MUC4-specific CTLs

To the editors: Wu et al.⁽¹⁾ reported cytotoxic T lymphocyte (CTL) recognition of LLGVGTFVV HLA-A*02 binding MUC4-derived peptide in healthy individuals, and MUC4 has been considered a target in cancer immunotherapy. MUC4 is a membrane-bound mucin that has been identified in various cancer forms, including pancreatic cancer, colon, esophageal and lung carcinoma but not in, for example, normal pancreatic tissue.⁽²⁾ Of interest is also the occurrence of novel tumor-associated antigenic epitopes due to aberrant glycosylation.

The approach used by Wu et al. was to generate mature monocyte-derived dendritic cells (DCs) pulsed with MUC4 peptides predicted and validated for binding to HLA-A*02. DCs were then cocultured with autologous purified CD8⁺ T cells obtained from healthy individuals. After 7 days, responder cells were collected and stimulated additional one to two times with peptide pulsed DCs and IL-2 for a total of 2–3 weeks. Expanded CTLs in coculture with peptide pulsed T2 cells where then used to test for IFN-γ secretion in an ELISPOT assay, or the CTLs were used in a cytotoxicity assay against relevant tumor cells. This procedure is indeed efficient for the identification of potential naive precursor cells in blood, but it carry the risk of DC-mediated priming of these rare cells in vitro. The experiments are informative, as they demonstrate that CTLs can be raised against the identified MUC4 epitopes from precursor cells in healthy individuals. However, for the description of frequencies of existing CTLs, for example in cancer patients it may overestimate the frequency of specific cells.

In contrast to the long expansion period of Wu et al., we pulsed peripheral blood mononuclear cells (PBMCs) for only 8–10 days with peptide and IL-2 before transfer to the ELISPOT assay. We investigated reactivity against four peptides predicted to bind HLA-A*02 (Table 1). We confirmed binding to HLA-A*02:01 and tested induction of peptide specific CTL in an ELISPOT assay in PBMCs from HLA-A*02-positive patients with colorectal cancer. However, we found no patients with responses to the LLGVGTFVV and only one response against the similar LLLGVGTFV peptide out of 15 patients and importantly this response was also identified in 3 out of 10 healthy individuals. Thus, we confirm that MUC4-specific CTLs can be raised in healthy donors, but they do not seem to be a dominant repertoire in colorectal cancer patients. In addition, the expanded CTL’s were not able to lyse T2 cells pulsed with the relevant peptide in a standard chromium release assay.

Table 1.  Half life and affinity to HLA-A*02:01.

Sequence	Half life (h)^a	Affinity (nM)^b
LLLGVGTFV	8.46	68
SMAEVNASV	29.20	134
LLSSPSYSV	3.45	281
LLGVGTFVV	2.20	59

^aStability was measured using radiolabelled β2m and biotinylated MHC heavy chain in a scintillation proximity assay.⁽³⁾

^bAffinity was measured using biotinylated MHC heavy chain and a conformational dependent antibody (W6/32) in a luminescent oxygen channeling immunoassay.⁽⁴⁾

Given these findings, it may be premature to conclude that MUC4-specific CTLs are important in immune responses against tumors and in immunotherapy. Indeed, we have data from ongoing studies that question the processing of glycosylated epitopes for presentation on major histocompatibility complex (MHC) class I molecules. In addition, MUC4-specific CTLs may even be more difficult to generate in cancer patients as compared to healthy controls due to tolerance induction. Finally, the epitopes tested by Wu et al. and in our lab represents unmodified peptides. In contrast, the peptide presented by tumor cells is likely to include posttranslational modifications. In addition, posttranslational modifications may inhibit peptide presentation on MHC in vivo in antigen presenting cells and tumor tartgets. Thus, the in vitro generated wt specificities may therefore in worst case be of limited relevance and identifiable due to lack of tolerance to this nonglycosylated in vitro epitope. Indeed, peptides LLLGVGTV, SMAEVNASV and LLGVGTFVV are predicted to contain O-glycosylation sites (http://www.cbs.dtu.dk/services/NetOGlyc/). However, we can not exclude that the wt peptides are presented and the study by Wu et al. are therefore still important as they demonstrate the possibility to break tolerance against these MUC4 epitopes which may also have more relevance in other cancer forms than colorectal cancer.

Declaration of interest

The authors report no conflicts of interest.