Abstract
The present invention relates to the field of tumor immunology. It provides a method for identifying mutation-related human CD8+ T cells, in particular, tumor-specific T cells of a human subject, comprising analyzing CD8+ T cells of the subject by analysing the expression of at least one marker selected from a first group consisting of CD82, CD194, CD244, CD28, CD62L and CD55, and preferably, a marker selected from a second group comprising CD11a or CD18 or CD43. A preferred marker for mutation-related CD8+ T cells is CD82, which may be analysed in combination, e.g., with CD11a. Without the need to identify any epitope to which T cells reacts, this method can advantageously be used to isolate the entire individual pool of mutation-related T cells, and, optionally, to identify the sequence of a mutation-related TCR, which allows for generation of transgenic T cells expressing the TCR. Compositions substantially comprising tumor-specific CD82.sup.hiCD8+ T cells and/or CD194.sup.hi, CD244.sup.−, CD28.sup.+, CD62L.sup.+ and/or CD55.sup.+ CD82.sup.hi CD8+ T cells can be used for treatment of a cancer patient, e.g., by adoptive T cell transfer. The method of the invention can also be used for diagnostic purposes to identify human mutation-related T cells or diagnosing a tumor disease or for testing responses of a cancer patient to an immune stimulatory therapy, preferably, a therapy with a checkpoint inhibitor.
Claims
1. A method for providing human mutation-related CD8+ T cells, comprising a) assaying CD8+ T cells from a subject for expression of at least one marker selected from a first group consisting of CD82, CD194, CD244, CD28, CD62L and CD55, wherein CD82.sup.hi cells, CD194.sup.hi cells, CD244− cells, CD28.sup.+ cells, CD62L.sup.+ cells, CD55.sup.+ cells, or any combination thereof are identified as mutation-related CD8+ T cells, and b) purifying the mutation-related CD8+ T cells.
2. The method of claim 1, comprising assaying for the expression of CD82 of CD8+ T cells of the subject, wherein CD82.sup.hi cells are identified as mutation-related CD8+ T cells.
3-7. (canceled)
8. The method of claim 1, further comprising analysing the expression of a marker selected from a second group consisting of CD11a, CD18 and CD43 of the CD8+ T cells, wherein cells identified as mutation-related CD8+ T cells according to at least one marker from the first group that are further CD11a.sup.int, CD18.sup.int, and/or CD43.sup.int are identified as mutation-related cells.
9. The method of claim 8, wherein cells that are both CD11a.sup.int cells and CD82.sup.hi, CD194.sup.hi, CD244, CD28.sup.+, CD62L.sup.+ cells, and/or CD55.sup.+ cells are identified as mutation-related cells.
10. The method of claim 9, comprising analysing the expression of CD11 a of the CD8+ T cells, wherein CD82.sup.hiCD11a.sup.int cells are identified as mutation-related cells.
11-15. (canceled)
16. The method of claim 8, wherein cells that are both CD18.sup.int cells and CD82.sup.hi, CD194.sup.hi, CD244−, CD28.sup.+, CD62L.sup.+ cells, and/or CD55.sup.+ cells are identified as mutation-related cells.
17-22. (canceled)
23. The method claim 8, wherein cells that are both CD43.sup.int cells and CD82.sup.hi, CD194.sup.hi, CD244−, CD28.sup.+, CD62L.sup.+ cells, and/or CD55.sup.+ cells are identified as mutation-related cells.
24-29. (canceled)
30. The method of claim 1, wherein the mutation-related T cells are selected from the group consisting of tumor-specific T cells, smoking-related T cells, and sunburn-related T cells.
31. The method of claim 1, wherein the subject is not a smoker and not a sunburned subject, wherein the mutation-related T cells are identified as tumor-specific T cells.
32. The method of claim 1, wherein the method comprises contacting cells with antibodies to CD82, CD194, CD244, CD28, CD62L and/or CD55.
33. The method of claim 32, wherein the method comprises contacting cells with antibodies to CD11a, CD18 and CD43.
34. The method of claim 32, wherein the antibodies are labelled, and wherein the method comprises flow cytometry.
35. (canceled)
36. A method for providing a nucleic acid encoding a TCR of a human mutation-related CD8+ T cell, comprising providing a human mutation-related CD8+ T cell according to the method of claim 1 and identifying the sequence of the TCR of the mutation-related CD8+ T-cell.
37. A method for providing a mutation-related CD8+ T cell, comprising providing a human mutation-related CD8+ T cell according to the method of claim 1 identifying the sequence of the TCR of the mutation-related CD8+ T-cell, and expressing the TCR in a CD8+ T cell.
38. The method of claim 1, comprising providing a composition comprising at least 90% of mutation-related human T cells.
39. The method of claim 1 further comprising activating the mutation-related CD8+ T cells.
40. A composition comprising at least 90% of human T cells that are a) CD8+ and b) CD82.sup.hi and/or CD194.sup.hi and/or CD244 and/or CD28.sup.+ and/or CD62L.sup.+ and/or CD55.sup.+ and c) CD11a.sup.int and/or CD18.sup.int and/or CD43.sup.int.
41. A method of treating a cancer patient, comprising administering an effective amount of the composition of claim 40 to the cancer patient, wherein the mutation-related CD8+ T cells are derived from the patient, wherein mutation-related CD8+ T cells isolated prior to treatment of the patient with a checkpoint inhibitor are adoptively transferred to the patient after development of hyperproliferative disease, and the treatment with the checkpoint inhibitor is halted.
42-43. (canceled)
44. A method for treating cancer in a subject, comprising a) treating a subject with a checkpoint inhibitor, b) assaying CD8+ T cells from the subject for the expression of at least one marker selected from a first group consisting of CD82, CD194, CD244, CD28, CD62L and CD55 of CD8+ T cells of the subject, wherein CD82hi cells, and/or CD194hi cells, and/or CD244− cells, and/or CD28+ cells, and/or CD62L+ cells, and/or CD55+ cells, or any combination thereof are identified as mutation-related CD8+ T cells, and c) continuing treatment with the checkpoint inhibitor if there is an expansion of mutation-related CD8+ T cells.
45. The method of claim 38, further comprising isolating a tumor-specific T cell clone.
46. The method of claim 1, comprising assaying the CD8+ T cells for the expression of CD244, wherein CD244− cells are identified as mutation-related CD8+ T cells.
47. The method of claim 1, wherein the CD8+ T cells are obtained from a blood sample from the subject.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0073] FIG. 1Aa shows vectors used in experiments and results already described in detail in Knocke at al. Cell Rep 2016; 17:2234-46. Ab depicts liver tumors expressing antigens included in said vectors in the absence of T cell tumor-immunogenicity, i.e., after injection of the control vector 1 (1/top)) and in the presence of potent T cell tumor-immunogenicity elicited by vector 2 (2/bottom), respectively. Ac shows dot plots of tumor-specific CD8 T cells of OVA and Spnb2-R913L elicited by vector 2. In Fig. Ad H and E (Hematoxylin and eosin) stained liver sections reveal massive lymphocytic tumor infiltration in tumors with potent immunogenicity elicited by vector 2. FIG. 1B shows differences in CD8 expression levels between murine tumor-specific CD8+ T cells elicited by vaccination with DC and by cancer immunosurveillance.
[0074] FIG. 2 shows differences in CD8 and CD11a expression levels on T cells derived from tumor-specific CD8+ T cells elicited by vaccination with DCs as described in Woller et al., J Clin Invest. 2011; 121(7):2570-2582, by virus-induced T cells, and by tumor-induced T cells. (A), CD8 and CD11a expression levels on T cells derived from a liver tumor in an overlay with tumor-specific T cells, which confirms the tumor specificity of the CD11a.sup.hi and CD8a.sup.hi T cells. (B), cytokine expression (C) and life cycle of these T cells (D). FIG. 2E confirms that the marker shown is also specific for pentamer-stained tumor-specific T cells. Thus, murine tumor-specific T cells have a unique phenotype.
[0075] FIG. 3 documents the experiment described in more detail below for marker-identification of tumor-specific human CD8+ T cells (A-D). FIG. 3E shows identification of a population of CD82.sup.hiCD11a.sup.inCD8+ T cells (bottom blots) specifically present in a regressing cancer patient, but not in a healthy donor, in comparison to the isotype control (for CD82 staining, top plots). The CD82.sup.lo, CD82.sup.in and CD82.sup.hi populations can be easily distinguished.
[0076] FIG. 4 shows the presence and frequency of CD82.sup.hiCD11a.sup.inCD8+ T cells in healthy control subjects (A, B) and in the regressing cancer patient (B). There are generally between 1-3% CD82.sup.hi positive CD8+ T cells in healthy donors.
[0077] FIG. 5 shows the frequency and presence of CD82.sup.hiCD11a.sup.inCD8+ T cells in smokers and non-smokers (A, B). C shows the exceptional presence of an untypically high frequency of CD82.sup.hiCD11a.sup.inCD8+ T cells in a non-smoker not diagnosed to have a tumor.
[0078] FIG. 6 compares cancer patients (no CPI treatment) with voluntary donors (smokers and non-smokers) with regard to CD82.sup.hiCD11a.sup.inCD8+ T cells.
[0079] FIG. 7 shows an increase in CD82.sup.hiCD11a.sup.inCD8+ T cells in cancer patients before and under CPI treatment, wherein the patients respond to said treatment (over plots: patient number+time of analysis relative to CPI treatment) (A). B shows the increase in marker-positive (CD82.sup.hiCD11a.sup.inCD8+) T cells, wherein the peak numbers of these cells after treatment assessed before staging (i.e., before CT assessment of tumor size) are compared to pre-treatment levels, for 10 liver cancer patients under CPI treatment. CPI treatment was second or third line treatment. As first line treatment, most patients had been treated with sorafenib. CPI treatment was with nivolumab (patients treated every 2 weeks) or with pembrolizumab (patients treated every 3 weeks). Patients are grouped for their response to treatment (NR=non-responder, MR=mixed response, SD=stable disease, R=responder (as found in CT, grouped according to response evaluation criteria in solid tumors (RECIST)). Only one patient was a responder, i.e., in this patient group, the overall response rate was 10%. This patient had an increase in marker-positive cells of more than 12%, in particular, 16.6%.
[0080] FIG. 8 shows the high frequency of CD82.sup.hiCD11a.sup.inCD8+ T cells in a 83 year old patient (Patient 43) which is shown to be stable over more than a year. The patient has an exceptional long survival time nearly 20 years post HCC diagnosis, as described in detail below.
[0081] FIG. 9A shows the relative stability of the high frequencies of CD82.sup.hiCD11a.sup.inCD8+ T cells seen in a smoker. In contrast, increased numbers of CD82.sup.hiCD11a.sup.inCD8+ T cells due to a dermatitis solaris rapidly contracted (FIG. 9B, C). UV-treatment in the absence of sunburn did not increase the level of CD82.sup.hiCD11a.sup.inCD8+ T cells.
[0082] FIG. 10 shows an analysis of the memory phenotypes of CD8+ T cells in the CD11a/CD82 plots.
[0083] FIG. 11A shows that mutation-related T cells identified according to the method of the invention have a different phenotype with regard to cytokine secretion. CD8+ T cells activated with PMA/lonomycin from a HCC patient could be separated into different groups. Naïve CD8+ T cells do not secrete tumor necrosis factor alpha (TNF-α) nor interferon gamma (IFN-γ). CD8+ CD11a.sup.hi CD82.sup.lo cells stain double positive in the intracellular cytokine staining for IFN-γ and TNF-α, whereas CD8 CD11a.sup.int CD82.sup.hi T cells are mostly TNF-α positive or double negative and secrete very little IFN-γ.
[0084] Another HCC patient showed a positive cytokine response to CMV (cytomegalovirus) (FIG. 11B). CD8+ T cells were stimulated with a CMV-peptide and gated for IFN-γ and TNF-α (left blot) and then analyzed for CD11a and CD82 (right top blot). Upon gating, the virus-specific cells positive for both IFN-γ and TNF-α were restricted to CD8 CD11a.sup.hi CD82.sup.lo section.
[0085] FIG. 12 shows contraction of the CD82.sup.hiCD11a.sup.inCD8+ T cells pool in a patient with hyperprogressive disease under CPI treatment.
[0086] FIG. 13 shows that CD194 also is a marker for mutation-related, in particular, tumor-specific CD8+ cells. Of the CD8+ cells gated, in the naïve gate, there are 7% CD194+ (or CD194.sup.hi) cells (A), in the pathogen-specific gate, there are 9.9% CD194+ cells (B), and in the tumor-specific gate, there are 84.8% CD194+ cells (C).
[0087] FIG. 14 shows that CD18 is coexpressed with CD11a, and can thus be used as an alternative marker. Of the CD8+ cells gated, in the naïve gate, there are 99.9% CD11.sup.in and CD18+ (i.e., CD18.sup.int) cells (MFI CD18: 3215, MFI CD11a: 1630) (A), in the pathogen-specific gate, there are 26.5% CD11.sup.in and CD18+ (i.e., CD18.sup.in) cells (MFI CD18: 22.256, MFI CD11a: 5956) (B), and in the tumor-specific gate, there are 98.1% CD11.sup.in and CD18+ (i.e., CD18.sup.in) cells (MFI CD18: 4950, MFI CD11a: 2481) (C). The ratios are CD11a: N (naïve)/P (pathogen-specific): 0.288, CD18: N/P 0.272, CD11a T (tumor-specific)/P: 0.417, CD18 T/P: 0.444. It is noted that both CD11a and CD18 are expressed at an intermediate level on the tumor-specific CD8+ T cells, while in the pathogen-specific CD8+ T cells, expression of both markers is high.
[0088] FIG. 15 shows expression of further markers for mutation-related, in particular, tumor-specific CD8+ cells. Expression is shown for each marker in naïve cells, pathogen-specific cells and tumor-specific cells. A: CD244, B: CD28, C: CD52L, D: CD55, E: CD43.
EXAMPLES
1. Animal Experiments
[0089] The mouse tumor model used for the analysis of T cell phenotypes has been described before [Knocke et al. Cell Rep 2016; 17:2234-46]. In brief, FIG. 1Aa shows transposon expression vectors to generate liver tumors that encoded epitope tags of highly immunogenic CD4 and CD8 T cell epitopes, such as Spnb2-R913L, SIINFEKL and both MHC-class II restricted OVA epitopes in combination with oncogenic Nras (including the G12V-mutation). Nras alone served as control. The tumor progression was severely inhibited in a T cell-dependent manner compared to the control group transfected with vector 1 (FIG. 1Ab). Tumor-specific T cells were detected with a pentamer (FIG. 1Ac). The HE histology shows signs of strong lymphocytic infiltration into the immunogenic tumors (FIG. 1Ac).
[0090] To further analyze tumor-specific T cell responses in this model with an emphasis of the T cell phenotype, a T cell response against Spnb2-R913L elicited by a DC vaccine (described in [Woller et al., J Clin Invest. 2011; 121(7):2570-2582]) was compared with a T cell response against the same specificity, but derived from a naturally elicited response in the mouse tumor model as described above (FIG. 2A). The results of flow cytometry using a Spnb2-R913L-specific pentamer show a decrease in CD8 expression in tumor-specific T cells when triggered by DC vaccination and unaltered expression levels of CD8 on these cells in mice with Spnb2-R913L-expressing tumors. Downregulation of CD8 in antigen-experienced T cells has been described by Ray et al. [J. Immunol. 2009; 183:7672-82]. The authors of this study used CD11a additionally to identify CD8.sup.low CD11a.sup.hi T cells as antigen-experienced cells. The present data, as shown in FIG. 2A, demonstrate that this phenotype complies with the phenotype of T cells derived from DC vaccination and virus-specific T cells. However, CD8+ T cells derived from immunogenic tumors display a different phenotype with CD8.sup.hi CD11a.sup.hi expression. Here, CD11a expression is reproducibly higher than CD11a expression on T cells derived from DC vaccination/viral infections. FIG. 2B shows a density plot in which CD8.sup.lo CD11a.sup.hi tumor-derived T cells are distinguishable from CD8.sup.lo CD11a.sup.hi pathogen-derived T cells, demonstrating that both populations do coexist and are not attributable to staining artifacts of flow cytometry (left plot). The plot on the right side proves that Spnb2-R913L-specific cells only occur in the CD8.sup.hi CD11a.sup.hi area (FIG. 2B).
[0091] Analysis of cytokine secretion of the respective T cells confirms the difference in the phenotype of T cells dependent on the kind of antigenic stimulation. Tumor-induced T-cells stain double positive for Interferon gamma and tumor-necrosis factor alpha at a significantly lower rate compared to virus-specific and DC-induced T cells. Each dot represents an individual mouse in the right graph of FIG. 2C.
[0092] Observing the kinetic of DC-induced T cells in a time course of 42 days showed that DC-induced T cells peak on day seven and contract rapidly until day 14. In contrast, the kinetics of tumor-induced T cells differ markedly. They appear to expand more slowly, but at a higher magnitude, and apparently more than one expansion is a naturally occurring phenomenon (FIG. 2D).
[0093] Thus, the data allows for the conclusion that tumor-induced T cells have a different phenotype than DC-induced T cells or pathogen-induced T cells.
[0094] To test if there are specific markers on CD8+ tumor-specific T cells to differentiate the phenotype of these cells from other CD8 T cells, a biomarker screen was performed. To this end, an antibody-based screen with flow cytometry of surface markers was performed. The source of screening material consisted of peripheral blood cells from mice that were injected with transposon plasmids as described above and that survived about 55 days. Those mice developed CD8.sup.hi CD11a.sup.hi T cells and CD8.sup.loCD11a.sup.hi T cells that were additionally stained with one of 243 PE-labelled murine screening antibodies (Legend Screen Kit (Biolegend)). The T cell population gated to CD8.sup.lo CD11a.sup.hi T cells, i.e., pathogen-specific CD8+ T cells, is depicted as a histogram of the corresponding screening marker with a grey line in the inlays. CD8.sup.hi CD11a.sup.hi T cells, i.e., tumor-specific T cells, are shown in the histogram with a black line. The histograms in FIG. 2E shows that the tumor-specific T cells stained by the pentamer are also positive for a marker not expressed by the pathogen-specific CD8+ T cells. Therefore, CD8 T cells from tumor-bearing mice were stained with a Spnb2-R913L-specific pentamer and the corresponding biomarker. As shown in FIG. 2D, cancer-specific CD8 T cells are sufficiently identified by a single staining with the biomarker alone, as double staining with the pentamer identifies tumor-specificity in each case. Thus, it is possible to identify tumor-specific T cells by reference to their phenotype based on surface protein expression in a TCR-agnostic manner.
2. Human Experiments—Marker Identification
[0095] After obtaining clear evidence for a specific phenotype of cancer-induced CD8+ T cells from a tumor mouse model, a similar screen was performed in human samples. A 40 years old male patient with a NSCLC (non-small cell lung cancer) stage IV suffering a progress after first line chemotherapy was treated with the PD-1 checkpoint inhibitor pembrolizumab. As shown in FIG. 3A, the primary tumor regressed markedly after four infusions of pembrolizumab. Additionally, bone-, liver-, and brain metastasis underwent a complete remission and tumor markers, such as CRP and LDH, normalized (FIG. 3B). At this time point, blood samples were obtained from the patient und compared with a healthy donor (FIG. 3C). The screening was done with 371 PE-labelled anti-human antibodies from the Legend Screen Kit (Biolegend) according to the manufacturer's recommendations (FIG. 3D). The following antibody panel from the same company was used: CD8a FITC (clone HIT8a), CD4 PerCP-Cy5.5 (clone OKT4), and CD11a APC (clone TS2/4). The screening revealed CD82 as one promising candidate among all screened surface markers. In FIG. 3E, the density plots of gated CD8+ T cells show CD11a on the X-axis and CD82 on the Y-axis. For CD82, a plot of an isotype is shown as well. The plot of the regressing cancer patient has an additional population with high expression of CD82 and intermediate expression of CD11a which is absent in the healthy control. Interestingly, in this population CD11a, with an intermediate expression level, also differs from the CD11a.sup.hi and CD11a.sup.lo population. Further characterization of CD11a.sup.lo T cells identified these cells as naïve T cells. Moreover, CD11a.sup.hi cells are antigen-experienced T cells (FIG. 10) and there is strong evidence that they are pathogen-specific as shown below. The CD8+ CD11a.sup.int CD82.sup.hi T cells are further characterized in the following experiments.
3. Analysis of Healthy Subjects and Cancer Patients
[0096] In the next step, a cohort of 70 healthy individuals volunteered in this study and donated blood to allow for characterization of expression patterns of CD8+ CD11a.sup.int CD82.sup.hi T cells. In FIG. 4A, representative density plots of these individuals are shown. In all figures, the plots are numbered in a consecutive manner for reference. Most healthy donors (59/70) had a CD8+ CD11a.sup.int CD82.sup.hi T cell population that accounted for three percent or less of the total CD8+ T cell pool. This is referenced as the normal value in healthy individuals. FIG. 4B shows the data compared to a regressing cancer patient for the purpose of comparison. Since 11 healthy donors had higher values, in part considerably higher, a subgroup analysis was performed.
[0097] In the subgroup analysis of healthy donors, non-smokers and smokers were separately investigated. There was a wide spread of individual smoking habits with regard of frequency and duration. Consequently, the smoking pack years (s.p.y.) were applied for each individual value as an indicator for the biological smoking burden (one s.p.y. equals one package of cigarettes per day for one year). Presenting the results in this way revealed a significant (p<0.0001) difference within these groups. 77% of smokers had elevated levels of CD8+ CD11a.sup.int CD82.sup.hi T cells, whereas <2% of the non-smoking arm was positive (NS=non-smoker) (FIG. 5A). This strongly indicates that consistent exposure to mutagens give rise to elevated levels of CD8 CD11a.sup.int CD82.sup.hi T cells. FIG. 5B shows representative CD82/CD11a blots for smokers. FIG. 5C shows the only non-smoker in this analysis having a proportion of more than 3% CD8+ CD11a.sup.int CD82.sup.hi T cells.
[0098] Next, blood from cancer patients was investigated to assess the frequency of CD8+ CD11.sup.int CD82.sup.hi T cells. These experiments were in accordance with the ethical guidelines of Hannover medical school. Patients, mainly suffering from hepatocellular carcinoma (HCC), donated a blood sample prior to treatment with checkpoint inhibitors (CPI) nivolumab or pembrolizumab (FIG. 6A and B).
[0099] Representative samples of patients treated with CPI are shown in FIG. 7A. Of three responding cancer patients (13, 37, 38) samples were unavailable prior to treatment. High levels of putatively cancer-specific T cells (based on the CD8+CD11a.sup.int CD82.sup.hi phenotype) were detectable in these patients. Patients 39 to 42 show an expansion of CD8 CD11a.sup.int CD82.sup.hi T cells following immunotherapy within a time frame of two to nine weeks. These results demonstrate that checkpoint inhibition leads to an expansion of these cells in cancer patients. The graph in FIG. 7B depicts the changes of tumor-specific CD8 T cells between prior to therapy and the highest value of a measurement in a temporal vicinity of the first staging during therapy, which was usually about 12 weeks. Results of 10 patients were available. The ORR in this study is 10%.
[0100] FIG. 8 illustrates the tumor-specific T cell measurement of an extraordinary case report of an 83 year old patient with an exceptional history of HCC. The median overall survival time of 518 untreated HCC patients was 3.6 months and survival times were 13.4, 9.5, 3.4, and 1.6 months for patients of Barcelona Clinic Liver Cancer stages 0/A, B, C, and D, respectively. (Khalaf at al., Clinical Gastroenterology and Hepatology 2017; 15:273-2). HCC has a 5-year survival rate of 18%. Unresectable HCC remains an incurable disease (Chen et al., Oncology Letters 15: 855-882, 2018)
[0101] The present patient was diagnosed in the year 2000 with HCC. The tumor was resected four month later. He left the clinic and returned almost seven years later with a relapse. This is very uncommon, since the 5-year survival rate of HCC is only 18%. The tumor was then treated with transarterial chemoembolization. Additionally, he was also enrolled into a study for a treatment with a peptide vaccine, which turned out to be negative. Again, close to seven years later he showed up with a relapse. The single tumor that progressed in the meantime was then treated with RFA (radiofrequency ablation, i.e. local boiling of tumor tissue with electromagnetic radiation). Two years later he had another relapse that was treated likewise. At that time point the first blood sample was available. One year later, the follow up identified small HCC lesions (time point of second blood sample). This patient is still alive and now approaches his 20.sup.th year post HCC diagnosis. Among oncologists, this has been described as a unique case.
[0102] The frequency of CD8+ CD11a.sup.int CD82.sup.hi T cells in the sample of this patient was found to be >65%, and it was stable over the period of one year between both blood withdrawals (FIG. 8). This result strongly correlates with the unusual survival of the patient and the high frequency of the putative cancer-specific T cells.
[0103] As described above, smokers have a higher rate of occurrence of CD8+ CD11a.sup.int CD82.sup.hi T cells. In one case it was possible to follow up this frequency throughout an extended period of time. Smoker 22, the volunteer with the highest rate of putative cancer-specific T cells, had a stable pool of these cells in a time frame of over three years (FIG. 9A). In contrast, two donors (donor 1 and 2) of the control arm of this study suffered a dermatitis solaris after extended sun exposure in Australia. Two weeks after this event, a blood sample was obtained and increased levels of CD8+ CD11a.sup.int CD82.sup.hi T cells were detected. Moreover, intensity of the sunburn correlated with the mutation-related T cell magnitude. Within a matter of weeks, these cells contracted and went back to normal levels.
[0104] Donor 3, a 9 year old girl, served as a control. This donor suffers a very rare pediatric disease of CD8 mycosis fungoides, a cutaneous T cell lymphoma that is treated with 311 nm narrow band UV radiation. The second sample of donor 3 was obtained immediately after 30 treatments of UV irradiation. As shown in FIG. 9B, UV treatment per se does not induce CD8 CD11a.sup.int CD82.sup.hi T cells. Experienced dermatologists scrupulously took care by adjusting the UV intensity, that dermatitis solaris was avoided during treatment to prevent long term complications. Of note, donor 3 was not treated by a checkpoint inhibitor therapy.
[0105] Taken together, these results show that sunburn induces a CD8+ CD11a.sup.int CD82.sup.hi T cell pool that rapidly contracts within weeks. This is an indication that mutagenic events caused by increased UV exposure with resulting dermatitis solaris specifically may trigger this cell type.
[0106] The memory phenotypes of CD8+ T cells in the CD11a/CD82 plots were investigated in FIG. 10. Expression levels of CD45RA and CCR7 are depicted in the upper right plot for all CD8 T cells. According to this analysis, T cells can be divided into naïve T cells (CD45RA+ CCR7+), central memory T cells (CD45RA− CCR7+), effector memory T cells (CD45RA− CCR7−), and effector memory RA+ T cells (CD45RA+ CCR7−). The source of this analysis is patient 13, and the results were validated in other patients and smokers. Expression of CD45RA and CD45RO are mutually exclusive, hence only one of these CD45 variants are expressed on the cell surface. CD8+ CD11a.sup.lo CD82.sup.lo T cells (lower left plot) express CD45RA+ CCR7+ and are thus naïve T cells by definition. CD8+ CD11a.sup.hi CD82.sup.lo split into two groups. In the central plot, these two groups can be distinguished as two separate populations. However, this resolution is not seen in every individual, as these populations mostly appear as one. However, the lower population is to a great extend CD45RA+ and thus CD45RO negative. With increasing CD82 expression levels, the count of CD45RA+ cells decreases in the upper population of CD8+CD11a.sup.hi CD82.sup.lo cells. Cells with highest expression levels of CD82 in cancer patients (the CD8+ CD11a.sup.int CD82.sup.hi pool) are negative for CD45RA. CD45RA has been described as a marker that correlates with the absence of the corresponding antigen on memory cells [Rovaris at al., Blood 2008; 108(9):2897-905].
[0107] This is additional evidence that CD8+ CD11a.sup.int CD82.sup.hi T cells are cancer-specific cells, because continuous presence of a T cell antigen in a cancer patient inhibits, as evident in this study, development of CD45RA expression. As shown in FIG. 9, rapid CD8+ CD11a.sup.int CD82.sup.hi T cell contraction was observed in individuals with solaris dermatitis. Additionally, absence of CD8+ CD11a.sup.int CD82.sup.hi T cells was found in a triple negative breast cancer patient 8 weeks after complete tumor remission following nivolumab treatment (data not shown), which indicates that T cells may rapidly contract after complete clearance of target cells. Until today, this woman is under frequent follow-up with CT scans and still found to be tumor free. These results indicate that tumor-specific T cells may contract upon clearance of the antigen, instead of generating a memory pool. This has important implications for the development of this method for diagnostic purposes, as it decreases the risk of false positives due to a past tumor disease already cleared by the immune system.
[0108] Activating CD8+ T cells with PMA and lonomycin leads to a general activation and cytokine secretion in an antigen-independent manner. Activating CD8+ T cells from a responding cancer patient allows for assessment of the cytokine profile of all three populations described above. FIG. 11 demonstrates that naïve CD8+ T cells, as expected, do not secrete tumor necrosis factor alpha (TNF-α) nor interferon gamma (IFN-γ). CD8+ CD11a.sup.hi CD82.sup.lo cells stain double positive in the intracellular cytokine staining for IFN-γ and TNF-α, whereas CD8 CD11a.sup.int CD82.sup.hi T cells are mostly TNF-α positive or double negative and secrete very little IFN-γ. This analysis show that CD8+ CD11a.sup.hi CD82.sup.lo and CD8 CD11a.sup.int CD82.sup.hi T cells differ greatly in their cytokine secretion profile (FIG. 11A).
[0109] In an antigen-specific approach, a HLA-A2+ HCC patient showed a positive cytokine response to CMV (cytomegalovirus) (FIG. 11B). CD8+ T cells were gated for IFN-γ and TNF-α (left blot) and then analyzed for CD11a and CD82 (right top blot). As shown by the comparison with the CD8+ T cells not gated for cytokine production (right bottom blot), the virus-specific cells producing both IFN-γ and TNF-α were only located in the CD8 CD11a.sup.hi CD82.sup.lo section and not in the CD8 CD11a.sup.int CDB2.sup.hi area. Hence, virus-specific CD8 T cells are excluded from the tumor-specific area in this experiment. This is further evidence that CD8 CD11a.sup.int CD82.sup.hi T cells account for tumor-specificity.
[0110] The plots of an HCC patient depicted in FIG. 12. It shows a case of hyperprogressive disease. Prior to CPI, an amount of about 16% of tumor-specific CD8 T cells were detectable. Upon treatment start, that pool contracted rapidly within the following weeks. She died prior to the 6a treatment week. In case of a contraction of the pool of tumor-specific CD8+ T cells under CPI therapy, the CPI treatment should be discontinued. If possible, tumor-specific T cells, preferably, isolated from the patient before start of the CPI treatment and cultivated in vitro may be useful in adoptive T cell therapy of such patients.
4. Analysis of Further Surface Makers for Mutation-Related Human CD8+ T Cells
[0111] In a further screen, as shown in FIG. 13, CD194 (ab clone L291H4) was identified as an alternative or additional marker for mutation-related human T cells. Further, FIG. 14 shows that CD11a and CD18 (ab done TS1/18) are substantially co-expressed, so that any of them can be used as an additional marker for mutation-related human T cells in combination with CD82 and/or CD194.
[0112] Further screening identified CD244 (ab done C1.7), CD28 (ab done CD28.2), CD82L (ab done DEREG-56) and CD55 (ab done JS11) as an alternative or additional marker for mutation-related human T cells. Additionally, CD43.sup.k expression (ab done CD43-10G7) was found to be an additional or alternative marker for mutation-related T cells, similar to CD11 or CD18 expression (FIG. 15).
