Adcitmer® , a new CD56-targeting monomethyl auristatin E-conjugated antibody, is a potential therapeutic approach in Merkel cell carcinoma

BACKGROUND

Merkel cell carcinoma (MCC) is an aggressive skin cancer, whose tumour cells often express CD56. While immune checkpoint inhibitors constitute a major advance for treating patients with MCC with advanced disease, new therapeutic options are still urgently required.

OBJECTIVES

To produce and evaluate the therapeutic performance of a new antibody-drug conjugate (Adcitmer^® ) targeting CD56 in preclinical models of MCC.

METHODS

CD56 expression was evaluated in a MCC cohort (immunohistochemistry on a tissue microarray of 90 tumour samples) and MCC cell lines. Interaction of an unconjugated CD56-targeting antibody with CD56⁺ MCC cell lines was investigated by immunohistochemistry and imaging flow cytometry. Adcitmer^® product was generated by the bioconjugation of CD56-targeting antibody to a cytotoxic drug (monomethyl auristatin E) using the McSAF Inside^® bioconjugation process. The chemical properties and homogeneity of Adcitmer^® were characterized by hydrophobic interaction chromatography. Adcitmer^® cytotoxicity was evaluated in vitro and in an MCC xenograft mice model.

RESULTS

Similar to previous reports, CD56 was expressed by 66% of MCC tumours in our cohort, confirming its relevance as a therapeutic target. Specific binding and internalization of the unconjugated CD56-targeting antibody was validated in MCC cell lines. The high homogeneity of the newly generated Adcitmer^® was confirmed by hydrophobic interaction chromatography. The CD56-mediated cytotoxicity of Adcitmer^® was demonstrated in vitro in MCC cell lines. Moreover, Adcitmer^® significantly reduced tumour growth in a MCC mouse model.

CONCLUSIONS

Our study suggests that Adcitmer^® should be further assessed as a therapeutic option in patients with MCC, as an alternative therapy or combined with immune checkpoint inhibitors.

[Adoptive transfer of T lymphocytes]

Within a few years, the success of treatments based on the use of T-cells armed with a chimeric T-receptor for the CD19 molecule (CAR-T CD19) has revolutionized the perception of adoptive transfer approaches. The levels of responses observed in acute leukemias, of the order of 70-90 % are indeed unprecedented. The medical and financial enthusiasm aroused by these results has led to the current situation where more than 300 clinical trials are under way, against some thirty different antigens. This enthusiasm, well justified by the first successes, must however be tempered by the difficulties associated with the use of these cells. Indeed, the management of patients is made very complex both for medical reasons, because the toxicities associated with these treatments are important, and for technical reasons, because the preparation of T lymphocytes for therapeutic use requires dedicated structures. During this same period, knowledge of the mechanisms of regulation of T lymphocytes and the possibilities offered by synthetic biology and techniques of genome engineering have progressed considerably. Combined, they allow envisaging a true "programming" of the T lymphocytes, intended to improve the efficiency of the treatments and the safety of the patients. Medical and industrial perspectives and the role of these approaches in the arsenal of cancer therapies will depend largely on two conditions: the emergence of a robust demonstration of their effectiveness in solid tumors, and the establishment of an acceptable production and distribution model 1.

Selection of Epstein-Barr virus specific cytotoxic T lymphocytes can be performed with B lymphoblastoid cell lines created in serum-free media

Epstein-Barr Virus (EBV)-transformed B lymphoblastoid cell lines (BLCL) are currently used for numerous applications in cellular immunology. Where protocols destined for clinical application are concerned, the final choice of assay is made according to a risk/benefit ratio analysis. In this balance the use of xenogenic or allogenic serum has always been a major concern, as it carries both an infectious and an immunological risk. So far, it is unknown whether serum can be omitted from the entire BLCL selection procedure. In addition, as BLCL have been described as heterogeneous, serum deprivation may affect their antigen-presenting capacity. In the present study, BLCL were generated in the absence or presence of fetal calf serum (referred to as BLCL0 or BLCL(FCS), respectively). Next, in order to assess the antigen-presenting capacity of these cells, we compared the ability of BLCL0 and BLCL(FCS) cells to stimulate the EBV-specific repertoire of the corresponding donor's peripheral blood mononuclear cells in vitro. Our results showed that addition of serum was not essential for BLCL infection and culture, and that as far as we could determine, BLCL0 cells were as effective as BLCL(FCS) in reactivating the EBV-specific T-cell repertoire in vitro. Notably, FCS-specific T-lymphocytes can be detected among the BLCL(FCS)-specific CD4+-CTL. Not only was this latter observation unexpected for an EBV-seropositive donor, but it implied that the BLCL had captured and processed the corresponding FCS-derived solubles antigens; taken together our results emphasized the interest of the possibility to generate BLCL0, both for research and for clinical applications.

Microchimerism and transmission of porcine endogenous retrovirus from a pig cell line or specific pathogen-free pig islets to mouse tissues and human cells during xenografts in nude mice

AIMS/HYPOTHESIS

Pig islets could transmit porcine endogenous retroviruses (PERV) to diabetic patients. Our previous work showed that pig islets expressed low levels of PERV mRNA and were not likely to transmit PERV to human cells in vitro. The real risk of infection during pig tissue xenografts can only be evaluated by in vivo experiments.

METHODS

Nude mice bearing tumours containing human 293 cells were grafted with specific pathogen-free pig islets or PERV-producing pig PK15 cells to determine whether pig cells could transmit PERV to mouse and human cells in vivo. Infection was monitored by PCR, long PCR, RT-PCR and long RT-PCR. As detection of PERV sequences could be due to the presence of residual pig cells, we looked for pig mitochondrial (mt) DNA. Quantitative PCR for PERV and pig mt DNA was done to compare the PERV-to-pig mt (P-to-M) ratio of each sample with the reference ratio for grafted pig cells.

RESULTS

Among 78 mouse tissues from PK15-grafted mice, 54 and 72 were positive for gag and pig mt DNA, respectively. Human tumours developed in these mice were positive for PERV (78%) and pig mt (89%). The P-to-M ratios for mouse tissues and PERV-positive human tumours from PK15-grafted mice were higher than the ratio in PK15 cells. Among 41 tissues from pig islet cell-grafted mice, 7 were positive for PERV (3 lymph nodes, 1 kidney, 2 salivary glands, 1 ovary), and 14 were positive for pig mt DNA. Three of these samples (1 lymph node, 1 kidney and 1 salivary gland) were positive for gag DNA, but negative for pig mt DNA. One human tumour in these mice was positive for PERV DNA. P-to-M reference ratio in grafted islet cells was 0.05+/-0.03. The three PERV-positive lymph nodes contained 78 gag/3 mt copies (P-to-M: 26), 101 gag/3 mt copies (P-to-M: 34), and 4 gag/0 mt copies. The two PERV-positive salivary glands contained 14 gag/1 mt copies, and 28 gag/0 mt copies. The ovary and the kidney contained 46 gag/3 mt and 69 gag/0 mt copies, respectively. The PERV-positive human tumour contained 47 gag/3 mt copies.

CONCLUSIONS/INTERPRETATION

Microchimerism and PERV transmission were frequently observed in both mouse and human tissues during grafting of pig PK15 cells into nude mice bearing human tumours, and sometimes during pig islet xenograft in this model. This strengthens the notion that there is a risk of transmitting PERV during pig islet xenograft.

Long-term follow-up failed to detect in vitro transmission of full-length porcine endogenous retroviruses from specific pathogen-free pig islets to human cells

AIMS/HYPOTHESIS

Islets from specific pathogen-free (SPF) pigs could prevent the transmission of conventional zoonosis, but not endogenous retroviruses (PERV), from pigs to diabetic patients. We previously reported that the pancreas showed the lowest expression of PERV mRNA among pig tissues intended for grafting. This study aimed to determine whether PERV from pig islets infect human cells during co-incubation.

METHODS

Human cells (including highly PERV-sensitive 293 cells) were incubated with SPF pig islet cells under conditions designed to increase contact (a high islet to human cell ratio, extended period of co-culture, and repeated contacts). PK15 and G2 retrovirus-producing pig cells were used in place of islet cells as "positive infection controls". Infection of human cells was monitored on cellular extracts and supernatants by PCR or long PCR, and RT-PCR or long RT-PCR, to detect PERV DNA and mRNA, respectively. Reverse-transcriptase activity was monitored by PERT.

RESULTS

Despite the presence of all PERV sequences in pig islet cells, including full-length inserts, no DNA or RNA for gag, pol, and the 3 env sub-types were detected in any human cell line or blood mononuclear cells incubated with pig islet cells, during an 18-week follow-up period. No PERV sequences or RT activity were detected in supernatants. PERV signals were negative even when the pig islet to human cell ratio was increased to 100:1, the time of co-culture was extended to 5 days and two sequential co-incubations were done. By contrast, all PERV DNA and mRNA were detected in all human cells co-incubated with PK15 or G2 cells. Depending on human cell types, productive or non-productive infections were obtained: full-length PERV RNA and RT activity in supernatants were detected or not; and PERV sequences to previously unexposed human cells by PERV-infected human cells were transmitted or not. Some human cells were not productively infected by PK15 cells but became productively infected after co-incubation with PERV-infected 293 cells.

CONCLUSION/INTERPRETATION

SPF pig islet cells, even with PERV inserts and transcripts, have very little probability of transmitting PERV to human cells during co-incubation. The sensitivity of human cells to stable and productive infection by PERV depends on the cell type. Human adaptation of PERV was observed.

In vitro co-incubation of pig islet cells with xenogeneic human blood mononuclear cells causes loss of insulin release during perifusion: involvement of non-T-cell- and T-cell-mediated mechanisms

Because the different steps of the human cellular immune rejection of pig islets are still poorly understood, our previous work concerned the intensity and mechanisms of the proliferation of human peripheral blood mononuclear cells (PBMC) to adult pig islet cells (PIC). As lymphocyte proliferation is not indicative of alteration of PIC, the present in vitro study evaluated cell-mediated immune effectors possibly involved in impairment of adult PIC. A test was thus developed, based on perifusion analysis of the alteration of insulin release from PIC incubated with different human cells. Compared to PIC incubation alone or with autologous pig splenocytes, seven-day co-incubation with whole human peripheral blood mononuclear cells (PBMC) (n = 18) led to almost complete abolition of basal and stimulated insulin releases (p < 0.0001). This effect could not be reversed by extensive sequential washes before perifusion of PIC, and the number of PIC was decreased by 78% after seven-day co-incubation with PBMC. PBMC are a complex mixture of cells involved in different xenogeneic mechanisms, and two components of this PIC impairment were then detected separately. First, the effect of PBMC against PIC was decreased (p < 0.0001) after removal of either MHC class II+ or CD14+ cells from PBMC. On the contrary, decreasing effect (p < 0.001) on insulin secretion was observed when only plastic-adherent or CD14+ cells were co-incubated with PIC. Additionally, alteration of insulin release from PIC cultured with PBMC or plastic-adherent cells was abolished dose-dependently (p < 0.0001 and p < 0.04, respectively) by gadolinium chloride (which inhibits macrophages), but not modified by cyclosporin A or mycophenolate mofetil which did not alter insulin release from PIC but blocked the proliferation of PBMC against PIC. A second mechanism was also detected, since co-incubation of PIC with purified human T cells remixed with antigen-presenting cells led to a decrease (p < 0.0001) of insulin release. This model based on the alteration of dynamic basal and stimulated insulin secretion provides detailed account of in vitro human cell-mediated impairment of PIC. It shows that the xenogeneic effect of whole mononuclear cells was strong and rapid. A crucial role was played by MHC class II+, CD14+, and plastic-adherent cells. Two mechanisms appear to be responsible for the role of these cells: 1) early direct effect, potentially involved in vivo in primary nonfunction of islets aggressed by monocytes/macrophages; and 2) the presentation of PIC xenoantigens leading to impairment by T lymphocytes, which may be involved in in vivo specific cellular rejection.

Porcine endogenous retroviral mRNAs in pancreas and a panel of tissues from specific pathogen-free pigs

Pigs are potential providers of donor tissues for xenotransplantation (e.g. of pancreatic islets) in Type 1 diabetes. In this context, our group has studied the use of islets from specific pathogen-free (SPF) pigs as a means of reducing the risks of "conventional zoonosis". Although this approach does not prevent the transmission of pig endogenous retrovirus (PERV) to humans, we attempted to determine the presence of C-type PERV mRNAs for gag, pol, and env subtypes as a first descriptive step in the retroviral characterisation of SPF pig tissues (especially pancreas). Using semiquantitative reverse-transcriptase polymer chain reaction with 18S rRNA and beta-actin as internal controls, PERV mRNA levels were compared in a large panel of tissues from SPF and conventional pigs. PERV mRNAs for gag, pol, env-A and env-B were present in all tissues studied from the nine SPF pigs tested. Signals for env-C mRNAs were of much lower intensity than those for env-A and B, and most often undetectable in pancreas. The mRNA levels for gag, pol, env-A, env-B and env-C mRNAs were lower in pancreas (p < 0.01) than in all other tissues. Among other porcine tissues likely to be grafted in man, the highest retroviral mRNA levels were detected in kidney (p < 0.01), followed by liver, lung and heart. Amplified PERV mRNA signals were about 17 times less frequent in pig pancreas than in the retroviral-producing porcine cell line G2, while kidney contained about 6 times more PERV mRNAs than pancreas. The levels of gag, pol, env-A, env-B, and env-C mRNAs also varied between tissues of conventional pigs: PERV mRNA levels were lowest in pancreas, and env-C mRNAs were most often undetectable. For all SPF tissues tested, pol, gag, env-A, env-B, and env-C mRNA levels were in the same range or slightly higher than in corresponding tissues of conventional pigs. In summary, this study of C-type PERV mRNAs in a large panel of tissues from SPF pigs, in the context of our strategy of quality assurance and sanitary control, indicated that PERV mRNA levels were in the same range in SPF and corresponding conventional pig tissues, confirming that the use of SPF pigs would not prevent the risk of PERV transmission to human recipients of xenografts. PERV-A and PERV-B may be mainly represented, and PERV-C much less, in these pig tissues (particularly pancreas). The fact that pancreas expressed the lowest PERV mRNA levels and kidney the highest, among porcine tissues likely to be grafted, could be of interest from a clinical point of view. Pig tissues may differ in their loads of PERV sequences, which could be a factor in the risk of PERV transmission during xenotransplantation.