotoxin, bacterial DNA, protein C, procalcitonin, binding protein LBP-LPS, fibrin degradation products, antithrombin III, dimer D, HLA-DR, CD-64, E-selectin, cortisol, ACTH, CD-14, sTNF-RI, sTNF-RII, TNF, IL-6, IL-8 and IL-10, D-dimer, prothrombin time, activated partial thromboplastin time, plasminogen activator inhibitor-1, soluble thrombomodulin, thrombin activatable fibrinolysis inhibitor, S protein, antithrombin and TNF-a.

51.

The method of Claim 1 comprising the steps of: measuring a plurality of amounts of a compound in accordance with formula (I)

image 1 or a salt or solvate thereof, where R is C10-C22 acyl, in the patient's fluid or tissue sample; and measure one or more clinical measurements of the patient.

52

The method of Claim 51 wherein the measurement (s) is clinical / s are selected from the group consisting of respiratory rate, temperature, heart rate, white blood cell count, monocyte count, lymphocyte count, granulocyte count, Neutrophil count, ratio of immature neutrophil and total neutrophil, platelet count, serum creatinine concentration, urea concentration, lactate concentration, excess bases, pO2 and HCO3 concentration.

53.

The method of Claim 51 wherein a clinical measurement is measured.

54

The method of Claim 53 wherein the respiratory rate is measured.

55.

The method of Claim 53 wherein the temperature is measured.

56.

The method of Claim 51 wherein it is measured more than a clinical measurement.

57.

The method of Claim 56 wherein the respiratory rate and temperature are measured.

58.

The method of claim 17 or Claim 51 wherein the patient is SIRS negative.

59.

The method of Claim 17 or Claim 51 wherein the patient is SIRS positive.

60

The method of Claim 17 or Claim 51 wherein the amount is measured by spectrometry, chromatography, immunoassay, electrophoresis or enzymatic assay.

61.

The method of Claim 17 or Claim 51 wherein the fluid or tissue is blood, plasma, saliva, serum, sputum, urine, cells, cell extract or tissue biopsy.

62

The method of claim 51 further comprising measuring one or more biomarkers in the patient's fluid or tissue sample, wherein the biomarkers are selected from the group consisting of endotoxin, bacterial DNA, protein C, procalcitonin, LBP binding protein - LPS, fibrin degradation products, antithrombin III, dimer D, HLA-DR, CD-64, E-selectin, cortisol, ACTH, CD-14, sTNF-RI, sTNF-RII, TNF, IL-6, IL- 8 and IL-10, D-dimer, prothrombin time, activated partial thromboplastin time, inhibitor-1 activator

plasminogen, soluble thrombomodulin, thrombin activatable fibrinolysis inhibitor, S protein, antithrombin and TNF-a