RU2626603C2 - Method for determination of risk of breast oncological diseases recurrence - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention is designed to determine the risk of breast cancer recurrence. The expression level of ELOVL5, IGFBP6, TXNDC9 marker genes is measured in the patient tumour tissue sample. The risk of recurrence is assessed by comparing the expression levels of marker genes to predetermined threshold values.

EFFECT: invention provides a simple, sensitive, reliable way of disease-free survival prediction.

14 cl, 6 dwg, 8 tbl, 3 ex

Description

FIELD OF THE INVENTION

The present invention relates to medicine, in particular to oncology and molecular biology, and can be used to assess the risk of breast cancer recurrence after surgical removal of the tumor.

The invention is based on the assessment of the expression level of 3 genes, the mRNA of which can be obtained from a patient’s tumor tissue sample. Moreover, the expression level can be measured both by known methods and by the method described in the present invention.

State of the art

Breast cancer (BC) is a malignant tumor of the glandular tissue of the breast. In the world, this is the most common form of cancer among women, affecting during life from 1:13 to 1: 9 women aged 13 to 90 years. It is also the second most common cancer after lung cancer in the general population (including the male population). The number of cases of breast cancer in developed countries has increased dramatically since the 1970s. For this phenomenon, the changed lifestyle of the population of developed countries is partially responsible (in particular, the fact that families have fewer children and breastfeeding periods have decreased) [Merabishvili V.M. (2011). Breast cancer: incidence, mortality, survival (population study). Oncology issues, 5, 609-615].

A new step in the treatment of cancer is the so-called “individual treatment selection”. It is known that chemotherapy occupies an important place in the treatment regimen for breast cancer patients. Previously, patients with tumor types similar in cell structure underwent chemotherapy according to the same scheme. Detailed studies of breast cancer transcript revealed that each tumor has its own characteristics that are unique to it. That is, with similar clinical manifestations and cell structure, different tumors can manifest themselves in different ways, including responding differently to ongoing drug therapy. This is determined by the molecular genetic characteristics of the tumor, including the peculiarities of the expression of certain genes in tumor cells.

Currently, several molecular genetic classes of breast cancer are distinguished. This classification involves the division of primary tumors and metastases depending on the presence and density on the surface of cancer cells of receptors for biologically active regulatory substances (hormones). Immunohistochemical typing allows you to select a group of patients who are shown treatment with antitumor endocrine drugs (drugs that block certain receptors of cancer cells, which leads to a decrease in their proliferative potential).

The problems of metastasis and relapse of breast cancer are some of the most difficult to predict in cancer practice and require the development of an integrated approach for early detection of the risk of metastasis.

Various methods for constructing such forecasts are known from the prior art, including those based on an assessment of the level of gene mRNA expression. In particular, when building forecasts, it is known to use the method based on real-time polymerase chain reaction (PCR-RV), which differs from traditional qualitative and semi-quantitative methods in that it allows you to quickly and accurately determine the number of copies of DNA or cDNA in a wide range of concentrations [ Aerts, J., Wynendaele, W., Paridaens, R., Christiaens, MR, Van Den Bogaert, W., Van Oosterom, A.T., & Vandekerckhove, F. (2001). A real-time quantitative reverse transcriptase polymerase chain reaction (RT-PCR) to detect breast carcinoma cells in peripheral blood. Annals of oncology official journal of the European Society for Medical Oncology ESMO (Vol. 12, pp. 39-46)]. The method is widely used in the world, both for scientific and clinical purposes, due to its high productivity: usually 96 reactions are carried out in one experiment, and in the latest models of devices - 384 reactions. In contrast to the usual polymerase chain reaction, when the amount of products is determined at the final stage of the reaction, PCR-PB allows you to monitor the accumulation of products in the exponential phase of the reaction. In this system, along with primers, a probe (probe) labeled with a fluorescent dye is used. A probe is an oligonucleotide that hybridizes to a DNA / cDNA sequence between two flanking primers. PCR-RV consists of two main stages: 1) hybridization of the studied matrix with the probe; and 2) PCR catalyzed by TaqDHK polymerase, which has 5'-exonuclease activity, splitting the probe, which leads to the appearance and increase of the fluorescent signal. Usually the reaction lasts about 2 hours (40 cycles), in new generation devices - only 40 minutes. Analysis results become available immediately after completion of the reaction.

Existing test systems allow you to clarify the diagnosis, predict the risk of metastasis and relapse, and the effectiveness of different types of therapy. However, Thomassen et al. (Thomassen M, Tan Q, Eiriksdottir F, Bak M, Cold S, Kruse TA. Prediction of metastasis from low-malignant breast cancer by gene expression profiling. Int J Cancer. 2007 Mar 1; 120 (5): 1070-5) It was shown that sets of genes in profiles practically do not intersect. In addition, they are useful for the analysis of certain types of breast cancer (estrogen receptor-sensitive LNN (ER +)) or for predicting the development of the disease in patients taking adjuvant therapy.

Group van't Veer et al. (van't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AAM, Mao M, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature [Internet]. Nature Publishing Group; 2002 Jan 31; 415 (6871): 530-6) used a Rosetta chip with 60-dimensional oligonucleotides and developed a 70-gene profile, with which it is possible to predict the formation of metastases over five years in patients with breast cancer type LNN (lymph node-negative) with greater accuracy than using the classic clinical and pathological approach. A proposed gene profile is described in US Pat. No. 7,514,209 B2 (MammaPrint).

In patent application US 2011/0145176 A1 (PAM50), a method, a set of reagents and microarrays for diagnosing breast cancer and predicting the effectiveness of therapy based on measurements of expression levels of 50 marker genes are described.

A set of primers for 14 genes and reagents for determining their expression by PCR-RV, as well as an algorithm for interpreting the expression of these genes, are proposed in US Pat. No. 7,695,915 B2 for predicting the formation of metastases in the luminal form of breast cancer.

The method and kit of reagents for predicting the development of breast cancer using six sets of marker genes are described in patent application US 2010/0035240 A1. The method is based on the assessment of the level of expression of 135 genes by the method of quantitative polymerase chain reaction in real time in freshly frozen breast cancer samples. The method allows you to assess the overall ten-year survival, the likelihood of distant metastases and lymph nodes.

US 7,622,251 B2 (Oncotype DX ™) provides numerical molecular indicators for predicting the likelihood of a successful response to an antiestrogen agent (tamoxifen) and chemotherapy for patients with breast cancer (ER +, LNN) without adjuvant therapy.

The closest technical result achieved is the Oncotype DX Breast Cancer Assay diagnostic test, used to assess the risk of breast cancer recurrence. Using this test, you can get additional information and decide on further treatment. The results of the study determine the aggressiveness of the tumor, and also assesses the risk of relapse and the need for chemotherapeutic treatment. Investigating the expression of the 21st gene (16 marker genes - Ki-67, STK15, Survivin, Cyclin B1, MYBL2, Stromelysin 3, Cathepsin L2, ER, PR, Bcl-2, SCUBE2, GRB7, HER2, GSTM1, Cd68, BAG1 , and 5 normalization genes - Beta-actin, GAPDH, RPLPO, GUS, TFRC), experts establish the likelihood of relapse within 10 years. The indicator itself is expressed by a number from 0 to 100. In this case, fragments of tumor tissue stabilized in formalin and fixed in paraffin are used for the test. An additional stage of sample preparation leads to an increase in test time and additional costs. The disadvantages include the need for microdissection of a sample of tumor tissue, which requires the involvement of highly qualified specialists.

An analysis of patent documentation indicates that the search for effective solutions aimed at making forecasts of overall and relapse-free survival, taking into account the individual molecular genetic characteristics of breast cancer, is an extremely urgent task.

Disclosure of invention

The objective of the proposed technical solution is the development of new diagnostic breast cancer markers that can reliably assess the risk of relapse after surgical treatment, as well as the development of a simple, sensitive, reliable method for assessing the risk of breast cancer relapse in clinical or outpatient medical institutions.

The claimed invention extends the arsenal of diagnostic tests by obtaining new diagnostic markers of breast cancer, allowing for comparable reliability and specificity to assess the risk of relapse after surgical treatment.

The problem is solved by a method for determining the risk of recurrence of breast cancer in a patient, including measuring the expression level (quantification of the representation) of marker genes ELOVL5, IGFBP6, TXNDC9 in a patient’s tumor tissue sample to obtain a risk of recurrence S _R and subsequent assessment of the risk of cancer recurrence breast diseases by comparing the obtained coefficient with predetermined threshold values.

The present invention in its first aspect relates to a method for determining the risk of recurrence of breast cancer in humans, including:

- obtaining a sample of tumor tissue;

- isolation and purification of mRNA from the sample;

- synthesis of single-stranded cDNA on mRNA using non-specific oligonucleotide primers with a random sequence of nucleotides;

- PCR of single-stranded cDNA with direct and reverse gene-specific primers and probes specific for the studied genes (marker genes) - ELOVL5, IGFBP6, TXNDC9, and with pairs of gene-specific primers and probes for normalization of PCR results, specific for FO 42 relative to , GUSB, HCFC1, RPLP0 and TPT1 with obtaining for each of the listed marker genes their unnormalized quantitative estimates of the representation of Q (g), their subsequent normalization and obtaining normalized quantitative estimates of the representation of Q _norm (g), you by calculating the risk factor for relapse (S _R ) according to the formula

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = - 0.72, w (IGFBP6 ) = - 0.65, w (TXNDC9) = 0.55, w ₀ = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized quantitative estimates of gene representation, and an assessment of the level of risk by comparing the risk factor for relapse with preset thresholds:

when S _{R is} greater than 0.2, a conclusion is made that there is a high risk of recurrence, when S _{R is} less than -0.2, a conclusion is also made that there is a low risk of recurrence, when S _{R is} obtained in the range from -0.2 to 0 , 2 - consider the risk uncertain.

Preferably, direct and reverse gene-specific primers with the sequences SEQ ID NO: 1-3 and SEQ ID NO: 4-6, respectively, and probes with the sequences SEQ ID NO: 7-9 are used for PCR.

It is preferable to normalize the results of PCR using pairs of gene-specific primers with sequences of SEQ ID NO: 10-19 and probes with sequences of SEQ ID NO: 20-24.

In another aspect, the invention relates to a kit of reagents for evaluating the expression of marker genes during PCR, including at least 3 pairs of primers for genes with the sequences SEQ ID NO: 1-3 and SEQ ID NO: 4-6 - direct and reverse , respectively, and probes with sequences of SEQ ID NO: 7-9, specific for the marker genes ELOVL5, IGFBP6, TXNDC9, presented in lyophilized form or in aqueous solution.

The invention also relates to a tablet with a set of reagents for PCR reaction in determining the risk of a relapse of breast cancer in humans, containing direct and reverse gene-specific oligonucleotide primers applied to the wells of a microplate with the sequences SEQ ID NO: 1-3 and SEQ ID NO: 4 -6, respectively, in lyophilized form (direct and reverse) and fluorescent probes with sequences of SEQ ID NO: 7-9.

Preferably, the primers are contained in a concentration of 250 nMol.

Preferably, the fluorescent probe contains a FAM dye with excitation parameters at a wavelength of 485 nm and emission at a wavelength of 520 nm, taken at a concentration of 250 nM.

In another aspect, the invention relates to a kit of reagents for determining the risk of a recurrence of breast cancer in humans, including reagents for the isolation and purification of RNA from a sample of tumor tissue; reagents for the synthesis of single-stranded cDNA on mRNA using oligonucleotide primers; reagents for evaluating gene expression; buffers for PCR, with a kit or plate used as reagents for assessing the expression of marker genes.

The invention also relates to a method for assessing the level of expression (quantification of representation) of marker genes ELOVL5, IGFBP6, TXNDC9 obtained by PCR to determine the risk of a relapse of breast cancer in humans, characterized in that threshold cycles for marker genes and normalization genes, they give non-standardized quantitative estimates of the representation of the marker genes Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) and then normalized quantitative estimates of the representation of Q _{n orm} (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9), and deriving, based on these estimates of the risk of relapse, S _R, and then constructing a forecast of the risk of relapse, R, by comparing the risk coefficient S _R with predetermined threshold values. Moreover, PCR is carried out using primers and probes specific for marker genes, primers with the sequences SEQ ID NO: 1-3 and SEQ ID NO: 3-6, and probes with the sequences SEQ ID NO: 7-9, respectively as well as primers and probes specific for the FBXO42, GUSB, HCFC1, RPLP0 and TPT1 genes for normalizing PCR results, primers with sequences SEQ ID NO: 10-19 and probes with sequences SEQ ID NO: 20-24, respectively, coefficient S _R is determined by the formula:

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = - 0.72, w (IGFBP6 ) = - 0.65, w (TXNDC9) = 0.55, w ₀ = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized quantitative estimates of the representation of marker genes;

in this case, two threshold values (0.2 and -0.2) are used as threshold values for constructing the risk of breast cancer recurrence, and when S _{R is} greater than 0.2, a conclusion is made that there is a high risk of relapse, with S _R less than -0 , 2 - conclude that there is a low risk of relapse, when receiving S _R values in the range from -0.2 to 0.2 - they consider the risk to be uncertain.

In another aspect, the invention relates to a method for evaluating the expression (quantification of representation) of marker genes ELOVL5, IGFBP6, TXNDC9 by microarray analysis to determine the risk of breast cancer recurrence in a patient, characterized in that based on microarray analysis of the tumor tissue, an irregular quantitative assessment of the representation of marker genes Q (ELOVL5), Q (IGFBP6), Q (TXNDC9), then normalized quantitative estimates of the representation of marker genes Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9), on the basis of which the risk of relapse risk S _R is determined based on which the risk of relapse _{R is} predicted by comparing the obtained risk coefficient of relapse S _R with predetermined threshold values. Moreover, when using microchip analysis, the data are preprocessed using the RMA method together with the normalization set or using the fRMA method, and the coefficient S _R is determined by the formula:

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = - 0.72, w (IGFBP6 ) = - 0.65, w (TXNDC9) = 0.55, w ₀ = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized quantitative estimates of the representation of marker genes;

in this case, two threshold values (0.2 and -0.2) are used as threshold values for constructing the risk of breast cancer recurrence, and when S _{R is} greater than 0.2, a conclusion is made that there is a high risk of relapse, with S _R less than -0 , 2 - conclude that there is a low risk of relapse, when receiving S _R values in the range from -0.2 to 0.2 - they consider the risk to be uncertain.

In another aspect, the invention relates to information carriers containing programs recorded thereon that implement methods for quantitatively assessing the representation of marker genes ELOVL5, IGFBP6, TXNDC9 by PCR-PB and microarray analysis.

Brief Description of the Drawings

In FIG. 1 shows the sequences of forward and reverse gene-specific primers (sequences of SEQ ID NO: 1-3 and SEQ ID NO: 4-6).

In FIG. 2 shows the sequences of probes (SEQ ID NO: 7-9).

In FIG. 3 shows the sequences of gene-specific primers for normalizing the results of PCR (SEQ ID NO: 10-19) and probes (SEQ ID NO: 20-24).

In FIG. 4 shows a diagram of the wells of the Tablet.

In FIG. Figure 5 presents a) Kaplan-Meyer curves and b) ROC curve constructed from a validation sample and confirming the operability of the claimed method.

The implementation of the invention

The invention provides new molecular genetic markers for predicting relapse-free survival after surgical removal of the primary tumor of the luminal form of breast cancer, based on quantification of the representation of 3 mRNA genes: ELOVL5, IGFBP6, TXNDC9. This is a simple, reliable way to predict relapse-free survival at various stages of the development of malignant transformation, including the initial ones. The functional of expression of 3 genes allows a high degree of certainty to assess the risk of relapse in the first five years after surgical removal of the primary tumor of the luminal form of breast cancer.

As samples for analysis, biopsy samples can be used, including material obtained by needle biopsy of a breast tumor.

Methods of isolating total RNA from mammalian tissue samples are well known to specialists and, as a rule, include the following stages: grinding of tumor and normal tissue samples in liquid nitrogen, cell lysis, RNA isolation and purification, RNA quality control by electrophoresis in 1% agarose gel in the presence of ethidium bromide dye or in a denaturing polyacrylamide gel, as well as spectrophotometric determination of the amount of RNA. Homogenization of tissue pieces is carried out manually, grinding with a pestle in a ceramic mortar, or using mechanical homogenizers, for example, Omni Mixer or Micro-Dismembrator U from Sartorius (Germany). For the isolation of RNA, various protocols are used that are widely known to those skilled in the art. In classical RNA isolation methods, strong chaotropic agents are used, such as guanidinochloride and guanidinisothiocyanate, dissolving proteins, and sequential extraction with phenol and chloroform to denature and remove proteins [Sambrook J., Fritsch E.F., Maniatis T., Molecular Cloning. A laboratory Manual 2nd Edition ed. 1989, Cold Spring Harbor: CSHL Press]. The Trizol reagent method [GIBCO / Life Technologies] is also widely used. In order to prevent RNA degradation by RNases, inhibitors are used, such as a RNAsin inhibitor of placental or recombinant origin, or, for example, a vanadyl-riboside complex, a non-specific broad-spectrum RNAse inhibitor.

Rapid and high-quality RNA isolation is carried out using a number of commercially available kits (Clonogen, St. Petersburg; RNeasy kits (Qiagen); SV Total RNA Isolation System, Promega (USA), etc.). The use of various devices, for example, QuickGene-810 (Life Science, Japan), eliminates the work with aggressive agents, accelerate and simplify the isolation of RNA. For the extraction of RNA, this device uses an 80 μm porous membrane, which is 12.5 times thinner than the glass filter (1000 μm) commonly used in such devices, which makes it possible to reduce RNA degradation and increase its yield.

The reverse transcription reaction, which synthesizes a single-stranded DNA chain on an RNA matrix, is carried out using commercially available reverse transcriptase preparations, such as Moloni mouse leukemia virus (M-MLV) reverse transcriptase, avian myeloblastosis virus (AMV), PowerScript (point mutation M-MLV-RT), C. Therm Polymerase and others, with which you can get amplification products up to several thousand and even several tens of thousands of nucleotide pairs (TPN). Thermostable Thermus thermophilus (Tth) DNA polymerase can be used, with reverse transcriptase activity in the presence of Mn ^{2+ ions} .

For reverse transcription, various primers were used, for example:

1) Oligo (dT) -containing primers bind to the endogenous polyA tail at the 3'-end of the mRNA. These primers are most often used to obtain full-size cDNA. To the oligo (dT) sequence, nucleotides A, C, or G are often added at the 3'-end to anchor the primer to the border of the transcript and poly-A tract;

2) Random hexanucleotide primers (statistical primers) hybridize with RNA in numerous sites. When reverse transcription with these primers receive short cDNA. Random hexamers are used to overcome difficulties associated with the secondary structure of RNA, they are more effective in reverse transcription of 5'-regions of mRNA;

3) Hexamers or other short oligonucleotides (10-12 nucleotides) of random composition can also be used in combination with oligo-HG-containing primers;

4) Specific oligonucleotide primers are used to transcribe the mRNA region of interest for the study. These primers are successfully used for diagnostic purposes.

The selection of specific primers and probes is carried out by a method well known to specialists in this field, for which they use the available programs, many of which are freely available on the Internet, for example, Genome Browser Gateway, BLAST, Primer3, etc.

The high specificity of PCR-RV is achieved through the use of a probe containing a fluorophore or fluorescent dye (in this case, FAM-6-carboxy-fluoroscein) at the 5'-end, and the so-called at the 3'-end a damper (in this case, RTQ1). During PCR, their interaction is disrupted due to the cleavage of the Taq probe with DNA polymerase due to its 5'-exonuclease activity, and fluorescence emission occurs. The selection of probes for PCR-RV is carried out in accordance with standard recommendations and method requirements.

In a preferred embodiment, primers and probes of SEQ ID NO: 1-24 (FIGS. 1, 2, and 3) may be used.

When assessing the level of mRNA of genes, the input receives the set of threshold cycle values C _p (p = 1, 2, ..., 24) corresponding to the results of one PCR study of a biological sample, where the value of the threshold cycle C _p corresponds to the sample (Planchet well) p, sample 1 -3 correspond to the ELOVL5 gene, samples 4-6 to the IGFBP6 gene, samples 7-9 to the TXNDC9 gene, samples 10-12 to the FBXO42 gene, samples 13-15 to the GUSB gene, samples 16-18 to the HCFC1 gene, samples 19-21 - the RPLP0 gene and samples 22-24 - the TPT1 gene. In this case, PCR efficiencies corresponding to the primers used are considered known: E (ELOVL5), E (IGFBP6), E (TXNDC9), E (FBXO42), E (GUSB), E (HCFC1), E (RPLP0) and E (TPT1). For primers SEQ ID NOs: 1-6, 10-19, these PCR efficiencies are 1.842, 1.914, 1.807, 1.885, 1.870, 1.904, 1.889, and 1.765, respectively.

пропорциональную количеству мРНК ELOVL5 в исследуемом образце, по формулеFirst, the values of the threshold cycles determine the number of corresponding transcripts individually for each sample corresponding to one of the marker genes ELOVL5, IGFBP6, TXNDC9. For this, for each of the samples p = 1, 2, 3, calculate the value

proportional to the amount of ELOVL5 mRNA in the test sample, according to the formula

пропорциональную количеству мРНК IGFBP6 в исследуемом образце по формулеSimilarly, for each of the samples p = 4, 5, 6, calculate the value

proportional to the amount of IGFBP6 mRNA in the test sample according to the formula

пропорциональную количеству мРНК TXNDC9 в исследуемом образце по формулеand for each of the samples p = 7, 8, 9 calculate the value

proportional to the amount of TXNDC9 mRNA in the test sample according to the formula

In the above formulas, normalization is based on the approach related to averaging using the geometric mean [D.V. Rebrikov, G.A. Samatov, D.Yu. Trofimov, P.A. Semenov, A.M. Savilova, I.A. Kofiadi, D.D. Abramov real-time PCR - Moscow: BINOM. Laboratory of Knowledge, 2009, p. 8.4; J. Vandesompele, K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A. De Paepe, F. Speleman Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes / / Genome Biol., 2002, 3 (7): RESEARCH0034]. Moreover, to normalize the samples associated with each of the marker genes, we use a subset of normalization genes with the amount of mRNA in the order closest to the average amount of mRNA of the corresponding marker gene.

Reusing averaging based on geometric mean, as an abnormal quantitative estimate of the representation of the marker gene ELOVL5 in a sample of biological material, take the geometric mean for all samples corresponding to this gene:

Similarly, as an irregular quantitative estimate of the presence of the marker gene IGFBP6 in a sample of biological material, the geometric mean is taken for all samples corresponding to this gene:

as an abnormal quantitative estimate of the representation of the marker gene TXNDC9 in the sample of biological material, take the geometric mean of all samples corresponding to this gene:

и Q(ELOVL5), Q(IGFBP6), Q(TXNDC9) могут осуществляться с использованием перехода в логарифмическую шкалу.To reduce computational errors

and Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) can be implemented using the transition to the logarithmic scale.

After the values of unnormalized quantitative estimates of the representation of Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) are calculated in the framework of the study of the sample of biological material, the relapse risk is predicted using a linear classifier. At the first step, a transition to the logarithmic scale and linear normalization are carried out: normalized quantitative estimates of representation are calculated

Q _norm (ELOVL5) = (log ₂ Q (ELOVL5) -b _ELOVL5 ) / a _ELOVL5 ;

Q _norm (IGFBP6) = (log ₂ Q (IGFBP6) -b _IGFBP6 ) / a _IGFBP6 ;

Q _norm (TXNDC9) = (log ₂ Q (TXNDC9) -b _TXNDC9 ) / a _TXNDC9 .

The normalization coefficients a and b are selected at the stage of system calibration in such a way that for each marker gene ELOVL5, IGFBP6, TXNDC9, the empirical mean value of Q _norm is zero, and the empirical standard deviation is unity. Values of normalization coefficients may depend on the equipment used during PCR-RV and the collection of samples, which were used to calculate the empirical mean and standard deviation. Possible values of normalization coefficients are given in table. one.

Next, the calculation of the risk factor for recurrence of breast cancer S _R :

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9).

The values of the weight parameters w ₀ , w (ELOVL5), w (IGFBP6), w (TXNDC9) are given in table. 2.

Based on the value of S _R , the final forecast of the risk of relapse _{R is} carried out: with a value of more than 0.2, a conclusion is made that the risk of relapse is high, with a value of less than -0.2 a conclusion is made that the risk of relapse is low, with a value between -0.2 and 0.2 consider the risk uncertain.

Formally, both steps taken in the transition from unnormalized quantitative estimates of the representation of Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) to the risk of relapse S _R can be combined into one, which consists in calculating a linear combination of the logarithms of these values. However, dividing into two steps allows, within the framework of the system under consideration, when changing the equipment used or even the method used to assess the quantitative representation of marker genes, only normalization parameters a _g and b _g are recalculated, while maintaining the values of the weight parameters w ₀ and w (ELOVL5), w (IGFBP6 ), w (TXNDC9).

This is relevant, in particular, when assessing the quantitative representation of marker genes using microarray analysis. In this case, only the method for calculating the normalized quantitative assessment of the representation of Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) changes, and the transition from these estimates to the risk factor of relapse S _R and the subsequent construction of the risk forecast R are carried out exactly in the way described above.

When assessing the number of transcripts corresponding to marker genes using microarray analysis, the microchip is pre-processed first, together with a standard normalizing set of microchips. Pretreatment is carried out using the RMA method (Irizarry R.A. et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data // Biostatistics. 2003. T. 4 C. 249-264). In the case of Affymetrix Human Genome U133A Array microarrays, the GSE17705 collection (Symmans WF et al. Genomic index of sensitivity to endocrine therapy for breast cancer // J. Clin. Oncol. 2010. T, 28, no. 27. C. 4111-4119. The use of a standard standardization set is necessary to bring the results of processing various microchips into a single scale. An alternative to using a standard standardization set of microarrays is the use of the fRMA method for preprocessing (McCall MN, Bolstad VM, Irizarry RA Frozen robust multiarray analysis (fRMA) // Biostatistics. 2010 T. 11, No. 2. C. 242-253.).

After preprocessing, Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) transcripts are taken as abnormal estimates of the quantitative representation of transcripts, and the values obtained as a result of preprocessing associated with the corresponding marker genes are selected from microarray sample sets. Values are taken on a standard (non-logarithmic) scale. Next, the transition to normalized quantitative estimates of the representation of marker genes Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9). The scheme of this transition is similar to that described above, the difference is in the coefficients used for normalization:

The values of the normalization coefficients corresponding to the Affymetrix Human Genome U133A Array microchip technology and the GSE17705 standard set are shown in Table. 3.

It should be noted that the selection of transcripts used directly for generating the prognosis was based on the analysis of data on the genome-wide transcript of more than 400 patients for whom there was information on the occurrence and time of diagnosis of relapse (GSE6532, GSE12093, GSE17705 data sets), as well as their own data genome-wide transcriptome analysis of more than 250 tissue samples affected by breast cancer. The selection was based on the construction of a binary (two-class) classifier that distinguishes between patients who have been diagnosed with relapse and patients who have not been diagnosed with relapse. At the same time, when constructing the training sample, the so-called “gray zone” was excluded from consideration: only patients in whom the relapse was diagnosed within the first five years after surgical treatment were included in the first class, and in the second patients, in whom the relapse-free period was not less than seven years.

For a fixed set of transcripts for the automated construction of a classifier using exactly this set of transcripts, the support vector method was used (SVM Supprot Vector Machine, see C. Cortes, V. Vapnik Support-vector networks // Machine Learning, Vol. 20 (1995) , 273-297) with a linear kernel and various penalties for errors of various kinds.

The formation of the final set of transcripts was carried out automatically using complete enumeration of pairs of transcripts and the subsequent completion of the most informative pairs to triples; completion was carried out using the so-called “greedy approach”, including adding one transcript to a pair and subsequent component-wise optimization (Galatenko VV et al. Highly informative marker sets consisting of genes with low individual degree of differential expression // Sci. Rep. 2015. T 5. C. 14967; Galatenko V.V. et al. On the construction of medical test systems using the greedy algorithm and the support vector method // Bulletin of Experimental Biology and Medicine. 2013. V. 156. No. 11. P. 654- 60).

Validation of the final set of transcripts and the classifier corresponding to this set was carried out on a validation sample, the samples of which were not used either in the selection of transcripts or when setting the classifier: training and validation samples did not intersect not only in patients, but also in clinics in which the biomaterial was collected, and in the laboratories that conducted research on this biomaterial.

The implementation of the invention

The invention is illustrated by the following examples of implementation.

Example 1

The example shows the implementation of a method for assessing the risk of breast cancer recurrence in a patient Inna U. based on PCR-RV.

RNA isolation:

a) took a sample of a biopsy of the tumor tissue of the mammary gland, preserved by the Reagent for stabilization of mRNA;

b) the analyzed fragment of the biopsy of the breast tumor tissue was removed with tweezers from the Reagent to stabilize the mRNA in a previously prepared chilled porcelain mortar. Re-cooled the mortar and pestle with liquid nitrogen to a state where the nitrogen is almost no longer evaporates;

c) carried out mechanical crushing / grinding of the sample in the mortar;

d) 700 μl of Buffer A and 20 ml of liquid nitrogen were added to the mortar. Grind to obtain a uniform pinkish powder. Upon completion, the mortar was allowed to warm to room temperature, and the sample melted;

d) transferred the entire contents of the mortar into a clean 1.5 ml tube and adjusted the homogenate volume to 1 ml by adding 300 μl of Buffer A.

e) carried out the stage of additional homogenization. For this, 500 μl of homogenate was transferred to a QIAAshredder (Qiagen) column or similar. Centrifuged for 2 minutes at a speed> 10,000 rpm. The solution passing through the column was carefully taken into a clean tube. The procedure was repeated for the remaining sample. The homogenized fractions were combined;

g) 200 μl of chloroform was added to the sample and thoroughly mixed for 15 seconds on a vortex. Incubated at room temperature for 2-3 minutes;

i) centrifuged at 4 ° C for 15 minutes at 12000 rpm;

j) transferred the aqueous phase (upper, unpainted) into a clean 1.5-2 ml tube and determined its volume;

l) 1.5 volumes from the transferred aqueous phase of 100% ethanol were added. Mix well with a pipette;

m) transferred 500 μl of the sample, including the precipitate, if precipitated, to an RNeasy mini spin column or similar. Centrifuged for 15 seconds at a speed of> 10,000 rpm. The solution passed through the column was removed;

m) repeat the previous step for the remaining solution (using the same column);

o) 350 μl of Buffer B was applied to the column and centrifuged for 15 seconds at a speed of> 10,000 rpm. The solution passed through the column was removed;

p) was treated with DNaseI. Transferred 80 μl of the prepared DNaseI solution to an RNeasy mini spin column or similar and incubated for 15 min at 20-30 ° C;

c) repeat point p);

t) was applied onto a column of 500 μl of Buffer B, centrifuged for 15 seconds at a speed of> 10,000 rpm. The solution passed through the column was removed;

s) 500 μl of Buffer B was added to the column and centrifuged for 2 min at a speed of> 10,000 rpm;

f) the column was carefully transferred to a clean 1.5 ml tube and centrifuged for another 1 min to completely remove the residues of Buffer B;

x) transferred the column to a clean 1.5 ml tube to elute the sample. 40 μl of deionized water was directly applied directly to the column membrane. Centrifuged for 1 min at a speed of> 10,000 rpm;

c) the eluted solution was collected and applied again to the membrane of the same RNeasy mini spin column or similar. Centrifuged for 1 min at a speed of> 10,000 rpm;

sh) the collected eluted solution was mixed, besieged by short centrifugation and divided into 2 aliquots. The resulting RNA solution was kept on ice;

y) the concentration of the isolated total RNA was measured using a NanoDrop spectrophotometer or analogue;

e) The integrity of the isolated total RNA was assessed using an Agilent 210 Bioanalyzer hardware-software platform or equivalent.

Carrying out reverse transcription (OT):

a) tubes with RT buffer, a mixture of RT primers and deionized water were thawed at room temperature. A reverse transcriptase tube was placed in ice;

b) two test tubes containing 2 μg of isolated RNA were prepared and added to each water deionized to a final volume of 12 μl;

c) incubated for 5 minutes at 65 ° C. Transferred to ice;

d) mixed with 11 μl of RT buffer, 1 μl of RT primer mixture, 1 μl of reverse transcriptase. Stirred and added 13 μl to 2 μg of RNA;

d) carried out the synthesis of cDNA in a programmable automatic thermal cycler according to the program presented in table. four

f) the obtained two cDNA samples were mixed in one tube. Kept on ice.

PCR-RV

a) 10-fold dilutions of cDNA solutions of the test sample, the RNA sample of the brain cells of the control and the sample of universal human RNA of the control were prepared. In the table. Figure 5 shows the calculation of the components of the reaction mixture for application to the Tablet with the applied reagents.

b) reaction mixtures for PCR-RV were prepared in accordance with the scheme presented in table. 6 (a calculation is given for PCR-RV in one of the 384-well plates, the composition of the mixtures is similar).

c) 12 μl of the reaction mixture was added to the corresponding wells of the Tablet (tablet well scheme in FIG. 4);

g) sealed the filled tablet with a transparent film for using the tablet. Shake the tablet on the vortex. The plate was centrifuged briefly in a centrifuge to deposit droplets from the walls and the film to the bottom of the tubes;

d) the tablet was placed in an amplifier for PCR-RV at 384 wells. The recommended program for amplification is presented in table. 7

After PCR-RV biopsy of the patient, Inna U., the values of the threshold cycles are shown in Table 8.

Performing calculations on a logarithmic scale and taking into account the PCR efficiencies E (ELOVL5) = 1.842, E (IGFBP6) = 1.914, E (TXNDC9) = 1.807, E (FBXO42) = 1.885, E (GUSB) = 1.870, E (HCFC1) = 1,904, E (RPLP0) = 1,889 and E (TPT1) = 1,765, lead to values

и, аналогично,

Соответственно,

откуда

Следует отметить, что значение Q(ELOVL5) приведено исключительно для полноты изложения, так как на следующем шаге (при вычислении Q_norm(ELOVL5)) используется значение именно log₂Q(ELOVL5), а не самой Q(ELOVL5).

and similarly

Respectively,

where from

It should be noted that the value of Q (ELOVL5) is given solely for completeness, since the next step (when calculating Q _norm (ELOVL5)) uses the value of log ₂ Q (ELOVL5), and not Q (ELOVL5) itself.

Similar calculations give

Next, normalized quantitative estimates of the representation of marker genes are calculated (see Table 1):

Q _norm (ELOVL5) = (log ₂ Q (ELOVL5) - (- 1.351)) / 0.934≈-0.163;

Q _norm (IGFBP6) = (log ₂ Q (IGFBP6) - (- 0.876)) / 0.752≈-1.014;

Q _norm (TXNDC9) = (log ₂ Q (TXNDC9) - (- 0.417)) / 0.629≈0.800.

Based on standardized quantitative estimates of the presence of marker genes, the risk of relapse risk S _{R is} calculated (see Table 2):

Since the obtained value of S _{R is} greater than 0.2, a conclusion is made about the high risk of relapse.

Example 2

The example shows the implementation of a method for assessing the risk of recurrence of breast cancer. The stage of interpretation of normalized quantitative estimates of gene expression of Q _norm . Cases with high and low risk of relapse are presented.

Calculation of unnormalized quantitative estimates of gene representation from threshold cycle values obtained by PCR-RV analysis of the sample.

Relapse risk assessment by standardized quantitative estimates of the representation of Q _norm genes.

2.1. After PCR-RV biopsy of patient Catherine A. and calculation of normalized quantitative estimates of gene representation, the following results were obtained: Q _norm (ELOVL5) = 0.418; Q _norm (IGFBP6) = 1.096; Q _norm (TXNDC) = - 0.602. In this case, the value of the risk factor for relapse S _R was (see Table 2)

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9) = 0.392-0.724⋅0.418-0.654⋅1.096 + 0.549⋅ (- 0.602) ≈ -0.958.

Since the obtained value of S _{R is} less than -0.2, it is concluded that there is a low risk of relapse.

2.2. After PCR-RV biopsy of patient Alexandra P. and calculation of normalized quantitative estimates of gene representation, the following results were obtained: Q _norm (ELOVL5) = - 0.545; Q _norm (IGFBP6) = 0.199; Q _norm (TXNDC) = 0.438. In this case, the value of the risk factor for relapse S _R was (see Table 2)

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9) = 0.392 + 0.724⋅0.545-0.654⋅0.199 + 0.549⋅0.438≈ 0.897.

Since the obtained value of S _{R is} greater than 0.2, a conclusion is made about the high risk of relapse.

Example 3

The example shows the implementation of a method for assessing the risk of breast cancer recurrence in a patient Daria P. based on a microarray analysis.

After a microarray biopsy analysis of the patient Daria P., the values associated with the ELOVL5, IGFBP6, and TXNDC9 genes were 1942.9, 149.0, and 223.1, respectively. The analysis was carried out on an Affymetrix Human Genome U133A Array microchip, GSE17705 collections were used as a normalizing set for preprocessing, as sample sets corresponding to marker genes, 208788_at, 203851_at and 211758_x_at, respectively.

Normalized quantitative estimates of the representation of marker genes (see Table 3) amounted to

Q _norm (ELOVL5) = (log ₂ 1942.9-10.427) /0.763≈0.651;

Q _norm (IGFBP6) = (log ₂ 149.0-7.379) / 0.640≈-0.250;

Q _norm (TXNDC9) = (log ₂ 223.1-8.316) / 0.581≈-0.885.

Accordingly, the risk factor for relapse S _R (see Table 2) was:

S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9) ≈ 0.392-0.724⋅0.651-0.654⋅ (-0.250) +0.549 ⋅ (-0.885) ≈ -0.402.

Since the obtained value of S _{R is} less than -0.2, it is concluded that there is a low risk of relapse.

Validation of Method Reliability

Validation of the reliability of the described methodology was carried out on our own data collection, to which PCR-RV analysis was applied, as well as on the microarray data set GSE3494 (Miller LD et al. An expression signature for p53 status in human breast cancer predicts mutation status, transcriptional effects, and patient survival // Proc Natl Acad Sci USA. 2005. T. 102, No. 38. C. 13550-13555), which were not used either when choosing marker genes used or when setting the classifier. Validation confirmed the high reliability of the method.

Thus, the Kaplan-Meyer curves and the ROC curve corresponding to microchip validation (201 samples, including 33 samples corresponding to the occurrence of relapse within five years, and 128 samples corresponding to relapse-free survival for at least 7 years) are shown in FIG. 5.

The relapse rates estimated for this sample in patients with a low, uncertain, and high risk of relapse were 0.094 (9.4%), 0.160 (16.0%), and 0.412 (41.2%), respectively. The proportion of patients in the class with uncertain risk was only 0.155 (15.5%). The values of sensitivity and specificity were 0.756 (75.6%) and 0.602 (60.2%), respectively, in the case of combining uncertain risk with high, and 0.636 (63.6%) and 0.766 (76.6%), respectively, in case of combining uncertain low risk. With the exclusion of samples with uncertain risk from consideration, the sensitivity and specificity were 0.724 (72.4%) and 0.720 (72.0%), respectively.

Thus, the experimental data show that the application of the claimed method allows with high sensitivity and specificity to assess the risk of relapse after surgical treatment using new diagnostic breast cancer markers, and also to develop a simple, sensitive, reliable method for predicting relapse-free survival on their basis.

Claims (23)

1. A method for determining the risk of relapse of breast cancer in a patient, comprising measuring the expression level of marker genes ELOVL5, IGFBP6, TXNDC9 in a patient’s tumor tissue sample and then assessing the risk of breast cancer recurrence by comparing expression levels of marker genes with predetermined threshold values. 2. The method according to p. 1, characterized in that as samples of the tumor tissue of the patient using freshly frozen tissue cells, or biopsy cells, or cells fixed in paraffin. 3. The method according to p. 1, characterized in that the measurement of the expression of marker genes is carried out by the method of microarray analysis. 4. The method according to p. 1, characterized in that the measurement of the expression of marker genes is carried out by PCR-RV, while for the measurement using single-stranded cDNA synthesized on mRNA isolated and purified from a patient’s tumor tissue sample. 5. The method according to p. 4, characterized in that the synthesis of cDNA is carried out using oligonucleotide nonspecific primers with a random sequence of nucleotides and with pairs of primers direct and reverse, and probes specific for marker genes, as well as with pairs of primers and probes specific for the FBXO42, GUSB, HCFC1, RPLP0, and TPT1 genes for normalizing PCR results, obtaining for each of the marker genes abnormal estimates of their expression levels Q (g), then normalizing them and obtaining values nd Q _norm (g) for each marker gene computation based Q _norm (g) S _R risk of relapse using the formula: S _R = w ₀ + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = -0.72, w (IGFBP6) = -0.65, w (TXNDC9) = 0.55, w ₀ = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized expression levels of marker genes. 6. The method according to p. 5, characterized in that at least two values are used as threshold values, upon receipt of S _R above the upper threshold value, a conclusion is made that there is a high risk of relapse, when receiving S _R below the lower threshold value, the conclusion about a low risk of relapse, when S _{R is} obtained from the interval of values between the indicated threshold values, they conclude that there is an indefinite risk. 7. The method according to p. 5, characterized in that as primers and probes specific for marker genes, use primers with the sequences SEQ ID NO: 1-3 and SEQ ID NO: 4-6 and probes with SEQ sequences ID NO: 7-9, respectively; as primers and probes specific for the FBXO42, GUSB, HCFC1, RPLP0 and TPT1 genes for normalizing PCR results, primers with the sequences SEQ ID NO: 10-19 and probes with the sequences SEQ ID NO: 20-24, respectively, are used. wherein the risk of relapse (S _R ) is assessed using the formula: S _R = w0 + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = -0.72, w (IGFBP6) = -0.65, w (TXNDC9) = 0.55, w0 = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized expression levels of marker genes, upon receipt of a value of S _R greater than 0.2 - conclude that there is a high risk of recurrence, when the value of S _{R is} less than minus 0.2 - conclude a low risk of recurrence, when a value of S _{R is} in the range from minus 0.2 to 0.2 - consider the risk uncertain. 8. The method according to p. 4, characterized in that the synthesis of cDNA for measuring the expression of marker genes is carried out using a set of reagents, including at least primers - direct and reverse, and probes specific for marker genes - ELOVL5 , IGFBP6, TXNDC9, presented in a lyophilized form or aqueous solution, or applied to the wells of a tablet. 9. The method according to p. 8, characterized in that the primers in the wells of the tablet are contained in a concentration of 250 nmol. 10. The method according to p. 9, characterized in that the probe in the wells of the tablet contains a FAM dye with excitation parameters at a wavelength of 485 nm and emission at a wavelength of 520 nm, taken at a concentration of 250 nMol. 11. The method according to p. 3, characterized in that the measurement of expression levels of marker genes ELOVL5, IGFBP6, TXNDC9 by microarray analysis is carried out by obtaining irregular estimates of expression levels of marker genes Q (ELOVL5), Q (IGFBP6), Q (TXNDC9) , then normalized estimates of the levels of marker genes Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9), are calculated _{R R} relapse risk assessment is performed using the formula: S _R = w0 + w (ELOVL5) Q _norm (ELOVL5) + w (IGFBP6) Q _norm (IGFBP6) + w (TXNDC9) Q _norm (TXNDC9), where w (ELOVL5) = -0.72, w (IGFBP6) = -0.65, w (TXNDC9) = 0.55, w0 = 0.39; Q _norm (ELOVL5), Q _norm (IGFBP6), Q _norm (TXNDC9) - normalized expression levels of marker genes. 12. The method according to p. 11, characterized in that at least two values are used as threshold values, upon receipt of S _R above the upper threshold value, a conclusion is made that there is a high risk of relapse, when receiving S _R below the lower threshold value, they are made the conclusion about a low risk of relapse, when S _{R is} received from the interval of values between the indicated threshold values, a conclusion is made about the uncertain risk. 13. The method according to p. 12, characterized in that upon receipt of a value of S _R greater than 0.2, they conclude that there is a high risk of recurrence, when a value of S _{R is} less than minus 0.2, they conclude that there is a low risk of recurrence values of S _R in the range of values from minus 0.2 to 0.2 - consider the risk as uncertain. 14. The method according to p. 11, characterized in that to obtain irregular estimates of the expression levels of marker genes, the data are preprocessed using the RMA method in conjunction with a normalization set or using the fRMA method.

Images