“Check”-ing the Data: A Review of Immune Checkpoint Inhibitor Biomarkers

David Hermel, MD
Resident Physician, Internal Medicine
University of Southern California, Los Angeles, CA

Darren Sigal, MD
Attending Physician, Division of Hematology/Oncology
Scripps Clinic Medical Group, San Diego, CA

Immunotherapy

William Coley’s late 19th-century observation of “spontaneous tumor regression” following injection of streptococcal organisms into the bloodstream of his patients set the stage for more than a century of public debate over the relationship between cancer and the immune system.¹ Only recently, with the success of immune checkpoint inhibitors in a variety of cancer subtypes, have we begun to appreciate the importance of therapeutic strategies targeting adaptive immune pathways exploited by tumor cells.² Monoclonal antibodies directed against anti–cytotoxic T-lymphocyte antigen 4 (CTLA-4) and/or the programmed death-1 (PD-1)/PD-1 ligand 1 (PD-L1) axis have shown efficacy in a range of malignancies, with the most mature data in melanoma, non–small cell lung cancer (NSCLC), and renal cell cancer.³

However, although durable outcomes have been documented in a subset of patients with checkpoint inhibitors, only a minority of all-comers benefit from this therapy. The lack of efficacy in many patients, coupled with the often delayed response to immune checkpoint blockade—with some responses documented up to 6 months after treatment initiation⁴—attests to the need for improved identification of those likely to benefit from immunotherapy. Furthermore, the exceedingly high cost of treatment makes careful utilization of these medications a priority. Therefore, there is great interest in identifying predictive biomarkers that can better assess the value of these novel agents in selected patients. In this review, we describe the landscape of emerging biomarkers that have been implicated as potential predictors of clinical outcome and response to checkpoint blockade.

PD-L1

Highly expressed by tumor cells and associated tumor antigen-presenting cells, PD-L1 (also known as B7-H1 or CD274) interacts with PD-1 receptors on the surface of cytotoxic T cells to downmodulate the T-cell response in tissues. PD-1 is one of many negative regulators of T-cell immunity that tumor cells co-opt to evade immune detection. Immunomodulatory checkpoint inhibitor antibodies, such as PD-1 antagonists nivolumab and pembrolizumab, are designed to directly enhance T-cell function by reversing this process.⁵

The role of PD-L1 as a potential biomarker emerged from initial phase 1 data from 296 patients with a variety of advanced solid tumors randomly assigned to doses of 0.1 to 10.0 mg/kg of nivolumab every 2 weeks in 8-week treatment cycles.⁶ Pretreatment tumor specimens from 42 patients were stained with murine antihuman PD-L1 monoclonal antibody 5H1; PD-L1 positivity was defined as greater than 5% tumor cell surface expression of any intensity in a section containing at least 100 cells. In a nonrandomized population subset, there was a descriptive association of PD-L1 positivity with an increased objective response rate (ORR)—9 of 25 (36%) PD-L1–positive patients responded to nivolumab versus none of the PD-L1–negative patients. This highlighted a possible link between tumor PD-L1 expression and immune checkpoint blockade efficacy.

Expanding on these initial findings, the predictive value of PD-L1 expression was assessed in a randomized, double-blind phase 3 study of 418 previously untreated patients with advanced melanoma assigned in a 1:1 fashion to either nivolumab 3 mg/kg every 2 weeks plus dacarbazine-matched placebo versus dacarbazine 1000 mg/m² plus nivolumab-matched placebo.⁷ PD-L1 positivity was defined as at least 5% of tumor cells exhibiting membrane PD-L1 staining, and PD-L1 expression was assessed by immunohistochemical (IHC) testing in formalin-fixed paraffin-embedded (FFPE) tumor specimens via a rabbit monoclonal antihuman PD-L1 antibody (clone 28-8) and an automated assay by Dako. In prespecified subgroup analysis, regardless of PD-L1 status, the nivolumab-treated patients had improved overall survival (OS) compared with the dacarbazine group, with a numerical trend favoring PD-L1–positive compared with PD-L1–negative nivolumab-treated patients. Although an improved ORR was seen in nivolumab- versus dacarbazine-treated patients in both PD-L1 subgroups, those with positive PD-L1 tumor expression had an ORR of 52.7% (95% CI, 40.8-64.3) compared with 33.1% (95% CI, 25.2-41.7) among those with negative or indeterminate PD-L1 expression. A parallel phase 3 trial in patients with previously treated metastatic melanoma reported similar outcomes (Table 1).⁸

 

Studies of combination immune checkpoint therapy have also assessed the predictive value of intratumoral PD-L1 expression. With similar criteria for PD-L1 positivity and IHC staining as the previous trials, a double-blind phase 3 study was conducted in 945 untreated metastatic melanoma patients randomized 1:1:1 to combination nivolumab and ipilimumab versus nivolumab versus ipilimumab monotherapy.⁹ Independent of PD-L1 tumor status, nivolumab-ipilimumab or nivolumab resulted in longer progression-free survival (PFS) and increased ORR than ipilimu­mab alone. However, in PD-L1–positive compared with PD-L1–negative patients, PFS was increased in the nivolumab group (14 months; 95% CI, 9.1-NR vs 5.3 months; 95% CI, 2.8-7.1) and the nivolumab-ipilimu­mab group (14 months; 95% CI, 9.7-NR vs 11.2 months; 95% CI, 8-NR), with a corresponding ORR increase in the nivolumab group (57.5%; 95% CI, 45.9-68.5 vs 41.3%; 95% CI, 34.6-48.4) and the nivolumab-ipilimumab group (72.1%; 95% CI, 59.9-82.3 vs 54.8%; 95% CI, 47.8-61.6). Interestingly, those with negative PD-L1 expression had an increased benefit from a combination rather than a monotherapy regimen.

In addition to advanced melanoma, the predictive utility of PD-L1 tumor status has been investigated in multiple large trials of PD-1 inhibitors in NSCLC. In previously treated nonsquamous NSCLC, an open-label phase 3 study randomized 582 patients to nivolumab 3 mg/kg every 2 weeks or docetaxel 75 mg/m2 every 3 weeks.¹⁰ An interim analysis revealed an overall significant population-wide increase in OS and ORR in the nivolumab group compared with the docetaxel group. In this study, PD-L1 status was identified after randomization without any prospective stratification for PD-L1 expression. Tumor PD-L1 positivity was defined according to prespecified membrane expression levels (≥1%, ≥5%, and ≥10%), with staining of archived or recent tissue with the Dako assay, using clone 28-8. Across all PD-L1 prespecified expression level cutoffs, nivolumab was associated with significantly increased PFS, ORR, and median duration of response when compared with docetaxel, with a strong predictive association between PD-L1 expression level and benefit from nivolumab. Among patients with PD-L1 <1%, there was no difference in OS between nivolumab and docetaxel; however, nivolumab had a better toxicity profile. In contrast, in a nearly methodologically identical phase 3, open-label study of 272 randomized patients with previously treated advanced squamous cell NSCLC, prespecified expression levels of PD-L1 were not predictive of nivolumab efficacy at any end point evaluated, with similar survival outcomes and ORR across PD-L1 subgroups.¹¹ Overall results favored the nivolumab cohort, consistent with the benefits of nivolumab treatment in the primary population.

In the first study to prospectively enroll patients based on PD-L1 expression, a phase 2/3 study randomized 1034 previously treated patients with advanced squamous and nonsquamous NSCLC and a tumor proportional score (or percentage of neoplastic cells staining positive for PD-L1) ≥1% in a 1:1:1 fashion to pembrolizumab 2 mg/kg, pembrolizumab 10 mg/kg, or docetaxel 75 mg/m² every 3 weeks.¹² The mouse anti–PD-L1 antibody 22C3 and a prototype IHC assay, validated in patients with tumors stained within 6 months of biopsy¹³ and approved by the FDA as a companion diagnostic, were utilized in archived as well as recent samples. In patients with a tumor proportional score ≥50%, OS, PFS, and ORR were statistically increased in low- and high-dose pembrolizumab groups (OS by 6.7-9.1 months; PFS by 0.9-1.1 months; and ORR by 22%-21%) compared with docetaxel. Although patients with a tumor proportional score <1% were not accrued in this study, past analyses of this subgroup¹³ have shown poor response rates (ORR of 10.7%; 95% CI, 2.3-28.2) especially when compared with those with a ≥50% score (ORR 45.2; 95% CI, 33.5-57.3). Table 1 describes additional select studies for PD-1/PD-L1 inhibitors in lung, renal cell, bladder, and gastric cancers.^14-17

Overall, current PD-L1 literature endorses its utility as a predictive biomarker for PD-1 agents with some important caveats. Unique IHC antibody stains and methodologies as well as intrinsic immunobiologic differences between histologies have led to inconsistencies between studies (Table 2). Unlike oncogenic driver mutations, PD-L1 is inducible, with many cytokines including interferon-gamma and treatment modalities such as radiation and chemotherapy able to upregulate expression.^5,18 In addition, PD-L1 status can be misclassified based on the heterogeneous nature of expression within the tumor microenvironment and is complicated by a higher expression rate in larger tumors, regional lymph node metastases, and poorly differentiated tumors.^19,20 Further complications arise from whether to measure PD-L1 expression on tumor cells or tumor-infiltrating lymphocytes (TILs), with some studies suggesting that assessment of both PD-L1 tumor expression and TIL recruitment may improve upon prediction estimations.²¹ Regardless, PD-L1 has been validated as a predictive biomarker for nonsquamous NSCLC, whereas squamous NSCLC with PD-L1 <1% benefited from PD-1 agents. In melanoma, PD-L1 positivity predicted significant benefit from either single-agent PD-1 or combination PD-1 and CTLA-4 therapy. However, PD-L1 negativity was predictive of benefit from the combination therapy. For now, it would be reasonable to consider PD-L1 assays in 2 settings: in nonsquamous NSCLC, where OS benefits of PD-1 agents over chemotherapy were found only among PD-L1–positive tumors; and in melanoma, where a combination of PD-1 plus CTLA-4 agents was statistically beneficial only when tumors were PD-L1 negative.

Tumor Mutational Load

The early efficacy of immune checkpoint therapy in melanoma and NSCLC, both associated with a high mutational burden and chronic exposure to environmental mutagens,²² spurred interest in the idea that tumor neoantigen formation from increased mutational load could facilitate a more robust response from immune checkpoint blockade. A clinically relevant surrogate marker for examining mutational burden in tumors is assessment of underlying mismatch repair (MMR) deficiency. Tumors with MMR deficiency are prone to frameshift mutations due to the accumulation of microsatellites, repetitive regions of DNA sequences, throughout the tumor genome.²³ As much as 50 times the amount of novel somatic mutations may be found in MMR-deficient versus MMR-proficient tumors.²⁴ These make MMR-deficient tumors especially vulnerable to CD8+ T-cell responses that recognize mutation-derived neoepitopes, which can serve to elicit a more robust response to immune checkpoint blockade.

A pivotal phase 2 clinical trial from Le et al evaluated pembrolizumab efficacy in patients with MMR-proficient colorectal cancer (CRC), MMR-deficient CRC, and MMR-deficient non-CRC.²⁵ MMR-deficient CRC and non-CRC cohorts had immune-related ORR of 62% (8 of 13) and 60% (6 of 10), respectively, compared with a 0% (0 of 25) ORR in the MMR-proficient CRC cohort. Moreover, mutation-associated neoantigens and mutational load correlated with improved PFS. Interestingly, investigators found PD-L1–expressing TILs at the tumor-invasive front only in the MMR-deficient cohorts (P = .04), suggesting a mechanistic link between mutational genomic landscape and PD-L1 and TIL expression. Updated results from the non-CRC MMR-deficient cohort were presented at the 2016 Gastrointestinal Cancers Symposium and remained impressive.²⁶

More fundamental analysis in mice identified particular tumor-specific antigens that mediated anti–PD-1 and/or anti–CTLA-4 tumor rejection. Incorporation of these epitopes into vaccines produced an identical response as checkpoint blockade.²⁷ This gave credence to the notion that somatic, nonsynonymous mutations and the resulting mutated proteins presented on cell surface major histocompatibility complex (MHC) molecules may provide the impetus for the amplified CD4+ and CD8+ T-cell response following immunotherapy. Multiple studies have investigated the effect of mutational load and corresponding neoantigen signature on checkpoint blockade efficacy. Whole exome sequencing of tumor DNA from a small subset of patients with metastatic melanoma revealed a significant correlation between long-term clinical benefit from anti–CTLA-4 therapy and increased mutational load.²⁸ This finding was corroborated by a study of pretreatment biopsies from 110 metastatic melanoma patients on ipilimumab in which benefit from ipilimumab significantly correlated with nonsynonymous mutational load.²⁹ Interestingly, in this study, mutational load was a better indicator of clinical response than neoantigen load, despite both being predictive of benefit. Likewise, in patients with NSCLC, whole exome sequencing of tumor DNA revealed a significant association between the median number of nonsynonymous mutations and durable clinical benefit from pembrolizumab.³⁰ Moreover, pembrolizumab efficacy was significantly associated with a higher candidate neoantigen burden, and dual analysis of nonsynonymous mutation burden and PD-L1 expression improved upon prediction of clinical benefit. Nonetheless, attempts to identify consensus neoantigen sequences with associated clinical benefit led to inconsistent results^28,29 as clinical benefit was found in patients with low mutational burden and atypical neoantigen patterns.

Further investigation is necessary to accurately characterize key cancer neoepitopes that emanate from nonsynonymous somatic mutations and promote an adaptive immune response. Additional work on understanding the tumor microenvironment, specifically clonal and nonclonal neoantigen intratumor hetero­geneity,³¹ as well as TIL content and PD-L1 expression, may help improve the predictive value of this biomarker. Future immune therapies may even be targeted specifically to cells that present these neoepitopes in an MHC-I context. Tumor genomic mutational load appears to serve as a surrogate for the presence of neo­antigens and as a potential predictive biomarker for immune therapies. However, even using clinically available massive parallel whole exome sequencing, no defined threshold exists for the number of mutations that predict immune therapy benefit. Presence of MMR deficiency is another surrogate marker for overall tumor mutational burden. Clinical investigation in patients with heavily pretreated cancer, and particularly recalcitrant malignancies, has shown remarkable and protracted responses with PD-1 blockade. Based on these studies, it is reasonable to already incorporate MMR status into clinical care as a predictive biomarker for PD-1 blockade.

TIL Content and the “Tumor Microenvironment”

A robust response to immune checkpoint blockade necessitates tumor lymphocytic upregulation and activation. This was illustrated in the first published clinical report of ipilimumab activity in humans, which documented extensive tumor necrosis and a dominant lymphocytic CD8+ infiltrate in 3 melanoma patients.³² A subsequent study of posttreatment biopsies from 7 melanoma patients treated with tremelimumab found CD8+ T-cell infiltration to be a hallmark of tumor regression, with this CD8+ accumulation presumed to be the main mechanism of the anti–CTLA-4 effect.³³ Further studies aimed to better characterize the presence, location, subtypes, and density of pretreatment TILs—in addition to the immunomodulatory components of the “tumor microenvironment”³⁴—to facilitate identification of the histologic determinants of clinical outcome are needed.

Support for intratumoral pretreatment CD8+ T cells as a biomarker for effective checkpoint inhibition originated from a study of pretreatment and posttreatment biopsies of 46 patients with metastatic melanoma enrolled in a larger phase 1 clinical trial assigned to varying pembrolizumab regimens: 2 mg/kg every 3 weeks, 10 mg/kg every 3 weeks, or 10 mg/kg every 2 weeks.³⁵ IHC analysis revealed that pretreatment CD8+ cell densities at the stromal-tumor margin and tumor parenchymal core were significantly increased in those with a response to therapy when compared with those who progressed during therapy, and CD8+ cell density at the stromal-tumor margin was the most accurate predictive parameter validated on pretreatment biopsies from 15 patients.

On the molecular level, increased expression of innate and adaptive immune system markers, including those associated with cytotoxic T-cell and Th1-cell production, has been correlated with clinical benefit in pre-ipilimumab metastatic tumor biopsies from 45 patients with advanced melanoma.³⁶

Further characterization of the pattern of lymphocytic infiltration and the concomitant expression of TIL and tumor PD-L1, PD-L2, and PD-1 expression was reported in 41 patients from a previously described study⁶ of nivolumab for treatment-refractory solid tumors.³⁷ Two pathologists reviewed FFPE sections from 68 archival or newly obtained tumor specimens, utilizing murine monoclonal antibody 5H1 to stain for PD-L1 expression and adhering to a cutoff of ≥5% membranous expression for positivity. Although tumor PD-L1 expression correlated significantly with ORR to nivolumab (P = .025) and expression was significantly correlated with infiltrating lymphocytes and histiocytes in 33 of 34 cases (P = .001), the same significance was not associated with TIL PD-L1 expression (P = .14). Interestingly, the presence of TILs as an independent factor did not correlate with clinical outcomes. Other components of the tumor microenvironment, such as the CD4:CD8 ratio, TIL or tumor cell expression of PD-L2, CD20+ B-cell presence, necrosis, or lymphoid aggregates did not correlate with treatment response.

Contrary to the previous findings, an early phase 1 study of 227 patients with advanced solid and hematologic malignancies in a dose-finding trial for atezolizu­mab (MPDL3280A) found a strong association between TIL PD-L1 expression on pretreatment FFPE tissue sections (stained with rabbit antihuman PD-L1 monoclonal antibody clone SP142) and ORR to atezolizumab (P = .007).³⁸ There was no significant association found with tumor cell PD-L1 expression. In addition, pretreatment CD8+ T cells in tumors failed to predict the response of atezolizumab, which may be the result of a “nonfunctional immune response” due to a high quantity of TILs lacking PD-L1 expression. Furthermore, gene expression analysis of pretreatment tumor samples found CTLA-4 expression and Th1 gene expression to correlate with clinical responsiveness. Serial biopsies on 28 patients noted that RNA isolated from regressing lesions displayed expression patterns indicative of a generalized activation of CD8+ and Th1 T-cell responses.

Whereas the dynamic nature of the host immune environment makes it difficult to identify consistent predictive biomarkers of checkpoint response, it appears that an inflammatory intratumoral milieu, associated with increased expression of immune pathway markers, is frequently tied to improved clinical response from checkpoint inhibitors. Evidence supports amplification of cytotoxic CD8+ cells, but not necessarily CD4+ cells, as predictors of response to these agents, although more data are needed to corroborate these findings. Overall results suggest that either tumor PD-L1 expression or TIL PD-L1 expression may enhance the predictive value of this biomarker. Future studies must examine multiple histologic features of tumors in unison to aid the predictive utility of TILs in a clinical setting.

Selected Novel Biomarkers

In addition to the putative immune checkpoint inhibitor predictive biomarkers we have reviewed, several other putative markers have accumulating data and are worth mentioning. Peripheral blood myeloid-derived suppressor cells (MDSCs) have been correlated with reduced efficacy of ipilimumab in melanoma patients. MDSCs exert an immunosuppressive effect through production of suppressive molecules, such as ARG1, cytokines, transforming growth factor–beta, or interleukin (IL)-10.³⁹ ICOS+ CD4+ T cells, important in the expansion and survival of T cells, have also been found on flow cytometry to prospectively correlate with improved survival after anti–CTLA-4 therapy.⁴⁰ Furthermore, patients seropositive for the cancer/testis antigen NY-ESO-1 have been shown to have greater clinical benefit from ipilimumab treatment than seronegative patients, suggesting a predictive value for this biomarker in upregulating the immune response to checkpoint blockade.⁴¹ Soluble CD25, which can neutralize IL-2 and/or its receptor subunits, has been shown in 262 metastatic melanoma patients to predict resistance to CTLA-4 blockade at elevated serum levels.⁴² Interestingly, even specific Bacteroides species (B thetaiotaomicron or B fragilis) in the gut microbiota have been shown to influence the efficacy of CTLA-4 therapy.⁴³ Likewise, specific for PD-1 antagonists, therapeutic response has been linked with MHC-II positivity, as assessed by HLA-DR antibodies, on tumor cells.⁴⁴ In addition, PD-1 and PD-L1 have downstream effects on mTOR signaling pathways, and the effector marker phospho (p)-S6 ribosomal protein has been shown to correlate with PFS and OS following anti–PD-1 therapy.⁴⁵

Conclusion

An increasing body of literature has revealed a variety of potential anti–PD-1 predictive biomarkers. In prospective trials, PD-L1 expression and microsatellite instability status provide predictive value for anti–PD-1 agents in a range of malignancies. However, the data for these markers as well as other putative markers continue to emerge. Several important challenges exist before these and other markers can be more broadly adopted in the clinic. First, harmonizing the results of various assays that utilize different IHC antibody stains, specimen preparation, and threshold positivity values will be important. Second, additional data need to be gathered in well-designed prospective or large retrospective analyses to validate the predictive role of mutational burden, TILs, and other novel markers. Finally, comprehensive analyses of these factors—and the interplay of these markers in combination—may help guide future therapeutic decision-making for those eligible for these promising immunomodulatory agents.

References

1. Hoption Cann SA, van Netten JP, van Netten C. Dr William Coley and tumour regression: a place in history or in the future. Postgrad Med J. 2003;79:672-680.
2. Ito A, Kondo S, Tada K, et al. Clinical development of immune checkpoint inhibitors. Biomed Res Int. 2015;2015:605478.
3. Khan H, Gucalp R, Shapira I. Evolving concepts: immunity in oncology from targets to treatments. J Oncol. 2015;2015:847383.
4. Pardoll DM. The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer. 2012;12:252-264.
5. Zou W, Wolchok JD, Chen L. PD-L1 (B7-H1) and PD-1 pathway blockade for cancer therapy: mechanisms, response biomarkers, and combinations. Sci Transl Med. 2016;8:328rv4.
6. Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. 2012;366:2443-2454.
7. Robert C, Long GV, Brady B, et al. Nivolumab in previously untreated melanoma without BRAF mutation. N Engl J Med. 2015;372:320-330.
8. Weber JS, D’Angelo SP, Minor D, et al. Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial. Lancet Oncol. 2015;16:375-384.
9. Larkin J, Hodi FS, Wolchok JD. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med. 2015;373:1270-1271.
10. Borghaei H, Paz-Ares L, Horn L, et al. Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer. N Engl J Med. 2015;373:1627-1639.
11. Brahmer J, Reckamp KL, Baas P, et al. Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer. N Engl J Med. 2015;373:123-135.
12. Herbst RS, Baas P, Kim DW, et al. Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet. 2016;387:1540-1550.
13. Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J Med. 2015;372:2018-2028.
14. Fehrenbacher L, Spira A, Ballinger M, et al. Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial [published online March 9, 2016]. Lancet.
15. Motzer RJ, Escudier B, McDermott DF, et al. Nivolumab versus everolimus in advanced renal-cell carcinoma. N Engl J Med. 2015;373:1803-1813.
16. Rosenberg JE, Hoffman-Censits J, Powles T, et al. Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial [published online March 14, 2016]. Lancet.
17. Muro K, Chung HC, Shankaran V, et al. Pembrolizumab for patients with PD-L1-positive advanced gastric cancer (KEYNOTE-012): a multicentre, open-label, phase 1b trial [published online May 3, 2016]. Lancet Oncol.
18. Peng J, Hamanishi J, Matsumura N, et al. Chemotherapy induces programmed cell death-ligand 1 overexpression via the nuclear factor-κB to foster an immunosuppressive tumor microenvironment in ovarian cancer. Cancer Res. 2015;75:5034-5045.
19. Mansfield AS, Murphy SJ, Peikert T, et al. Heterogeneity of programmed cell death ligand 1 expression in multifocal lung cancer. Clin Cancer Res. 2016;22:2177-2182.
20. Lin Z, Xu Y, Zhang Y, et al. The prevalence and clinicopathological features of programmed death-ligand 1 (PD-L1) expression: a pooled analysis of literatures [published online February 22, 2016]. Oncotarget.
21. Ock CY, Keam B, Kim S, et al. Pan-cancer immunogenomic perspective on the tumor microenvironment based on PD-L1 and CD8 T-cell infiltration. Clin Cancer Res. 2016;22:2261-2270.
22. Lawrence MS, Stojanov P, Polak P, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature. 2013;499:214-218.
23. Vormehr M, Diken M, Boegel S, et al. Mutanome directed cancer immunotherapy. Curr Opin Immunol. 2016;39:14-22.
24. Zhao H, Thienpont B, Yesilyurt BT, et al. Mismatch repair deficiency endows tumors with a unique mutation signature and sensitivity to DNA double-strand breaks. Elife. 2014;3:e02725.
25. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med. 2015;372:2509-2520.
26. Le DT, Uram JN, Wang H, et al. PD-1 blockade in mismatch repair deficient non-colorectal gastrointestinal cancers. J Clin Oncol. 2016;34(suppl 4S). Abstract 195.
27. Gubin MM, Zhang X, Schuster H, et al. Checkpoint blockade cancer immunotherapy targets tumour-specific mutant antigens. Nature. 2014;515:577-581.
28. Snyder A, Makarov V, Merghoub T, et al. Genetic basis for clinical response to CTLA-4 blockade in melanoma. N Engl J Med. 2014;371:2189-2199.
29. Van Allen EM, Miao D, Schilling B, et al. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science. 2015;350:207-211.
30. Rizvi NA, Hellmann MD, Snyder A, et al. Mutational landscape determines sensitivity to PD-1 blockade in non–small cell lung cancer. Science. 2015;348:124-128.
31. McGranahan N, Furness AJ, Rosenthal R, et al. Clonal neoantigens elicit T cell immunoreactivity and sensitivity to immune checkpoint blockade. Science. 2016;351:1463-1469.
32. Hodi FS, Mihm MC, Soiffer RJ, et al. Biologic activity of cytotoxic T lymphocyte-associated antigen 4 antibody blockade in previously vaccinated metastatic melanoma and ovarian carcinoma patients. Proc Natl Acad Sci U S A. 2003;100:4712-4717.
33. Ribas A, Comin-Anduix B, Economou JS, et al. Intratumoral immune cell infiltrates, FoxP3, and indoleamine 2,3-dioxygenase in patients with melanoma undergoing CTLA4 blockade. Clin Cancer Res. 2009;15:390-399.
34. Pitt JM, Marabelle A, Eggermont A, et al. Targeting the tumor microenvironment: removing obstruction to anticancer immune responses and immunotherapy [published online April 10, 2016]. Ann Oncol. 2016;27:1482-1492.
35. Tumeh PC, Harview CL, Yearley JH, et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature. 2014;515:568-571.
36. Ji RR, Chasalow SD, Wang L, et al. An immune-active tumor microenvironment favors clinical response to ipilimumab. Cancer Immunol Immunother. 2012;61:1019-1031.
37. Taube JM, Klein A, Brahmer JR, et al. Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to anti-PD-1 therapy. Clin Cancer Res. 2014;20:5064-5074.
38. Herbst RS, Soria JC, Kowanetz M, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature. 2014;515:563-567.
39. Meyer C, Cagnon L, Costa-Nunes CM, et al. Frequencies of circulating MDSC correlate with clinical outcome of melanoma patients treated with ipilimumab. Cancer Immunol Immunother. 2014;63:247-257.
40. 4Ng Tang D, Shen Y, Sun J, et al. Increased frequency of ICOS+ CD4 T cells as a pharmacodynamic biomarker for anti-CTLA-4 therapy. Cancer Immunol Res. 2013;1:229-234.
41. Yuan J, Adamow M, Ginsberg BA, et al. Integrated NY-ESO-1 antibody and CD8+ T-cell responses correlate with clinical benefit in advanced melanoma patients treated with ipilimumab. Proc Natl Acad Sci U S A. 2011;108:16723-16728.
42. Hannani D, Vétizou M, Enot D, et al. Erratum: anticancer immunotherapy by CTLA-4 blockade: obligatory contribution of IL-2 receptors and negative prognostic impact of soluble CD25. Cell Res. 2015;25:399-400.
43. Vétizou M, Pitt JM, Daillère R, et al. Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota. Science. 2015;350:1079-1084.
44. Johnson DB, Estrada MV, Salgado R, et al. Melanoma-specific MHC-II expression represents a tumour-autonomous phenotype and predicts response to anti-PD-1/PD-L1 therapy. Nat Commun. 2016;7:10582.
45. Kleffel S, Posch C, Barthel SR, et al. Melanoma cell-intrinsic PD-1 receptor functions promote tumor growth. Cell. 2015;162:1242-1256.

Related Articles