ASSAY FOR DETERMINING THE TYPE AND/OR STATUS OF A CELL BASED ON THE EPIGENETIC PATTERN AND THE CHROMATIN STRUCTURE

Abstract

The present invention relates to a method for identifying a specific type and/or state of a mammalian cell in a sample obtained from a mammal, comprising a) analyzing the relative amount of accessible chromatin in regions that are specific for a cell-type and/or cellular state in the genome of said cell, b) comparing said relative amount of accessible chromatin said in regions with the relative amount of accessible chromatin in regions in the genome of said cell that are unspecific for a cell-type and/or cellular state, and c) deducing the specific type and/or state of said mammalian cell in said sample based on said comparison. Preferably, said identifying further comprises a relative quantification of said specific cell type and/or state based on said comparison. The method can further comprise a diagnosis of a predisposition to a disease or a disease based on said identification. Kits and certain markers in regions of accessible chromatin in the genome are described, too.

Claims

1. A method for identifying a specific type and/or state of a mammalian cell in a sample obtained from a mammal, comprising a) analyzing the relative amount of accessible chromatin in regions that are specific for a cell-type and/or cellular state in the genome of said cell, b) comparing said relative amount of accessible chromatin in said regions with the relative amount of accessible chromatin in regions in the genome of said cell that are unspecific for a cell-type and/or cellular state, c) optionally, normalizing the relative amount of said regions that are specific for a cell-type and/or cellular state and said regions in the genome of said cell that are unspecific for a cell-type and/or cellular state using a control plasmid, and d) deducing the specific type and/or state of said mammalian cell in said sample based on said comparison.

2. The method according to claim 1, wherein said identifying further comprises a relative quantification of said specific cell type and/or state based on said comparison.

3. The method according to claim 1, further comprising a step of determining a specific cell-type and/or cellular state comprising measuring the relative amount of accessible chromatin in the genome of a cell having a known specific cell-type and/or cellular state prior to step a).

4. The method according to claim 3, further comprising generating a knowledge base comprising information on the relative amount of accessible chromatin in the genome of cells having a known specific cell-type and/or cellular state.

5. The method according to claim 1, wherein analysis comprises measuring the relative amount of accessible chromatin with an assay comprising DNAse I digestion, ChIP Chip®, chromatin immunoprecipitation microarray, quantitative PCR analysis based on bisulfite converted or not, selective precipitation and/or conversion of cytosines with bisulfite.

6. The method according to claim 1, wherein said regions that are specific for a cell-type and/or cellular state in the genome of said cell are selected from regions comprising a gene selected from FOXP3, GNLY, CD3, platelet glycoprotein IX (GP9); low affinity immunoglobulin epsilon Fc receptor (FCER2); protein S100-P (S100 calcium-binding protein P); homeodomain-interacting protein kinase 3 (HIPK3); transmembrane 4 L6 family member 19 (TM4SF19); CD160 antigen precursor (Natural killer cell receptor BY55) (CD160); and LIM domain-binding protein 2 (LDB2).

7. The method according to claim 1, wherein said regions that are unspecific for a cell-type and/or cellular state are selected from regions comprising a housekeeping gene.

8. The method according to claim 1, wherein said cell type is selected from an immune cell; kidney cell; bone cell; neuronal cell; blood cell; lung cell; colon cell; and a precursor of any of these, excluding human embryonic stem cells.

9. The method according to claim 1, further comprising a diagnosis of a predisposition to a disease or a disease based on said identification.

10. The method of claim 9, wherein the disease is selected from the group consisting of immune diseases or conditions, cancer, birth defects, mental retardation, obesity, neurological disease, diabetes, and gestational diabetes.

11. A method for monitoring the effect of a drug on the relative amount of a specific type and/or the state of a mammalian cell in a sample obtained from a mammal, comprising performing the method according to claim 1 on a sample obtained from a mammal treated with said drug, and comparing the relative amount of said specific type and/or the state of said mammalian cell with an untreated sample.

12. The method according to claim 1, wherein the sample is selected from the group consisting of blood or fractions thereof, saliva, buccal, tears, semen, urine, sweat, faecal material, skin and hair.

13. The method according to claim 11, wherein said cell type is selected from an immune cell; kidney cell; bone cell; neuronal cell; blood cell; and a precursor of any of these, excluding human embryonic stem cells.

14. The method according to claim 11, wherein said treatment is for a disease or condition selected from the group consisting of immune diseases or conditions, cancer, birth defects, mental retardation, obesity, neurological disease, diabetes, and gestational diabetes.

15. A diagnostic kit comprising materials for performing the method according to claim 1.

16. The method, according to claim 7, wherein said housekeeping gene is GAPDM.

17. The method, according to claim 8, wherein said immune cell is selected from the group consisting of CD19+ B cells, CD3+CD8+ cytotoxic T cells, CD15+ granulocytes, CD14+ monocytes, CD56+ natural killer cells, and CD4+ helper T cells.

18. The method, according to claim 10, wherein said cancer is selected from the group consisting of colorectal cancer, oesophageal cancer, stomach cancer, leukaemia/lymphoma, lung cancer, prostate cancer, uterine cancer, breast cancer, skin cancer, endocrine cancer, urinary cancer, pancreatic cancer, other gastrointestinal cancer, ovarian cancer, cervical cancer, head cancer, neck cancer, and adenomas.

19. The method, according to claim 13, wherein said immune cell is selected from the group consisting of CD19+ B cells, CD3+CD8+ cytotoxic T cells, CD15+ granulocytes, CD14+ monocytes, CD56+ natural killer cells, and CD4+ helper T cells.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0085] FIG. 1 shows the A) Genomic organization and localisation of the genes CD3, GNLY and GAPDH. Transcripts are shown depending on the direction above or below the chromosomal bar. Amplicons aligning to the various gene regions are indicated. B) Epigenetic profiling of selected amplicons. The results from bisulfite sequencing are indicated in a coded matrix, where each line represents the sequencing data in a cell type and each block represents an amplicon. Black corresponds to DNA that is inaccessible to conversion by bisulfite and gray corresponds to DNA that is accessible to bisulfite conversion.

[0086] FIG. 2 shows amplification profiles of the specific RT PCR assays. Each time, in the left panel the PCR system using the primers and probes for fully bisulfite converted DNA is shown, while in the right panel the version for DNA with CpGs inert to bisulfite conversion is demonstrated. Linearity of all PCR systems is shown inside of each graph by plotting measured CP values over the log concentration of template used.

[0087] FIG. 3 shows a preferred embodiment of the fully bisulfite converted control plasmid according to the examples below.

[0088] FIG. 4 shows the ratio of GAPDH fully bisulfite converted DNA versus only partially bisulfite converted DNA in a variety of cells and tissues (1. Granulocytes, 2. Monocytes, 3. NK cells, 4. CD4 naïve cells, 5. CD8 cells, L. Lung, C. colon, U. uterine tissue, B. breast tissue. The inventors could never detected a signal for only partially converted DNA. Therefore, all cells appear to contain only fully accessible DNA, i.e. open chromatin.

[0089] FIG. 5 shows the methylation analysis-based chromatin accessibility analysis of certain preferred cell type and/or status specific genes. The cell types are BCL05: CD19+ B cells; CTL05: CD3+CD8+ cytotoxic T cells; GRC01: CD15+ granulocytes; MOC02: CD14+ monocytes; NKCO2: CD56+ natural killer cells; and THC04: CD3+CD4+ helper T cells. The regions/genes tested are

TABLE-US-00001 AMP-ID/SEQ ID No. Gene Name/Gene-ID 1583/16 Platelet glycoprotein IX/ENSG00000169704 (GP9) 1584/17 Low affinity immunoglobulin epsilon Fc receptor/ENSG00000104921 (FCER2) 1588/18 Protein S100-P (S100 calcium-binding protein P)/ENSG00000163993 1589/19 Homeodomain-interacting protein kinase 3/ ENSG00000110422 (HiPK3) 1594/20 Transmembrane 4 L6 family member 19/ ENSG00000145107 (TM4SF19) 1599/21 CD160 antigen Precursor (Natural killer cell receptor BY55)/ENSG00000117281 1601/22 LIM domain-binding protein 2/ENSG00000169744 (LDB2)

[0090] The CpG positions as analyzed in the actual amplicon are indicated by the numbers following the amplicon number.

[0091] FIG. 6 shows the analysis of peripheral blood from ovarian cancer patients before and after treatment with Catumaxumab (see examples, below). Measurements were conducted for all available blood samples using FACS sorting selecting either for CD3+ cells versus all nucleated cells (A), CD4+CD25+CD127- versus all nucleated cells (B), CD4+CD25+CD127- versus CD3+ cells (C) and CD56+ cells versus all nucleated cells (D). FACS results in percent [%] are plotted on the X-axis and compared to the percentage that resulted from the epigenetic analysis of cells by means of qPCR which is plotted on the Y-axis. R indicates the Pearson correlation coefficient for each FACS to the epigenetic measurement. The p-values indicate the statistical significance of the correlations.

[0092] FIG. 7 shows tissue infiltrating lymphocytes in healthy and cancerous tissue. Boxplots showing the relative abundance in percent of the total cell count of A) Tregs, B) CD3.sup.+ T cells, C) GNLY.sup.+ cells and D) Tregs within the CD3 compartment in healthy and cancerous ovarian (OC), lung (BCa) and colorectal (CRC) tissues. N indicates the number of patients included in each boxplot. The box in the middle depicts 50% of the distribution. The central line in the box represents the median of the distribution, and the whiskers cover 95% of all measured data. Outliers from this distribution are indicated by circles. The indicated p-value were obtained from the two-sided, heteroskedastic students t-Test.

BRIEF DESCRIPTION OF SEQUENCES

[0093] SEQ ID NO: 1 shows the DNA sequence of the insert as cloned into the plasmid as used in the examples.

[0094] SEQ ID NOs: 2 to 15 show sequences of the primers as used in the experiments that are specific for bisulfite-converted DNA.

[0095] SEQ ID NOs: 16 to 22 show sequences of the amplicons as analyzed in FIG. 5.

[0096] SEQ ID NOs: 23 to 29 show sequences of the regions of interest which also can be analyzed in the context of the invention. The ROIs correspond to AMP1583, AMP1584, AMP1588, AMP1589, AMP1594, AMP1599, and AMP1601, respectively.

Examples

[0097] Materials and Methods

[0098] Abbreviations:

[0099] Amp, amplicon; CD3D, T-cell surface glycoprotein CD3 delta chain; CD3G, T-cell surface glycoprotein CD3 gamma chain; FOXP3, Forkhead box protein P3; GAPDH, Glyceraldehyde-3-phosphate dehydrogenase; GNLY, Granulysin.

[0100] Cells and Tissue Samples

[0101] Formalin-fixed and paraffin-embedded tissue samples were retrieved from the archives of the Institute of Pathology, Charité-Universitaitsmedizin Berlin, Campus Benjamin Franklin. Representative paraffin blocks of tumour and normal tissue were selected and tissue microarray (TMA) of the colorectal or bronchial carcinoma specimens with the corresponding normal parenchyma were constructed using cores of 1 mm in diameter. Fresh frozen ovarian tissue samples and blood were retrieved from the tumour bank ovarian cancer, Charite-Universitaitsmedizin Berlin, Campus Virchow.

[0102] Isolation of Genomic DNA

[0103] For purification of genomic DNA from human blood, tissues, the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) was used. In case of ascites the inventors followed the protocol for isolation of total DNA from cultured cells. Genomic DNA from formalin-fixed paraffin-embedded (FFPE-) tissues samples was isolated using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). Paraffin blocks were trimmed to remove excess of paraffin and tissue section thickness was adjusted to 10 um. Each reaction was carried out using 10 tissue sections.

[0104] Sodium Bisulfite Conversion of Genomic DNA

[0105] Sodium bisulfite-mediated conversion was performed applying the EpiTect Bisulfite Kit (Qiagen, Hilden, Germany) and reactions were carried out using 0.5 to 1 ug of purified genomic DNA. In brief, thermal cycling of genomic DNA under high bisulfite salt concentrations and low pH lead to the conversion of unmethylated cytosine residues into uracil (which is replicated as thymine in a subsequent PCR). As under these conditions methylated cytosines (as found in the context of CpG dinucleotides) remain unchanged, the treatment translates epigenetic marks into sequence information.

[0106] Oligonucleotides

[0107] Oligonucleotides such as amplification primers and hydrolysis probes used in this work are indicated by their chromosomal positions relative to the assembly of the human genome GRCh37 (e!Ensemble release 56; September 2009).

[0108] Oligonucleotides for quantitative bisulfite sequencing: a) intergenic CD3G (ENSG00000160654)/CD3D (ENSG00000167286) region: Amplicon No. 1, forward primer: 11:118213200-118213221:1, reverse primer: 11:118213616-118213637:1; amplicon No. 2, forward primer: 11:118214271-118214292:1, reverse primer: 11:118214685-118214705:1; amplicon No. 3, forward primer: 11:118214702-118214723:1, reverse primer: 11:118215151-118215173:1; b) GNLY (ENSG00000115523) gene region: Amplicon No. 1, forward primer: 2:85921382-85921404:1, reverse primer: 2:85921742-85921763:1; amplicon No. 2, forward primer: 2:85921807-85921828:1, reverse primer: 2:85922259-85922279:1; amplicon No. 3, forward primer: 2:85922895-85922916:1, reverse primer: 2:85923327-85923348:1; c) GAPDH (ENSG00000111640) CpG island: Amplicon No. 1, forward primer: 12:6644119-6644135:1, reverse primer: 12:6644635-6644656:1; amplicon No. 2, forward primer: 12:6643586-6643604:1, reverse primer: 12:6643990-6644011:1.

[0109] Oligonucleotides for real-time PCR based assays: a) FOXP3 (ENSG00000049768) TSDR: methylation-specific PCR: forward primer: X:49117219-49117246:1, reverse primer: X:49117283-49117307:1, probe: X:49117256-49117273:1; demethylation-specific PCR: forward primer: X:49117219-49117246:1, reverse primer: X:49117283-49117307:1, probe: X:49117256-49117278:1. b) CD3: methylation-specific PCR: forward primer: 11:118213633-118213653:1, reverse primer: 11:118213686-118213707:1, probe: 11:118213670-118213687:1; demethylation-specific PCR: forward primer: 11:118213632-118213653:1, reverse primer: 11:118213686-118213709:1, probe: 11:118213664-118213690:1. c) GNLY: methylation-specific PCR: forward primer: 2:85921878-85921895:1, reverse primer: 2:85921964-85921992:1, probe: 2:85921918-85921943:1; demethylation-specific PCR: forward primer: 2:85921877-85921895:1, reverse primer: 2:85921964-85921992:1, probe: 2:85921911-85921939:1. d) GAPDH: demethylation-specific PCR: forward primer: 12:6644378-6644399:1, reverse primer: 12:6644456-6644476:1, probe: 12:6644429-6644457:1.

[0110] The sequences of the primer as used specific for bisulfite-converted DNA are as follows (see, for example, FIG. 5):

TABLE-US-00002 Amplicon/ SEQ ID NO. Orientation Sequence AMP1583-2 Reverse CTTCTCTAAACCCAACATCAAT AMP1583-3 Forward GAATTTAGGAGGTAGAGGTGGT AMP1584-4 Reverse GGATATTTGATTTGGGAGTTTA AMP1584-5 Forward AACCACTAACAACTTCTATTTTCA AMP1588-6 Reverse TTTTGTGTTAATATGAGGTTGTTT AMP1588-7 Forward ACCCTCTCCCTACTCAAATACT AMP1589-8 Reverse AGTGGTATAATTTTGTTTTGATTTT AMP1589-9 Forward AAATTCTCATCCTCCCACTAA AMP1594-10 Reverse ACCCACAAACCTACATTAAAAA AMP1594-11 Forward GTAAGGAGAGTGATGAGGAAAA AMP1599-12 Reverse CAATTCACAAATCCCATAAATA AMP1599-13 Forward TTGTTTAGGTGAGGATAGGTTT AMP1601-14 Reverse AGGTATTTTAAGGGTTTGAATG AMP1601-15 Forward TCTCCTCACAATCTAACAAAAA

[0111] Quantitative Bisulphite Sequencing

[0112] Targeted regions were pre-amplified from 7 ng sodium bisulfite converted genomic DNA using bisulfite-conversion specific primers. PCR was performed in a final volume of 25 μl containing 1×PCR Buffer, 1 U Taq DNA polymerase (Qiagen, Hilden, Germany), 200 μM dNTP, 12.5 pmol each of forward and reverse primers. Thermocycling was performed at 95° C. for 15 min, followed by 40 cycles of 95° C. for 1 min, 55° C. for 45 s and 72° C. for 1 min, and a final extension step of 10 min at 72° C. The PCR product was purified using ExoSAP-IT (USB Corp.) and directly sequenced applying the amplification primers and the ABI Big Dye Terminator v1.1 chemistry (Applied Biosystems). Products were purified by Ethanol precipitation, dissolved in IM betain and subjected to capillary electrophoresis on an ABI 3100 genetic analyzer. AB1 files were interpreted using ESME, which normalizes sequence traces, corrects for incomplete bisulphite conversion and allows for quantification of methylation signals at CpG sites.

[0113] Real-Time PCR

[0114] Real-time PCR was performed using Roche LightCycler 480 Probes Master chemistry (Roche Diagnostics) in a final reaction volume of 20 μl containing 30 pmol each of methylation- or demethylation-specific forward and reverse primers, 5 pmol of hydrolysis probe, 50 ng of 2-phage DNA (New England Biolabs), and 60 ng of bisulfite-converted genomic DNA template or a respective amount of plasmid standard. Each sample was analyzed in triplicates using a LightCycler 480 System (Roche). For all assay systems cycling conditions included a 95° C. preheating step for 10 min followed by 50 cycles of 95° C. for 15 s and 1 min at 61° C. CP (“crossing point”) values were computed by the second-derivative maximum method applying the LC480 analysis software and template copy numbers were calculated from calibration curves (using serial dilutions of appropriate plasmid-based standards) by linear regression.

[0115] Plasmid Standard

[0116] Bisulfite-converted methylated, and bisulfite-converted demethylated target regions for the various real-time PCR based assays were designed in silico, synthesized (Genscript Inc.) and fragments were inserted into plasmid pUC57. Recombinant plasmids were linearized and serially diluted in 10 ng/μl of 2-phage DNA (New England Biolabs) to obtain standards for real-time PCR based assays with final concentrations of 12,500, 2500, 500, 100, 20 and 4 template copies per reaction.

[0117] Cell Sorting of Major Peripheral Blood Leukocyte Population

[0118] Peripheral blood samples were obtained from healthy donors after informed consent in accordance with local ethical committee approval. Fractionation of blood samples into different leukocyte populations such as granulocytes (CD15+), monocytes (CD14+), CD4+ T cells (CD3+CD4+), Treg (CD4+CD25.sup.highCD45RA−), B cells (CD19+), NK cells (CD56+, CD56.sup.bright, CD56.sup.dim), naïve CD8+ T cells (CD3+CD8+CD45RA+CD127+) and memory CD8+ T cells (CD3+CD8+CD45RA-CD127+/−) was performed as described previously (Baron et al. Eur J Immunol). Purities of sorted cells were >97% as determined by flow cytometry and viabilities were always >99%.

[0119] Statistical Analysis

[0120] Amounts of methylated (CpG variant) and unmethylated (TpG variant) DNA were estimated from calibration curves by linear regression on crossing points from the second-derivative maximum method. The median was used to aggregate triplicate measurements of the tested samples. The proportion of gene specific DNA was computed as the ratio of the gene specifically TpG variant DNA and either the sum of the TpG and CpG variants of this same gene or the number of GAPDH TpG variant copies. Cumulative survival was calculated by the Kaplan Meier method using SPSS. The univariate comparison between groups, statistical significance was assessed using the Cox-Mantel test. For correlation analysis, Pearson's product moment coefficient, or Spearman rank correlation and t test statistics were used. All P values are two-sided.

[0121] Results

[0122] Certain specific genes as mentioned herein were analyzed that were identified by the present inventors as fully converted by bisulfite, and thus indicating an accessible chromatin structure (see FIGS. 1 to 4). Using specific RT-PCR settings, the inventors could show that amplification of DNA only occurs when a fully bisulfite converted region is present. In case of a non-converted region, no amplification products are observed.

[0123] The data in the Figures show that the established PCR systems exclusively amplify either the entirely converted, or the entirely unconverted DNA species. No cross contamination between the two species was observed.

[0124] In order to test the reverse specificity, plasmid control systems were designed that mimic bisulfite conversion, as shown in FIG. 3. The plasmid system contained all required components to quantify the amount of copies of fully bisulfite converted CD3, FOXP3, GNLY and GAPDH gene regions as analyzed in this particular setting.

[0125] The plasmid was constructed by introducing the sequence *gcggccgc*CCTAAACACTACCACATCT*CA*AAACCCCTTAAAAAAA AC*CA*T*CA*ACCCCATAA*CA*CAAAC*CA*TAACAACTAAATTTCT*gatc*GTTTT*TG*ATTTG TTTAGATTTTTT*TG*TTATTGA*TG*TTATGG*TG*GT*TG*GATG*TG*T*TG*GGT TTTAT*TG*ATATTA*TG*GAGGAAGAGAAGAGG*c**tcgac*CCAAACCCCTACCTC*CA*CATCTA*CA*TAATAAAAACCATTAACCCTCAT*CA*ATAAATCTA*CA*TTT CCT*CA*AACCTACACTATCTAAAATTATA*CA*AAACTAATAAAAAAACAAAAT CTCTTCTATATTC*agtc*GGAATAGAGGAGAAGAGAGAGTTT*CA*TTTTTTTGGTT TTTTAGAAGGAA*CA*TGAGAATA*CA*TGTTTGTGTTGAGAGTGGGTTAGAG*CA*GTTTTAGGGTAAAGTATGTGGATA*agtc**G*GTTTT*TG*GTAT*TG*TAGGTTT*T G*GGATGTTAGTG*TG*TAG*TG*GGTGTATTTTTGTT*TG*GATGTTG*TG*TTTG*T G*GTAGAG*TG*GT*TG*TTATGTTGTAAT*TG*G*agtc*GTTTTTTTTAAAGAGTGTT TTTGATAGGGATTGTTTTAGGAATTAGGTAGGAGAGAAGGGAGTGTGAGAGGTG AAAGTTATTATTATT*ctcgag* (SEQ ID NO:1) into a pUC 57 plasmid background by using the NotI-XhoI restriction sites. The asterisks designate the potential methylated sites, small letters indicate borders of the general structure of the construct NotI-CD3-FOXP3-NKII-NKIII-GAPDH-CFF-XhoI

[0126] The quantification for a real time PCR assay is achieved by providing a standardizing plasmid, which is quantified by absorption measurement in nanodrop or alternative methods such as UVette analysis or Quibit system (Invitrogen), and the determination of the optical density.

[0127] Based on this measurement, a concentration of the plasmid is determined and a standard measurement row is made by the application of a serial dilution of the measured plasmid. By this means, a standard is prepared and determined (provided) that is exactly equimolar for all genes on the plasmid. While this absolute equimolarity is a preferred embodiment, and the inventors propose to use this standardization system for all samples, an analysis is also envisaged with a similar system, if various different standards are employed, which might be on different plasmids or even do not consist of plasmid or DNA standards.

[0128] Then biological samples were analyzed after initially detecting the fully bisulfite converted fraction of CD3, FOXP3, GLNY and GAPDH in the plasmid system.

[0129] Living cells are defined by the activity of so called house keeping genes, thus, these genes by definition must be active in all cells. It was shown in various experiments that all cells have a fully bisulfite accessible GAPDH locus (i.e. active). For this, the inventors analyzed granulocyte cells, monocyte cells, NK cells, CD4 naïve cells, CD8 cells as well as tissue from lung, uterus, breast and colon and showed that all the loci in the cells were entirely accessible to bisulfite conversion.

[0130] In order to test for cells that had accessible chromatin, the inventors analyzed the above PCR plasmid system that recognizes only fully bisulfite converted DNA. To show that no residual cells were present that have restricted access to bisulfite conversion, the inventors tried to amplify these cells with a system that was specific for non-fully converted DNA, and could not detect any signal in any sample.

[0131] Since approximately all possible DNA signals are derived from either fully converted or fully unconverted DNA, the total number of non apoptotic, non-necrotic cells can be reliably determined by measuring fully accessible GAPDH.

[0132] Next, DNA fragments were analyzed that are only transcriptionally active in particular cells. Again, the inventors analyzed the fully accessible DNA at these regions, and related them then to the amount of accessible GAPDH in the test plasmid construct.

[0133] To check for the accuracy of the data the inventors compared the two following analysis with each other:

% SCT1=copy FBC SPG/copy FBC GAPDH
to the result of the measurement/calculation
% SCT2=copy FBC SPG/(copy FBC SPG+ copy NBCSPG)
wherein
% SCT1 is the amount of the specific cell type as determined by the first method, and
% SCT2 is the amount of the specific cell type as determined by the second method.
copy FBC SPG is the copy number of the fully bisulfite converted DNA of the specific gene, copy FBC GAPDH is the copy number of the fully bisulfite converted DNA of GAPDH, and copy NBCSPG is the copy number of the non bisulfite converted DNA of the specific gene.

[0134] The analyses were repeated on whole blood samples, and gave the following data:

TABLE-US-00003 NK [%] when normalized to NK[%] normalized to NBC NK FBC GAPDH 3.5 3.6 5.5 5.3 6.4 6.4 7.2 7.6 5.3 5.6 7.1 7.2

[0135] Establishment of Cell Type Specific Gene Regions Susceptible for Complete Bisulfite Conversion and qPCR Assay Design

[0136] Bisulfite-conversion accessibility of CpG dinucleotides in the intergenic control region of the CD3D and CD3G (chr.11q23.3) genes, granulysin gene region (Chr.2 p11.2) and the CpG island in the GAPDH gene (Chr. 12 p130.31) was tested by means of bisulfite sequencing. It was found that in the CD3 proximate region all cytosines were completely converted in naïve CD4.sup.+ and CD8.sup.+ T-lymphocytes (FIG. 1B) resulting in the TpG variant only. The same region is not bisulfite converted in the other tested cell types, including granulocytes, monocytes, B-lymphocytes and NK cells resulting in the “CpG variant”. For the granulysin, it was found that the analyzed gene region exists exclusively in the CpG variant in naïve CD4 and CD8 T lymphocytes, monocytes, granulocytes and B-cells, while it appears to exist in the TpG variant in natural killer cells (FIG. 1B). The inventors exclusively found the TpG variant in all tested cell types in the analysed GAPDH region (FIG. 1A). Based on these data, PCR amplicons for the analyzed loci of the CD3, GNLY and GAPDH regions were designed. For each region, one PCR system was designed that exclusively recognizes the TpG variant template, and one PCR system that is specific for the CpG variant template, including a variant-specific fluorescence labelled detection probe for each assay (FIG. 2). Also a plasmid system for each of the three loci was constructed that corresponded to the TpG and CpG variants. The inventors showed high linearity of amplification over orders of magnitude (amplification efficiency ranged between 1.95 and 2). Also a high specificity was shown, since cross-reactivity with each TpG- and CpG-variant specific PCR system with the mutually opposite template was detected, even when tested at unphysiologically high concentration (copy numbers ranged from 20 to 12500 copies of plasmid DNA) (FIG. 2).

[0137] Characterization of Main Blood Cell Fractions with qPCR Assays for CD3, GLNY and GAPDH

[0138] The PCR systems for CD3, GNLY and GAPDH was tested on blood cell fractions that were purified according to the separation scheme published recently (Baron, U., et al., DNA demethylation in the human FOXP3 locus discriminates regulatory T cells from activated FOXP3(+) conventional T cells. Eur J Immunol, 2007. 37(9): p. 2378-89). Using serial dilutions of purified plasmids containing the equivalent of the genomic, bisulfite converted DNA regions as standard, the number of DNA copies consisting of the TpG variants was determined. As control, the CpG template variant was measured in a separate reaction. The ratio of the copies of both fractions for each gene region was calculated and is shown in Table 1.

TABLE-US-00004 TABLE 1 CD3-Assay GLNY-Assay GAPDH-Assay Copy Ratio [%] Copy Ratio [%] Copy Ratio [%] number TpG/ number TpG/ number TpG/ Immune Cell Type TpG CpG TpG + CpG TpG CpG TpG + CpG TpG CpG TpG + CpG Granulocytes CD15+ 4.3 1513.8 0.3 6 1120 0.5 0.5 100 Monocytes CD14+ 0 873 0 4 645.6 0.6 0 100 NK cells CD56+ 2.3 218.8 1 240.7 6.1 97.5 0 100 CD16+ CD3− NK T-cells CD56+ n.d. n.d. n.d. 150.5 6.3 96 232.1 0 100 CD8− CD3+ Th cells CD4+ 1009.4 1.6 99.8 11.7 435.7 2.6 772.7 0 100 Regulatory CD4+ 2726.2 3.6 99.9 0 1162.7 0 2197.4 0 100 T-cells CD25+ FOXP3+ Memory CD8+ 488.6 0.8 99.8 0 220 0 384.5 0 100 cytotoxic T- CD45RA− lymphocytes CCR7− Naive CD8+ 1859.9 4.3 99.8 159.4 614.8 20.6 1475.1 0.4 100 cytotoxic T- CD45RA+ lymphocytes CCR7+ B- CD19+ 0.6 161.9 0.4 5.8 131.9 4.2 224.1 0 100 lymphocytes

[0139] The results indicated that CD8.sup.+ and CD4.sup.+ T lymphocytes contain above 99% TpG variant for the CD3 position, while CD19.sup.+ B-cells, CD15+ granulocytes, CD14.sup.+ monocytes and CD3.sup.− CD56.sup.+ natural killer cells contain below 1% of the TpG variant and consist exclusively of the CpG variant. When an equivalent analysis was performed at the GNLY region, more than 95% of the TpG variant was observed for both CD3-CD56.sup.+ NK and CD3.sup.+ CD56.sup.+ NKT cells. CD19.sup.+ B-cells, CD15.sup.+ granulocytes, CD14.sup.+ monocytes consist exclusively of the CpG variant. 2.6% of CD4.sup.+ cells and 20.6% of CD8 memory T cells were detected as TpG variant for the GNLY locus.

[0140] Finally, the GAPDH gene region was tested for bisulfite-conversion accessibility in all named cell types. Here, specific amplification of the CpG variant failed completely in all tissues and cell types. This is consistent with the bisulphite sequencing data and the notion that this region must always be fully transcriptionally active. The data showed an efficient amplification of the TpG DNA variant for GAPDH in purified cell types (Table 1). The inventors therefore assume that this gene is optimally suitable for determination of the whole cell count in any given sample, measuring a biologically required, fully unmodified DNA stretch for controlling cell numbers.

[0141] Based on these data, the inventors intended to further prove the technical accuracy of the various qPCR systems. To do so, FACS purified regulatory T cells were selected, whose bisulfite converted DNA was shown to consist of 99.9% of the TpG variant in the CD3 locus. Granulocytes were shown to be completely inaccessible to bisulfite conversion and consisted to 99.7% of the CpG variant in the CD3 analyses. Then 40, 20, 10, 5, 3, 2 and 1% of CD3 positive regulatory T cells were spiked into a background of granulocytes, and the share of TpG variant at the CD3 locus in the background of the CpG variant was determined. As shown in Table 2, a strict correlation between the spiked samples and the CD3 PCR measurements (Pearson r=0.998) was found, corresponding with the strict correlation that was observed for the FOXP3 PCR (Pearson r=0.998) used as control comparison. The spiking experiment was performed using the GNLY assay, using purified CD3-CD56.sup.+ NK cells as TpG variant and granulocytes as CpG variant. As with the CD3 assay, the ratio of TpG variants corresponded well (Pearson r=0.98) to the expected ratios from the spiking experiment (Table 2).

TABLE-US-00005 TABLE 2 Target Calculated ratio [%] cells FOXP3-Assay CD3-Assay GNLY-Assay spiked in TpG/TpG TpG/TpG TpG/TpG [%] TpG/TpG + CpG GAPDH/ TpG/TpG + CpG GAPDH/ TpG/TpG + CpG GAPDH/ 0 0 0 0.6 0.9 0.1 0.1 1 1 1.4 1.4 1.7 0.4 0.3 2 1.7 2.5 2.5 3 0.6 0.5 3 2.6 3.9 3.6 4.5 0.6 0.5 5 4.3 6.1 5.9 6.9 1.5 1.1 10  8.7 12.2 11.4 13.4 3.1 2 20  20.5 25.8 22.1 26.5 5.6 4.2 40  37.2 46.5 42.4 47.8 13.6 9.3 Pearson 0.9985 0.9987 0.9998 0.9989 0.9958 0.9979 Correl. Coeff. [r]

[0142] Next, a plasmid was designed that contained sequences corresponding to the TpG versions of the regions in CD3, FOXP3, GNLY and GAPDH (FIG. 3). This construct is considered as the ultimate standard for quantification as it harbours all target regions in an equimolar stoichometry. Using this plasmid for normalization, the relative amount of CD3, FOXP3 and GNLY TpG variants compared to the overall cell count as determined by GAPDH TpG variant (Table 2) was re-quantified. It was shown that the results are in very good agreement with the quantification by the internal standard as well as the original dilution of the cells.

[0143] Analysis of Treg, general T lymphocytes and NK cells in whole blood samples applying epigenetic qPCR

[0144] In order to test the applicability of this method, the inventors tested whole blood samples from ovarian cancer patients that were enrolled in a catumaxumab trial, and compared the results with data obtained from flow cytometric analysis. The proportion of CD3.sup.+ cells as determined by FACS analysis and the proportion of the TpG variant as determined by CD3 qPCR, the inventors showed Spearman rank correlation and high statistical significance (R=0.80; p=7.25E-5) (Table 2). Similarly, comparison of the proportion of CD4.sup.+ CD25.sup.+ Cd127.sup.− cells obtained from FACS measurement with the proportion of TpG DNA found for the Foxp3 locus showed a strong and statistically highly significant correlation (R=0.84, p=2.13E-5). The ratio of Foxp3 to CD3 cells measured by either FACS or epigenetic analysis were also strongly correlated (R=0.7, p=0.00098). The inventors thus found a lower, but solid correlation between flow cytometrically measured CD3-CD56.sup.+ and CD3+CD56.sup.+ (NK and NKT) cells to the ratio of the TpG variant found in the granulysin locus (R=0.59, p=0.006).

[0145] QPCR Analysis of Foxp3 TSDR, CD3 and Granulysin in Solid Tumours

[0146] To provide a fully quantitative evaluation of tissue-infiltrating Foxp.sup.3+ Tregs, CD3.sup.+ T-lymphocytes and granulysin positive cytotoxic cells in healthy tissue and tumour microenvironment, fresh frozen ovarian cancer samples (n=86) were compared with healthy ovarian tissue and benign cysts (n=15) from independent donors. A statistically significant (p=2.34E-11) increase of Foxp3.sup.+ cells in the tumour (median: 1. 28%) compared to healthy controls (0.12%) was observed. Furthermore, the inventors observed a higher amount of CD3+ T cells in the tumour (median: 7.76%, n=84) than in healthy tissue (4.27%, n=15). This increase was statistically significant with p=2.56E-7. For the ratio between Foxp3 Tregs and the overall T lymphocyte count, the inventors observed a median change from 3.38% Treg of the overall T cell count in healthy tissue (n=15) to app. 19.7% in tumour tissue (n=84) (p=8.06E-07). No meaningful changes were observed for granulysin expressing cells. For bronchial carcinoma (BC), the inventors were able to compare formalin fixed, paraffin embedded (FFPE) patient-matched healthy and tumour samples. The inventors observed a strong increase of Tregs in tumour (median: 4.2%) compared to healthy tissue (mean: 2.0%). This increase was statistically highly significant for both pairwise (n=52, p=2.76E-2) and non-pairwise (n.sub.healthy=52, n.sub.tumour=91, p=9.14E-3) comparison. Here, the inventors observed a pronounced decrease of CD3 T lymphocytes in tumour (median: 22.3%) compared to normal tissue (median: 29.6%). This change was statistically significant in both pairwise (n=76, p=3.36E-3) and non-pairwise (n.sub.tumour=87, n.sub.healthy=76, p=3.4E-3) analysis. The ratio of Treg to overall CD3 cells was increased in the tumour (median: 18.3%) compared to healthy samples (median: 7.6%). This change was statistically significant in pairwise (n=48, p=6.35E-11) and non-pairwise comparisons (n.sub.tumour=86, n.sub.healthy=48, p=1.59E-14). The inventors also tested granulysin positive cells and found a pronounced decrease of these cells in tumour (median: 1.9%) compared to healthy (median: 4.8%) tissues. This difference is statistically significant, both in pairwise (n.sub.tumour=89 n.sub.normal=46, p=1.59E-14) and in unpaired comparisons (n=46, p=9.77E-9). Finally, the inventors analysed the same epigenetic parameters in colorectal cancer (CRC) samples and their adjacent healthy control tissue. The inventors observed a statistically highly significant increase of Treg counts in tumour (median: 4.2%) versus healthy (median: 1.78%) tissue (pairwise comparison: n=49, p=6.29E-4; unpaired: n.sub.tumour=49, n.sub.healthy=52, p=5.9E-4). The inventors also found a statistically significant reduction of the median overall T lymphocyte count, which was at 24.9% in the tumour and 32.3% in healthy tissue (pairwise comparison: n=61, p=2.59E-2; unpaired: n.sub.tumour=61, n.sub.healthy=69, p=2.75E-2). The median increase of the Treg to overall T-lymphocyte was highly significant (pairwise comparison: n=49, p=4.75E-7; unpaired: n.sub.tumour=52, n.sub.healthy=49, p=5.1E-7). In healthy tissue, the median number of Tregs in T-lymphocytes is at 7.8%, while in CRC tissue this ratio jumps up to 21.8%. The inventors also tested granulysin as a marker for cytotoxic immune cells, and the inventors observed a trendwise decrease of granulysin positive cells, with a median of 2.31% GNLY+ cells in tumour versus 3.0% in healthy tissue. This trend, however, was not statistically significant (pairwise comparison: n=50, p=8.43E-2; unpaired: n.sub.tumour=50, n.sub.healthy=58, p=8.04E-2).

[0147] Since Tregs are also CD3 positive, the inventors wanted to understand if there is a correlation between Foxp3 to CD3 cells in blood and healthy and tumour tissue. With not sufficient healthy ovarian tissue available for this analysis, the inventors find Spearman rank correlation rho equalling 0.47 (p=0.000, N=124), 0.68 (p=0.000, N=48) and 0.55 (p=0.000, N=49) for healthy blood, lung and colorectal tissue respectively (Table 3) between Foxp3 and CD3 TpG. Similarly, the inventors observe a correlation of rho=0.48 (p=0.000, N=86), 0.325 (p=0.325, p=0.019) and 0.76 (p=0.000, N=107) for bronchial, colorectal and ovarian cancers. According to the inventors' findings, the number of cells with an accessible granulysin locus does not significantly correlate to patient prognosis. The inventors thus conclude that the amount of regulatory T cells in the tumour microenvironment depends on the number of overall CD3 cells.

[0148] Correlating Disease Prognosis with Intra-Tumoral Immune Cell Counts

[0149] For colorectal and ovarian cancer patients follow-up data were available. Hence, the inventors tested if the measured immune cell counts within tumour microenvironment at diagnosis and surgery correlated with the prognosis of the patients. In agreement with data shown by Gallon et al., the inventors observed a statistically significant survival advantage for patients with high compared to low CD3 counts in colorectal cancer patients. For this analysis, the inventors distributed patients in two groups, one containing patients with CD3 counts below the median of 23.9% CD3 cells and the other with CD3 counts above the median. A mean survival of 75-99 months compared to 50-73 months in the 95% confidence interval and a hazard ratio of 0.58 was observed. Survival analysis for ovarian cancer patients yielded a strong, but statistically non-significant trend towards better survival for patients within the group with high CD3 counts (above 7.76%) versus those with a lower CD3 count (below 7.76%). The inventors' data indicate a statistically non-significant association of increasing Treg numbers with improved survival. Instead, the inventors show a direct linear correlation between Treg numbers and overall CD3 cell count.