Immunotherapy Resistance Remains a Puzzle

Antoni Ribas, MD, PhD

Checkpoint blockade immunotherapy has been hailed as a significant advance in anticancer treatment. Yet only a subset of patients experience long-term cancer remission as a result of these therapies, because a significant number of those who initially respond eventually develop resistance.

Why do some patients and not others develop resistance? How can primary and acquired resistance to immunotherapy be overcome? These questions are drawing intense interest from the research community. As investigators seek to unravel the mechanisms of acquired resistance, they also hope to shed light on the components of primary resistance.

The biological and genetic processes that may contribute to relapses were a lively area of inquiry during the 32nd Annual Meeting of the Society for Immunotherapy of Cancer (SITC), held November 8-12 in National Harbor, Maryland. Investigators reported findings about a variety of diagnostic tools to monitor immune responses throughout the treatment cycle in preclinical settings, and about the results of explorations into novel therapies and combinations.

"There’s very little so far that we understand about mechanisms of acquired resistance to immune checkpoint inhibitors,” said Katerina Politi, PhD, an associate professor in Yale University’s Department of Pathology and the Yale Cancer Center, in an interview with OncologyLive®.

Since these therapies are a relatively new modality in clinical practice, the prevalence of acquired resistance among patients treated with these agents is unclear. “I’m not sure we know how many patients are affected by acquired resistance,” Politi said. “It might be different for different cancers.”

Politi, who served as a co-chair of a session on immunotherapy resistance at the SITC conference, has used genetically engineered mouse models to study genetic changes in lung cancer that cause resistance to targeted therapies; now, she is exploring cellular changes that may cause or contribute to resistance to immunotherapy.

Definition and Magnitude of Resistance

Politi’s team is studying tumor specimens before treatment is administered and after resistance develops to determine differences at the DNA or RNA level, and in proteins and immune cells. They are using patient-derived tumor xenografts to understand the impact of the alterations they identify. “We want to be able to validate them functionally,” she said.Antoni Ribas, MD, PhD, a leading authority on immunotherapy and a Giants of Cancer Care® awardwinner, and colleagues have defined 3 types of resistance to immunotherapy that have been manifested in clinical scenarios: (1) primary resistance, in which the cancer fails to respond to therapy; (2) adaptive immune resistance, in which cancer cells are recognized by the immune system but develop ways to protect themselves from attack (may be primary resistance, mixed response, or acquired resistance); and (3) acquired resistance, in which a cancer responds to immunotherapy but progresses after a period of time.¹

The mechanisms that cause these responses are dynamic; they can evolve in response to environmental and genetic factors and to treatment interventions including surgery, chemotherapy, radiation therapy, and immunotherapy. Notably, they may exist prior to the administration of immunotherapy.¹

As of November 12, the FDA had approved 6 checkpoint blockade immunotherapy agents: ipilimumab (Yervoy), an anti—CTLA-4 antibody; nivolumab (Opdivo) and pembrolizumab (Keytruda), both anti–PD-1 antibodies; and atezolizumab (Tecentriq), avelumab (Bavencio), and durvalumab (Imfinzi), all anti–PD-L1 antibodies. These agents were approved based on clinical trials that demonstrated overall or objective response rates ranging from approximately 11% to nearly 70%; in many of the studies, the median duration of response was not reached (Table 1).²

Table. Checkpoint Immunotherapy Response Rates and Duration in Pivotal Trials²

Long-term survival rates for patients with melanoma or non—small cell lung cancer (NSCLC) who were treated with anti–PD-1 inhibitors in early clinical trials exceed what would have been expected with standard therapies in these disease settings. Forty percent of patients with advanced melanoma who started taking pembrolizumab during the clinical trial that led to its initial approval were alive after 3 years, researchers reported at the 2016 American Association of Clinical Oncology Annual Meeting.³

The results from the KEYNOTE-001 study showed that 15% of the patients experienced complete remissions according to immune-related criteria (irRC) after taking the PD-1 inhibitor and that nearly 90% of these responders remain in remission. The median overall survival (OS) was 24.2 months.

At the 2017 American Association for Cancer Research Annual Meeting in April, investigators presented updated findings from CA209-003 that they described as the first report of long-term OS rates in patients with metastatic NSCLC treated with an immune checkpoint inhibitor. After Katerina Politi, PhD approximately 5 years of follow-up (minimum follow-up, 58.25 months) the OS rates for patients treated with nivolumab were 42% at 1 year, 24% at 2 years, 18% at 3 years, and 16% at 5 years.⁴ Researchers said those rates were about 4 times higher than the survival that would have been expected with docetaxel.

Nevertheless, acquired resistance has manifested itself in a significant subset. “It is becoming clear that approximately one-fourth to one-third of patients with metastatic melanoma who have objective responses to checkpoint blockade therapy with anti—CTLA-4 or anti–PD-1 will relapse over time, even despite receiving continued therapy,” Ribas and colleagues wrote.¹

During a presentation at the SITC conference, Valsamo Anagnostou, MD, PhD, included an analysis of progression after response among patients with NSCLC who received checkpoint blockade immunotherapy during 4 clinical trials (Table 2).5 The rates of progression among responders ranged from 19.4% to 48%. Notably, initial response and later progression rates were markedly lower among patients who had received docetaxel, the standard comparator in each trial.

Table 2. Progression Among Responders in Selected NSCLC Trials⁵

Mechanisms of Resistance

Strategies to improve primary response rates and the duration of responses to checkpoint blockage immunotherapies are needed, said Anagnostou, an oncology instructor and researcher at the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University in Baltimore, Maryland. “Despite the clinical benefit, a fraction of patients respond to these therapies and, among the patients who do respond, the majority of them develop acquired resistance. There is therefore an unmet need for understanding mechanisms of response and resistance to these therapies,” she said.Appropriately selected patients who do not experience tumor shrinkage or regression after initial treatment with an immunotherapy exhibit primary resistance. “In that case, it’s most likely that the cancer is not immunogenic enough to have induced an immune response, so we cannot use its checkpoints to turn it on,” said Ribas, director of the Tumor Immunology Program at the Jonsson Comprehensive Cancer Center at UCLA, in an interview. Factors intrinsic and extrinsic to the tumor cell have been identified in ongoing studies as contributing to primary and adaptive resistance mechanisms to immunotherapy.¹

Intrinsic factors include an absence of tumor antigens and genetic factors that prevent immune cell infiltration, while extrinsic factors may be found in the tumor microenvironment; these include myeloid-derived suppressor cells and inhibitory immune checkpoints.1 In acquired resistance, possible mechanisms that have been explored include the loss of T-cell function, lack of T-cell recognition by downregulation of tumor antigen presentation, and development of escape mutation variants in the cancer.1 Ribas and colleagues postulate that each of these mechanism may promote acquired resistance to checkpoint blockade antibodies and to adoptive cell transfer therapies, a category that includes chimeric antigen receptor (CAR) and T-cell receptor therapies.¹

Researchers are discovering possible overlap of mechanisms of primary and acquired resistance to immunotherapy. “The studies analyzing resistance to immunotherapy identified several genetic alterations that are shared between tumors that are primary resistant and those that acquire resistance under therapy,” said Annette Paschen, MD, an associate professor in the Department of Dermatology at the University Duisburg-Essen and German Cancer Consortium, in an email interview. She also discussed resistance mechanisms at the SITC conference.

Lack of T-Cell Recognition

One important mechanism of acquired resistance appears to be genomic changes in T-cell receptor clonality that prevent continued recognition of tumor antigens after an initial response, according to research Anagnostou and colleagues conducted at Johns Hopkins.

Investigators analyzed the genomic evolution of 4 patients with NSCLC who had initially responded to anti—PD-1 or anti–PD-1/CTLA-4 therapy and then developed resistance.6 To do so, they performed genome-wide sequencing of protein coding genes and T-cell receptor clonotype analysis on paired tumor specimens, followed by evaluation of eliminated neoantigens in peripheral blood T cells. They also analyzed neoantigens in 1 patient with head and neck squamous cell carcinoma who had been treated with a combination of PD-1 and KIR blockade therapy.

The investigators found that NSCLC tumors that acquired resistance shed between 7 and 18 mutation-associated neoantigens (MANAs) through the elimination of tumor subclones or through the chromosomal loss of truncal alterations.6 Some of the MANAs that were eliminated were in genes that encode peptides recognized by T cells in peripheral blood. This “therapy-induced immune editing of MANAs” appeared to allow the tumor to avoid attack, the investigators wrote, although they noted that larger studies would have to be conducted at various times of relapse and response to confirm these findings.

The focus on neoantigen changes during therapy is consistent with prior studies that have described tumor mutational load as an indicator of response to checkpoint blockade antibodies, Anagnostou noted during her SITC presentation.

“Responses are expected in tumors with a very high mutational load such as mismatch repair— deficient colon cancers as well as in tumors with known environmental exposures such as lung cancer and melanoma, whereas responses are less expected in tumors with a low mutational load such as [pediatric] neuroblastoma,” she said. The next step for the research team was to determine whether changes in the neoantigen landscape could be accurately assessed through circulating cell-free DNA (cfDNA) isolated from the plasma of patients at any stage in the treatment timeline. Investigators used a method called targeted error correction sequencing, or TEC-Seq, that was developed at Johns Hopkins to evaluate sequencing changes in cfDNA.

By comparing cfDNA and TCR dynamics during and after treatment, investigators delineated patterns of response. “For a patient with an ongoing response, there is early clearance of cfDNA [by week 4] and evidence of expansion of TCR clones in the circulation,” said Anagnostou. “For a patient with primary resistance, there is no clearance of cfDNA and no change in productive frequencies.” For patients who initially respond but develop acquired resistance, there is a clearance of cfDNA at the time of response followed by outgrowth at the time of resistance. “There is some expansion of TCR clones at time of response followed by a decrease to pretreatment levels at the time of therapeutic resistance,” said Anagnostou.

She said the findings have implications for the development of immunotherapies that target multiple clonal neoantigens to prevent immune escape. Additionally, she noted that liquid biopsies should be incorporated into the design of future clinical trials.

Genetic Changes

The evidence thus far suggests that tumors that develop resistance to checkpoint immunotherapy are highly genetically similar to tumor samples taken prior to the development of resistance. A study published in Clinical Cancer Research used genetic and immunologic analysis to study tumor evolution over 5 years in a patient with cutaneous melanoma.⁷

The 60-year-old male patient received anti— PD-1 therapy from the time of metastasis to death; at that time, some metastases were responding to therapy, while others were progressing. Researchers conducted whole-genome sequencing on 26 responsive and nonresponsive tumor specimens (4 premortem, 22 postmortem), but they found few genetic differences.

“At the genomic level, the tumors did not display a significant number of genetic subclones with respect to the cancer cell fraction,” wrote Ascierto and collaborators. “…We found remarkable genetic similarities among regressive/progressing cutaneous metastases, indicating that all were derived from a single subclone that had not diverged significantly.” Investigators concluded that the patient’s disease was driven by a functional loss of NF1, which has been associated in earlier studies with RAS activation in melanoma cells and with MAPK pathway reactivation in resistance to BRAF inhibitors.⁸

Notably, they did not find significant differences in progressing versus regressing metastases either in PD-L1 expression on tumor cells or infiltrating immune cells or in CD8 T-cell densities in intratumoral or peritumoral locations. However, global gene expression profiling revealed distinct gene signatures in progressing and regressing tissue samples. The most differentially expressed gene, LAMA3, a subunit of laminin-5, was also overexpressed at the protein level in progressing metastases. Laminin-5 is associated with epithelial-to-mesenchymal transition and poor prognosis in lung cancer and with the development of aggressive melanoma in culture.⁷

Interruptions in the Interferon-Gamma Pathway

The interferon-gamma (IFNγ) signaling pathway is a critical part of the immune response, and multiple research studies now suggest that defects along the pathway play a key role in immunotherapy resistance. Notably, JAK1/2 mutations, which affect IFNγ signaling, have been found in melanoma samples exhibiting primary resistance, as well as in tumors that developed acquired resistance.

A study published in the New England Journal of Medicine analyzed paired tumor samples of 4 patients with metastatic melanoma who relapsed while receiving pembrolizumab. (The mean time to relapse was 624 days, with a range of 419-888 days). The researchers compared tissue samples from baseline and relapsing lesions through whole-exome sequencing. In the relapse samples of 2 patients, they identified acquired loss-of-function mutations in JAK1 and JAK2, resulting in a lack of response to IFNγ. A truncating mutation in B2M, which encodes the antigen-presenting protein beta-2-microglobulin, was identified in a third patient.⁸

Paschen and colleagues analyzed exome data derived from the cell lines of patients with metastatic melanoma to demonstrate a “genetic evolution” of resistance to IFNγ. 9 By screening cultured melanoma cell lines for genes in the IFNγ signaling pathway, investigators found mutations in JAK1, JAK2, and STAT1 that undermined the immune response and resulted in T-cell—resistant human leukocyte antigen (HLA) class I–negative lesions.

“Mutations abrogating IFNγ signaling in tumor cells protect from cytokine-induced cell death and cell cycle arrest, and allow the tumor cells to maintain an HLA class I—low phenotype, which can protect against T-cell recognition. PD-L1 expression is no longer inducible by IFNγ, which makes the tumor cells resistant to anti–PD-1 therapy,” Paschen said in the interview.

Additionally, the study suggested that JAK mutations may occur early on—or even prior to—the development of resistance. Paschen and colleagues identified 10 JAK1/2 mutations in 59 tissue samples from patients undergoing anti-PD1 therapy; investigators also identified mutations in IFN signaling pathway genes in 10 of 110 pretreatment biopsies from patients who went on to receive anti—CTLA-4 therapy.⁹

CRISPR Gene Editing Proves Useful

Although the authors “could not detect significant differences in responses to anti—PD1 or anti-CTLA-4 treatment between patients with and without mutations,” they noted that 2 of 3 patients with “clearly inactivating mutations” experienced cancer progression while undergoing anti–PD-1 immunotherapy. The third patient showed a partial response to treatment. Importantly, investigators determined that melanoma cells with JAK1/2 mutations “progress to a ‘higher level’ of immunotherapy resistance,” evolving from IFNγ-resistant melanoma into HLA class I–negative lesions with complete resistance to CD8-positive T-cells.Novel research into mechanisms of resistance to immunotherapies presented at SITC included findings about the dual role of type II IFNγ signaling in tumor cells and in the tumor microenvironment. Investigators from the University of Chicago explored the role of mutations in the type II interferon pathway affecting IFNγR2 and JAK1 activity.10 They identified these mutations for study through the use of CRISPR gene-editing technology; they then generated B16.SIY cells that selectively lacked IFNγR2 or JAK1 and implanted these cells into mice.

Although these aberrations are associated with resistance to immunotherapy in vitro, the mutant tumors were controlled better when implanted into mice. The investigators repeated the experiment with multiple IFNγR2- and JAK1-deleted cell lines using different guide-RNAs and with another tumor cell line with similar results. The team also reintroduced IFNγR2 and found that tumor growth was restored and antigen-specific CD8-positive T cells increased in the tumor microenvironment.

The investigators theorize that IFNγ produced by CD8-positive T cells in the tumor microenvironment “might have a dominant negative effect through upregulation of inhibitory factors such as PD-L1 on tumor cells,” they wrote in their conference abstract. They observed that tumor cells that were positive for IFNγR2 and JAK1 in vivo reduced expression of PD-L1 and IDO, a tumor suppressor checkpoint, compared with wild-type tumors and that transduction to express PD-L1 restored the ability of IFNγR2-negative tumors to subvert the host immune response.

Strategies to Overcome Resistance

They concluded that whether IFNγ on tumor cells has a positive or negative impact on tumor control “may depend on whether inhibitory factors like PD-L1 are dominantly expressed and functional through tumor cells versus host immune cells.”¹¹As researchers learn more about the underlying causes of immunotherapy resistance, they are also working to develop strategies to overcome resistance.

“The data suggest that if we were better at treating cancer up-front, there would be less acquired resistance,” Ribas said. “I think an important strategy will be to kill more cancer cells with more immunotherapy. If there are no cancer cells left, none can become resistant.”

Exactly how to do so is a question that is being studied. “That [strategy] may be different from one person to another,” Ribas said. In their paper, Ribas and colleagues cited a broad range of strategies being considered.¹

These include partnering a checkpoint blockade antibody with other inhibitory and stimulatory immune-targeting agents, emerging strategies such as adoptive T-cell therapy, and standard chemotherapy or radiotherapy. “Several combination therapies are now being studied in clinical trials,” Paschen said, noting that results so far indicate that “it seems to be more efficient to attack the tumor from ‘multiple sides’ to prevent or overcome resistance.”

As more is learned about genetic changes related to resistance, “it might be possible to identify the patient population that has an enhanced risk of developing JAK1/2 deficiency based on their chromosomal alterations,” Paschen said. Ultimately, “it should become routine to take biopsies from patient lesions in order to define the molecular characteristics of tumor cells and the cellular composition of the tumor microenvironment,” she noted.

“Based on this knowledge, a personalized treatment strategy could then be developed,” Paschen said. “If a tumor shows an irreversible loss of the surface expression of HLA class I antigen, then therapeutic approaches that are dependent on the activity of CD8-positive T-cells—recognizing tumor antigens are no longer applicable. But these tumor cells still could be killed by CAR-modified CD8-positive T cells, and therapeutic approaches based on CD4-positive T-cells recognizing tumor antigens in the context of HLA class II molecules could be followed.”

For now, clinicians can help by obtaining biopsy samples before, during ,and after treatment. These samples will help researchers clarify mechanisms of resistance and advance understanding of acquired resistance.

“We’re going to need a lot more data to be able to really tackle and understand the frequency of different mechanisms of resistance, and then [understand]how to treat them,” Ribas said.

References

1. Sharma P, Hu-Lieskovan S, Wargo JA, Ribas A. Primary, adaptive, and acquired resistance to cancer immunotherapy. Cell. 2017;168(4);707- 723. doi:10.1016/j.cell.2017.01.017.
2. Drugs@FDA: FDA approved drug products. FDA website. accessdata.fda.gov/scripts/cder/daf. Accessed October 29, 2017
3. Robert C, Ribas A, Hamid O, et al. 3-year overall survival for patients with advanced melanoma treated with pembrolizumab in KEYNOTE-001. J Clin Oncol . 2016;34(suppl; abstr 9503). meetinglibrary. asco.org/record/123428/abstract.
4. Brahmer JR, Horn L, Jackman D, et al. Five-year follow-up from the CA209-003 study of nivolumab in previously treated advanced non-small cell lung cancer (NSCLC): clinical characteristics of long-term survivors. Presented at: 2017 American Association for Cancer Research Annual Meeting; April 1-5, 2017; Washington, DC. Abstract CT077. cancerres.aacrjournals.org/ content/77/13_Supplement/CT077
5. Anagnostou V. Evolution of neoantigen landscapes. Presented at: 32nd Annual Meeting of the Society for Immunotherapy of Cancer; November 9-12, 2017; National Harbor, MD.
6. Anagnostou V, Smith KN, Forde PM, et al. (2016). Evolution of neoantigen landscape during immune checkpoint blockade in non—small cell lung cancer. Cancer Discov. 2016;7(3):264-276. doi:10.1158/2159-8290.cd-16-0828.
7. Ascierto ML, Makohon-Moore A, Lipson EJ, et al. Transcriptional mechanisms of resistance to anti—PD-1 therapy. Clin Cancer Res. 2017;23(12): 3168-3180. doi:10.1158/1078-0432.ccr-17-0270.
8. Zaretsky JM, Garcia-Diaz A, Shin DS, et al. Mutations associated with acquired resistance to PD-1 blockade in melanoma. N Engl J Med. 2016;375(9): 819-829. doi: 10.1056/NEJMoa1604958.
9. Sucker A, Zhao F, Pieper N, et al. Acquired IFNγ resistance impairs anti-tumor immunity and gives rise to T-cell-resistant melanoma lesions. Nat Commun. 2017;31:15440. doi: 10.1038/ncomms15440.
10. . Williams J, Li S, Gajewski T, Huang H. A dual and opposing role for tumor intrinsic type-II interferon sensing in tumor establishment and acquired resistance. 32nd Annual Meeting and Pre-Conference Programs of the Society for Immunotherapy of Cancer (SITC 2017): Part One. J Immunother Cancer. 2017;5(suppl 2):86. Abstract O29. doi: 10.1186/s40425-017-0289-3.