EP4069869A1 - Method to predict the response to cancer treatment with anti-pd1 immunotherapy - Google Patents

Abstract

Biomarkers, in vitro methods and kit or dispositive of predicting or prognosticating the response of a human subject to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy.

Description

Method to predict the response to cancer treatment with anti-PD1 immunotherapy.

DESCRIPTION

TECHNICAL FIELD OF THE INVENTION

The present invention relates to the field of medicine, and more specifically to the field of precision medicine. It refers to a series of functional biomarkers for response to anti-PD1 treatment, which can be implemented to make therapeutic decisions, in a method to predict the response to said treatment.

BACKGROUND OF THE INVENTION

Immune checkpoint blockade (ICB) has been clinically applied in multiple cancers and revolutionized the field of cancer treatment. It has demonstrated acceptable toxicity and durable response in responders. Programmed death 1 (PD1) blockade is a common ICB treatment in advanced melanoma patients, as second or third line treatment. PD1 is expressed on T lymphocytes and is the dominant inhibitory immune checkpoint for maintaining the self-tolerance of T cells. Activated PD1 signalling restrains the T cell cytotoxicity towards cancer cells (Ribas, 2012). It negatively affects chemokine and cytokine production, as well as the proliferation of CD4+ and CD8+ T cells. Treatment with Nivolumab, a PD1 antibody, yields 20% higher objective response rate than chemotherapy in advanced melanoma patients with better tolerability (Falchook, 2015).

In recent pivotal trials, immune checkpoint blockade (ICB) demonstrated durable responses and acceptable toxicity, resulting in the regulatory approval of 8 checkpoint inhibitors to date for 15 cancer indications. However, up to -85% of patients present with innate or acquired resistance to ICB, limiting its clinical utility.

In addition to the subsequent impact on patient survival, the identification of resistance biomarkers for patient selection is critical given the escalating costs of this type of treatment, which currently ranges from 50,000 to 150,000$/quality-adjusted life year (Verma et al. , 2018). So far, only PD-L1, quantified by immunohistochemistry stands as an FDA/EMA approved biomarker of response to ICB, for selection of candidates to be treated with monotherapy anti-PD1 in NSCLC and recurrent or metastatic squamous cell carcinoma. But, even though there are 4 IHC assays approved by FDA, protein PD-L1 expression fails to accurately predict the response to ICB in some cases (Friedrich, 2019). The exploration of other molecular and cellular biomarkers of response is still limited, with relatively modest and retrospective cohorts. Thus, identification of novel, more predictive biomarkers that could identify patients who would benefit from ICB constitutes one of the most important areas of immunotherapy research

However, a portion of patients are insensitive or present even with hyper- progressive phenotype due to the primary or developed resistance (Sharma et al., 2017), and hence the rational use of PD1 blockade in melanoma is limited by the lack of robust biomarkers.

DESCRIPTION OF THE FIGURES

Figure 1. Association of gene expression levels with overall survival in cutaneous melanoma. Kaplan-Meier analysis of overall survival (B) in 16 patients with metastatic cutaneous melanoma treated with Nivolumab. The most significant associations of gene expression levels and overall survival are shown.

Figure 2. Tumour infiltrating cell component analysis by CiberSortX. Out of all cellular types present in the tumor stroma of the 16 metastatic melanoma patients, good and bad responders presented different distributions. Good responders were significantly enriched in plasmablasten, while bad responders were significantly enriched in CD8+ exhausted T cells, and BC1 subtype of B cells.

Figure 3. Heatmap that represent the expression of the different genes differentially expressed according to the study sample. As we can see in Figure 3A, in which the p value adjusted is less than 0.05, there is a large number of DE genes and a certain grouping can be seen in the expression of some genes in some non-responders and responders, but it is not described clearly. However, in Figure 3B, in which the adjusted p-value has been limited to less than 0.01, an expression pattern is clearly seen in some subsets of genes (orange boxes) in a group of non-responders (green box) and responders (purple box).

Figure 4. The total number of receptor clonotypes is associated to a good response to Nivolumab. A) Quantification of the sum of BCR and TCR clonotypes in good vs bad responders. B) Stratification by type of cell.

Figure 5. The abundance in HLA loci is associated to a good response to Nivolumab.

A) Quantification of the HLA loci in general in good vs bad responders. B) Stratification by type of Class I, Class II, and Non-Class loci. Figure 6. HLA, TCR and BCR abundance and diversity is higher in good responders to Nivolumab. Representative depictions of the abundance and diversity of HLA, TCR, and BCR, based on bulk RNA-seq data. Each concentric circle and colour represents a variant, and the covered angle of the circumference indicates the amount of the specific variant. A) HLA diversity and abundance; B) Comparative abundance and diversity of TCR vs BCR in good responders; C) Comparative abundance and diversity of TCR vs BCR in bad responders.

Figure 7. The integration of bulk RNA-seq data and the BCR and HLA abundance.

Circle plot with the heatmap of the DE expressed genes in good versus bad responders (external circle), the total BCR clonotypes sum (yellow squares), and the heatmap depicting the amount of HLA loci in each patient (internal circle).

Figure 8. Validation of 35 out of the 140 DE genes that constitute our firm in another cohort of melanoma patients treated with anti-PD1. In yellow, those genes that also form part of the prognostic signature. In red, a gene that is among the most common BCR clonotypes in good responders.

Figure 9. Multispectral fluorescence validation of the B cell CD19 biomarkers as a response marker to anti-PD1. A) CD 19 is significantly associated with response in a cohort subgroup of 12 patients, specifically plasmablasts (CD19+, CD20-, CD138-) show a trends towards significant association with responders. B) Lymphoid structures containing B cells, CD8 T cells, macrophages and myeloid cells are identified in the tumors of some good responders (representative image).

Figure 10. Technical validation of the RNA-seq by using RQ-PCR. A. Pearson correlation between TaqMan target genes. B. Scatterplot with correlations between RQ-PCR and TPM (RNA-seq) results. Samples with low expression in RQ-PCR have in most cases resulted in low expression in TPM as well.

DESCRIPTION OF THE INVENTION

The present invention shows the potential biomarkers for patient selection and therapy stratification in melanoma patients.

In this work, inventors used coding and non-coding transcriptome analysis in bulk tumour samples from melanoma patients treated with the anti-PD1 agent Nivolumab to identify a signature of 140 genes associated with response of which 58 were also prognostic, that unmasked a pattern of high B lymphocyte activity under response. The integration with single-cell transcriptome datasets from melanoma patients treated with ICB (anti-PD1, anti- CTLA4, and anti-CTLA4+ anti-PD1), served for biological validation of 35 out of the 140 genes, and led to the refinement of the T and B lineage-related subtypes associated with the treatment: CD8 exhausted T cells associated with resistance, plasmablasts associated with response, and the BC1 subtype (activation of B cells, antigen presentation, proliferation of naive T cells) associated with resistance. This results will pave the way for de-complexing the interaction among the tumour and the different immune cells implicated in the reactivation of the antitumor response associated to ICB.

RNA-Seq was employed in the anti-PD1 therapy discovery cohort (n=21) for selecting those genes that could bear genetic variants relevant to the resistance phenotype. The expression of the genes related to the responding phenotype was associated with the overall survival (OS).

This study addresses the lack of definite biomarkers of response to ICB in metastatic melanoma patients, by studying the differential gene expression of the responders versus the non-responders to Nivolumab using a transcriptomic approach that includes coding and non coding transcripts.

The gene expression results, explored through pathway analyses and deconvolution of the immune cell populations, highlight a putative involvement of B cells in the response to immunotherapy. The inventors show that many of the dysregulated genes in responders are enriched in pathways related to B cell receptor signaling pathway, immunoglobuline production and other immune response associated mechanisms (Table 4). Furthermore, the perturbation in the relative abundances of immune populations that could mirror a more “immune ignorant” environment in the non responders, presents a substantial variation of different B cell populations between the two groups of patients. This is particularly intriguing giving that evidence of both pro- and anti-immune roles have been reported for these cells. In addition, it is important to note that most of studies related to ICB are evaluating the biological significance of tumor infiltrating lymphocytes (TILs) focused on T cells and tumor neoantigens (Gubin M.M, et al. ,2014)( Le D.T., et al. , 2015), (Linnemann C, et al., 2015), (Stronen, et al., 2016), (E.M. Verdegaal, et al., 2016), while less attention has been devoted towards B cells.

In conclusion, the immune and particular B cell specific transcriptomic signature associated with the responder phenotype allows to envisage a novel mechanism of resistance where regulatory T lymphocytes and B lymphocytes, among several other immune populations that infiltrate the tumor, are disbalanced in a context of resistance to treatment with Nivolumab. In addition, various genes that conform the transcriptomic signature of the response, are able to predict overall survival and constitute good candidates as biomarkers. All in all, the present invention represent a step-forward in cancer precision immunotherapy, and a base for understanding the complexity of the immune system-tumor interactions that trigger a resistance phenotype in melanoma patients treated with ICB.

The initial cohort is 21 patients, in which there are 3 of the uveal subtype, 2 of the mucous subtype and 16 patients of the cutaneous subtype.

Among these 21 patients we have 11 responders and 10 non-responders to immunotherapy treatment. Responding or non-responding criteria include:

- Non-respondent: progression in less than 3 months from the beginning of immunotherapy.

• Severe: patients not responding to treatment who present hyperprogression after treatment with immunotherapy.

• Non-severe: those non-responders who previously had a poor prognosis due to immunotherapy treatment due to present some adverse symptoms such as brain metastasis or an "animal-like" tumor, so the non-response to immunotherapy may not be related to the treatment but to the tumor profile of the patient.

- Respondent: patients with partial (PR) or complete (CR) response for about a year or at least who have remained in treatment for one year, following the criteria for partial and complete response recorded in the RECIST v1.1 guide (“ Response Evaluation Criteria in Solid Tumors”) of evaluation of response criteria in solid tumors.

Then, in the present invention the term "responder" refers to those patients who have a complete or partial response to the drug and with the term "nonresponder" to those who do not respond to it so that the tumor continues to progress in the short term after the beginning of immunotherapeutic treatment.

Then, one aspect of the invention refers to the use of the 140 genes described in Table 5 or any combinations thereof for prognosticating or predicting the response to anti-PD1 treatment of a subject suffering from a cancer disease.

In a preferred embodiment, the invention, relates to the simultaneous use of the 140 genes described in Table 5 for predicting the response to anti-PD1 treatment of a human subject suffering from a cancer disease. Out of the 140 genes, 58 also serve for prognosticating the overall survival of the patients

(Table 6).

Then, the use of the 140 genes described in Table 5, or the use of the 58 genes described in Table 6, can be independent or in any combination thereof, or can be used simultaneously.

METHODS OF THE INVENTION

Other aspect of the invention refers to an in vitro method of predicting or prognosticating the response of a human subject to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy, hereinafter first method of the invention, wherein the subject is suffering from a cancer disease, and wherein the method comprises using, as an indicator a) expression levels of the genes of Table 5 to predict the response, b) expression levels of the genes of Table 6 (selected from Table 5) to prognosticate the response, and wherein the result is indicative of a positive response if the expression levels of genes highlighted in grey (shadowed) in Tables 5 are over-expressed while the genes in white in Tables 5 are infra-expressed.

More preferably, the /firstmethod of the invention, also comprising determine the protein CD19, and wherein the result is indicative of a positive response if the expression levels of protein CD19 is over-expressed.

Another aspect of the invention relates to a method for obtaining useful data, hereinafter the first method of the invention, for prognosticating or predicting the response to anti-PD1 treatment of a subject suffering from a cancer disease, wherein said method comprises using, as an indicator, expression levels of the 140 genes described in Table 5.

Another aspect of the invention refers to a method of prognosticating or predicting the response to anti-PD1 treatment of a subject suffering from a cancer disease, wherein said method comprises using, as an indicator, expression levels of the 140 genes described in

Table 5.

Preferably the cancer disease is selected from the list consistin on melanoma, lung cancer, renal cell carcinoma, Hodgkin lymphoma, head and neck cancer, colon cancer, liver cancer, or combinations thereof. More preferably the cancer disease is melanoma, and more preferably metastatic cutaneous melanoma.

In another preferred embodiment, the anti-PD1 treatment is an anti-PD1 antibody, more preferably the anti-PD1 antibody is selected from Pembrolizumab and/or Nivolumab, and most preferably is Nivolumab.

The term "subject", as used in the description, refers to animals, preferably mammals, and more preferably, humans. Then, subject is preferred a human subject, and is not intended to be limiting in any respect; it may be of any age, sex of physical condition.

The methods of the present invention may be applied with samples from individuals of either sex, i.e. men or women, and at any age. The profile determined by the present invention is predictive and prognostic.

In the context of the present invention “Response” refers to the clinical outcome of the subject, “Response” may be expressed as overall survival or progression-free survival. Survival of cancer patients is generally suitably expressed by Kaplan-Meier curves, named after Edward L. Kaplan and Paul Meier who first described it (Kaplan, Meier: Amer. Statist. Assn. 53:457-481). The Kaplan-Meier estimator is also known as the product limit estimator. It serves for estimating the survival function from life-time data. A plot of the Kaplan-Meier estimate of the survival function is a series of horizontal steps of declining magnitude which, when a large enough sample is taken, approaches the true survival function for that population. The value of the survival function between successive distinct sampled observations is assumed to be constant. With respect to the present invention, the Kaplan- Meier estimator may be used to measure the fraction of patients living for a certain amount of time after beginning of chemotherapy and/or radiotherapy. The clinical outcome predicted may be the (overall/progression-free) survival in months/years from the time point of taking the sample. It may be survival for a certain period from taking the sample, such as of six months or more, one year or more, two years or more, three years or more, four years or more, five years or more, six years or more. In each case, “survival” may refer to “overall survival” or “progression-free-survival”.

Thus, in one embodiment, the response is clinical outcome, which is “overall survival” (OS). “Overall survival” denotes the chances of a patient of staying alive for a group of individuals suffering from a cancer. The decisive question is whether the individual is dead or alive at a given time point. For the in vitro determination of the expression levels is carried out by using any of the known techniques from the state of the art. Preferably, the techniques are selected from the list consisting on: a. a gene profiling method, such as a microarray, or a Next Generation Sequencing panel and/or b. a method comprising PCR, such as real time PCR; and/or c. northern Blot and/or d. an immunohistochemistry method; and/or e. an elisa-based method

In order to technically validate the DE genes, the inventors employed real time quantitative PCR (RQ-PCR) to assess the gene expression of selected genes form our signature of 140 genes.

RQ-PCR is a sensitive and reproducible gene expression quantification technique which can particularly be used to profile mRNA expression in cells and tissues. Any method for evaluation of RT-PCR results may be used, and and the AACt-method may be preferred (Livak et al. Methods 2001, 25:402-408.) (Ct = Cycle threshold values). The AACt-method will involve a control sample and a treatment sample. For each sample, a target gene and an endogenous control (as described below) gene are included for PCR amplification from (typically serially diluted) aliquots. Typically several replicates are used for each diluted concentration to derive amplification efficiency. PCR amplification efficiency can be defined as percentage amplification (from 0 to 1). During the PPCR reaction, a software typically measures for each sample the cycle number at which the fluorescence (indicator of PCR amplification) crosses an arbitrary line, the threshold. This crossing point is the Ct value. More dilute samples will cross at later Ct values. To quantify mRNA gene expression, the Ct for an RNA or DNA from the mRNA gene of interest is divided by Ct of nucleic acid from the endogenous control, such as non-tumoral tissue, to normalize for variation in the amount and quality of RNA between different samples. This normalization procedure is commonly called the AACt-method (Schefe et al., 2006, J. Mol. Med. 84: 901-10). AACt calculations express data in the context of test sample (here: mRNA) versus calibrator (endogenous control). If the AACt calculation is positive (for example +2.0), then: 2-AACt = 2-(2.0) = 0.25. The amount of target, normalized to an endogenous reference and relative to a calibrator, is given by: 2-AACt. Details of the AACt calculation method can be found in: Applied Biosystems user Bulletin No. 2 (P/N 4303859). Without prejudice of the method used to determine the response (RQ-PCR, immunohistochemistry, an elisa-based method etc..), in the context of the present invention, we firstly establish a relative expression of a selection of patients that are controls for the high or low expression of the tested genes against the housekeeping gene ACTB. This is followed by a Pearson correlation between the gene expression values generated by RNA- seq and the ones generated by RQ-PCR. In our case, the expression was technically validated with an overall correlation coefficient of 0.7 (p <0.001) (Figure 10).

In a particularly preferred aspect of the present invention the response is determined by using multispectral immunofluorescence. In a still more preferred aspect of the invention, the used of this technique entails the utilisation of the random trees algorithm classifier.

In the context of the present invention, they investigated the myeloid and lymphocytic contexture of 16 melanomas in formalin-fixed paraffin-embedded (FFPE) tissue samples. Two complementary multiplex panels were used to enable the simultaneous examination of several cellular markers. The myeloid and lymphoid cell panel targeted the myeloid marker CD11b, the phagocytic cell marker CD68 of macrophages, CD3+ and CD8+ T cells, CD20+ B lymphocytes, and the melanoma marker MELAN-A. The melanoma-associated B cells panel included the CD19+ and CD20+ B cells, CD138+ plasma cells.

Multiplex immunofluorescence (IF) development and validation workflow and protocols have been implemented as previously described (PMID: 32591586; PMID: 30742120;). Briefly, 4- micron sections of formalin-fixed paraffin-embedded (FFPE) tissue from 16 melanomas were deparaffinised and antigen retrieval was performed using DAKO PT-Link heat induced antigen retrieval with low pH (pH6) or high pH (pH 9) target retrieval solution (DAKO).

A random trees algorithm classifier was trained separately for each cell marker by an experienced pathologist (CEA) annotating the tumour regions. Interactive feedback on cell classification performance is provided during training in the form of markup image, improving significantly the accuracy of machine learning-based phenotyping. (PMID: 29203879, PMID: 32591586). All phenotyping and subsequent quantifications were performed blinded to the sample identity. Cells close to the border of the images were removed to reduce the risk of artifacts.

Based on the fluorescence panels, cells were further subclassified as CD11b+, CD68+, CD3+, CD8+, and CD20+. For the myeloid and lymphoid cell panel, CD4+ T cells were defined as CD3+ CD8-. MELAN-A was used to visualize the melanoma cells. For the melanoma-associated B cells panel, subpopulations were then classified as: i) total B cells (CD19+ CD20-, CD19+ CD20+ and CD19- CD20+ cells), ii) plasmablasts (CD19+ CD20- CD138- cells), iii) plasma cell-like (CD19+ CD20- CD138+ cells), and iv) mature plasma cell (CD19- or + CD20- CD138+ cells), as previously described (PMID: 31519915). Cells negative for these markers were defined as ‘other cell types’.

In the context of the present invention illustrative non-limiting examples of a biological sample include different types of samples from tissues, as well as from biological fluids, such as blood, serum, plasma, cerebrospinal fluid, peritoneal fluid, faeces. Preferably, said samples are samples from tissues and most preferably, said samples of tissues originate from tumour tissue of the individual the response of which is to be predicted, and may originate from biopsies.

The cancer could be any immunogenic type of cancer that is susceptible to clinical benefit from immunotherapy. In a further preferred embodiment of the invention, the cancer disease as defined in any of the methods of the invention is melanoma, non-small cell lung cancer, head and neck cancer or combinations thereof..

Prognosis depends on the stage of the cancer and, in this sense, it is important to find good prognosis markers for survival after treatment for this specific disease and thereby the usefulness of the biomarkers of the present invention in the prognosis of this disease.

Another aspect of the present invention refers to anyone of the methods of the invention, wherein the method is a drug response predictive method which is performed in vitro using a biological sample originating from the human subject, and wherein at the time point of taking the sample from the human subject, the human subject has not been treated yet with anti- PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy. The outcome for evaluation of the response is the clinical response using the RECIST criteria at a specific time after the beginning of the treatment.

Another aspect of the present invention refers to anyone of the methods of the invention, wherein the method is a prognostic method which is performed in vitro using a biological sample originating from the human subject, and wherein at the time point of taking the sample from the human subject, the human subject has not been treated yet with anti-PD1 immune and/or anti PD-L1 checkpoint inhibition immunotherapy. The outcome for evaluation of the prognosis is the overall survival.

Another aspect of the invention refers to a pharmaceutical composition comprising with anti- PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy for treating a human subject of group 1 as identifiable by any of the methods of the invention. Preferably, the anti-PD1 immune checkpoint inhibition immunotherapy is selected from the list consisting on: Pembrolizumab (Keytruda), Nivolumab (Opdivo), Cemiplimab (Libtayo) or combinations thereof.

In another preferred embodiment, the anti PD-L1 immune checkpoint inhibition immunotherapy is selected from the list consisting on: Atezolizumab (Tecentriq), Avelumab (Bavencio), Durvalumab (Imfinzi) or combinations thereof.

KIT OR DEVICE FROM THE INVENTION

The present invention also provides a kit or device suitable to put into practice the method of the invention. The kit, comprises at least one oligonucleotide(s) capable of hybridizing with the mRNAs of any of the 140 genes of the drug response predictor signature of Table 5, and/or 58 genes of the prognosis predictor signature of Table 6.

The kit or device is based on the predictive power of the method of the present invention. In the particular case of the kit, the reference value indicative for non-response (and/or a reference value indicative for response) may be provided with the kit. With the help of the kit, the expression of each target gene can be calculated, i.e. relative to, such as the endogenous control samples exemplified above. The endogenous control can thus also be comprised within the kit.

The kit may further include, with no type of limitation, buffers, agents to prevent contamination, protein degradation inhibitors, etc. Therefore, the kit may include all the supports and receptacles necessary for its implementation and optimization. Preferably, the kit further comprises the instructions for carrying out any of the methods of the invention.

In particular embodiments, the kit is selected from (a) a kit suitable for PCR, (b) a kit suitable for Northern Blot, (c) a kit suitable for microarray analysis, (d) a kit suitable for Next Generation Sequencing, and a kit suitable for immunohistochemistry (CD19). Any two or more of these embodiments may also be combined, so that the kit may comprise, for example both (a) and (c).

As regards (a) a kit suitable for PCR, this PCR is typically real-time quantitative PCR (RQ- PCR), a sensitive and reproducible gene expression quantification technique. A Northern Blot involves the use of electrophoresis to separate RNA samples by size and subsequent detection with an oligonucleotide(s) (hybridization probe) complementary to (part of) the target sequence of the RNA of interest.

It is also possible that the oligonucleotide(s) are immobilized in spots on a (preferably solid) surface.

In one embodiment thereof, the kit comprises a microarray. An RNA microarray is an array on a solid substrate (usually a glass slide or silicon thin-film cell) that assays large amounts of different RNAs which are detectable via specific probes immobilized on spots on the solid substrate. Each spot contains a specific nucleic acid sequence, typically a DNA sequence, known as probes (or reporters). While the number of spots is not as such limited, there is a preferred embodiment in which the microarray is customized to the methods of the invention. In one embodiment, such a customized microarray comprises fifty spots or less, such as thirty spots or less, including twenty spots or less.

On another embodiment of the invention, the kit comprises a number of capture probes for specific genes that are hybridised in suspension and subsequently amplified by PCR and sequenced in a low performance sequencer since the total number of reads needed for the targeted sequencing is low; therefore, it can be implemented in clinical routine laboratories.

Finally, the detection of CD19 protein by antibody staining is simple and affordable by any clinical or Anatomy Pathology Laboratory.

A further embodiment of the invention refers to a kit suitable for detecting the level of expression of the genes of Table 5 and/or Table 6, or the protein CD19 which comprises a media having affixed thereto a capture antibody capable of complexing with any of biomarker proteins encoding by the genes of Table 5 and/or Table 6, or the protein CD19, or a fragment thereof and an assay for the detection of a complex of the biomarker and the capture antibody.

The kit may be used and the use is not particularly limited, although use in the method of the invention in any of its embodiments is preferred.

The kit may also be automated, or can be incorporated in devices capable of carrying out the methods of the invention automatically.

COMPUTER IMPLEMENTED INVENTION Another aspect of the invention relates to computer-readable storage means comprising program instructions capable of making a computer perform the steps of any of the methods of the invention

The invention also extends to computer programs adapted so that any processing means can carry out the methods of the invention. Such programs may take the form of source code, object code, an intermediate source of code, and object code, for example, as in partially compiled form, or in any other form suitable for use in implementing the processes according to the invention. Computer programs also encompass cloud applications based on this procedure.

In particular, the invention encompasses computer programs arranged on or within a carrier. The carrier can be any entity or device capable of supporting the program. When the program is incorporated in a signal that can be directly transported by a cable or other device or medium, the carrier may be made up of said cable or another device or medium. As a variant, the carrier could be an integrated circuit in which the program is included, the integrated circuit being adapted to execute, or to be used in the execution of, the corresponding processes.

For example, the programs could be embedded in a storage medium, such as a ROM, a CD ROM or a semiconductor ROM, a USB memory, or a magnetic recording medium, for example, a floppy disk or a disk. Lasted. Alternatively, the programs could be supported on a transmittable carrier signal. For example, it could be an electrical or optical signal that could be transported through an electrical or optical cable, by radio or by any other means.

The invention also extends to computer programs adapted so that any processing means can carry out the methods of the invention. Such programs may take the form of source code, object code, an intermediate source of code, and object code, for example, as in partially compiled form, or in any other form suitable for use in implementing the processes according to the invention. . Computer programs also encompass cloud applications based on this procedure.

Therefore, another aspect of the invention relates to a computer-readable storage medium comprising program instructions capable of causing a computer to carry out the steps of any of the methods of the invention.

Another aspect of the invention relates to a transmittable signal comprising program instructions capable of causing a computer to carry out the steps of any of the methods of the invention. In the context of the present invention, the terms “subject”, “patient” or “individual”' are used herein interchangeably to refer to all the animals classified as mammals and includes but is not limited to domestic and farm animals, primates and humans, for example, human beings, non-human primates, cows, horses, pigs, sheep, goats, dogs, cats, or rodents. Preferably, the subject is a male or female human being of any age or race.

Along the description and claims, the word "comprises" and variants thereof do not intend to exclude other technical features, supplements, components or steps. For persons skilled in the art, other objects, advantages and features of the invention will be understood in part from the description and in part from the practice of the invention. The following examples and drawings are provided by way of illustration and they are not meant to limit the present invention.

EXAMPLES OF THE INVENTION

The following specific examples provided in this patent document serve to illustrate the nature of the present invention. These examples are included only for illustrative purposes and must not be interpreted as limiting to the invention which claimed herein. Therefore, the examples described above illustrate the invention without limiting the field of application thereof.

Materials and Methods Subjects

Subjects

A Discovery cohort of 21 metastatic melanoma (16 cutaneous, 3 uveal, 2 mucosal) patients treated with Nivolumab donated FFPE tumour biopsy samples that were collected pre treatment (Supplementary Table 1).

Patient biopsies were selected to exclude lymph node metastases and to include only samples with availability of clinical data and information on progression after treatment. The cohort is distributed in responders (11) and non-responders (10), where the distinction criteria is described as follows:

- Non-responders: progression in less than 3 months from the start of immunotherapy. Of them, a subgroup of “severe” non-responders are define as those who progressed in less than 60 days. - Responders: patients with maintained partial or complete response for a year or in treatment during at least one year.

Response to Nivolumab was assessed according to the “Response Evaluation Criteria in Solid Tumors” criteria (RECIST v1.1 guide). The study follows the Declaration of Helsinki and has been vetted and aproved by the Ethical Committee of Malaga. All patients signed an Informed Consent to participate in the study.

As Validation cohort, we employed the cohort of 32 melanoma patients treated with anti-PD1 and analysed with single cell RNA-seq reported in Sade-Feldman, M. et al. , 2018 (Sade- Feldman et al., 2018).

Nucleic acid extraction

The tumour-specific area in FFPE melanoma samples was predefined by a pathologist. Two to four 10 pm slides were dissected for nucleic acid extraction, using the microtome HM 340E (Thermo Scientific). RNA was extracted with the RNeasy FFPE kit following the manufacturer instructions (Qiagen; Ref. 73504).

Next Generation Sequencing

RNA-Seq libraries were prepared using TruSeq Stranded Total RNA Gold (lllumina; Ref.20020598) and indexed by IDT for lllumina - TruSeq RNA UD Indexes (lllumina; Ref. 20020591). These libraries include coding and non-coding RNA by ribosomal RNA depletion. In order to obtain a better exclusion of ribosomal RNA, the manufacturer protocol was modified including a double-depletion. Libraries concentration (0.1-1 micrograms) was determined by Qubit dsDNA BR kit, and the size distribution was examined by Agilent Tapestation 2200. Each libraries contained 0.1 -Paired-end reads (75bp ^c 2) were acquired from the lllumina NextSeq 550 platform according to the corresponding protocol.

Technical validation

The technical validation of the All RNA-seq data was perfomed using real-time quantitative PCR (RT-qPCR) of the following selected genes based on the gene expression patterns and the representation of tumor and immune system candidate biomarkers: TNFRSF11B, IGLV6- 57, IGHA1, and GRIA1. RNA from selected samples of Discovery cohort presenting with high and low expression for the genes of study was retrotranscribed into cDNA and subjected to RT-qPCR. The housekeeping control gene ACTB was used for normalization.

Multispectral immunofluorescence Depending on the multiplex immunofluorescence protocols, each tissue section was subjected to three or six successive rounds of antibody staining, each round consisting of protein blocking with 20% normal goat serum (Dako) in phosphate-buffered saline (PBS), incubation with primary antibody, biotinylated anti-mouse/rabbit secondary antibodies and Streptavidin-HRP (Dako), followed by TSA visualization with Opal fluorophores (Akoya Biosciences) diluted in 1X Plus Amplification Diluent (Akoya Biosciences).

The myeloid and lymphoid cell panel included: CD11b (Rabbit monoclonal, clone EPR1344, 1:1000, Abeam, product number ab133357), CD68 (Mouse monoclonal, clone PG-M1, ready-to-use, Agilent, product number IR613), CD3 (Rabbit polyclonal, IgG, ready-to-use, Agilent, product number IR503), CD8 (Mouse monoclonal, clone C8/144B, ready-to-use, Agilent, product number GA62361-2), CD20 (Mouse monoclonal, lgG2a, clone L26, ready- to-use, Agilent, product number GA604), and MELAN-A (Mouse monoclonal, clone A103, ready-to-use, Agilent, product number IR63361).

The melanoma-associated B cells panel included: CD19 (Mouse monoclonal, clone LE- CD19, ready-to-use, Agilent, product number GA656), CD20 (Mouse monoclonal, lgG2a, clone L26, ready-to-use, Agilent, product number GA604), CD138 (Mouse monoclonal, lgG1, clone M 115, 1:100, Agilent, product number M7228).

Thus, in the last round, nuclei were counterstained with spectral DAPI (Akoya Biosciences) and sections mounted with Faramount Aqueous Mounting Medium (Dako).

Tissue Imaging, Spectral Unmixing and Phenotyping

Each whole-tissue section was scanned on a Vectra-Polaris Automated Quantitative Pathology Imaging System (Akoya Biosciences). Tissue imaging and spectral unmixing were performed using inForm software (version 2.4.8, Akoya Biosciences), as previously described (PMID: 32591586; + diego’s paper - no PMID yet). Image analysis was then performed in the whole-tumour area (referred to the total melanoma area) using the open- source digital pathology software QuPath version 0.2.3, as previously described (PMID: 32591586). In short, cell segmentation based on nuclear detection was performed using StarDist 2D algorithm, a method that localises nuclei via star-convex polygons, incorporated into QuPath software by scripting. Invasive margin was defined as a region of 100pm width assessed in the interface tumour and stromal compartments (determined by cell marker MELAN-A). Bioinformatic analysis

Fastq quality control was performed with FastQC. Fastq files were trimmed using tool HISAT2(v 2.1.0) with a customized index built using combined rRNA data from HGNC, ENA, SILVA, and additional manually curated sequences from NCBI. Trimmed fastq files were mapped against reference (genome build GRCh38) using STAR (v 2.5.1b) and read quantification was done with the same tool. % of uniquely mapped reads and M mapped reads were computed with Qualimap.

To perform the normalization and test for differential expressed (DE) genes, we used the Bioconductor package DESeq2. A genes was considered as DE the baseMean count was >10, the absolute Fold Change (FC) was >1.5, and the adjusted p value was < 0.05.

Pathway analysis was done with R in-house scripts using two different approaches: gene set enrichment analysis (GSEA and DAVID) and network based pathway analysis. Packages used were STRINGdb, clusterprofiler, pathfindR. MPC Counter was used to infer the abundance of different immune cells populations of the tumor infiltrate using the normalized counts of RNA-seq. We utilized MIXCR and Seq2HLA for HLA, TCR and BCR profiling, and GATK for calculation of the Tumour Mutational Burden (TMB) from our bulk RNA-seq dataset. Also, 30 COSMIC mutational signatures we evaluated in our cohort. Survival analysis was performed using the Kaplan-Meier method. For validation in a scRNA-seq cohort, the dataset of anti-PD-1 treated melanoma patients that conformed our Validation cohort was integrated at the Seurat environment, normalized and pre-processed. Following, we used PCA (Principal Component Analysis) in combination with tSNE (t- Distributed Stochastic Neighbour Embedding), together with the PanglaoDB markers to identify and distribute the cell populations defined by the scRNA-seq built expression dataset. Finally, we localized the 140 DE genes of our bulk-seq data within the generated cell type map, and tested their overlap with the response-associated genes of the scRNA-seq dataset.

Results

Gene expression signatures of the response to PD1 blockade

The coding and non coding transcriptome of 21 melanoma samples of different subtypes (16 cutaneous, 3 uveal and 2 mucosal melanomas respectively;) was generated using RNA sequencing with ribosomal depletion with RNA-Seq. Given the known molecular differences of the different subtypes of melanoma probably due to diverse carcinogenic events (Hayward et al., 2017), we explored common and subtype-specific response expression signatures in our cohort. In all melanoma subtypes, 22 genes conformed a common signature that features the response to anti-PD1 (Table 1). Importantly, the expression signature of the responding patients with cutaneous melanoma was more extensive, with 140 differentially expressed genes, of which 11 genes overlapped the general signature (Table 2). In non responders, 108 genes were down-regulated while 32 genes were up-regulated compared to responders. The downregulated 108 genes are enriched in Cytokine- cytokine receptor interaction (P = 4.39 e-3), while the upregulated genes interestingly were enriched in Calcium signalling pathway (P = 2.3 e-2) (Table 3).

Interestingly, 73 of 140 DE genes were Immunoglobulin-related, which were presumably B lymphocytes related. The KEGG pathway analysis of 140 DE genes revealed cytokine- cytokine receptor interaction significantly enriched. Notably, IGHV1-69-2 was differentially expressed in both comparisons. Interestingly, a similar functional profile was obtained excluding the samples of ganglionic origin, discarding therefore the enrichment of the B cell- related signature due to an eventual overestimation of the B cell population. Taken together, these results revealed that the immunoglobulins and cytokine-cytokine receptor interaction may underlies behind the mechanism of response in ICB.

With the aim of exploring the prognosis and predictive value of the identified expression signatures, we performed overall survival analysis with log rank test. For all subtypes of melanoma, the levels of expression of 9 genes of the common signature correlated with overall survival ( P < 0.05). In cutaneous melanoma, 58 genes conformed the specific expression profile associated with OS ( P < 0.05), among which 33 were immunoglobulin- related; the most relevant ones, including the receptor of B lymphocytes CD22, a gene coding for the 3-21 V region of the variable domain of immunoglobulin heavy chains, IGHV3- 21, and the gene coding for the cell adhesion molecule LRFN5 are represented in Figure 1.

For the subgroup of non-ganglionic cutaneous melanomas, expression of 88 genes marked the prolonged overall survival of the patients after immunotherapy ( P < 0.05)] out of them 23 genes were immunoglobulin-related, and 32 genes conformed in their majority by IG-related genes were present in the cutaneous general profile associated with OS.

Underlying Molecular mechanisms of PD 1 blockade response

Based on the substantial number of DE genes and their prognosis value of ICB response, we explored deeper regarding the gene expression. For 140 DE genes identified in 16 cutaneous samples, 82 genes were expressed in a similar pattern to at least one other genes. In addition, 23 of 140 DE genes were regulated by super- enhancers. Then we continued to mine the possible mechanisms of response in PD1 blockade in advanced melanoma. Gene Ontology (GO) (Mi et al. , 2017) and analysis were performed with GSEA. For responders and non-responders with 16 cutaneous melanoma, 22 biological process (BP) terms, 9 GO molecular function (MF) terms and 15 cellular component (CC) terms were significantly enriched in GO analysis (Table 4). The BP enrichment analysis demonstrated the dramatic difference of B lymphocytes- related biological processes including phagocytosis, B cell/complement activation, and immunoglobulin production among PD1 blockade responders and non responders. Similarly, immunoglobulin receptor binding and monomeric IgA immunoglobulin complex, were discovered in MF and CC enrichment analysis.

Since our results suggest that the presence of B lymphocytes in the tumour microenvironment is involved in the response to the PD-1 therapy, we investigated the abundance of each immune cell type among our patients with CibersortX (Newman et al., 2019), in a set of cell types defined by single cell RNA-seq of patients with melanoma treated with ICB (Figure 2). The deconvolution of the stromal cells of the tumour microenvironment from our RNA-seq dataset in metastatic cutaneous melanoma reinforces the critical contribution of the B cell lineage in the response to Nivolumab among all cell types including T cells. Interestingly, the T cells that were significantly associated to the treatment (resistance) were the exhausted CD8 T cells, while circumscribing our analysis to the B cell lineage, plasmablasts and the subtype of B cells BC1 were significantly enriched in good and bad responders, respectively, refining the characterisation of the B cells-related subtypes involved in the efficacy of Nivolumab.

Given that the involvement of the B cell lineage in the reactivation of the antitumour immune response has been outlined in multiple dimensions of our analysis, we decided to characterise the abundance and diversity of the BCR, TCR clonotypes, and HLA loci aiming at confirming the potential functional translation of an enhanced transcriptomic signature of the B cells. Consistently, we identified that the number of clonotypes in general (combined count of BCR and TCR) was significantly higher in good versus bad responders (Figure 4A); in addition, the major contribution of B cells compared to T cells was again highlighted when we stratified by BCR and TCR (Figure 4B).

In addition, we have identified those BCR clonotypes that are more different between good and bad responders. Interestingly, all of them include the constant region of immunoglobulin light chains IGKC, which is part of the transcriptomic signature of response to Nivolumab

(Table 7).

With regards to HLA loci abundance, we also evidence a role for in the resistance mechanism in our patients, given the difference in predicted loci between responders and non-responders, particularly significant for the Class II loci (Figure 5).

When assessing the diversity in BCR, TCR, and HLA, we evidenced that, particularly of BCR, it is proportional to therapeutic response. In addition, we identified a higher abundance and diversity in TCR compared to BCR, in line with the prominent implication that we have observed of B lymphocytes in the response to Nivolumab in our cohort. This pattern is exacerbated with regards response, where non responders express very low levels of TCR, and less abundant and diverse BCR that responders (Figure 6).

Finally, the integration of the trancriptomic signature of response to Nimoluvab with the BCR and HLA abundance indicates that BCR function and antigen presentation are amongst the processes that are critical for the efficacy of the treatment (Figure 7).

Given that our data was produced using bulk sequencing, we aimed at validating the B cell enrichment and our 140 genes signature in independent datasets of single-cell RNA- sequencing (scRNA-seq) from tumor samples of metastatic melanoma patients treated with anti-PD1 (Sade-Feldman et al. , 2018). Using the sets that were specifically treated with anti- PD1, and the celular subtypes that were associated with response in our dataset, we validated 35 genes out of the 140, of which 17 were among the ones associated with survival (Figure 8).

In addition, we have validated the CD19 B cell marker as a protein biomarker of response to anti-PD-1 in melanoma, identifying a trend towards statistical association of plasmablasts (Figure 9).

Combined, the results above demonstrate the dominant presence of B lymphocytes and their possible importance regarding ICB response. Mutation burden positively predicts survival as a biomarker for PD1 blockade in melanoma (Rizvi et al., 2015). The tumor mutation burden (TMB) was defined as the ratio of the number of mutations and gene length. However, the TMB was similar in both responders and non-responders ( P>0.05 ). Subsequently, we then use the COSMIC mutational signatures (https://cancer.sanqer.ac.uk/cosmic/siqnatures) to annotate 21 melanoma samples. Overall, no specific mutational signature pattern significantly differentiate responders and non responders. Interestingly, 4 melanoma samples (2 responders and 2 non responders) contained no mutation signature 1 but harbored signature 9 for compensation. Tables

Table 1. Differentially expressed genes in responders of ICB for all subtypes of melanoma. In green, the upregulated genes.

Table 2. genes among all-subtypes comparison and cutaneous sample comparisons. Table 3. KEGG pathway analysis of differentially expressed genes

Table 4. Gene Ontology enrichment of the responding and non-responding patients with metastatic cutaneous melanoma.

GO biological process Count Adj p value complement activation, classical pathway 61 3.21 E-88 regulation of complement activation 41 1.33E-56 B cell receptor signaling pathway 41 3.6E-55 phagocytosis, recognition 39 3.84E-54

Fc-gamma receptor signaling pathway involved in phagocytosis 40 1.64E-52 phagocytosis, engulfment 39 1.25E-51 positive regulation of B cell activation 39 1.05E-47

Fc-epsilon receptor signaling pathway 37 7.9E-44 receptor-mediated endocytosis 41 5.28E-43 leukocyte migration 44 5.81 E-43 immunoglobulin production 28 4.07E-31 innate immune response 43 1.99E-29 organelle organization 2 0.00018 glomerular filtration 4 0.000181 antibacterial humoral response 5 0.000623 positive regulation of respiratory burst 3 0.00366 regulation of gene expression 8 0.00696 nucleic acid metabolic process 1 0.00782 regulation of protein oligomerization 4 0.0211 gene expression 1 0.0393 regulation of RNA metabolic process 7 0.0452 protein transport 0 0.0477

GO molecular function Count Adj p value antigen binding 68 7.9E-95 immunoglobulin receptor binding 40 1.25E-56 signalling receptor binding 51 2E-22 binding 106 7.85E-05 organic cyclic compound binding 10 0.000131 heterocyclic compound binding 10 0.000175 nucleic acid binding 4 0.000452 metal ion binding 6 0.00487 cation binding 8 0.0438

GO cellular component Count Adj p value external side of plasma membrane 41 1.36E-36 blood microparticle 21 5.75E-21 membrane-enclosed lumen 5 5.26E-07 intracellular organelle lumen 5 5.49E-07 organelle lumen 5 5.74E-07 monomeric IgA immunoglobulin complex 4 4.04E-06 cytosol 6 7.65E-06 nuclear lumen 3 8.82E-06 nucleoplasm 2 2.55E-05 secretory dimeric IgA immunoglobulin complex 3 0.00014 pentameric IgM immunoglobulin complex 3 0.000151 integral component of postsynaptic density membrane 4 0.0103 extracellular exosome 24 0.0156 cytoskeletal part 1 0.0471

Table 5. List of the 140 differentially expressed genes of the 16 samples of cutaneous metastatic melanoma that are significant by restricting a set value of less than 0.05, an absolute value of FoldChange> 1.5 and the base value Mean> 10. In grey (shaded), the upregulated genes.

Table 6. List of 58 genes of the transcriptomic response signature of the metastatic cutaneous cohort associated with OS

Table 7. Ranking of counts of the 26 clonotypes that are enriched in good responders to Nivolumab.

Sample ID Response Clonotypes counts

IMK31 Good 4659

IMK26 Good 3075

IMK36 Bad 899

IMK30 Good 591

IMK38 Good 491

IMK27 Good 236

IMK32 Good 235

IMK48 Good 189

IMK23 Good 187

IMK24 Good 106

IMK22 Bad 100

IMK35 Bad 80

IMK34 Bad 39

37IMK Bad-severe 16

39IMK Bad-severe 15 Supplementary table 1. Clinical characteristics of melanoma patients treated with immunotherapy.

References

1. Duruisseaux, et al. (2018). Epigenetic prediction of response to anti-PD-1 treatment in non- small-cell lung cancer: a multicentre, retrospective analysis. Lancet Respir. Med. 6, 771- 781.

2. Falchook, G. (2015). Nivolumab: another weapon in the growing immunotherapy arsenal. Lancet Oncol. 16, 350-351.

3. Hayward, N.K., , et al. (2017). Whole-genome landscapes of major melanoma subtypes. Nature 545, 175-180.

4. Gubin, M.M. et al., (2014), Checkpoint blockade cancer immunotherapy targets tumour specific mutant antigens Nature 515577-581

5. Ghoneim, H.E., et al. (2017). De Novo Epigenetic Programs Inhibit PD-1 Blockade- Mediated T Cell Rejuvenation. Cell 170, 142-157. e19.

6. Le, D.T., et al., (2015) PD-1 blockade in tumors with mismatch-repair deficiency, N. Engl. J. Med. 372. 2509-2520.

7. Linnemann, C et al., (2015)High-throughput epitope discovery reveals frequent recognition of neo-antigens by CD4+ T cells in human melanoma. Nat. Med. 21 81- 85.

8. Meshcheryakova A et al,. 2014, B cells and ectopic follicular structures: novel players in anti-tumor programming with prognostic power for patients with metastatic colorectal cancer., PLoS One. 2014, 6;9(6)

9. Mi, H., et al . (2017). PANTHER version 11: expanded annotation data from Gene Ontology and Reactome pathways, and data analysis tool enhancements. Nucleic Acids Res. 45, D183-D189.

10. McKenna, A., et al. (2010). The Genome Analysis Toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297-1303.

11. Newman, A.M., et al (2015). Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods 12, 453-457.

12. Guan H, et al, 2016 An adaptive immune response driven by mature, antigen- experienced T and B cells within the microenvironment of oral squamous cell carcinoma, Int J Cancer. 2016 15;138(12):2952-62. 13. Ribas, A. (2012). Tumor Immunotherapy Directed at PD-1. N. Engl. J. Med. 366 , 2517-2519.

14. Sharma, P.,et al. (2017). Primary, Adaptive, and Acquired Resistance to Cancer Immunotherapy. Cell 168, 707-723. 15. Stronen, E. et al., Targeting of cancer neoantigens with donor-derived T cell receptor repertoires, (Science 352 (2016) 1337-1341

16. Topalian, S.L.,et al. (2016). Mechanism-driven biomarkers to guide immune checkpoint blockade in cancer therapy. Nat. Rev. Cancer 16, 275-287.

17. Poplin, et al. (2018). Scaling accurate genetic variant discovery to tens of thousands of samples. BioRxiv.

18. Rizvi, N.A., et al. (2015). Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science 348, 124-128.

19. Schwartz, M., et al(2016). B cell regulation of the anti-tumor response and role in carcinogenesis. Journal for immunotherapy of cancer, 4, 40. doi:10.1186/s40425-016-0145- x.

20. Verdegaal EM , et al., (2016) Neoantigen landscape dynamics during human melanoma-T cell interactions, Nature 53691-95. ).

Claims

1 The simultaneous use of the genes of Table 5 to predict or prognosticate the response to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy.

2.- The use according to claim 1 , wherein from the genes of Table 5, the genes from Table 6 are selected to prognosticate the response to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy

3.- The use according to anyone of claims 1-2, also comprising the use of the protein CD19.

4 - An in vitro method of predicting or prognosticating the response of a human subject to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy, wherein the subject is suffering from a cancer disease, and wherein the method comprises using, as an indicator a) expression levels of the genes of Table 5 to predict the response, b) expression levels of the genes of Table 6 (selected from Table 5) to prognosticate the response, and wherein the result is indicative of a positive response if the expression levels of genes highlighted in grey (shadowed) in Tables 5 are over-expressed while the genes in white in Tables 5 are infra-expressed.

5.- The in vitro method according to claim 4, also comprising determine the protein CD19, and wherein the result is indicative of a positive response if the expression levels of protein CD19 is over-expressed.

6.- The in vitro method according to any one of claims 4-5, wherein protein or mRNA that encodes the genes of Table 5 and/or Table 6 is used as an indicator.

7.- The in vitro method according to any one of claims 4-5, wherein for predicting the response of a human subject to anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy, wherein the subject is suffering from a cancer disease, and wherein the method comprises using, as an indicator, expression levels of the genes of Table 5 (140 genes).

8.- The method according to any one of claims 4-5, wherein for prognosticating the prognosis of a human subject treated with anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy, wherein the subject is suffering from a cancer disease, and wherein the method comprises using, as an indicator, expression levels of the genes of Table 6 (58 genes).

9.- The method according to any one of claims 4-8, wherein the in vitro determination of the expression levels is carried out by: f. a gene profiling method, such as a microarray, or a Next Generation Sequencing panel and/or g. a method comprising PCR, such as real time PCR; and/or h. northern Blot and/or i. an immunohistochemistry method; and/or j. an elisa-based method

10.- The method according to any one of claims 4-9, wherein over-expressed is defined as a level of expression greater than 1/3 of the maximum score achieved in normal human tissue sampes.

11.- The method according to any one of claims 4-10, wherein the response refers to the overall survival rate.

12.- The method according to any one of claims 4-11 , wherein the cancer disease is selected form the list consisting on: melanoma, non-small cell lung cancer, head and neck cancer or combinations thereof.

13.- The method according to any one of claims 4-12, wherein the biological sample is fresh tissue, paraffin embed tissue or RNA extracted from a tissue from a patient with cancer.

14.- The method according to any one of claims 4-13, wherein the anti-PD1 immune checkpoint inhibition immunotherapy is selected from the list consisting on: Pembrolizumab (Keytruda), Nivolumab (Opdivo), Cemiplimab (Libtayo) or combinations thereof.

15.- The method according to any one of claims 4-13, wherein the anti PD-L1 immune checkpoint inhibition immunotherapy is selected from the list consisting on: Atezolizumab (Tecentriq), Avelumab (Bavencio), Durvalumab (Imfinzi) or combinations thereof.

16.- A method for allocating a human subject suffering from cancer in one of two groups, wherein group 1 comprises subjects identifiable by the method according to any one of claims 4-15; and wherein group 2 represents the remaining subjects.

17. A pharmaceutical composition comprising anti-PD1 and/or anti PD-L1 immune checkpoint inhibition immunotherapy according to anyone of claims 15-16, for treating a human subject of group 1 as identifiable by the method of claim 16.

18.- A kit or dispositive suitable for the perform of the methods of any of the preceding claims, comprising at least one oligonucleotide(s) capable of hybridizing with the mRNAs of any of genes of Table 5 and/or Table 6, or the protein CD19.