ORLANDO – Patients with metastatic melanoma who have high blood levels of the soluble form of the programmed death-ligand 1 (sPD-L1) have poor clinical outcomes, decreased overall survival, and disease that is resistant to PD-L1 checkpoint inhibitors, compared with patients with low levels of sPD-L1, investigators have found.
High sPD-L1 levels are also associated with an immunosuppressive disease phenotype and with higher levels of pro-inflammatory cytokines, said Roxana S. Dronca, MD, from the Mayo Clinic in Rochester, Minn.
“Measuring soluble PD-L1 levels at baseline could therefore identify patients who have primary resistance to anti-PD-1, or possibly anti-PD-L1, if this molecule is to act, for instance, as a sink for the therapeutic anti-PD-L1. Therefore, it gives us an opportunity to develop a priori combinatorial therapeutic approaches to sensitize resistant patients,” she said at the ASCO-SITC Clinical Immuno-Oncology Symposium, jointly sponsored by the American Society of Clinical Oncology and the Society for Immunotherapy of Cancer.Tumor-induced immune suppression
Membrane-bound, tumor associated PD-L1 has been shown to play a key role in tumor-induced immunosuppression in melanoma and many other malignancies. Expression of PD-L1 on tumors has been shown to be associated with more aggressive tumor biology and with decreased survival in various tumor types, and it was previously thought to be prognostic, she said.
“However, other investigators more recently have found that expression of PD-L1, for instance in metastatic melanoma, is associated with improved survival, possibly reflective of endogenous anti-tumor immunity. So, therefore, the prognostic role of tumor associated PD-L1 is unclear. And also, PD-L1 has been found to be a suboptimal predictive biomarker for response to PD-1 blockade, likely due to heterogeneous and dynamic expression in the tumor tissues, which really cannot be captured with a single-time-point, random tumor biopsy,” she added.
In 2011, Mayo investigators reported on the presence of sPD-L1 (then called B7-H1) in the sera of patients with advanced renal-cell carcinoma and that it was associated with advanced tumor stage and negative clinicopathologic tumor characteristics.
“It seems that the molecule is biologically able to engage PD-1 on circulating T cells, and therefore, it may represent an unanticipated contributing factor to immune homeostasis beyond the tumor microenvironment,” Dr. Dronca said.
Higher levels correlate with outcomes
To see whether sPD-L1 levels are related to outcome and response to immune checkpoint inhibitor therapy in patients with metastatic melanoma, the investigators collected baseline peripheral blood samples from 276 patients with advanced melanoma prior to enrollment in nonimmunotherapy clinical trials, as well as samples from 36 healthy blood donors at their center.
They also evaluated samples from 80 patients who were undergoing anti-PD-1 based immunotherapy, with peripheral blood collected at baseline and each subsequent radiographic tumor evaluation, and serial monthly blood samples from healthy pregnant women (number not specified), with samples taken at 2 hours and at 6 weeks post delivery. Levels of PD-L1 were measured by enzyme-linked immunosorbent assay.
The investigators first observed that sPD-L1 levels rose steadily during pregnancy then fell sharply after delivery, showing the presence of PD-L1 levels in healthy subjects and in a normal model of immune tolerance (that is, pregnancy). This finding is not especially surprising given that PD-L1 was first cloned from human placentas, where it is present in abundant levels and forms a barrier at the fetal-maternal interface, Dr. Dronca said.
They also found that sPD-L1 was significantly higher among melanoma patients than among controls, with a mean level of 1.73 ng/mL, compared with 0.77 ng/mL in controls.
Using receiver operating characteristic analysis, the researchers determined a cutoff value of 0.239 ng/mL to distinguish between low and high levels of sPD-L1.
They found that melanoma patients with levels above 0.293 ng/mL had a median overall survival of 11.3 months, compared with 14.8 months for those with levels of 0.293 ng/mL or lower (P = .04).
They also found that high sPD-L1 levels were associated with resistance to anti-PD-1 therapy. Patients who had complete or partial objective responses had a mean level of 0.3 ng/mL, whereas patients who had unequivocal disease progression at 12 weeks had levels 7.5 times higher.
“Interestingly, at 12 weeks the levels were actually quite stable, both in responders and progressors, suggesting that, maybe, soluble PD-L1 is not only a direct reflection of the tumor load, but as mentioned, it can be released by other immune cells and is possibly a more global marker of immune dysfunction,” Dr. Dronca said.
‘A little bit curious’
Douglas G. McNeel, MD, PhD, from the University of Wisconsin–Madison, the invited discussant, commended the authors for their study and noted that it raises important questions about the role of PD-L1 in healthy and malignant cells.
“Given that PD-L1 expression in most studies seems to be associated with response to anti-PD1, it’s a little bit curious that the soluble PD-L1 would give the opposite finding,” he said.He added that it’s still unclear, but worth pursuing, whether measuring sPD-L1 levels can identify patients who may benefit from anti-PD1 monotherapy versus combinatorial strategies and agrees with the authors’ conclusion that larger studies are needed to establish whether sPD-L1 can be a prognostic or predictive biomarker.
The study was supported by grants from the National Institutes of Health, Mayo Clinic, and Fraternal Order of Eagles Cancer Research Fund. Dr. Dronca disclosed institution research funding from Merck Sharp & Dohme, and other financial relationship with Elsevier. Dr. McNeel disclosed leadership, stock ownership, and consulting with Madison Vaccines, and consulting and/or institutional research funding from Bristol-Myers Squibb, Dendreon, Janssen, Madison Vaccines, and Medivation.