SAN FRANCISCO – , and a recent analysis of MHC-II–positive tumor features provided some insight into the evolution of that response.
The analysis, which involved RNA sequencing on 58 patients with anti–programmed cell death-1 (PD-1)–treated melanoma and lung tumors and on a subset of matched pretreatment specimens at acquired resistance, also highlighted the Fc-receptor–like 6 (FCRL6) molecule as a potential novel immunotherapy target, Justin M. Balko, PharmD, PhD, of Vanderbilt University Medical Center, Nashville, Tenn., reported at the ASCO-SITC Clinical Immuno-Oncology
“MHC-II functions to present class-II restricted antigens to CD4+ T cells, especially T helper cells,” he said, explaining that the expression is typically confined to the professional antigen presenting cell (pAPC) population, but has also previously been shown to be both constitutively and dynamically expressed on tumor cells.
He and his colleagues showed in athat MHC-II expression on tumor cells had potential as a biomarker for anti-PD-1 response.
The current study was undertaken to further explore the biological implications of MHC-II expression in tumor cells.
“Importantly here, instead of using mRNA for MHC-II, which could be confounded by other cells in the stroma or microenvironment, we performed immunohistochemistry (IHC) for MHC-II specifically on the tumor compartment within these samples,” Dr. Balko said, noting that he and his colleagues were specifically looking for what was different in gene expression patterns in the MHC-II+ tumor cells.
They compared the gene sets that were enriched in HLA-DR+, or MHC-II+, tumor cells within human tumors with those from melanoma cell lines grown ex vivo in culture (which eliminated any confounding factors of RNA data from contaminating stroma or immune cells), and found substantial gene set overlap.
“These were signatures of innate autoimmunity or inflammation, including those describing allograft rejection gene sets, viral myocarditis, and asthma, suggesting there’s a tumor-intrinsic inflammation signal associated with class II expression on tumor cells,” he said. “We also previously showed in the melanoma data set that HLA-DR expression specifically on tumor cells had a strong association with CD4 infiltrate, and a slightly weaker association with the degree of CD8 infiltration within the tumors.”
Similarly, quantitative immunofluorescence of MHC-II expression in 100 triple negative breast cancer tumors showed that those tumors with HLA-DR or MHC-II expression on tumor cells had a greater degree of CD4 infiltrate than did the negative tumors. CD8 infiltrate was also increased, but enrichment was greater toward the CD4 compartment – an interesting finding given that MHC-II presents antigen to T helper cells, Dr. Balko noted.
A closer look at individual genes that were different between class II–negative and positive tumors showed that LAG-3 mRNA was more enriched in the HLA-DR+ tumors, and also in patients who experienced a significant response to anti-PD-1 therapy.
“We also had a small population of samples that were derived from relapsed specimens,” he said. “We performed IHC within a small subset where we had paired tumors from pre-PD-1 response and relapse [to look at] LAG-3+ lymphocytes in the tumor ... and saw significant enrichment of LAG-3 infiltrate in the relapsed specimens. Importantly, all of these tumors were MHC-II+.”
Findings in a mouse model
The functional significance of this was explored using an MHC-II–negative orthotopic model cell line unlikely to induce expression of MHC-II when treated with interferon-gamma in culture; MHC-class-II transactivator (CIITA), the master regulator of MHC-II, was used to transduce the cells, resulting in cells that were “constitutively 100% class II+.”
Immunocompetent mice injected with these cells rejected tumors at a much higher rate, but IHC showed more nonregulatory CD4 cells in mice that did not reject the MHC-II+ tumors, Dr. Balko noted.
Gene expression analysis of the rejection-escaped tumors showed more mRNA for PD-1 and LAG-3, similar to what was seen in the study subjects.
“To see if the effect was truly an increase in PD-1 and LAG-3 on lymphocytes within the tumor microenvironment or in lymphoid tissues, we injected immunocompetent mice with either control or CIITA-positive tumors, and then at 7 days harvested either the contralateral lymph node, the spleen, or the proximal or tumor-draining lymph node,” he said.
This showed increased amounts of LAG-3 and PD-1-positive CD4 and CD8 cells in the tumor-draining lymph node, and more LAG-3 PD-1-positive CD8 cells within the tumor itself.
“To perform a therapeutic study, but also to eliminate any confounding factors of the rejecting mice, we waited 14 days after injection of the tumor cells and only enrolled mice with actively growing tumors. We randomized the mice to treatment with either IgG vehicle control, or anti-PD-1, or the combination of anti-PD1 plus LAG-3, and we had a very substantial [75%] complete response rate in the mice with class II–positive tumors treated with the combined PD-1 and LAG-3,” he said. “Importantly, all of the mice in this study were reinoculated with the [MHC-II–negative] cell line and had complete rejection of any subsequent injection of tumor cells.”
To assess whether any other MHC-II receptors could be expressed in the tumor microenvironment, Dr. Balko and his colleagues turned their attention to the FCRL6 molecule, which has previously been shown to be an MHC-II receptor that is expressed on cytolytic cells.