Community Research and Development Information Service - CORDIS

Final Report Summary - IMMOMEC (IMmune MOdulating strategies for treatment of MErkel cell Carcinoma)

Executive Summary:
1.1. Executive Summary
Epidemiologic data suggest that there are approximately 2500 new Merkel cell carcinoma (MCC) cases per year within the EU; approximately 1000 of these patients will die from their disease. Although MCC is 40 times less common than melanoma, MCC has a dramatically higher mortality rate than melanoma rendering MCC as the most lethal skin cancer (37 versus 15 percent). This high mortality rate is largely due to the fact that until recently none of the available therapeutic interventions improved the overall survival of patients suffering from metastatic disease. Consequently, new therapeutic strategies were needed for metastatic MCC. Since several lines of evidence indicate the outstanding immunogenicity of MCC, immune modulating treatment strategies are of particular interest.
IMMOMEC was a project at the frontline of therapeutic, translational and medical research. The project tried to establish a novel immunotherapy taking advantage of the innovative concept of antibody-targeted cytokines to enrich the respective immune modulating agent at the tumor site. Until recently, treatment of MCC patients was solely based on anecdotal observations and case series, which implies a very low level of evidence. The prospectively randomized European multicenter clinical trial conducted within IMMOMEC aimed at establishing the clinical impact and immunological effects following therapeutically intended immune modulation by targeted interleukin-2 (IL2) (WP1).
However, the therapeutic landscape for advanced MCC changed dramatically and unexpectedly since the initiation of IMMOMEC in the beginning of 2011. At that time no competing therapeutic trials for advanced MCC patients existed. Early after initiation of the IMMOMEC trial, 3 competing trials had been activated which were designed for the same target population of metastatic MCC. Once preliminary data from these trials suggested a high efficacy of PD-1/PD-L1 blocking antibodies, recruitment into our trial comparing an immune intervention arm to a chemotherapy arm had become virtually impossible. Thus, while this was very fortunate for the patients suffering from advanced MCC, it was unfortunate for the IMMOMEC trial, which subsequently did not reach sufficient recruitment to provide any significant results with respect to the clinical efficacy of targeted cytokine therapy.
Nevertheless, IMMOMEC did not only aim at a new therapeutic option for MCC patients, but also at establishing new tools to monitor patients receiving immune modulating therapies in MCC (WP3, WP4, and WP5) as well as to compile prognostic and predictive biomarkers of MCC helping to individualize therapeutic indications (WP2 and WP5). To this end the IMMOMEC project was very successful as the consortium reached the objectives strived for:
(i) to identify MCC-specific T-cell epitopes (WP3),
(ii) to establish the logistics to collect high quality peripheral blood lymphocytes (PBL) from patients across Europe recruited into the IMMOMEC trial, which are suitable for detailed characterization of specific T-cell responses (WP4)
(iii) to identify a blood-based surrogate biomarker for tumor burden of MCC (WP2)
(iv) to identify and characterize immune escape mechanisms of MCC and means to counteract them (WP2 and WP5)
(v) to establish methods to scrutinize tumor tissues with respect to the inflammatory infiltrate both at a morphological and molecular level including the T-cell receptor repertoire usage (WP2 and WP5)
However, due to the low recruitment into the trial and thus the limited number of patients receiving targeted IL2 as immunotherapy, no expedient correlative studies of the results obtained by the analyses addressing the systemic and localized immune responses could be performed. Instead we focused on comparative studies on these systemic and localized immune responses (WP5).
Since the IMMOMEC clinical trial was open for recruitment and ongoing therapy close until the end of the project, only some of the results have been published to date. However, the consortium has a specific list of publication plans for the next months.
Some of the translational data and results from IMMOMEC were presented at an open IMMOMEC International Symposium in Maastricht, The Netherlands, November 16th, 2016. On this occasion the stakeholders, IMMOMEC consortium members, invited guests as well as other participants were involved in fruitful discussions on how to use the obtained results from IMMOMEC, together with evidence from the literature and personal experiences to improve the clinical management of advanced MCC patients.
IMMOMEC has established a webpage at, which provides further information on the project and its results. The consortium aims at maintaining and updating this website to allow communication for future projects addressing immune modulating strategies for the treatment of Merkel cell carcinoma.

Project Context and Objectives:
1.2. Project Context and Objectives
1.2.1. Project context
IMMOMEC represents a collaborative research effort of 9 academic and 2 industrial partners (SMEs) from 8 European countries.
MCC is a highly aggressive and often lethal neuroendocrine cancer of the skin, associated with the recently discovered common Merkel cell polyomavirus (MCPyV) or with chronic UV exposure. Epidemiologic data suggest that there are approximately 2500 new MCC cases per year within the EU; approximately 1000 of these patients will die from their disease. The incidence of MCC is considerably increasing: The reported incidence has more than tripled over the past 20 years. This increase can partially be explained by the demographic development, since MCC usually affects the elderly. The median age at diagnosis lies in the 8th decade of life, and there is a 5- to 10-fold increase in incidence after age 70 as compared with an age less than 60 years. Thus, it is likely that in an ageing European population the impact of this deadly cancer will further increase continuously. However, preliminary data from a MCC registry created within IMMOMEC suggest that besides an increasing incidence, the age distribution of patients is slowly shifting towards younger patients.
Notably, MCC has a dramatically higher mortality rate than melanoma, rendering MCC as the most lethal skin cancer. This high mortality rate is largely due to the fact that until recently none of the available therapeutics improved the overall survival of patients suffering from metastatic disease. Consequently, new therapeutic strategies were needed for metastatic MCC.
Since several lines of evidence indicate the outstanding immunogenicity of MCC, immune modulating treatment strategies are particularly attractive. To this end, the fundaments of the IMMOMEC project were based on a prospectively randomized phase II trial investigating the safety and efficacy of an innovative immunotherapy of advanced MCC (WP1). In detail, the trial was set up to compare the clinical efficacy of an - at that time - common chemotherapeutic intervention with a combination of this chemotherapeutic with an immune modulating therapy, i.e. the targeted delivery of IL2 to the tumor microenvironment by the tenascin C-reactive immunocytokine F16-IL2. The primary endpoint of this trial was overall survival; secondary endpoints included safety and the induction/boost of MCC-specific cellular immune responses. The translational research program, however, was not restricted to the identification and monitoring of MCC-specific T cell epitopes (WP3 and WP4), but also included the identification of blood and tissue based biomarkers of MCC as well as the detailed analysis of possible immune escape mechanisms (WP2 and WP5). Finally, two work packages covered coordination and management of the consortium (WP6), as well as dissemination and exploitation of the results (WP7).
All work packages with the exception of WP5 were run in parallel. Thus, the three main research areas of IMMOMEC comprised clinical testing, immune monitoring and biomarker identification.
1.2.2. Project Objectives
Specific objectives of IMMOMEC, which are also reflected in the respective work-package structure, include:
I. Establish an effective therapy for MCC evaluated in a multicentre randomized clinical phase II trial, thereby demonstrating the feasibility of immunotherapy for solid cancers (WP1 Randomized phase II trial paclitaxel alone versus paclitaxel in combination with F16-IL2)
II. Identification of prognostic and predictive biomarkers (WP2 Identification of biomarkers predicting patient prognosis, treatment outcome and immune response)
III. Identification and characterization of HLA-restricted immunodominant T cell epitopes specific for MCC to monitor the immune modulating effect and to develop specific therapeutics (WP3 Identification and characterization of HLA-restricted immunodominant MCC-associated T cell epitopes; WP4 Immune monitoring of MCC patients under therapy & WP5 Correlation of systemic immunological responses and clinical benefit: impact of the microenvironment, immune escape mechanisms and regulatory circuits)
IV. Establish a European network for research and therapy of MCC (WP6 Management of the project and the consortium & WP7 Exploitation and dissemination)

The detailed objectives of the respective work packages were:

• Implementation and execution of a phase-II clinical trial investigating the efficacy of the immunocytokine F16-IL2 plus paclitaxel versus paclitaxel alone for the therapy of patients with metastatic Merkel cell carcinoma (MCC)
WP2 • Identification of surrogate, prognostic and/or predictive biomarkers of metastatic MCC
WP3 • Identification and characterization of immunodominant T cell epitopes restricted by HLA-A1, A2, A3, A24 derived from the MCPYV T antigens and novel MCC-associated novel antigens
• Monitoring of immune responses against the identified peptides in normal donors and MCC patients
WP4 • Standardized collection of high quality PBMC patient samples in the multi-center trial by a proven European laboratory network.
• Determination of immunological efficacy (proof of concept) of F16-IL2 by detecting and characterizing T cell activation
• Validation of the relevance of novel MCPYV epitopes identified in WP3 will establish these as parameter for monitoring or even therapeutic applications in future MCC trials
• Identification of new cellular markers and potential surrogates for immunological efficacy of F16-IL2 to discover patient populations with a better chance of clinical response
WP5 • Correlation of immunological and clinical responses to F16-IL2 immunotherapy
• Correlation of systemic and intratumoral immune responses
• Identification and characterization of immune escape mechanisms of MCC
• Establish the impact of the microenvironment on the immune-mediated tumor eradication

Within WP6 and WP7 the Medical University of Graz (MUG) coordinated and managed IMMOMEC. The coordinator and his team managed the effective implementation of the ambitious work plan to ensure coordinated progress towards the projects aims and the optimal dissemination of its expected results to obtain optimal visibility. In addition, dissemination activities included the set-up and maintenance of the project website (, two public Workshops and a public International Symposium with invited external experts, in which IMMOMEC’s results were discussed in the context of current evidence from the literature and personal experiences of experts in MCC to improve the clinical management of advanced MCC patients.
These seven WPs are closely interrelated (Figure 1). The consortium executing all tasks of the respective WPs comprised the following beneficiaries (Table 1), covering the different European countries to a large extent (Figure 2).

Figure 1. IMMOMEC work packages and their interdependencies
Table 1. List of Beneficiaries
Participant Number Participant name Participant short name Country Date enter project Date exit project
7 CHARITE - BERLIN Charite Germany 1 60

Figure 2: IMMOMEC Partners. Geographical distribution of the participating trial centers across Europe
1.3. The IMMOMEC Work Packages and Major Scientific Results

Project Results:
1.3. The IMMOMEC Work Packages and Major Scientific Results
1.3.1. WP1 - Randomized phase II trial of paclitaxel alone versus paclitaxel in combination with F16-IL2 in order to establish the therapeutic efficacy of targeted IL2 therapy
The IMMOMEC trial was a phase II study of the tumour-targeting human F16-IL2 monoclonal antibody-cytokine fusion protein in combination with paclitaxel versus paclitaxel alone in patients with Merkel cell carcinoma.
This study was thought to explore an innovative immunochemotherapy for MCC which is based on the targeted delivery of interleukin-2 (IL2) to the tumour microenvironment via the armed antibody F16-IL2.
Study design and management of regulatory issues.
This phase II study was aimed at determining the anti-cancer activity of F16-IL2 in combination with paclitaxel in patients with advanced MCC, as measured by overall survival (OS) rate at 12 months after beginning of study treatment (primary objective). As secondary objectives, the study aimed at investigating safety and tolerability of the combination treatment of F16-IL2 and paclitaxel as well as efficacy in terms of response to treatment measured as overall response rate (ORR, rate of complete plus partial responses) or disease control rate (DCR, rate of complete plus partial responses plus disease stabilizations).
The preparation of the Clinical Study Protocol was an important commitment for the Sponsor, and for the Coordinator of IMMOMEC, together with all the Principal Investigators that were in charge of the trial. An unanimous agreement on the protocol was difficult, as the sponsor for safety reasons demanded an age limitation, which, however, was regarded by the clinical partners as a restrictive condition for a sufficient recruitment; a compromise was achieved in October 2012.
The Clinical Trial was planned to take place at nine different sites in seven countries:
1. Medizinische Universität Graz - MUG - Austria – Prof. J.C. Becker
2. Region Hovedstaden / CCIT Denmark – Dr. A. Krarup-Hansen
3. Assistance Publique - Hopitaux De Paris –APHP- France- Prof. C. Lebbè
4. Fundacio Privada Clinic Per A La Recerca Biomedica –FCRB- Spain – Prof. S. Puig
5. Charite - Universitätsmedizin Berlin - Germany – Dr. F. Kiecker
6. Eberhard-Karls-Universität Tübingen -CDO - Germany - Prof. C. Garbe
7. University of Nottingham – UNOTT - United Kingdom – Prof. P. Patel
8. Universitätsklinikum Essen –UKEssen - Germany – Prof. D. Schadendorf
9. Maria Sklodowska-Curie Memorial Cancer Center - MCMCC – Prof. P. Rutkowski
As a start, the approval of the different regulatory authorities of each country in which the trial was going to be managed was needed. An important number of documents were prepared and submitted by the Sponsor to the Competent Authorities (CA) and Ethics Committees (EC) involved. When preparing the proposal, the time frame to obtain the approval was unpredictable, due to the number of CAs and ECs involved. Preparation and collection of documents from investigators and submissions of documents was done as scheduled, nevertheless ECs and CAs had different timelines in answering and requesting modifications or new documents. Approval timescales did not depend on the beneficiaries but on the authorities. The duration of the procedure varied from country to country and from EC to EC within the same country. In some countries, such as Spain and France, more details and documents were asked, which made regulatory processes more time consuming when compared to other countries such as e.g. Austria.
In the first 18 months, applications to Competent Authorities had been finalized in Germany (PEI), Austria (BASG), UK (MHRA), France (AFSSAPS), Spain and Denmark. MHRA and BASG had already approved the study in UK and in Austria, respectively. The submission to the Ethics Committees were completed in Graz (Austria), Berlin, Tübingen and Essen (Germany), Nottingham (UK), Barcelona (Spain) and Paris (France). Approval of the Clinical Trial was achieved from Competent Authorities in UK, Austria and Germany and from the Ethics Committees in Graz, Tübingen, Berlin, Barcelona.
In the second reporting period, i.e. month 19 to 36, the submission to Competent Authority in Poland and to Ethics Committees in Poland and Denmark was completed. We received approvals from the Competent Authorities in Germany, France, Spain, Denmark and Poland. The Ethics Committees of Paris, Essen, Copenhagen and Warsaw, besides, gave favorable opinion to the Study.
Basically, we received full approvals in seven sites which actually recruited into the Clinical Trial (Table 2). Two centers that were initially planned, Graz and Nottingham, did not take part in the multicenter study, due to internal difficulties. The fully authorized centers subsequently started recruiting patients.
Table 2. Summary of site initiation dates

F16-IL2 is a novel bio-therapeutic agent, which had been tested so far up to a maximum dosage of 45 Mio IU. F16-IL2 is a recombinant fusion protein composed of a fully human recombinant monoclonal antibody (F16), and human recombinant interleukin-2 (IL2) (Figure 3). Clinical-grade F16-IL2 was manufactured in Philogen’s GMP facility in sufficient amount for the execution of the clinical trial. Philogen owns a 2’000 m² GMP production facility, authorized to produce antibodies in mammalian cells. The production of biologic medicines is more complex and more variable than the production of chemical drugs. For such reason, manufacturing of biologics does not only require more steps, but also more stringent regulation and control of the processes. Critical quality attributes define the product properties of safety and efficacy, based on a thorough understanding of the product, including molecular properties and experience from pre-clinical and clinical evaluations. Quality controls were set in place to verify that manufacturing processes result in products of desired quality.

Figure 3. The immunocytokine F16-IL2. (A) F16-IL2 binds to tenascin C expressed by stromal cells of MCC tumors, and boost cellular immune responses (B) F16-IL2 consists of a stable non-covalent homodimer of the scFv F16 sequentially fused to human IL2, a strong pro-inflammatory cytokine. The N- and C-termini of the molecule as well as the VH and VL domains of the scFv moieties are indicated
Regulatory GMP guidelines ensure that personnel, infrastructure and logistics of manufacturing facilities provide optimal conditions for the implementation and validation of production processes. The study drug, F16-IL2, was manufactured, identified by a lot number, and shipped to the different clinical sites according to the requirements of Directive 2001/20/EC. F16-IL2 was supplied in pyrogen-free 10 ml glass containers as liquid, sterile, apyrogenous solution. All study drug supplies had to be stored at -80°C until use. Investigators were aware of and well instructed during the Site Initiation Visit about the storage of the IMP, and knew that study drug doses that were stored differently from the recommendations were no to be used and had to be replaced with fresh doses.
Seventeen patients were screened for the Clinical Trial, of which 13 were enrolled and treated (summarized in Table 3). Seven patients were randomized in Arm A (F16-IL2 + paclitaxel), and six patients in Arm B (paclitaxel). All patients completed the first cycle of treatment. Five patients received 2 complete treatment cycles, the others discontinued previously, three of them due to progression of disease. The distribution of patients in the seven centers, the randomization in Arms A and B, and the number of treated patients per cycle are reported in Table 3.
Table 3. Patient enrollment and treatment summary
Patients enrolled
Tübingen 4 (30.77 %)
Berlin 3 (23.08 %)
Barcelona 1 (7.69 %)
Paris 1 (7.69 %)
Essen 2 (15.38 %)
Copenhagen 1 (7.69 %)
Warsaw 1 (7.69 %)
Patient randomization Arm A (F16-IL2 + PACLITAXEL) 7 (53.85 %)
Arm B (PACLITAXEL) 6 (46.15 %)
Treatment by cycle Cycle 1 13 (100.00 %)
Cycle 2 8 (61.54 %)
Cycle 3 4 (30.77 %)
Cycle 4 2 (15.38 %)
Cycle 5 1 (7.69 %)
Cycle 6 1 (7.69 %)
Patients who completed 2 cycles 5 (38.46 %)
Reason for NOT Adverse Event 1 (12.50 %)
completing the treatment Progressive Disease 3 (37.50 %)
Serious Adverse Event 2 (25.00 %)
Withdrawal of Informed Consent 2 (25.00 %)

Monitoring visits were programmed and scheduled in accordance to the needs of all clinical centers and depending on enrollment rate. The activities during the visit was done in compliance with the International Conference on Harmonization (ICH) guidelines for GCP and following Philogen Standard Operating Procedures. Aim of the Monitoring Visits was to review the progress of the clinical study, hence to: (i) Ensure protocol adherence, (ii) assure accuracy of data, and (iii) assure regulatory compliance
Monitoring Visits were done by the Clinical Research Associate (CRA) of Philogen or by a local CRO (Warsaw and Copenhagen, as required by Ethics Committees to overcome language issues). All personnel involved in the trial had to be present (PIs, AIs, Research Nurses, Trial Coordinators, Data Managers, Pharmacists, Monitors).
The CRA or CRO ensured that (i) all participant original informed consents were filed in the medical record, signed and dated; (ii) medical records contained laboratory reports, X-ray, scan reports, physician notes, nursing notes, procedures documenting study parameters reported in Case Report Forms (CRFs), and (iii) all CRF’s were complete, accurate, and up to date. At the end of the visit the monitor collected site queries/clarifications and prepared the monitoring visit report that was then placed in the Trial Master File.
Efficacy evaluation.
Of the 17 screened patients, 13 were enrolled and treated, with 8 of them evaluable, i.e. receiving at least two cycles of therapy, for primary efficacy outcome. Overall survival was calculated as the time between randomization and death due to any cause. Patients who were lost to follow-up were censored at the time when last known to be alive (Figure 4).
Overall survival (OS) was 100% at 2 months, 75% at 4 months, and 50% at 1 year for Arm A (F16-IL2 plus Paclitaxel) compared to 67% at 2 and 4 months, and 0% at 1 year for Arm B (Paclitaxel alone) (Figure 3). The analysis for differences in OS between the two arms was not significant due to low numbers (log rank test p=0.60). The median OS was equal to 303.5 days (95% C.I: 89-NE, upper limit not estimable) and 304 days (95% C.I: 44-304), respectively, in F16-IL2 plus Paclitaxel and Paclitaxel alone.

Figure 4. Overall survival in the per-protocol population
Response to treatment was evaluated in 8-week intervals (from week 8 up to week 24 or end of treatment) according to RECIST for measurable disease (target lesions), non-measurable disease (non-target lesions) and new lesions, and according to immune-related response criteria (irRC) for index and non-index lesions.
None of the patients treated in the study achieved a complete or partial response, neither in Arm A nor in Arm B, according to RECIST v. 1.1 or irRC at week 8, 16 or end of treatment. Figure 5 shows the maximum change of the sum of diameters of target lesions and the best objective response recorded for each study patient evaluable for efficacy during the study. The vertical bars indicate the most favorable RECIST score/irRC variation as a percent of the value recorded at baseline.

Figure 5. Objective responses. Change in diameters of target lesions and tumor response
Safety Evaluation.
Overall, all thirteen patients treated within the trial were evaluable for the secondary study endpoint safety (Figure 6). Seven patients were exposed to F16-IL2 plus Paclitaxel at least for one week of treatment, and six patients were exposed to Paclitaxel alone for at least one week of treatment. Overall, 12 out of 13 treated patients (92.3%) experienced at least one adverse event, among which in 3 (23.1%) patients these were classified as a severe adverse events were observed. Notably, all serious adverse events occurred in Arm A,
The most commonly involved system organ class for treatment-related AEs (any grade) were blood and lymphatic system disorders, general disorders, administration site conditions, and skin or subcutaneous tissue disorders.

Figure 6. Safety. Number of patients showing adverse events, by system organ class (SOC) and total (overall).
In conclusion, the exiguous number of patients that could be recruited within the timeframe of the study implies that these results have to be considered preliminary have to be confirmed. Notably, patients treated with F16-IL2 plus Paclitaxel appeared to have a favorable outcome with 50% of patients alive after one year compared to 0% of patients treated with Paclitaxel alone.

1.3.2. WP2 - Identification of biomarkers predicting patient prognosis, treatment outcome and immune responses
Over the past decade, biomarkers and targeted therapies have been essentially important to many of the major success stories in cancer treatment. Moreover, therapies are rarely efficient for all patients of a certain cancer entity. Therefore, the inclusion of biomarker analyses into the context of clinical trials is desirable. The identification of biomarkers does, however, not only allow the selection of patients with a high probability to benefit from a specific therapy more accurately, but also improves the understanding of how the respective therapeutic works.
From large retrospective studies of MCC patients it is known that the survival curve of patients with advanced disease initially drops very fast, but stagnates at about 20% relative survival after two years. Thus, it seems that a subpopulation of patients is endowed with a constitutively better survival, or shows a better response to therapy.
One pre-requisite for studies on tissue- or blood-derived biomarkers is the existence of a set of biological specimens annotated with the respective clinical data. Thus, it was essential to establish a data base linked to biobanked patient materials. Long-lasting discussions within the IMMOMEC consortium, with the sponsor of the clinical trial, and the respective data protection officers revealed that the originally planned modus operandi of a Web-based open accessible data base would not sufficiently protect the privacy of the patients. An alternative, sufficiently secure version would have been possible, but beyond the planned budget as it would have to be adapted to the respective firewalls of the partner institutions. Consequently, we decided to establish a central data base with access rights for the partners. In addition, we established a web-based clinical registry for concurrent and retrospective MCC cases.
As a discovery set to identify a panel of candidate biomarkers we used a set of annotated biological samples from 77 advanced MCC patients. As the first step we determined the quality of the respective materials by H&E staining, demonstrating that 12 of these cases had insufficient quality for immunohistochemistry analyses. For 61 cases we could determine the mitotic rate, which was stratified into categories ranging for 0-1 to >10 mitoses/mm². In only 1 case we observed 0-1, the majority of cases, i.e. 37, had 1-5, 6 cases had 5-10, and 17 cases had >10 mitoses/mm². Next, the inflammatory infiltrate was analyzed. The quantity of infiltration was categorized from absent = 0 to strong = ++. In addition, the infiltrate was distinguished qualitatively between peritumoral and intratumoral distribution pattern. This discrimination was automated using the InFormTM software.
Since the status of the Merkel cell polyomavirus (MCPyV) has been suggested to be a prognostic marker, we evaluated the MCPyV status in our discovery sample set both by real time PCR and immunohistochemistry. MCPyV positivity by PCR was defined when at least two of three analyzed primer sets gave a positive result, i.e. a relative value presence of at least 0.01 compared to the MCPyV+ MCC cell line WaGa. Applying this definition, from the evaluable cases 29 were regarded as negative while 35 were MCPyV positive. Notably, among the 29 cases in which we could not clearly detect MCPyV DNA, 10 stained positive for MCPyV large T antigen in immunohistochemistry. Notably, for twelve cases we were able to only get either a PCR (six cases) or an immunohistochemistry result. Thus, from the cases analysed for MCPyV status, 21 were negative (17 in both methods), 24 were positive in one assay (8 cases only analyzed by one method), and in 25 cases we obtained positive results by both methods. In total, 70% of the analyzed cases were positive for MCPyV.
Up to date, almost all studies on MCPyV prevalence are based on PCR techniques, which can not reveal the nature of the MCPyV DNA, i.e. episomal or integrated. Therefore, we performed MCPyV fluorescence in situ hybridization (FISH) on MCC tissue samples to determine whether this technique allows gathering information about the quality of the viral presence on the single cell level. To this end, MCPyV FISH was performed on tissue microarrays containing 62 tissue samples of 42 patients including all tumor grades. The hybridization patterns were correlated to the qPCR data determined on corresponding whole tissue sections. Importantly, MCPyV FISH and qPCR data were highly correlated, i.e. 83% concordance for FISH-positive and 93% concordance for FISH-negative cores. Interestingly, two hybridization patterns were distinguishable in the MCPyV FISH specimen: a punctate pattern (85%) indicating viral integration correlating with a moderate viral abundance, and a combination of the punctate with a diffuse pattern (15%) indicating the coexistence of integrated and episomal virus DNA which was associated with very high qPCR values. Thus, MCPyV FISH seems to be useful to further elucidate MCPyV related carcinogenesis by adding important information about viral integration or episomal presence on the single cell level within the histomorphological context.
Nevertheless, since the presence of viral antigens in MCC cases should render this tumor specifically immunogenic, the tumor has to escape immune surveillance. This hypothesis is already sustained by the heterogeneous infiltration of MCC tissues with lymphocytes. To address this notion, we established a multiplexed immunofluorescence staining based on the OpalTM system, which allows for the detection of multiple antigens in a single tissue section. In order to use the full possibilities of this technique, a multispectral image acquisition platform is required. To this end, we used the Mantra System from PerkinElmer, which is an integrated workstation incorporating multispectral imaging technology, novel image acquisition and inForm analysis software. As exemplified in Figure 7, this method allowed us to scrutinize the lymphocytic and myeloid infiltrate in MCC in detail. The depicted example visualized the presence of tertiary lymphoid tissue, which has been described to be associated with an improved prognosis in other skin cancers.

Figure 7. Multiplexed immunofluorescence. FFPE tissue sections of an MCC tumor were stained for T cell subtypes (CD4+ T helper cells, CD8+ cytotoxic T cells, FoxP3+ regulatory Tcells), B cells (CD20+), and macrophages (CD68+). Cytokeratin 20 was used to detect the MCC cells.
However, even by this means we could not detect a correlation between MCPyV status and TILs or clinical course. . This is on one hand explained by the recently reported high mutational load in UV-associated MCC, but on the other hand, both MCPyV+ as well as MCPyV- MCC could either be characterized by a brisk, dim or absent T cell infiltrate. Consequently, we next tested for possible immune escape mechanisms. As we established that MCC tumors in situ do not express the NKG2D ligands MICA and MICB, which act activating on NK cells and costimulatory on T cells, we analyzed the expression of these ligands in MCC cell lines in vitro.
In contrast to tissue analytics, blood-based biomarkers as a surrogate of tumor burden allow to be repeatedly checked to monitor the clinical course of cancer patients. Consequently, the American Joint Committee on Cancer (AJCC) incorporated several blood-based biomarkers into their staging systems, e.g., prostate-specific antigen (PSA) and precursor PSA (proPSA) for prostate cancer, or α-fetoprotein and β-HCG for testicular cancer. Unfortunately, for the majority of solid cancers no reliable blood-based biomarkers have been established yet, including MCC. In order to establish a reliable blood-based biomarker for MCC, we first focused on the presence of neuron-specific enolase (NSE), which has been established as a surrogate marker for other neuroendocrine cancers. Unfortunately, however, we could not detect a significant correlation of NSE serum concentration with the tumor burden of MCC patients.
Advances in genomic technologies have boosted the number of candidate DNA- and RNA-based biomarkers. In this respect, circulating free (cf) DNA has been successfully demonstrated as a serum-derived biomarker for tumor burden in Langerhans cell histiocytosis. Although cancer in general is associated with increased serum DNA concentrations, most approaches using cf DNA as a blood based biomarker rely on tumor-specific DNA mutations, such as the BRAFV600E mutation in the given example. Unfortunately, due to the absence of specific hotspot mutations, this strategy is not applicable for MCC. An alternative approach takes advantage of cf miRNAs specifically overexpressed in certain cancer types. Since miRNAs are very resistant to degradation, sera from cancer patients contain large amounts of miRNAs derived from tumor cells. Indeed, miRNAs have been recognized as biomarkers for a variety of different cancer entities, such as miR-205 in breast cancer, or miR-19 in colorectal cancer. Since the analysis of MCC cells revealed a high expression of a set of specific microRNAs, we analyzed the respective microRNA expression in MCC cases. To this end, expression of this set of specific microRNAs normalized to the U6 snRNA was much higher in MCC cases compared to melanoma samples or the lung adenocarcinoma epithelial cell line A549. In a xenotransplantation model for MCC recently established by us, we could clearly observe an increase of a set of MCC-specific microRNAs in mouse serum from a mean relative expression of 2.4 in 5 mice without tumor to 13.2 in 5 mice with tumor. Indeed, we not only demonstrated the abundant expression of this miRNA in MCC cell lines and tissues and its extracellular presence in MCC cell culture supernatants and sera of tumor-bearing preclinical animal models, but subsequently, we demonstrated the value of this cf miRNA as a surrogate marker of tumor burden in MCC patients. Notably, this cf miRNA not only discriminated patients with or without evidence of disease, but also correlated with the stage of disease. Additional studies including the comparison with other potential surrogate markers are needed to ascertain whether monitoring this cf miRNA levels in individual patients will be a useful tool for the early detection of disease recurrence or progression.
We established a pseudonymized online registry for MCC patients meant for, but not restricted to the IMMOMEC consortium. In fact, the input boxes are open to any interested participant ( Methods for data validation and safeguarding the uniqueness of each data set had been developed and implemented. Based on data from this registry, we were able to demonstrate that the MCPyV status of MCC is in general not associated with the clinical course of the disease. However, in a subgroup of patients in which MCPyV appears to be initially involved in a hit-and-run carcinogenesis, but is subsequently lost, absence of MCPyV appears to be associated with an impaired prognosis. Indeed, several reports in 2016 confirmed that in MCPyV-associated MCC there is a very low mutational burden; thus there are very few neo-epitopes apart from those derived from the virally encoded proteins. These observations emphasize the importance of re-inducing MHC class I expression without impairing the expression of the viral proteins. Retrospective analyses of data concerning the somatostatin receptor (SSTR) expression in MCC measured by PET revealed SSTR as a valuable target for molecular imaging of MCC.
Data from the MCC registry which now comprises almost 1000 cases was also used in the development of the European guidelines for diagnosis and treatment of MCC. Furthermore, we reported observations concerning the biology and clinical course of MCC. , , ,
1.3.3. WP3 - Identification and characterization of HLA-restricted immunodominant T cell epitopes derived from the MCPYV T antigens
Discovery of CD8+ T cell epitopes in MCC is important for understanding the immune recognition of cancer cells. This cellular component plays a major role in mediating tumor cell eradication and should help us to understand the induction of T cell reactivity following F16-IL2 therapy. Indeed, MCPyV-encoded proteins are likely targets for cytotoxic immune responses to MCC as they are both foreign to the host and necessary to maintain the oncogenic phenotype. However, prior to the IMMOMEC project only a single MCPyV-derived CD8 T-cell epitope has been described, thus impeding specific monitoring of T-cell responses to MCC. Consequently, the purpose of this study was to identify and validate a sufficient number of MCPyV-derived T-cell epitopes for future characterization of spontaneous, modulated, or induced immune responses to MCC.
Virus antigens are degraded intracellulary, and thereafter processed and presented on the cell surface by MHC class I molecules - the classical pathway for the immune system to survey changes in intracellular compartments. T cells of the host will recognize fragments (peptides) from the virus proteins on the surface on tumor cells embedded in the MHC class I molecules, dependent on the patients’ HLA type. T cell epitopes from MCPyV-encoded proteins are ideal targets to kill cancer cells, since these are recognized as “foreign” by the host immune system and not subjected to tolerance mechanism as observed with most self-antigens. Thus, the goal of this work package was to identify MHC class I restricted T-cell epitopes derived from MCPyV-encoded proteins (Figure 8).

Figure 8. T-cell mediated tumor cell killing. Cytotoxic T cells require T-cell receptor mediated recognition of specific epitopes presented by MHC class I on the targeted cell for cognate killing.
This goal was achieved by scanning the MCPyV oncoprotein large T (LTA) and small T antigens (STA) and the virus capsid protein (VP1) for potential T-cell epitopes, and testing for MHC class I affinity. We confirmed the relevance of these epitopes using a high-throughput platform for T-cell epitope identification to characterize MCPyV-specific CD8 T-cell epitopes in LTA, STA, and VP1. This platform is based on a parallel enrichment of peptide–MHC-reactive T cells and the detection of specific responses by combinatorial encoding with peptide–MHC multimers, enabling multi-epitope identification using limited patient material. , Using this approach, we detected T-cell responses among 398 predicted T-cell epitopes restricted to HLA-A1,-A2, -A3, -A11, and -B7. In total, 56 MCPyV-specific T-cell responses were detected in 38 individuals, representing 35 different specificities (Figure 9).

Figure 9. T-cell responses against MCPyV in MCC patients and healthy donors (adapted from 14 and 16). (A) Schematic overview of the applied methodology (B) Specific responses were detected using flow cytometry and combinatorial encoded MHC- multimers followed by verification with either a 2nd enrichment (top) or a 2nd MHC-multimer detection (bottom). (C) Number of MCPyV specific T-cell responses per individual in MCC patients and healthy donors. (D) MCC patients or healthy donors with LTA/STA or VP1 specific T-cell responses. Asterisk indicate significant differences: *: p<0.05, **: p < 0.01, *** p < 0,001.

In parallel, we attempted to identify T cell epitopes by mass spectrometry through the elution of peptides presented in MHC class I on the surface of three MCC tumor cell lines. However, this analysis did not reveal any MCPyV-derived epitopes.
Strikingly, although T cells specific for VP1-derived epitopes were detected both in patients with MCC and healthy donors, T-cell responses against oncoproteins were exclusively present in patients with MCC, but not in healthy donors. We further demonstrated both the processing and presentation of the oncoprotein-derived epitopes, as well as the lytic activity of oncoprotein-specific T cells toward MHC-matched MCC cells. Demonstrating the presence of oncoprotein-specific T cells among tumor-infiltrating lymphocytes further substantiated the relevance of the identified epitopes.
We have observed T-cell based recognition in MCC patients of both the MCPyV-derived oncoproteins and the capside protein VP1. Importantly T cells directed against the oncoproteins are exclusively found in MCC patients and not healthy virus-infected individuals, indicating that these epitopes are exclusively found on cancer cells. Furthermore, we have shown for a number of these epitopes that T cell recognition can lead to killing of MCC cells. Thus, the host T cells do recognize the virus-derived T cell epitopes, but they are obviously not sufficiently effective in inducing cancer cell killing in vivo, but can be boosted with therapeutic measures. MCC tumor infiltrating lymphocytes do express a series of markers indicating T cell exhaustion and an immunosuppressive environment within the tumor. However, very relevant for the current application, the observation that MCC develops more frequently in immuno-compromised individuals suggests that the boosting of immune responses in these individuals will have a clinical benefit.

1.3.4. WP4 - Immune monitoring of MCC patients under therapy to correlate therapeutic effects with therapy-induced immune responses
In the IMMOMEC clinical trial standardized and centralized analysis of specific and functional T lymphocytes was performed in order to monitor each patient’s immune competence during successful F16-IL2 delivery. To this end, a network of 7 specialized cell culture laboratories had been successfully established and trained for the standardized and GCP-compliant collection and isolation of patients’ peripheral blood mononuclear cells (PBMC).
As the available PBMC cell numbers per patient were expected to be limited and as the number of immune-dominant T-cell epitopes identified from MCPyV and MCC tumor antigens identified in WP3 were supposed to vary, Immatics’ established immune monitoring assays were adapted to include important cell-saving technologies in preparation for the IMMOMEC clinical trial: Successful implementation, validation and improvement of the Multimer Multiplexing technology followed by benchmarking the sensitivity of this new technology to Immatics’ previously used standard Multimer technology. Additionally, Immatics has implemented further cell-saving technologies (e.g. ex vivo analysis of rare cells as compared to methods using an in vitro amplification step), with the objective that a maximal data set can be generated out of the inherently limited biological material within the IMMOMEC clinical trial. For the characterization of changes in additional immunological parameters during F16-IL2 therapy, such as other relevant cell populations like immunosuppressive Foxp3+ regulatory T cells (Tregs) or myeloid derived suppressor cells (MDSCs), Immatics developed and established a highly standardized proprietary flow cytometry detection panel for detailed characterization and quantification of numerous cell populations contained within patient PBMC samples.
Within the IMMOMEC clinical trial a total of 26 PBMC samples from 16 patients at 7 participating clinical centers have been successfully collected. Of these 16 patients, 7 patients were enrolled in the experimental arm (arm A: paclitaxel plus F16-IL2) and 6 patients were enrolled in the control arm (arm B: paclitaxel only). Moreover, 4 baseline samples were taken from MCC patients who were finally not enrolled into the trial and PBMC samples of 5 healthy donors were included in the baseline Treg and MDSC analyses.
F16-IL2 administration is expected to lead to an activation of MCPyV- and MCC-specific cytotoxic T cells (CTLs) directed towards naturally occurring epitopes, and to the proliferation and acquisition of effector functions in these CTLs. Thus, T-cell responses to published MCC peptide antigens and peptide antigens identified in WP3 were monitored in detail. MCC-specific CD8+ T cells were assessed ex vivo to monitor specific T cell response profiles before and after application of F16-IL2. These analyses enabled the validation of the relevance of novel MCPyV epitopes and may establish these novel targets as MCC antigens for monitoring or even therapeutic applications in other MCC trials. 11 of the enrolled patients exhibited a suitable HLA-type (HLA-A*02, -B*07 or -A*24) and were evaluable for determination of MCC-specific peptide reactivities. Of these patients 55% (6/11) showed CD8+ T-cell responses at baseline to at least one of the MCC-specific peptides (exemplarily shown in Figure 10, for one MCC-specific antigen and HIV as control antigen).

Figure 10. Ex vivo measured, MCC-peptide specific T-cell response and negative control.

MCC-specific responses are in the range of <0.001% - 0.024% of CD8+ T cells. However, neither for the 3 patients of the experimental arm nor for the 2 patients of the control arm, who were evaluable for pre- and post-treatment CD8+ T cell reactivities, a treatment-induced increase in MCC-specific CD8+ T cells was detected.
Moreover, immune cell populations influencing specific T cell responses were assessed in order to compare pre-treatment Treg and MDSC levels of healthy donors and cancer patients, to assess Treg and MDSC levels before and after application of F16-IL2, as well as to identify Treg or MDSC populations that are potentially predictive biomarkers for the clinical outcome or for the immune responses.
All 11 Treg populations analyzed had comparable baseline levels in MCC patients and healthy donors (data not shown). However, during treatment with F16-IL2 plus paclitaxel (experimental arm) Treg levels strongly increased in most of the Treg populations analyzed, while treatment with paclitaxel only (control arm) had no influence on Treg levels. This is exemplarily shown in Figure 11 for one Treg population at baseline and at week 6 (and week 24) for patients evaluable in Arm A (A) and Arm B (B). Even though the effect of IL-2 on Tregs is well known, the strength of the induction is impressive.
In addition, 6 different MDSC populations were analyzed in detail (for details about the different populations, please refer Table 4). While baseline levels of MDSC populations 1, 2 and 4 seemed to be comparable between MCC patients and healthy donors, levels of the MDSC populations 2, 5 and 6 tended to be numerically higher in MCC patients (Figure 12). However, due to the limited sample numbers these differences did not reach statistical significance. Because of the small patient number and the patient variability no clear trend over time could be observed for the different MDSC populations at baseline, week 6 (W6) and week 24 (W24).
In summary, cellular biomarker analyses showed 100 % evaluability rate for Treg and MDSC analyses. 65% of patients (11/17) were evaluable for CD8 MCC-specific T cell responses on the basis of their HLA restriction.

Figure 11: Quantification of regulatory T cells (Tregs). Percentage of Tregs 1 (CD3+ CD4+ CD8- CD25high CD127low Foxp3+) before and after treatment in patients of Arm A (F16-IL2 plus paclitaxel) and Arm B (paclitaxel alone).

Figure 12. Quantification of Myeloid-derived Suppressor Cells (MDSCs). Percentage of MDSCs 2, 5 and 6 (normalized to lymphocytes) in PBMCs of MCC patients and healthy donors.

Table 4: Prospectively defined MDSC populations within multicolour panel
Population Phenotype Type Reference
MDSC 1 CD14+ CD124+ monocytic J Immunol. 182, 6562-6568

MDSC 2 CD15+ CD124+ granulocytic J Immunol. 182, 6562-6568

MDSC 3 Lin- HLA-DR- CD33+ other Cancer Res 66, 9299-9307

MDSC 4 CD14+ HLA-DR- SSCim monocytic J Clin Oncol 25, 2546-2553
MDSC 5 CD11b+ CD14- CD15+ granulocytic Cancer Res. 65, 3044-3048
MDSC 6 CD15+ FSClo SSChi granulocytic Sci. Rep. 5, 15179

1.3.5. WP5 - Correlation of systemic immunological responses and clinical benefit: impact of the microenvironment, immune escape mechanisms and regulatory circuits
Given the difficulty in obtaining biological specimens from MCC patients, it is obligatory that specimens are not squandered on unfocused exploratory studies, or evaluated using assays that are seriously lacking in robustness and reproducibility. Consequently, innovative methods and the respective standard operating procedures (SOPs) had been established within the IMMOMEC project.
While the sample acquisition for FFPE blocks occurred within the anticipated extent, the acquisition of cryopreserved tissue samples was frustratingly slow and low. Thus, we decided to focus on the FFPE tissue samples and to establish the respective analyses planned for the cryopreserved tissue samples on FFPE samples, which was possible by the advent of new technologies such as gene expression profiling by the nCounter Analysis System (nanoString Techologies), TCR clonotype mapping by immunoSEQ (Adaptive Biotechnologies) or immune phenotyping by the Opal System (PerkinElmer). For example, next generation sequencing (NGS)-based T-cell repertoire analyses by ImmunoSeq allows the comprehensive characterization of FFPE tumor tissues, which is the prerequisite for T cell clonality analyses and clonotype tracking, which was formerly only possible if fresh frozen (aka cryoconserved) tumor tissue was available. By comparing 2 cases for which both FFPE and cryopreserved tissue samples were available, we confirmed that these analyses indeed yield comparable results. Unfortunately, however, in situ peptide/MHC-multimer staining to detect specific T cells in tissue, a technique we had previously established for cryopreserved tissue sections, did not work at all in FFPE sections.
With respect to the characterization of MCC- and MCPyV-specific T-cell responses, peripheral blood lymphocytes obtained at screening were available for all 17 patients screened for the IMMOMEC clinical trial. However, only 11 of these samples could be evaluated for specific T-cell responses. This restriction is based on the fact, that only these 11 patients had an HLA-phenotype for which specific MCC- and MCPyV-epitopes are known. For 5 patients MCC- and MCPyV-specific T-cell responses were also evaluated at post-screening.
Over the past year it became increasingly obvious that cancers contain a heterogeneous and dynamic microenvironment communicating with the immune system. The immune contexture of the tumor microenvironment showed to be able to influence the course of the disease. Hence, pre-existing immunity is determining the fate and survival of the patient and the likelihood of response to immunotherapy. Indeed, several reports suggest in different tumor entities ranging from colorectal cancer to melanoma, that the course of cancer disease is controlled by the host’s immune system underlying the importance of including immunological biomarkers for the prediction of prognosis and response to therapy. Quantification of immune cell densities revealed the major positive role of infiltrating T cells for patient’s survival. These reports, however, also revealed that the localisation of the immune infiltrate is equally important than the mere quantification. These findings became the foundation of a new concept: immune contexture, which also takes functional orientation, density, and location within distinct tumor regions of a natural in situ immune reaction into account. The evaluation of the immune contexture, however, is a complex challenge. In our studies we took advantage of the InForm software, which can be trained to quantify not only cells based on their nuclei, their specific form, but also their immunohistological or immunofluorescent staining pattern (e.g. expression of cytokeratin 20 for MCC tumor cells). Moreover, the counted cells can be categorized according to their localisation, e.g. in the stroma surrounding or within the tumor or the tumor itself. Such an analysis is exemplified in Figure 13 for CD8+ T cells infiltrating MCC tumors.

Figure 13. Quantification and categorization of CD8+ T cells in MCC. (A) Low stroma and tumor infiltration, (B) low tumor, dim stroma infiltration, (C) mostly tumor border infiltration, and (D) brisk tumor infiltration.

Using this approach, we scrutinized the lymphocytic (i.e. CD4 – T helper and regulatory T cells; CD8 – cytotoxic T cells; PD-1 – exhausted T cells; FoxP3 – regulatory T cells; CD20 – B cells) and the myeloid infiltrate (CD68 – macrophages; CD163 – tumor associated macrophages/M2 macrophages; arginase – tumor associated macrophages; PD-L1 – presumably tumor associated macrophages), as well as the expression of MICA, MHC class I and PD-L1 on CK20+ tumor cells (MICA – NKG2D ligand; HLA-A – MHC class I; PD-L1 – PD-1 ligand; CK20 – MCC tumor cells).
These analyses demonstrated a very heterogeneous presence and composition of infiltrating lymphoid or myeloid cells. Interestingly, there were a number of trends with respect how these infiltrates were composed: (i) a strong infiltration of M2 polarized PD-L1 expressing macrophages is inversely correlated to infiltration of CD4 and CD8 T cells; (ii) strong infiltration with CD8+ T cells was correlated with the presence of FoxP3 positive regulatory T cells; and (iii) MICA and HLA-A expression on tumor cells was correlated with a stronger CD8+ T-cell infiltration, which was associated with the PD-L1 expression on tumor cells (but not PD-L1 expression on stroma cells). However, probably due to the high number of tested variables with regard to the limited number of analysed samples, none of these trends reached statistical significance.
These morphological characterizations were expanded by a NanoString nCounter platform-based molecular analyses for an immune-related gene expression profile previously described to be a predictive marker for the clinical outcome of an immune checkpoint blocking antibody treatment such as pembrolizumab or nivolumab. The original codeSet was supplemented by 4 genes specifically expressed in MCC cells, i.e. CK20, neuron-specific enolase and the two MCPyV-encoded early genes. After normalization for the tumor content and expression of classical housekeeping genes as well as CK20 and neuron-specific enolase, a hierarchical clustering using either Euclidean distance or Spearman correlation was performed. While neither clustering method clearly separated MCC lesions characterized by brisk or dim CD8+ infiltrate, individual genes did. As exemplified for mRNA expression encoding the non-classical MHC molecule HLA-E, granzyme B and interferon gamma, these were either inversely (HLA-E) or directly (granzyme B and interferon gamma) correlated with the degree of the CD8+ T-cell infiltrate.
The next series of experiments addressed the T-cell receptor repertoire usage of the T-cell infiltrate in MCC lesions by ImmunoSeq technique. Both the number of TCR transcripts as well as the number of overexpressed TCR clonotypes varied substantially in the analysed samples. Notably, the TCR high clonality group was associated with a brisk CD8+ infiltrate, higher numbers of FoxP3+ cells as well as a low numbers of M2-polarized macrophages. Again, these correlations did not reach statistically significance, probably due to the high number of tested variables in regards to the limited number of analysed samples. When we correlated the clonality of the TCR repertoire usage in the MCC lesions with the immune-related gene expression profile described above, none of the applied hierarchical clustering methods clearly separated MCC lesions characterized by high or low TCR BV clonality. However, similar to the brisk and dim CD8+ T cell infiltrate the same individual genes did.
When we set obtained results of the characterization of systemic and localized immune responses into an associative context, a number of conclusions became obvious: (i) systemic and localized T cell responses are comparable if the tumor cells express MHC class I molecules and there is a low abundance of M2-polarized macrophages in the tumor, (ii) a high PD-L1 expression on tumor cells is associated with both systemic and in situ T cell responses, and (iii) a TCR repertoire usage of TIL characterized by a high clonality seems to be associated with systemic T cell responses.
Based on the disease-specific mortality rate, MCC is more lethal than melanoma. Still, spontaneous remissions of both primary MCC as well as metastatic lesions are frequently reported and explained by adaptive immune responses; thus, MCC appears to be a prime candidate for immunotherapy. Indeed, recent clinical trials demonstrated the efficacy of immune checkpoint blocking antibodies. However, at least half of the patients in these trials were characterized by a primary resistance to checkpoint blockade, and a relevant proportion of the responding patients developed secondary resistance. Characterization of the mechanisms underlying immune-resistant cancer progression may contribute to the rational design of strategies to improve the efficacy of immunotherapy in patients suffering from advanced MCC.
Natural Killer group 2D (NKG2D) is a lectin-like type 2 transmembrane receptor encoded by the gene Klrk1 (killer cell lectin-like receptor subfamily member 1), and is part of a critical pathway signaling cellular stresses to the innate and adaptive immune system. Charged residues in the transmembrane region enable NKG2D to pair with the signaling adaptor protein DAP10, which is essential for NKG2D surface expression and downstream signaling to PI3K and GRB2. These signaling molecules then stimulate proliferation, cytokine production, immune cell activation, and cytotoxic potential of NK and T cells. A recent study suggested a link between the Natural Killer group 2D (NKG2D) receptor system and up-regulation of immune responses to MCC. Specifically, transcriptional analyses of MCC tumors revealed that NKG2D was among the highest expressed mRNAs in tumors obtained from patients with a good prognosis. However, these tumors represented a minority of patients, suggesting most MCCs evade NKG2D signaling as a means of immune escape.
The NKG2D ligands include UL16-binding proteins (ULBPs) as well as the MHC class I chain-related protein (MIC) A and B family. MICA and MICB are present at low to undetectable levels in normal cells, but are induced by cellular stresses including infectious agents and neoplastic transformation. Indeed, MICA and MICB are highly expressed in a number of solid tumors like carcinomas of the breast, colon, kidney, ovary, or prostate, as well as in melanoma. However, NKG2D expression renders tumor cells more susceptible to elimination by the immune system. The importance of MICA and MICB induced NKG2D-signaling for immune surveillance of virally infected and transformed cells is highlighted by the fact that viruses and cancer cells have developed mechanisms to interfere with this interaction. These mechanisms include shedding of surface-expressed molecules, binding and retaining of MICA and MICB proteins in the cytoplasm, over-expression of MICA and MICB mRNA-targeting microRNAs, as well as other epigenetic mechanisms such as chromatin remodeling.
MICA and MICB are not or only slightly expressed by MCC tumors in situ and completely absent on MCC cell lines in vitro. We demonstrated that the lack of MICA and MICB mRNA and protein expression in MCC is largely due to epigenetic silencing via histone hypo-acetylation in their promoter region. This epigenetic silencing is very robust even in the presence of several well-established stress factors known to induce their expression. However, this silencing can be abrogated by treatment with HDAC inhibitors both in vitro and in vivo. Since the ultimate goal of our studies was to establish a therapeutic approach for MCC, we used a clinically relevant concentration, i.e. concentrations attained in patients treated with the FDA approved HDAC inhibitor vorinostat (ZolinzaTM). Although this concentration of vorinostat increased histone acetylation at the MICA and MICB promoter as well as subsequent mRNA and protein surface expression in MCC cell lines, the effects were not very strong. Classical MCC cell lines grow as 3D-cultures in large spheroids and therefore represent the in vivo situation of a solid tumor much closer than other cancer cell lines. To increase the susceptibility of cancers to HDAC inhibitors, they are frequently combined with other drugs. Mithramycin A is a gene-selective Sp1 inhibitor, which has been reported to potentiate HDAC inhibitor-induced transcriptional activation. To this end, promoter acetylation of MICA and MICB genes and subsequent mRNA and protein expression were markedly enhanced in MCC cells upon this combined treatment, which resulted in an increased susceptibility to cellular cytotoxicity (Figure 14).

Figure 14. Inhibition of HDACs in MCC cell lines increased their susceptibility to LAK cell mediated lysis, which is is subdued by MICA/B blockade. (A) The gating strategy of the flow cytometry based cytotoxicity assay is illustrated for untreated and treated BroLi cells; target cells were gated as CFSE positive cells in an FSC/CFSE plot, lysed target cells were defined as 7AAD/CFSE double positive cells. (B) Untreated (grey), vorinostat (V, light blue), mithramycin A (MA, turquoise), or the combination thereof (V+MA, dark blue) treated BroLi and MKL-2 cells served as target cells for LAK cells in a 4h cytotoxicity assay.

Potential Impact:
1.4. Potential Impact
The overall aim of IMMOMEC was an increase in public health with respect to a rare, but highly aggressive skin cancer, i.e. MCC, affecting mostly the elderly, not only in the member states, but worldwide. Moreover, we aimed at strengthening the competitiveness of biomedical research in Europe with respect to both clinical as well as translational research. The ambition was to provide a proof-of-concept study of an innovative form of immunotherapy using antibody-targeted interleukin-2 in a well characterized, highly immunogenic cancer entity. The clinical trial was accompanied by a comprehensive translational research program comprising newly established immune monitoring techniques allowing the analyses of localized immune responses in the tumor microenvironment and systemic responses in the circulating blood. We further assumed that the obtained results would open new avenues for a generic adaptation of this concept of immunotherapy to other cancer entities.
1.4.1. Potential impact of the results of WP1
Unfortunately, due to the reasons described above, which were at least in part beyond the potential of the consortium, the clinical trial implemented within IMMOMEC did not reach sufficient recruitment to provide significant results. A cornerstone of the Lisbon Strategy is to improve the competitiveness and increased productivity of the European research community and industry. Even though the IMMOMEC consortium despite all efforts was not able to recruit a sufficient number of patients into the clinical trial to deliver informative evidence of the clinical efficacy of an antibody-targeted interleukin-2 treatment, it still had a substantial impact on the standing of the consortium and thus the European Community in the field of clinical and translational research and development in rare cancers and tumor immunology. Most European centers caring for patients with MCC are organized within the EORTC. However, no multicenter trials for MCC had been organized by this European organization in the past. The direct interactions of European sites with respect to clinical and translational research in MCC have now been established as IMMOMEC has intensified the cooperation of centers from all over Europe with regard to both MCC patient care and translational research. The implementation of such a tightly organized network had strengthened the importance of European researchers in the field of MCC. Until recently, most clinical trials initiated by large pharmaceuticals companies (‘big pharma’), even those with a European background, are coordinated by the US. The demonstration of a well functioning European network conducting a randomized clinical trial in a very rare cancer improved the standing of European groups to such a degree, that it was subsequently possible to launch an European investigator initiated trial (IIT) for the adjuvant treatment of MCC using the CTLA-4 blocking antibody Ipilimumab (ADMEC, EudraCT 2013-000043-78). This trial is conducted largely with participation of the IMMOMEC consortium funded by an educational grant provided by Bristol-Myers Squibb. Notably, this trial would not have been possible without the network established in IMMOMEC and the experience the consortium accumulated during the term of the project. To date more than 40 patients had been screened and almost 35 enrolled into the ADMEC clinical trial. Similar to IMMOMEC, ADMEC is accompanied by an innovative translational research program based on the established techniques, results and experience obtained within IMMOMEC – however, lacking the same generous financial support by the EC, less comprehensive.
1.4.2. Potential impact of the results of WP2
Major advances in cancer therapy over the past years were due to a better understanding of the tumor biology and immunology, thereby allowing a more personalized therapy improving the lifespan and quality of life for the concerned patients and their relatives. Within IMMOMEC we analysed an archived set of clinically annotated tumor tissues and serum samples, thereby we identified a number of biomarkers serving as surrogate for tumor burden and in part predicting patient prognosis. Moreover, we established the relevance of the biologic pathways underlying these biomarkers. However, retrospective analyses of archived bio-samples, even collected in a prospective observational registry, have to be confirmed in a prospective study. While this validation could not be sufficiently done using the prospectively collected bio-samples within the IMMOMEC trial, this validation is planed. Indeed, within the ADMEC trial, which is active in several of the IMMOMEC centers, tumor samples are prospectively collected.
We established a pseudonymized online registry for MCC patients meant for, but not restricted to the IMMOMEC consortium. In fact, the input boxes are open to any interested participant. Methods for data validation and safeguarding the uniqueness of the data set had been developed and implemented. Data from the MCC registry which now comprise almost 1000 cases was used in the development of the European guidelines for diagnosis and treatment of MCC. In the future this database is likely to be used as a benchmark to compare new approaches in the clinical management of MCC patients. These approaches are not necessarily restricted to immunotherapy of advanced disease, but also comprise the initial therapy of primary or locally advanced MCC such as safety margins, sentinel lymph node dissection or adjuvant radiation.
We established several techniques for the comprehensive immune profiling using FFPE samples. These techniques include complex gene expression analyses, T-cell receptor repertoire analyses, and multiplexed immunofluorescence for immune infiltrate characterization. These techniques are currently used to scrutinize MCC lesions to dissect possible predictive biomarkers as well as escape mechanisms to immune checkpoint blocking antibodies. Furthermore, the detailed SOPs will be published and made available via as soon as the respective publications have been accepted for publication.
1.4.3. Potential impact of the results of WP3
MCC is one among several cancers that are known to have a viral origin. Worldwide, the WHO International Agency for Research on Cancer estimated that in 2002 17.8% of human cancers were caused by infection, with 11.9% being caused by one of seven different viruses. The oncogenic process in most cases originates from the integration and expression of oncogenes in the host genome – taking the first step towards the development of a cancer cell. This is of essential importance as far as these cancers hold viral components that are ideal targets for immune recognition, since these are both foreign to the immune system and essential for cancer cell survival. Thus, immunological recognition of the integrated oncogenes was shown to be a very relevant strategy for cancer therapy of e.g. cervical cancer induced by human papilloma virus, and with novel knowledge about immune recognition of MCPyV from our laboratory, we hypothesize that also this virus-induced cancer can be efficiently targeted by immunotherapeutic strategies. Especially in the light of a number of new immune modulatory drugs (antibodies blocking CTLA4, PD1, PD-L1, and other immune checkpoints) that have recently been approved for other types of solid cancer. These immunomodulating agents may induce a shift towards an immune-stimulatory environment in the patient – and together with specific targeting of T cells towards MCPyV-encoded antigens enable the generation of efficient anti-tumor immune responses against MCC.
Understanding the precise peptide-MHC targets on tumor cell that are relevant for immune-cells recognition and killing will allow the monitoring of immunotherapeutic strategies to understand and foresee the efficacy. Further, such therapies, like F16-IL2 may be combined with specific immune cell targeting toward MCC-derived epitopes, through vaccination strategies or adoptive cell therapy.
1.4.4. Potential impact of the results of WP4
IMMOMEC has a substantial impact on the standing of the European Community in the field of immune monitoring. For the IMMOMEC clinical trial, we showed that it is feasible to establish a network of specialized cell culture laboratories for the standardized and GCP-compliant collection and associated logistics of patients’ PBMC with a success rate of 100%. Such validated and highly standardized PBMC collection is a prerequisite for every clinical trial including elaborated immune response analyses, and the associated laboratories are now trained to collect PBMCs of the same quality in future trials.
Because the available PBMC numbers per patient were limited (due to limited availability of blood for each blood drawing time point during immunotherapeutic intervention), the establishment of cell saving techniques was an important step to generate a maximal data set out of the inherently limited biological material in the IMMOMEC trial. These new tools will be used in future clinical trials to monitor patients.
Moreover, the standardized and centralized T-cell immune monitoring in this multi-center, randomized Phase II clinical trial using cutting edge methods define a new state-of-the-art and enabled the collection of proof-of-concept data of the immune competence in MCC patients before and after treatment with paclitaxel alone or in combination with F16-IL2. One of the findings of the immune monitoring was the strong and consistent increase of Tregs after treatment with F16-IL2. Interestingly, even if the prognostic value of regulatory T cells in cancer remains controversial, high Treg infiltration seems to be significantly associated with shorter overall survival in the majority of solid tumors. . However, Sihto et al. published that in a cohort of more than 100 Finnish MCC patients with MCPyV DNA-positive cancer FoxP3+ tumor-associated lymphocytes were correlated with better prognosis. Considering the F16-IL2 induced increase in Tregs observed in the IMMOMEC trial, this publication gains in importance even if it must be considered that for IMMOMEC only peripheral Tregs were measured.
Unfortunately, the number of samples collected for IMMOMEC was small, and thus the possibilities for comparison of immune cell populations between both treatment arms were limited. However, the pre-treatment analyses of the blood cell population were helpful to get more insights into the basic principles of MCC and may help to develop new therapy options. For example, one option could be a peptide vaccination using peptides for which pre-existing T cell reactivities were found in the MCC patients analyzed. As more than 50% of the evaluable patients showed at baseline MCC-specific T-cell responses to at least one of the MCC-specific peptides, this may lead to a strategy to identify patients which are more likely to respond to vaccination with MCC-specific peptides. The feasibility of such a concept of pre-immunogenicity testing has recently been shown within GAPVAC, another EU FP7-funded project. In addition, the immunogenic MCC epitopes identified could be used for immune monitoring in other MCC trials.
In summary, work performed in WP4 of the IMMOMEC project has a potential impact on technical advances in cutting edge immune monitoring that may be beneficial for many future trials in the growing field of cancer immunology as well as on our understanding of the immune biology of MCC and potential future immunotherapeutic approaches for this disease.
1.4.5. Potential impact of the results of WP5
Recently, many exciting developments have led to new, effective cancer immunotherapies. Immune checkpoint blockade, cytokines with and without tumor targeting, as well as adoptive T-cell transfer with and without chimeric antigen receptors results in objective, long lasting clinical responses with response rates, speed and depth even in advanced tumor stages. However, a majority of patients still do not benefit from therapy. Predictive biomarkers for response to immunotherapy are immune response gene signatures or the presence of clonally expanded CD8+ T cells within the tumor. Unfortunately, only 20% of the patients’ MCC lesions are characterized by such a favorable immune signature.
As reported from WP4, MCPyV-specific CD8+ T cells are present in the peripheral blood of more than half of the MCC patients tested, and the intra-tumoral infiltration of CD8+ lymphocytes is a positive prognostic marker for these patients. Unfortunately, MCPyV-reactive CD8+ T cells are not fully functional in most MCC patients. This exhausted phenotype was associated with expression of PD-1 on the MCPyV-reactive T cells; notably, PD-L1 expression has been reported for both MCC cells and myeloid cells infiltrating the tumor microenvironment. Signaling via NKG2D may prevent the exhaustion of MCPyV-reactive CD8+ T cells. In addition to restoring pre-existing T-cell responses, induction of NKG2D ligand expression on MCC cells is likely to trigger new T-cell responses. Activation of NK and γδ T cells via NKG2D increases tumor cell killing and thus cross-presentation of antigens, as well as production of chemokines and cytokines attracting and activating CD8+ T cells. Furthermore, naïve CD8+ T cells express NKG2D as a co-activating receptor and binding to NKG2D ligands boosts their activation. In pre-clinical models it is well established that NKG2D ligand over-expression on tumor cells results in an increased priming and activation of tumor-specific CD8+ T cells and long lasting T-cell memory responses even against NKG2D-negative tumor cells: (i) Induction of NKG2D ligands on carcinoma cells boosts anti-tumor effects of CTLA-4 blockade, and (ii) treatment with immune stimulating cytokines such as IL-2 and IL-12 is more effective against NKG2D ligand expressing tumors.
The lack of MICA and MICB expression on MCC cells is likely to contribute to this immunological state as the re-induction of these NKG2D ligands by HDAC inhibition restores the susceptibility of MCC cells to cytotoxic lymphocytes. Thus, HDAC inhibitor-mediated MICA and MICB induction in MCC is likely to enhance the effects of immune therapeutic approaches currently tested in the clinic: (i) autologous MCPyV specific CD8+ T cell transfer (NCT01758458), (ii) the CTLA-4 blocking antibody ipilimumab (NCT02196961), (iii) the PD-L1 blocking antibody MSB0010718C (NCT02155647), or cytokine based therapies using (iv) tumor-stroma targeting antibody-IL2 fusion proteins (NCT02054884) or (v) IL12-encoding plasmids delivered by electroporation (NCT01440816). Consequently, “epigenetic priming” of cancer cells for immune recognition appears to be a valuable addition to current immune therapeutic interventions in MCC.

List of Websites:

Related information

Reported by

Medizinische Universitaet Graz


Life Sciences
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top