In a landmark paper, Balkwill, et al, evaluated the immune system responses of 54 high-grade serous tubo-ovarian cancer patients undergoing platinum-based neoadjuvant chemotherapy (NACT) (Clin Cancer Res 2016;22(12):3025-3036). In their studies, the post-NACT biopsies had CD4+ T cells that displayed elevated levels of interferon-g (IFN-g) and Th1 gene signatures relative to those samples taken prior to NACT. Although the levels of CD8+ T cells and CD45RO+ memory cells remained fairly constant in the post-NACT tumor microenvironment (TME), fewer FoxP3+ T-regulatory cells (Treg) cells were found in those biopsies showing good therapeutic response. Despite these positive findings for immune response post-NACT, there was also indication of significantly increased expression of immune checkpoint molecules within the TME, such as PD-1, CTLA-4, and PD-L1.
Zhang, et al, found that 102 of 186 frozen ovarian tumor samples contained tumor-infiltrating CD3+ T cells. Of the remaining samples, 72 showed no evidence of T-cell infiltration and 12 were not evaluable (N Engl J Med 2003;348:203-213). Those patients having T-cell infiltration had a 5-year survival rate of 38 percent; however, the corresponding rate for those patients without T cells in their tumor samples was 4.5 percent. Among the 74 patients having a complete response after cytoreduction surgery and platinum-based adjuvant chemotherapy, those having tumor-infiltrating T cells had a 5-year survival rate of 73.9 percent, while for those that did not, the rate was 11.9 percent. In multivariate analysis, the presence of T-cell tumor infiltration was independently correlated delayed death or delayed disease recurrence; additionally, this presence was also associated with elevated expression of intereleukin-2, IFN-g, and other relevant chemokines. Interestingly, elevated levels of vascular endothelial growth factor were associated with the absence of T-cell tumor infiltration.
A 2014 study by Tumeh, et al, evaluated the responses of 46 metastatic melanoma patients to pembrolizumab (an anti-PD-1 monoclonal antibody) therapy (Nature 2014;515(7528):568-571). Their findings showed a direct correlation between the proliferation of CD8+ T cells within the tumor and radiologically-confirmed tumor shrinkage. Additionally, those patients showing good therapeutic responses tended to have higher degrees of clonality in their T-cell receptor (TCR) repertoires prior to treatment.
The present study, which was led by Alexandra Snyder, MD, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York City, sought to determine how carboplatin and taxane NACT affected the TCR repertoires of high-grade serous ovarian cancer (HGSOC) patients. Snyder, when discussing the findings by Tumeh, et al, clarified the concept of clonality, “In their study, they found that the patients that tended to have better responses to the pembrolizumab therapy had higher clonality; that means that they had a less diverse TCR repertoire. Conversely, when we say that clonality is lower, we have a more diverse TCR repertoire.” The data obtained from this study was presented at the Society of Gynecologic Oncology's 2017 Annual Meeting on Women's Cancer (Abstract 36).
In this study, 152 HGSOC patients underwent NACT between 2008 and 2013, and of these, 149 had subsequent cytoredution surgery. “Typically, these patients had three or more rounds of standard carboplatin and taxane chemotherapy prior to their debulking surgeries,” Snyder noted. Of these patients, there were 48 matching samples taken from patients before NACT (often initial biopsies) and then at the time of de-bulking surgery. Of these 48 paired samples, 37 pairs yielded usable data.
Additionally, of the 37 samples providing usable data, 16 were paired samples that were site-matched; that is, the samples came from the same tumor location both before and after NACT. When describing the sequence in which the samples were obtained, Snyder explained, “Biopsies were taken prior to NACT (first sample), then the patient underwent three or more rounds of carboplatin and taxane NACT, and after that, the second tumor sample taken was obtained at the time of the debulking surgery for comparison.”
Tumor content for the samples was confirmed visually by a pathologist. Sequencing was performed on the variableβ region of the CDR3 portion of the TCR. “We sequenced the Vβ region of the CDR3 because that is the most variable part of the TCR, and thus serves as a unique barcode or fingerprint for that TCR,” Snyder explained.
In the specimens gathered, the median number of cells analyzed per sample was 64,011 cells, with a range from 90 to 587,466 cells. Of the cells analyzed, the median number of T cells per sample was 4,009, with a range of three to 124,737.
When looking at the clonality across all tumor sites for initial and post-NACT samples, 73 percent of samples (27 of 37) showed an increase in clonality, while 27 percent (10 of 37) showed decreased clonality. When explaining these findings, Snyder noted, “These global counts were somewhat scattered; however, a different picture emerges when we looked at direct site-to-site comparisons for levels pre- and post-NACT.”
In direct comparisons of levels from the same sites pre- and post-NACT, 88 percent of these samples (10 of 16) showed an increase in clonality while 12 percent (two of 16) showed a decrease in clonality.
When evaluating the data for T-cell infiltration across all sites, Snyder offered the following observations: “The T-cell levels within the tumors across all sites tended to be all over the map, which is similar to what was observed in Balkwill's study; however, as we noted before in our study, when comparing site-specific T-cell levels pre- and post-NACT, a clearer picture emerges. For the direct site-to-site comparison, 69 percent of the samples (11 of 16) showed increased T-cell levels after NACT while 31 percent showed stable or decreased levels post-NACT.”
When discussing these findings, Snyder commented: “The carboplatin and taxane NACT regimen we utilized clearly showed an increase in clonality overall and especially when considering the direct site-to-site sample comparisons.
“Although globally, we did not observe a predominant effect on T-cell infiltration post-NACT, we did see a slight increase in matched site samples. The T-cell levels within a patient at any one time will vary from site to site, thus, we feel the best data is obtained by comparing results from specific matched sites.
“The increase in clonality observed here may have implications for future therapies; as such, changes have been associated with increased responses to immune checkpoint blockade therapy,” she further added.
When asked to explain the implications of their research, Snyder commented, “Based on these findings, we feel that the use of immune checkpoint inhibitors may be warranted after a patient has undergone debulking surgery, as there was a clear increase of clonality for matched sites after the carboplatin and taxane-based NACT utilized here.
“This use of a checkpoint inhibitor could be included at the neoadjuvant and/or the adjuvant setting,” Snyder concluded. “In a planned clinical trial, HGSOC patients will receive the checkpoint inhibitor in combination with neoadjuvant dose-dense paclitaxel and carboplatin for three or more cycles followed by debulking surgery; after cytoreduction, once again, the checkpoint inhibitor will be administered with three or more cycles of adjuvant dose-dense paclitaxel and carboplatin.”
Richard Simoneaux is a contributing writer.