CA1196861A - Immunoassay for oncofetal antigen carried by lymphocytes - Google Patents

Abstract

Abstract of the Disclosure An immunoassay, to be carried out in a semiquantitative or quantitative manner, to determine the amount of ferritin associated with human peripheral blood mononuclear cells, in which a competitive binding immunoassay for ferritin is carried out using labeled antibody and wherein analogous mononuclear cells carrying a defined amount of ferritin are used as a positive quality control reagent. An immunochemical test kit comprising the necessary reagents is also disclosed.

Description

i6~

IMMUNOASSAY FOR FERRITIN
CARRIED ON HUMA~ MONONUCLEAR WHITE BLOOD CELLS

Background of the ~nvention The present invention relates to procedures and clinical test kits for determining the presence of ferritin on the surface of human mononuclear white ~lood cells ~PMC's). More particularly, the present invention relates to improved immunoassay procedures of the antigen-antibody type wherein a unique positive reference is employed to provide a mononuclear white blood cell control means Eor checking the viability of rea~ents and to verify test reliability. Semiquantitative and quantitative embodiments are disclosed.
U.S. Patent No. 4,116,77b to Dalbow et al. discloses a diagnostic proceduxe for determining the presence of human chorionic gonadotropin (hCG~ on lymphocytes. The procedure is directed to an improvement in the detection of hCG in men and nonpregnant women as a test for cancer. Dalbow et al.
note the conventional radioimmunoassay test for the presence of hCG in blood serum and then ~tate that they have found that the host's lymphocytes collect and bind the hCG secreted by malignant neoplastic cells and, because of this, their test is characterized by a sensitivity not found in the conventional serum test. After removing sialic acid residues from -the hCG molecules bound to the iymphocytes, Dalbow et al. expose the lymphocytes to labeled antiserum specific to the beta-subunit of hCG and, after equilibrium is reached, measure the concentration of hCG by measuring the amount of label, such as fluorescence or radioactivity. Negative controls, such as blood from known healthy individuals or a selected nonspecific antiserum attached to the same la~el, are described to be normally employed in the Dalbow et al.
assay.
U.S. Patent No. 3,988,115 to Modabber discloses a diagnostic procedure for determining the amount of lymphocytes capable of binding to a ligandO The test is ~ased on a ligand interacting with lymphocytes having receptors specific to the ligand, and in this manner stimulating the proliferation of only the lymphocytes having the receptors specific to the ligand. Labeled ligand is used to detect the quantity of the lymphocytes having the r~ceptors specific to the ligand.
One advantage of the test is stated to ~e its usefulness prior to the presence of a significant amount of antibody in the blood stream. Blood from a normal individual or particles of predetermined ligand binding activity, such as antibody absorbed onto gel beads, can be used as a control.
Rockoff et al., "Sensitive and Convenient Quantitation of Antibody Binding to Cellular Antigens Using Glutaraldehyde Preserved Cells", Journal of Immunological Methods, 26, 369 (1979), describe a procedure to detect and quantify antibodies specific to cancers using glutaraldehyde-fixed target cells.
The controls are normal serum and liquid medium.
Tomasi, Jr., "Structure and Functions of Alp~a-feto-protein", A . Rev. Med., 1977, 28:453, descri~es a high_ incidence of alpha-fetoprotein (AFP, an oncofetal antigen) on the surface of peripheral blood mononuclear cells of so-called cancer-prone families. The present inventors have been informed that Tomasi, Jr. used a fluorescent immunoassay technique where a monospecific fragment of anti-AFP antibody was chemically coupled to fluorescein and then incubated with patient lymphocytes, followed by counting positive cells under a microscope.
Naughton et al., "Localization of the ~ Chain on Human Chorionic Gonadotropin on Human Tumor Cells and Placental Cells", Cancer Research, 35, 1887 (1975), located hCG on the surfaces of human tumor cells using enzyme-labeled hCG
antibody.
Ferritin, another oncofetal protein, has been described as a potential tumor marker. Marcus et al., "Serum Ferritin Levels in Patients with Breast Cancer", Clin. Res., 23, 447 (1975) measured serum ferritin using a double-anti~ody radioimmunoassay technique. On the other hand, more recent work by Moroz et al. indicates that a more specific malignant tumor marker is the presence of significant ferritin-bearing lymphocytes ~"Lymphocytes Bearing Surface Ferritin in Patients with Hodgkin's Disease and Breast Cancer", The New England Journal of Medicine, correspondence, 276, No. 20, 1172 (1977) and "Ferritin-Bearing Lymphocytes and T-Cell Levels in Peripheral Blood of Patients with Breast Cancer", Cancer Immunol. Immunother., 3, 101 (1977)]. Lymphocytes are a -subpopulation of the mononuclear white blood cell population.
Moroz et al. used a cytotoxicity test to determine the ferritin-bearing cells wherein after antibody is added to the patient's lymphocytes, complement is added to lyse cells which have antibody on the surface and then the number 6~

of dead cells is determined using a standard dye of the type which can't be excluded by dead cells.
Other references of interest include the following:
Rinke, "CEA: Still Trying to Find a Niche", Journal of the American Medical Association, 244, 2707 (1380), points out some limitations on the use of carcinoembryonic antigen (CEA) levels as a tumor marker, but also clearly identifies situations in which the knowledge of such levels is clinically significant and useful.
Keller et al., "Alpha-Fetoprotein Synthesis by Murine Lymphoid Cells in Allogeneic Reactions," The Journal of Experimental Medicine, 143, 1140 (1976), describes work directed at identifying those cells responsible for synthesizing alpha-fetoprotein (AFP). The procedures employed included the estimation of AFP on the surfaces of murine spleen or lymph node cells by, for example, observing surface fluorescence for AFP after reacting the cells with monospecific anti-AFP
antiserum in the presence of mouse amniotic fluid as a source of AFP. No attempt was made to relate AFP synthesis with any cancerous disease state, and peripheral mononuclear cells were not involved in the experimental work.
Japanese Published Patent Application (Kokai) 79,130,093 (Shitsuzaki et al.) ~Chem. Abstr., 92, 18402s (1980)], discloses a solid anti-ferritin antibody-carrier reagent suitable for use in a radioimmunoassay for soluble ferritin in serum. In general, serum ferritin levels primarily are of interest in the diagnosis of anemic conditions, particularly those resulting from iron deficiencies. Thus, the disclosure has no relationship to the present invention.

Japanese Published Patent Application (Kokai) 79,130,094 (Shitsuzaki et al.) [Chem. Abstr., ~ 4alr (1980)], __ discloses a radioimmunoassay for soluble ferritin in serum, using the above-described reagent. Again, the disclosure has no relationship to the present invention.
U.S. Patent No. 4,144,031 to Acevedo et al. describes a diagnostic test and kit for determining whether or not a patient has cancer. The test measures the presence or absence of the hormone human chorionic gonadotropin (hCG);
neoplastic cells have hCG on their surfaces, while normal cells do not. The marker ~hCG) is not disease specific, i.e., it does not appear to ~e capa~le of discriminating among the various types of cancers. The present assay measures a marker bound to a subpopulation of normal T
lymphocytes, not tumor cells. Moreover, the present assay involves direct measurements rather than the indirect measurements employed by Acevedo et al. In addition, the present assay provides a means for validating assay results, a means lacking in the disclosed test.
Gleich et al., "Measurement of IgE in Normal and Allergic Serum by ~adioimmunoassay," J. Lab. Clin. Med., 77, 690 (1971), discloses the measurement of immunoglobulin E (IgE) in serum bv means of a sensitive specific immunoassay. The work reported was directed to the role of IgE in allergic reactions. The procedure is unrelated to cancerous disease states and does not involve the measurement of any type of a marker on cell surfaces.

Summary of the Invention It is an object of this invention to provide a clinical blood test for detecting ferritin on peripheral mononuclear white blood cells (PMC's).

6i~

A further object of this invention is to provide a clinical testiny procedure for quantifying the subpopulation of PMC's carrying ferritin.
Still another object of this invention is to provide an improved immunoassay label-t~pe procedure for detenmining the presence of ferritin on PMC's in a quantitative or semiquantitative fashion w~erein an artificially positive cellular reference control consisting of ferritin attached to PMC's grown in culture is used to validate the test system.
Another object of this invention is to provide an immunoassay label~type procedure for use in the early detection, therapeutic monitoring, and prognosis of at least certain types of cancer.
A further object of this invention is to provide clinical test kits for carrying out the above-described types of procedures.
It has now been found that ferritin present on the surfaces of peripheral blood mononuclear cells, e.g., lymphocytes, can be detected in a semiquantitative or quantitative manner by carrying out an immunoassay procedure wherein (A~ an aliquot of a patient's perpiheral mononuclear cells is admixed with (1) labeled antibody for ferritin and a second aliquot of the same cell sample is admixed with (2j the labeled antibody plus a competitive ~inding inhibiting amount of purified ferritin and ~B) duplicating the above Procedure (A) with artificially prepared peripheral blood mononuclear-type cells coated with ferritin being substituted for the patient's test samples. ~hereafter, appropriate calculat~ons are c~rried out, as will be explained more fully hereinbelow.

In a preferred embodiment of the invention, where a quantitative assay is to be carried out, a fourth reagent, namely, immobilized antibody to human ferritin, is used in combination with the labeled antibody, prefera~ly labeled with radioactive iodine such as I125, and the purified human ferritin to develop a series of standard curves.
The pr~sent invention ~lso provides reagent kits containing in separate fashion each of the reagents disclosed h~rein, with the immobilized anti~ody ~eing included where standard curves for quantitative testing are to be developed.
Thus the present in~ention provides a method for detecting ferritin on peripheral blood mononuclear cells which comprises:
(A) carrying out a first test by admixing an aliquot of a white blood cell sample comprising peripheral blood mononuclear cells with an ali~uot ~f l~eled anti-ferritin antibody, removing the anti~ody which does not attach to said cells and determining the xelative amount of antihody attached to said cells through the use of 6aid la~el;
(B) carrying out a second test ~y admixing an aliquot of said white blood ~ell sample with an aliquot o labeled anti-human ferritin antibody and a co~petitive binding inhibiting amount of ferriti~, removing the antibody which is not attached to said cells and determining the relative amount of antibody attached to said cells through the use of said label;
(C) determining the relative amount of ferritin carried by said cells from the values obtained in steps lA~ and (B);
(D) repeating steps (A)-(C) but with su~stituting peripheral blood mononuclear-type cells carrying a known quantity of ferritin for the blood sample; and ,~

~ E) checking the ~alidity of the method and of the labeled antibody reagent from the re~ults of step ~D).
In a preferred embodiment the invention provides such a method wherein a quantitative detection method is carried out which comprises, in addition, to said steps (A)- (E):
(F) preparing a standard curve by admixing increasiny amo~nts of ferritin with a defined amount of immo~ ed ~ntibody and a defined amount of labeled antl~ody;
(G) determining specific binding for each ferritin concentration through the use of said label; and (H) comparing the result obtained in step (C~ with said standard curve to determi~e the amount of ferritin corresponding to said value obtained in step ~C~.
In another aspect the invention provides a method for detecting the presence of cancer in a patient which comprises carrying out the method as above and then comparing the value obtained in step (C) with the range of values obtained with healthy, nonmalignant individuals.
In still another aspect the invention provides a method for detecting the presence of cancer in a patient which comprises carrying out the method as abo~e and then comparing the value obtained in step (H) with the range of values obtained with healthy, nonmalignant individuals.
In still another aspect the invention provides an immunochemical test kit useful for detecting ferritin on peripheral blood mononuclear cells, said kit comprising as separate reagentsO
(1) labeled anti-human ferritin antibody for said ferritin;

- 7a -6~

(2) ~id ferritin; and

(3) peripheral blood morlonuclear-t~pe cells carrying a known quantity of erritin.

Detailed Description of the_Invention In the past few years, an increasing amount of ~iochemical work has been carried out with the objective of detecting and quantifying oncofetal antigens. Oncofetal antigens normally axe present in signi~icant quantities only in fetal life, or in adults during certain nonmalignant physiological conditions, such as the presence of hCG in normal pregnant women. However, there is increasing evidence that oncofetal antigens also are present in adults exhibiting or having a predisposition for cancer. A pro~lem is that indicated above, namely, the presence of oncofetal antigens in adults due to other physiological conditions, e.gO, hCG in the serum of nonpregnant women having inflammatory diseases of the digestive tract, alpha~fetoprotein in the serum of patients with acute hepatitis and chronic liver disease, and so on.
The most extensively studied of these oncofetal antigens is carcinoembryonic antigen ~CEA), a soluble antigen found in the liquid component of whole blood of many cancer patients.

- 7b -While originally believed to be specific for colorectal cancer diagnosls, extensive clinical studies over the past ten years have revealed that serum CEA levels may be elevated in a variety of malignant and nonmalignant disorders, thereby rendering invalid the use of such levels in early diagnostic procedures. Despite this lack of specificity, the measurement of CEA levels has found a valuable and growing place in clinical medicine in the evaluation, therapeutic management, and prognostic monitoring of the cancer patient. See, e.g~, Fuchs et al., "Theoretical and Practical Considerations of the Utility of the Radioimmunoassay for Carcinoem~ryonic Antigen (CEA) in Clinical Medicine" in Stuart Sell, Editor, "Cancer Markers - Diagnostic and Developmental Significance,"
Humana Press, Clifton, New Jersey, 1980, p. 315; and Zamcheck and Kupchik, "~ummary of Clinical Use and Limitations of the Carcinoembryonic Antigen Assay and Some Methodological Considerations" in Noel Rose and Herman Freedman, Editors, "Manual of Clinical Immunology," American Society for Micro-biology, Washington, D.C., 1976, p. 753.
The measurement of CEA levels has proven of value as a monitor of therapy in advanced cancer [Mayer et al., "Carcin-oembryonic Antigen (CEA) as a Monitor of Chemotherapy in Disseminated Colorectal Cancer," Cancer, 42, 1428 (1978)].
Increases in CEA levels in serially monitored colorectal carcinoma patients have ~een shown to preceed clinical signs of recurrence by up to 26 montns [Mach et al., "Long-Term Follow-up of Colorectal Carcinoma Patients by Repeated CEA
Radioimmunoassay," Cancer, 42, 1439 (1978)]. It is clear 3~

from these reports that CEA lev~l measurements, when utilized under an atmosphere of proper and prudent clinical interpretation, are of extreme value in the management of the cancer patient.
The 1980 consensus meeting on CEA held that CEA measurement i5 the best presently available noninvasive technique for postoperative surveillance of patients to detect recurrent and/or metastatic colorectal cancer. C. Rinke, "CEA: Still Trying to Find a Niche," Journal of the American Medical Association, 244, 2707 ~1980); and "Carcinoembryonic Antigen:
Its Role As a Marker in the Management o~ Cancer - National Institutes of Health Consensus De~elopment Conference Statement,"
Cance Research, ~1, 2017 (1981). Thus, a clinical test_ system for a given oncofetal antigen need not be of a~solute specificity to be of utility.
Recent work reported ln the literature has distinguished between serum oncofetal antigen and oncofetal antigen ~ound to mononuclear peripheral blood cells. It has been described in the literature that significant amounts of oncofetal antigen related to a malignant state, or even a predisposition thereto, are found on the surfaces of the mononuclear peripheral blood cells, instead of in the serum. Furthermore, the literature strongly suggests that the subpopulation of mononuclear peripheral blood cells carrying oncofetal antigen is independent of the serum 'level of oncofetal antigen, but is related to the presence of a predisposition to malignancy.
A number of researchers have focused upon ferritin and its possible relationship to normal and diseased states in man. Ferritin is a normal iron storage protein with a molecular weight of a~out 450,000. Ferritin in actuality is composed of a plurality of subunits, designated as isoferritins i86~

and being of about 18,000 molecular weiqht each, which are noncovalently bound to one a~other to constitute the ferritin molecule. Separation of at least some isoferritins ~rom one another can be accomplished by procedures which take advantage of differences in isoelectric points. A particularly interesting range of isoferritins are those having isoelectric points within about 4.7 to 5.6. see ~lpert et al., "Carcino-Foetal Human Liver Ferritins", Nature, 242, 194 (March 16, 1973~.
Alpha-2-H-globulin, an acidic isoferritin found primarily in the fetal liver (classified as an oncofetal antigen), also is found in adults with various malignancies such as hepatoma and teratoma. Immunologically, the various isoferritins have not been distinguished from normal human ferritins, although it is contemplated that through immunological antibody-producing techniques, it will be possible to produce antibody for specific isoferritins, such as the above-mentioned alpha-2-H-globulin, and isoferritins having isoelectric points within the 4.7 to 5.6 range.
It is difficult to use serum ferritin levels as an accurate diagnostic tool in vie~ of the a~ove-descri~ed naturally occurring biological phenomena, namely, the presence of ferritins in the normal blood serum over a fairly wide concentration range, inability at this time to immunologically distinguish between different isoferritins, and, in addition, the presence of elevated serum ferritin levels in some non-malignant disease states, such as a nonmalignant anemia. In one study, about 70 percent of the patients tested and having malignancy exhibited elevated serum ferritin levels.
It also has been reported that in early ~reast cancer serum ferritin levels overlap the normal serum ferriti~ concentration range, although extremely elevated levels may be found in women with recurrent or metastatic carcinoma of the breast.
Jacobs et al., "Serum Ferritin Concentration in Early Breast Cancer", Br. Cancer, 34(3), 286 ~1976) and "Serum Ferritin Concentration in Untreated Hodgkin's Disease", ibid, 34(3), 162 (1976); Sarcinone et al., "Increased Ferritin Synthesis and Release by Hodgkin's Disease in Peripheral Blood Lymphocytes", Int. J. Cancer, 20, 339 (1977) and Niitsu et al., "Radio-immunoassay of Serum Ferritin in Patients wit~, Malignancy", Ann. N. Y. Acad. Sci., 259, 450 (1975).
Some of the work of Moroz et al. has ~een referred to hexeinbefore. In addition, Moroz et al. determined that apoferritln acts as a blocking substance Eor T lymphocytes in Hodgkin's lymphoma, thereby reduclng the tendency of such cells to spontaneously form rosettes with sheep red ~lood cells [Moroz et al., "Ferritin on the Surface of Lymphocytes in Hodgkin's Disease Patients - a Possi~le Blocking Substance Removed by Levamisole", Clin. ~. Immunol., 29, 30 (1977)].
This may have led to their work investigating the quantifying of ferritin-bearing lymphocytes a~ indicative of breast cancer, using the cytotoxicity test described hereinbefore.
One of the more pronounced drawbacks of the cytotoxicity test is the need to separately evaluate the effects of each reagent used on the patient's cells.
The present invention provides a rapid, noninvasive method to detect and quantify ferritin-bearing PMC's. In addition to experimental laboratory usage for ferritin detection per se, the testing procedures of this invention have been shown to be of value in detecting and evaluating treatment of breast cancer. One of the major advantages of the present invention is the development and use of human mononuclear white blood cells grown in tissue culture to which soluble human ferritin has been chemically bound.
Because of the marked stability of this reagent in retaining the ferritin on the cell surfaces and because of khe similar surface characteristics of these cells grown in culture to patient PMC's, such reagent is suitable for use as quality control reference cells having a stable ferritin value which can be used to monitor daily assay validity, interassay variability, and the proper calibration of other reagents.
Descriptions of the three ~semiquantitative test~ or four (quantitative test~ reagents required in the present invention will now be set forth.

(1) Labeled Anti-Human Ferr_tin Anti~ody The ferritin antibody used in the studies described hereinbelow was monospecific anti-human ferritin antibody.
It is contemplated that an antibody of even greater specificity, that is, an isoferritin-specific antibody, could ~e prepared by substituting at least one isoferritin species or an isoelectric-determined range thereof for normal human ferritin in the procedure described below to prepare the la~eled anti-human ferritin antibody.
Rabbits are immunized with human ferritin using standard immunological procedures. The following afinity purification procedure then can be carried out to yield the monospecific antibody. Equivalent procedures will ~e apparent to the skilled artisan.
Arylamine controlled-pore (1350A) glass, 200 to 400 mesh particle size, is activated with glutaraldehyde (5 ml/gm glass). After washing to remove any non-reacted ~61~

glutaraldehyde, l mg human fexritin per gram of glass is gently admixed with the activated glass for about 2 hours at room temperature. Thereafter, the glass is successively washed with 1 M sodium chloride, 0.1 M glycine, and then 3 to 5 washes with 0.1 M phosphate buffer. The arylamine glass is readily prepared by known procedures; see, for example, Weetall, Science, 166, 615 (1969), and the references cited therein.
The ferritin-coupled glass is then gently admixed with the collected ferritin antiserum, about 1 ml antiserum per gram of ferritin-coupled glass. After incubation at 22C
overnight, the glass is washed successively with 0.1 M
sodium carbonate, 0~15 M sodiu~ chloride, p~ 8.0; 0.1 M
sodium acetate; 0.15 M sodium chloride, pH 5.0; and two washes of distilled water adjusted to pH 4.4 with l M hydrochloric acid. The monospecific antibody is eluted with distilled water, adjusted to pH 2.2. The ferritin-coupled glass then is filtered into 0.3 M sodium phosphate, pH 7.35, and the filtrate concentrated by membrane ultrafiltration. Antibody can be stored in a solution of 0.1 M sodium phosphate with 0.02 percent azide.
The antibody can be coupled to any of the conventional labels, such as radioactive element, enzyme, or fluorescing agent. In the studies described herein, the monospecific anti-human ferritin antibody was radiolabeled with Il25 in accordance with the well-known chloramine-T procedure.

(2) Ferritin Positive Mononuclear Cells This reagent is used as a positive quality control to monitor the stabil}~y and adjust the concentration of reagent 1, above, to validate the assay pxocedure, and to assure that test sample results generated on a day-to-day basis (interassay variability) are correct.

The mononuclear cells used in prepariny this reagent are usually culture-grown lymphocytes, so as to avoid fluctuations of the ferritin level which can occur with naturally occurring human lymphocytes. The work descri~ed herein used acute human lymphoblastic leukemia cells (CCL-119, ATCC); however, other cultured lymphocytes can ~e substituted therefor, such as murine lymphocytes, CCL-86 or CCL-126 lines, or even normal subject peripheral blood lymphocytes if adequate screening is employed. Such mononuclear cells, then, are peripheral blood mononuclear-type cells.
Such cells need not be naturally occurring peripheral blood mononuclear cells as long as such cells exhibit me~brane surface characteristics which are similar to naturally occurring peripheral blood mononuclear cells.
In these studies, the positive quality control reagent cells were prepared using acute lymphoblastic leukemia cells (CCl-ll9) grown in RPMI-1640 medium (Grand Island Biological Co., Grand Island, N.Y.) supplemented with 10 percent fetal calf serum (FCS, Grand Island Biological Co.). These cells can be grown in continuous culture using spinner techniques and are suitable for large-scale pr~paration. The CCl-ll9 lymphocytes were harvested by centrifugation at 200 to 300xg for 15 minutes and washed two times in phosphate-buffered saline (PBS, 0.01 M phosphate, 0.15 M sodium chloride, pH
7.4). The cells were then reconstituted to a concentration of 2 to 3 x 107 cells/ml in PBS. Cell concentration was measured by counting in a hemocytometer. Sufficient glutaraldehyd solution of 25 percent (w/w) was added to make the final glutaraldehyde concentration app~oximately 1 percent.
Activation at cell densities above this concentration resulted -- 1~ --in extensive cell aggregation. Cells were incubated with continuous rocking for 30 minutes at 22C follo~7ed ~y harvesting and washing b~ centrifugation. Purified human ferritin obtained from JBL Chemical Company, San Luis Obispo, California was diluted with PBS to a final concentration of 200 to 300 ~g/ml and added to the cell pellet in an amount sufficient to give a final cell concentration of 2 to 3 x 107 cells/ml.
The cells were incuba~ed with rocking for 30 minutes at 22C and then at 4C for further incu~ation for an additional 16 to 18 hours. The cells were then washed as described above and suspended in PBS containing 1 percent bovine serum albumin (BSA) and 0.2 percent sodium azide. The average recovery of derivatized cells from the starting population was 40 to 50 percent. These cells were stored at 4C and showed only a 10% decrease in activity after one year when estimated by specific binding of I125-labeled anti-human ferritin antibody. The approximate loading of ferritin on these cells was determined to be 5 to 20 ng per 2.5 x 105 cells.
(3) Purified Human Ferritin This reagent is commercially available and is used in both the semiquantitative and quantitative test modes. In the semiquanti~ative mode it is used in high concentration to competitively block labeled anti-ferritin antibody binding to ferritin present on the patient's lymphocytes in order to assess specific labeled antibody binding. In the quantitative mode it is also used in variou~ concentrations in conjunction with the immunoradiometric assay (IRMA) in order to develop a standard curve from which absolute values of PMC-bound ferritin may be determined.

~3~

(4) Immobilized Anti-Human Ferritin Antiserum This reagent is used in conjunction with reagents l and 3 to generate a standard curve from which values for PMC-bound ferritin may be extrapolated and consists of anti-human ferritin antiserum covalently coupled to a substrate such as controlled-pore glass.
In the tests described herein, this reagent was prepared using l ~ particle size, 550A arylamine controlled-pore glass.
This glass was first activated at 0C with 2 N hydrochloric acid and sodium nitrite. After 20 minutes, the glass was washed extensively with distilled water. One ml of whole anti-human ferritin antiserum per gram of glass then was added. The pH was adjusted to 8.4 and the reaction allowed to proceed overnight at 4C. The immobilized antiserum was washed extensively in 0.l M phosphate buffer, pH 7.0, and stored at 4C at a final concentration of 20 mg glass/ml in 0.0l M phosphate buffer, pH 7.0, containing 0.l percent BSA.
Detailed descriptions of test modes using the above-~0 described reagents to carry out the assays of this inventionwill be-disclosed hereinbelow. In general, the first three of the reagents listed above ~labeled anti-human ferritin antibody, ferritin positive PMC's, and purified human ferritin) are used to perform a semiquantitative assay while all four reagents (including the immobilized anti-human ferritin antiserum) are used for a quantitative determination of PMC-bound ferritin.
Regardless of whether quantitative or semiquantitative tests are desired, the test component will be a sample of the patient's peripheral mononuclear blood cells, particularly the lymphocytes. Although it is believed that with careful analytical technic~ues, a sample of all of the patient's white blood cells can be used (such as a buffy coat), it is recommended at this time that the mononuclear peripheral blood cells be separated from other blood components as a fixst step. Techniques known to the skilled artisan are employed to separate the mononuclear peripheral blood cells from other blood components, such as a mononuclear cell separator ~Luderer et al. t "Rapid, Quantitative Human L~nphocyte Separation and Purification in a Closed System", Molecular Immunol., _, 621 (1979)] or the Ficoll-Hypaque technique of Boyum [Scand .J. Immunol., 5 (suppl. 5), 9 ~1976)]. Conventional washing such as by centrifugation and adjustment to a standardized cell number, such as by counting in a hemocytometer or automatic blood cell counter, will complete the preparatory phase of the assay. Although not necessary, the lymphocyte subfraction of the mononuclear peripheral blood cells can be obtained by a known procedure and used as the test sample.
In the semiquantitative mode, an aliquot of the patient's PMC's is admixed with an aliquot of the labeled anti-human ferritin antibody, and, preferably, an aliquot of buffer.
Separately, the patient's PMC's, the labeled anti-human ferritin antibody, and an aliquot of purified human ferritin, the latter also taking the place of the ~uffer where used, are admixed to prov1de the usual nonspecific binding control, since the human ferritin is added in high concentration to competitively block any specific binding of the labeled antibody to the test cells. The second part of the test involves dupllcating the two separate steps above, but with the substitution of the ferritin positive PMC's for the test cells to monitor the functio~ality of the assay system and the stability of the labeled antibody. As an optional step, 3~

if desired as a further check, the second part of the test can be duplicated, but with the substitution of a control negative cell for the ferritin positive PMC's. The control negative cell is one carrying a protein which will not react with the labeled antibody, such as CCL-ll9 lymphocytes carrying bovine serum albumin.
After a suitable incubation period, the cells are washed with PBS and centrifuged. The supernatant is recovered and the pelleted cells assessed for bound labelO If I1~5-i0 labeled antibody is used, pellets are counted in a gammaspectrometer. Semiquantitative results reported as percent specific binding are determined by subtracting counts per minute (cpm) bound to cells in the presence of excess soluble human ferritin (X) from cpm bound in the absence of added excess human ferritin (B), divided by the cpm of the total I125-labeled antibody added to the assay system (T) minus cpm bound to cells in the presence of added excess human ferritin (X). Added excess human ferritin is defined as the amount of soluble ferritin added to the assay system that will completely block specific labeled anti~ody ~inding to the cell.
Thus, % Specific Binding = B - X x 100 Generally, duplicate analyses are tested both with ~X) and without (B) excess ferritin and the averages or means of these values are used in the specific binding calculation.
Examples of typical values are set forth in Table 1.

Tabl~ 1 CPM Mean CPM
B 15,650 14,985 15,317 X 4,20~
4,500 4,350 T 100,500 99,400 99,950 Consequently, % Specific Binding = 19,317 4'355- x 100 ~ 11.5%
As already mentioned, the ferritin-positive reference control PMC's are used to monitor the validity of the patient test result generated in any individual assay as these cells have an invariant % specific binding value. Such cells are used in parallel and at the same time patient samples are assayed. If the % specific bindlng result generated with the ferritin-positive PMC's is below or above the expected or known value within the limits of some approved statistical evaluation system, such as + two standard deviations of the mean (~ 2S.D.), then one can assume that the test results for that individual patient assay are invalid, due to, for example, experimental errors or deteriorating reagents.
Likewise, a comparison of % specific binding values for ferritin-positive PMC's obtained in conjunction with patient assays carried out at different times allows the user to accept such patient test data as valid. Test values from different patients then can be accumulated and compared over time.
An accumulated quality control chart of ferritin-positive PMC % specific binding values, o~tained as described .

6~

above in conjunction with Example 1, i9 set forth in Figure 1 which demonstrates that three of 26 assays were invalid (assays represented by points A, B, and C).
Generally, speciic binding is evaluated with at l~ast two different cell concentrations. Results are reported as specific cpm or percent bound to peripheral blood mononuclear cells per specified ~umber of cells.
In the quantitative assay, patient and quality control PMC's are assayed as above, but a standard curve also is prepared by incubating increasing amounts of human ferritin with a fixed concentration of immo~ilized anti-human ferritin antiserum (IMA). In this system, IMA is first mixed with buffer alone or varying concentrations of human ferritin.
At a fixed time interval or concurrently, la~eled antibody is added and the incubation allowed to proceed to a final time interval equal to that used in analyzing he test cells. The IMA is washed with buffer and pelleted by centrifugation. The supernatants are decanted and the pellets assessed for radioactivity. Specific binding is determined for each human ferritin concentration in a manner similar to that described above except that cpm bound to IMA
in the presence of buffer alone (Bo) is substracted from both the total cpm added (T) and cpm bound with increasing amounts of added soluble human ferritin (BFJ. A standard curve is prepared by semilogarithmic plotting of specific binding (linear scale) versus the amount of ferritin added (log scale). By extrapolating the specific binding by patient cells to the log scale, a quantitative measure of ferritin bound to a patient's peripheral blood mononuclear cells is made.

3~

As with the analysis of patient cells and ferritin-positive reference control cells, it ls preferred that duplicate analyses be conducted at each level of ferritin employed to generate the standard curve.
For the quantitative embodiment standard curve, the following equation would be used to determine the ~ specific binding:

% Specific Binding = BF ~ Bo 0 x 1 0 T - B
Examples of typical data obtained for the preparation of a standard curve are sur~marized in Ta~le 2.
Table 2 Standard Curve Data for Determination of PMC-Bound Ferritin B BF Mean, ~ Specific Ferritin,ng CP~ CPM CPM Binding 0 4,200 4,125 - 4,162 0.1 - 5,210 - 4,695 4,952 0.8 0.25 - 5,880

- 5,672 5,776 1.7 0.5 - 10,241 - 9,844 10,042 6.1 1.0 - 13~010 - 12,875 12,g42 9.2 2.5 - 19,921 - 19,8~0 19,860 16.3 5.0 - 37,374 ~ - 36,900 37,137 34.4 T, CPM: 100,500 and 99,400 Mean T, CPM: 99,950 i8~i~

Using the data from Tab]e 2, a standard curve was constructed as shown in Figure 2.
By direct extrapolation of the semiquantitative %
binding values for patient cells and the ferritin-positive PMC's by means of the standard curve of Figure 2, a quantitative value, expressed in nanograms ferritin per unit number of cells, is derived. Some representative values for ~oth patient cells and the ferritin-positive P~C's are set forth in Table 3.
Table 3 Sample ~ Specific Binding ng Ferritin/106 Cells Ferritin-Positive PMC's 14.0 1.8 Patient A 18.0 2.65 Patient B 3.0 0.3 Patient C 8.0 0.8 In order for patient values to remain valid from lot to lot of reagents, new batches of reagents, specifically reagent (1), labeled anti-human ferritin antibody, and reagent (4), immobilized anti-human erritin antiserum, must be calibrated and concentrations adjusted so that they give the expected values when used with the ferritin-positive PMC's. Similarly, new preparations of reagent (2), ferritin-positive mononuclear cells, are adjusted to the appropriate concentration by comparing % specific ~inding values to the known values of older preparations of ferritin-positive PMC's. Thus, the calibration of new lots or preparations of reagents is yet another use for the ferritin-positive reference control c~lls.
At the present time, it is recommended that the standard curve be developed at the time of patient testing, using the same human ferritin to be employed in the patient testing.

Ç~l As standaxdization and shelf life, i.e., stability, are improved, standard curve generation may be carried out at infrequent intervals, realizing of course that where radioactive isotopes are employed, the standard curve should always be freshly generated.
of course, in each of the above modes, the identical calculations are run using the results obtained in the second part of the assay with the ferritin-positive lymphocytes to validate the reagents and test results.
As will be illustrated more fully ~ereinbelow, it is believed that the testing procedure described herein can be used as an effective tool in the early diagnosis of cancer, to monitor the effectiveness of treatment, and also to obtain evidence of disease recurrence. Normal healthy individuals will exhibit essentially no difference, or a small difference, in bound label measured in the presence and absence of the purified human ferritin, while the PMC's from patients with early (preclinical) cancer or advanced cancer show increased bound label in the system in which the human ferritin has been omitted.
A more detailed description of an assay procedure using the reagents specifically described hereinbefore is presented below, as exemplifying the assay techniques of this invention.
Variations in operating procedure will be obvious to the skilled artisan.
Anticoagulated blood specimens were su~jected to the technique of Boyum mentioned earlier for isolation of PMC's.
This involved a 1:2 dilution of the blood with physiologic phosphate buffered saline, followed by gently layering the blood onto a solution of Ficoll-Hypaque. Following centriguation at 400 xg for 30 minutes, the P~C's banded at a defined interface and were removed with a Pasteur pipette, washed two times in PBS by yentle centrifuyation (200 xg, lO minutes), and suspended in PBS-0.1% BSA-0.024 sodium azide or RPMI
culture medium at a concentration of 1 x 107 cells/ml. The control positive cells (PMC's bearing ~erritinj, mixed with control negative cells in a ratio of l part positive cells to 9 parts negative cells, were also suspended at this concentration.
Four 12 x 75 mm plastic test tubes were allocated for each sample tested at each cell concentration. Reactions were set up in a 4C water bath. 100 ~l of cells was added to each test tube. The first two tubes of each series contained an additional lO0 ~1 of the PBS-BSA-aZide solution while the second two tubes contained 100 ~l of the human ferritin in PBS-BSA-aZide added at a concentration of lO to 50 ~g/ml (excess HF). These concentrations were sufficient to completely block the specific binding of I125-labeled antibody to pure posltive control cells at a dose of 3 x 106 cells. After addition of 10 ~ 25-labeled antibody (50,000 to 100,-000 cpm), the tubes were transferred to a shaker bath and incubated overnight at 4C (16 to l~ hours) to allow equilibrium binding to be reached. After incubation, the cells were washed with 3 ml of PBS and centrifuged at 3,000 rpm for 15 minutes. Supernatants were aspirated and cell pellets counted for bound radioactivity. Duplicates were averaged and values for specific binding calculated as described above.
The nearly complete blocking of binding by the human ferritin establishes the specificity of the labeled anti-- ferritin antibody for the ferritin bound to the test cells and the low level of nonspecific adsorp~ion to cell surfaces.

~ . . .

The extremely good sensitivity of the assay of this invention is discussed below.
The lower limit of sensitivity for detection of ferritin on PMC's was determined by competitive binding using increasing concentrations of human ferritin, a fixed concentration of radiolabeled anti-ferritin antibody and 2.5 x 105 positive control cells (CCl-ll9 bearing a defined amount of ferritin3.
A lower limit of detection of 1.0 x 10 10 g or 2.2 x 10-7 nanomoles was found. This limit was based on the minimal amount of human ferritin which would cause a reproducible inhibition of specific binding of radiolabeled anti-ferritin antibody to positive control cells.
The minimum detectable number of ferritin-bearing cells in a population of negative cells was studied using control positive cells mixed with negative cells at a final cell population of 2.S x 105 cells. At this cell density, 2,500 cells could be detected at a discriminating level of 1.0 percent specific binding. This is significant since the literature has suggested that 16.6 percent ferritin positive cells of the total lymphocyte population is indicative of breast cancer. lMoroz et al., Cancer Immunol. Immunother., _, 101 (1977)].
In the quantitative assay, a typical method to develop the standard curve is as follows:
To duplicate 12 x 75 mm plastic tubes there is added 100 ~1 of the anti-human ferritin antiserum (IMA) previously diluted 1:10 with PBS-BSA azide solution from a stock IMA of 20 mg glass per ml. To each set of duplicates there is added, in 100 ~1, buffer alone or human ferritin diluted to various amounts with buffer. This assay is set up in a 4DC

. .

water bath. Radiolabeled anti-ferritin antibody (190 ~1) is added to each tube containing IMA and incubation allowed to proceed overnight at 4C in a shaker bath. The tubes are washed and centrifuged in a fashion identical to patient or positive reference control samples.
When specific binding for positive control or patient cells is extrapolated from the standard curve, the absolute amount of ferritin bound can be determined.
Table 4 sets forth testing results with a num~er of patients using the assay of this invention in a semiquantitative mode. The percent accuracy was developed from other diagnostic procedures, such as biopsy, which also determined the patient's category.

Table 4 ~umber~ Number*
Number Positive Negative Category Patients Patients Patients ~ Accuracy Metastatic Disease 23 12 11 52 (active) 20 Metastatic Disease 25 0 25 100 (remission) Pre-operative 16 4 12 25 (primary) Benign Breast 36 1 35 97 Hodgkin's Disease 3 2 1 67 *based on the PMC-ferritin assay procedure.

In evaluating the results ta~ulated in Table 4, it is apparent that the test of this invention has a unique value in screening out nonmalignant patients and in indicating the effectiveness of treatment.

1~36~

The data in Table 4 were obtained in the laboratories of Corning Glass Works, using patient samples which had been shipped to Corning, New York. Consequently, the patient samples typically were at least 24 hours old when assayed.
A more recent clinical patient study then was carried out at Downstate Medical Cen~er, Bxooklyn, New York. The assays were conducted in the Center's own laboratories on freshly drawn samples. It should be noted that ~oth studies were done blind, i.e., without knowledge of patient status prior to the assays. The results of this second study are summarized in Table 5.

Table 5 Number* Num~er*
Number Positive Negative Category Patients Patients Patients % Accurac~
Non Breast Cancer Normal 12 l ll 92 Benign Breast 36 3 33 92 Other Diseases 52 10 42 81 Breast Cancer _ Pre-Operative Primary Breast 10 5 5 50 Cancer-Stages O-III

Metastatic Breast Cancer 10 10 0 100 Stage IV

Post-Operative Primary Breast Cancer-No 13 l 12 92 Evidence of Metastases It should be noted that the efforts of Moroz et al.
were directed to the detection of the early stages of cancer, e.g., Stage I of breast cancer [see, for example, Moroz et al., The New ~ Journal of Medicine, 276, No. 20, 1172 (1977), described hereinbefore]. Significantly and unexpectedly, the present invention identifies more advanced cases, e.g., Stage IV of breast cancer. Indeed, it is seen from Ta~les 4 and 5 that the ferritin levels on lymphocyte surfaces as detected by the present invention are not sufficiently specific to serve as a marker for the earlier stages of breast cancer, i.e., Stages O-III, inclusive.
Table 6 sets forth actual mean values of ferritin-positive lymphocytes using the semiquantitative procedure at 4C with 16-18 hour incubation using radiolabeled (I-125) anti-human ferritin antibody. These values, although based on a limited number of patients, show the relative values between malignant and non-malignant states.
Table 6 1. Normal women bind a mean of 4.0 percent.
2. Benign breast diseases, including galactorrhea, fibrocystic disease and fibroadenoma, bind a mean of 3.g percent 3. Patients with various metabolic diseases bind a mean of 4.5 percent.
4. Women who have undergone mastectomy and show no evidence of recurrent disease or are responsive to radiation and chemotherapy have reduced binding values (mean 4.0 percent) relative to those women with metastatic disease who do not respond to therapy (mean 10.5 percent).

5. Half of all women with brea~t cancer, prior to undergoiny surgery, show elevated binding levels (6.8 to 14.3 percent, mean 10.5 percent).
~ ariations of the invention will be apparent to the skilled artisan. For example, various coupling agents, such as carbodiimide or diisothiocyanates could be used in place of the glutaraldehyde; other immobili7ation substrates such as starch or gel beads could be employed in place of the glass; and other conventional standard curve development procedures could be employed.

Claims (20)

WE CLAIM:

1. A method for detecting ferritin on peripheral blood mononuclear cells which comprises:
(A) carrying out a first test by admixing an aliquot of a white blood cell sample comprising peripheral blood mononuclear cells with an aliquot of labeled anti-ferritin antibody, removing the antibody which does not attach to said cells and determining the relative amount of antibody attached to said cells through the use of said label;
(B) carrying out a second test by admixing an aliquot of said white blood cell sample with an aliquot of labeled anti-human ferritin antibody and a competitive binding inhibiting amount of ferritin, removing the antibody which is not attached to said cells and determining the relative amount of antibody attached to said cells through the use of said label;
(C) determining the relative amount of ferritin carried by said cells from the values obtained in steps (A) and (B);
(D) repeating steps (A)-(C) but with substituting peripheral blood mononuclear-type cells carrying a known quantity of ferritin for the blood sample; and (E) checking the validity of the method and of the labeled antibody reagent from the results of step (D).

2. The method of claim 1 wherein a quantitative detection method is carried out which comprises, in addition, to said steps (A)-(E):

(F) preparing a standard curve by admixing increasing amounts of ferritin with a defined amount of immobilized antibody and a defined amount of labeled antibody;
(G) determining specific binding fox each ferritin concentration through the use of said label; and (H) comparing the result obtained in step (C) with said standard curve to determine the amount of ferritin corresponding to said value obtained in step (C).

3. The method of claim 1 wherein the peripheral blood mononuclear cells are separated from said white blood cell sample and used to carry out said method.

4. The method of claim 2 wherein the peripheral blood mononuclear cells are separated from said white blood cell sample and used to carry out said method.

5. The method of claim 1 wherein the peripheral blood lymphocytes are separated from said white blood cell sample and used to carry out said method.

6. The method of claim 2 wherein the peripheral blood lymphocytes are separated from said white blood cell sample and used to carry out said method.

7. The method of claim 3 wherein the label is radioactive, enzymatic, or fluorescent.

8. The method of claim 4 wherein the label is radioactive, enzymatic, or fluorescent.

9. A method for detecting the presence of cancer in a patient which comprises carrying out the method of claim 1 and then comparing the value obtained in step (C) with the range of values obtained with healthy, nonmalignant individuals.

10. A method for detecting the presence of cancer in a patient which comprises carrying out the method of claim 2 and then comparing the value obtained in step (H) with the range of values obtained with healthy, nonmalignant individuals.

11. A method for monitoring the effectiveness of cancer therapy which comprises carrying out the method of claim 1 and then comparing the value obtained in step (C) with the range of values obtained with healthy, nonmalignant individuals.

12. A method for monitoring the effectiveness of cancer therapy which comprises carrying out the method of claim 2 and then comparing the value obtained in step (H) with the range of values obtained with healthy, nonmalignant individuals.

13. A method for the long-term monitoring of post-operative cancer patients for the early detection of the recurrence of cancer which comprises carrying out the method of claim 1 and then comparing the value obtained in step (C) with the range of values obtained with healthy, nonmalignant individuals.

14. A method for the long-term monitoring of post-operative cancer patients for the early detection of the recurrence of cancer which comprises carrying out the method of claim 2 and then comparing the value obtained in step (H) with the range of values obtained with healthy, nonmalignant individuals.

15. An immunochemical test kit useful for detecting ferritin on peripheral blood mononuclear cells, said kit comprising as separate reagents:
(1) labeled anti-human ferritin antibody for said ferritin;
(2) said ferritin; and (3) peripheral blood mononuclear-type cells carrying a known quantity of ferritin.

16. The diagnostic kit of claim 15 which, in addition, includes said antibody immobilized on a substrate.

17. The kit of claim 15 or claim 16 wherein the label is radioactive, enzymatic, or fluorescent.

18. The kit of claim 15 or claim 16 wherein reagent (3) is cultured lymphocytes carrying ferritin on the surface thereof.

19. The method of claim 1 wherein, in addition to said steps (A)-(E), the following step (F) is carried out:
(F) repeating steps (A)-(E) but with substituting peripheral blood mononuclear-type cells carrying a protein substance known to produce a negative reaction with the anti-ferritin antibody.

20. The method of claim 2 wherein, in addition to said steps (A)-(H), the following step (I) is carried out:
(I) repeating steps (A)-(H) but with substituting peripheral blood mononuclear-type cells carrying a protein substance known to produce a negative reaction with the anti-ferritin antibody.