Rapid and Complete Response to Combination Anti-CTLA-4 and Anti-PD-1 Checkpoint Inhibitor Therapy in a Patient With Stage IV Refractory End-stage Epithelioid Sarcoma: A Case Report

Summary: Epithelioid sarcoma, in the relapse-refractory setting, has limited expected survival. SMARCB1 inactivation, common in epi- thelioid sarcoma, causes loss of INI1 protein expression and over- expression of the cancer cell growth promoting methyltransferase enzyme, EZH2. We treated a 19-year-old male with stage IV SMARCB1 inacti- vated epithelioid sarcoma presenting with recurrent end stage (Eastern Cooperative Oncology Group Performance Status 4) rapidly progressing bulky disease with combination ipilimumab and nivolumab. He failed standard therapy and an EZH2 inhibitor (tazemetostat). He presented (May 13, 2019) with a large (16.1×18.6 cm) soft tissue back mass extending from T10 to L3. Complete clinical regression of the back mass occurred within 2 weeks (May 28, 2019) of cycle 1 of combined check- point inhibition therapy followed by a positron emission tomography– negative complete remission (October 11, 2019). After a second negative positron emission tomography/computed tomography scan (January 13, 2020), checkpoint inhibition therapy was discontinued. He has returned to normal activities with a normal physical examination and Eastern Cooperative Oncology Group Performance Status of 0 at his last visit (June 29, 2020). In conclusion, combined checkpoint inhibition therapy warrants further study in the salvage setting in patients with epithelioid and other INI1 protein–deficient sarcomas seemingly regardless of prior therapy, extent of disease, and performance status.

Key Words: sarcoma, immunotherapy, gene expression profiling, case report

BACKGROUND

Effective salvage therapy is unavailable for most histo- logic variants of soft tissue sarcoma when standard primary surgery, radiation therapy, and chemotherapy fail.1 While up to 50% of patients with soft tissue sarcoma have identified actionable mutations on next-generation sequencing (NGS), there is little evidence that NGS-guided targeted therapy can induce complete and durable remissions.2 Moreover, pazopa- nib, the first tyrosine kinase inhibitor approved in soft tissue sarcoma for use in the salvage setting, results in significant improvement in progression-free survival in a variety of sar- coma subtypes, but no improvement in overall survival and no reported durable complete responses.3 Progression or lack of response to salvage NGS-guided targeted therapy and pazo- panib in all cases is associated with limited survival (months).4 Epithelioid sarcoma, a rare variant of soft tissue sar- coma, until recently had no effective care options in the relapse or refractory setting with limited expected survival.5 Inactivation of SMARCB1, a core subunit of the SWI/SNF chromatin remodeling complex, results in loss of INI1 protein expression and is commonly found ( > 90%) in epi- thelioid sarcoma.6 INI1 loss results in overexpression of EZH2 methyltransferase, an enzyme that inhibits tumor suppressor gene expression, thus promoting cancer cell growth. A targeted inhibitor of EZH2 methyltransferase (tazemetostat) received Food and Drug Administration (FDA) approval for patients with unresectable epithelioid sarcoma with metastatic or locally advanced disease, based on a single trial of 62 patients (61% received prior chemo- therapy with 15% overall response rate; complete response 1.6%, partial response 13%; duration of response 67% of patients > 6 mo; range: 3.7–24+ mo).7 Immune-based ther- apy, through inhibition of immune checkpoints, is dramat- ically altering clinical cancer care delivery and improving on expected outcomes in the primary and salvage settings, for a broad array of cancer histologic subtypes.8–10 Preclinical and early clinical experience suggests the potential for effi- cacy of checkpoint inhibitors in patients with INI1-deficient sarcoma, warranting further evaluation.11–13

CASE PRESENTATION

A 17-year-old Hispanic male presented with rapidly evolving back pain and progressing arm and leg weakness with urinary and fecal incontinence. Magnetic resonance imaging (MRI) on March 6, 2017 (Fig. 1) showed a paraspinal mass extending from the cervical to thoracic spine with cord compression. Pathology revealed an epithelioid sarcoma. NGS testing revealed a SMARCB1 splice site variant (501-2A > G), microsatellite status of stable, low tumor mutational burden (TMB-2Muts/Mb), and a PTCH1 alteration (R441H). Loss of INI1 expression was confirmed by immunohistochemistry. Positron emission tomography (PET)-computed tomography (CT) showed dis- ease along the entire spine and a metastatic left eighth rib lesion. Between March 2017 and June 2017 he received 5 cycles of adriamycin (37.5 mg daily ×2) and ifosfamide (3 g/m2 daily ×3) every 21 days. He achieved a partial remission on PET-CT scan (May 31, 2017) so had the involved rib removed surgically and received involved field radiation therapy to all areas of spine disease. He had resolution of pain and weakness, but a follow-up PET-CT revealed PET-avid residual disease (October 31, 2017). First symptomatic clinical disease progression (January 2018) included new neck and back pain and bilateral lower extremity weakness. MRI confirmed (January 12, 2018) disease pro- gression in areas of prior radiation therapy. Following second decom- pression surgery, a clinical trial (COG-APEC1621C) of tazemetostat, started and he achieved a second partial response that lasted 9 months until October 25, 2018 when an MRI revealed progression at T12 and a new L2 lesion. A third decompression surgery was performed in January 2019 followed by radiation therapy to the new disease site. PET-CT done on April 2, 2019 showed new PET-avid disease in C7 and T3 spine and a new extensive lobulated mass in the paraspinal erector back muscle region (Fig. 2). Salvage pazopanib (800 mg daily) was started but after 3 weeks of therapy an MRI (April 27, 2019) showed a marked increase in the paraspinal soft tissue mass (16.1×18.6 cm T10–L3).

He presented to our center in severe pain. He was in a fetal position on a stretcher unable to walk and with new complaints of diffuse muscle weakness. Physical examination revealed a cachectic male with temporal wasting, Eastern Cooperative Oncology Group Performance Status 4, a large (18 cm) mass fixed to his spine raised off his back (10 cm) and diffuse upper (3+/5+) and lower extremity (2+/5+) muscle weakness. After obtaining compassionate use informed written consent, he received cycle 1 (May 13, 2019) of combination ipilimumab (3 mg/kg) and nivolumab (1 mg/kg). The following day he was admitted for fever and hyperbilirubinemia and transaminitis. Pazopanib was discontinued and his labs normalized over 2 weeks without corticosteroid administration.

Complete clinical regression of the back mass occurred within 2 weeks (May 28, 2019) of cycle 1. Combined therapy continued every
3 weeks for 4 total cycles followed by nivolumab (240 mg fixed dose) every 2 weeks and ipilimumab (1 mg/kg) every 6 weeks. Pazopanib (200 mg raised to 400 mg daily) was restarted 10 weeks (July 23, 2019) after the first dose of combined immunotherapy (May 13, 2019). A follow-up PET-CT scan first showed a partial remission on July 5, 2019. A PET-negative complete remission was achieved on October 11, 2019 (Fig. 3). During his course of therapy, he experienced grade 1 adverse immune events (skin and liver function test) but did not require delay of therapy or use of steroids. A second negative PET-CT scan (January 2, 2020) was observed so checkpoint inhibition therapy was discontinued, and the patient returned to normal activities. Pazopanib was discontinued on May 4, 2020 due to persistence of his complete remission status on examination and CT scans. He was asymptomatic and had a normal physical examination with an Eastern Cooperative Oncology Group Performance Status of 0 on his most recent visit (June 29, 2020).

DISCUSSION

To our knowledge, this is the first reported case of a complete and durable response to combination anti– cytotoxic T-lymphocyte–associated protein 4 (CTLA-4) (ipilimumab) and anti–programmed death-ligand 1 (PD-L1) (nivolumab) therapy in a patient with epithelioid sarcoma refractory to standard and NGS-guided targeted therapy. The World Health Organization defines over 100 variants of soft tissue sarcoma, of which epithelioid sarcoma, is rare accounting for <1%. Chemotherapy, including histology specific, in the salvage setting is marginally effective, with limited objective responses usually of short duration with an expected median overall survival of 8–16 months.14 After primary debulking surgery, our patient achieved only a partial response, to adriamycin and ifosfamide chemo- therapy and involved field radiation therapy, of 10-month duration before his first progression. Patients with tumors deficient in INI1, may respond to EZH2 methyltransferase inhibition, however, as in our case, objective responses to date have mostly been partial (13% partial response rate) and transient.8 Pazopanib, a drug among the class of angiogenesis inhibitors has broad biologic effect inhibit- ing vascular endothelial growth factors 1, 2, and 3, platelet-derived growth factor receptors alpha and beta, fibroblast growth factor receptor 1 and 3, and several kin- ases. In a phase III trial (PALETTE) pazopanib resulted in a significant prolongation of progression-free survival (4.6 vs. 1.6 mo; P < 0.001) in patients with soft tissue sarcoma progressing after standard chemotherapy. However, there was no difference in overall survival compared with placebo and there were limited objective responses (partial response rate 6% and no complete response).15 After a 10-month partial response to tazemetostat, our patient received pazopanib (800 mg daily) but did not respond and had rapid life-threatening bulky disease progression.

Checkpoint inhibitors are effective in a variety of solid and liquid tumors in the relapsed and chemotherapy refractory set- ting, potentially including sarcoma. The Alliance A 091401 study conducted in patients with advanced sarcoma having failed at least 1 prior therapy, showed potential efficacy with a 5% response rate to single agent anti–programmed cell death protein 1 (PD-1) therapy (nivolumab) and a 16% response rate to combined anti-PD-1 and anti-CTLA-4 therapy (ipilimumab).16 Tumor mutation burden (TMB), a manifestation of genomic instability, when “high,” is associated with response and improved clinical outcomes in patients with melanoma, lung cancer and urothelial cancers treated with checkpoint inhibitors.17 Similarly, in colorectal and other cancers, patients with tumors associated with “microsatellite instability–high” disease, experience high rates of objective and durable responses (objective response rates > 50%) to checkpoint inhibitors.18 Data of TMB and microsatellite instability in sarcoma while limited suggest that TMB and microsatellite instability are expected to be low and stable, respectively.19 Similarly, as in our case, INI1-deficient cancers are associated with low TMB and microsatellite stable disease and accordingly previously not considered candidates for checkpoint inhibitor therapy.20
PD-L1 expression was not measured in our patient. How- ever, preclinical and more recent clinical data revealed that despite low TMB and microsatellite stability, INI1-deficient tumors significantly express PD-L1 as demonstrated in 1 series of 16 patients with INI1-deficient malignant rhabdoid tumors where 50% of patients (8/16) had membranous expression of PD-L1 in 10%–70% of tumors of tumor cells assessed.21 These INI1-defi- cient rhabdoid tumors also have extensive cytotoxic T-cell infil- tration, predictive of responsiveness to checkpoint inhibitor therapy.11–13 Mechanistically, mutations in the SWI/SNF com- plex observed in our case are found to significantly increase INI1-deficient tumor immunogenicity through “viral mimicry” by spontaneously activating intratumoral interferon signaling because of loss of interferon gene regulation.11

Several case reports involving patients with a variety of INI1-deficient tumor types treated with checkpoint inhibitors have reported objective responses including, in a 24-year-old with metastatic epithelioid sarcoma, a partial response to single agent nivolumab administered concurrently with pazo- panib after progression on pazopanib.22 These observations, “the know increased objective response rate and treatment-free survival duration of combined versus single agent checkpoint inhibition in a variety of cancer types” and the Alliance study conclusion that nivolumab single agent salvage therapy should not be pursued in sarcoma led to our patient being offered compassionate combined checkpoint inhibitor therapy with nivolumab and ipilimumab therapy.23

The speed, extent and duration of response to combined checkpoint inhibitor therapy observed in our patient was unexpected, given the extent of his prior treatments and his presenting rapidity of disease progression and poor perform- ance status. Several unanswered questions remain including what if any role prior treatment with the EZH2 methyl- transferase inhibitor tazemetostat played in the dramatic response we observed because EZH2 inhibitors have been shown to increase tumor cell PD-L1 expression and down- regulate intratumoral regulatory T cells?23,24 In addition, did prior treatment with pazopanib contribute to the partial response observed in our case? Also, pazopanbib was resumed 10 weeks after the first dose of immunotherapy. It was resumed because of prior experience with continuing pazopanib when checkpoint inhibitor therapy is started after pazopanib failure.22 Whether resuming pazopanib contributed to the ultimate complete and durable response achieved remains unknown. Moreover, the effect of the PTCH1 alteration observed on his disease course and response to checkpoint inhibitors is also unknown. PTCH1 is a tumor suppressor gene that encodes a protein that inhibits the release of smoothened and downstream hedgehog signaling, and when altered, as in our case, cancer cell growth occurs because of excess Hedgehog signaling. Furthermore, single-agent checkpoint anti-PD-1 therapy has been shown to result in durable objective responses in 3 of 4 patients with basal cell cancer with PTCH1 alterations who failed prior Hedgehog inhibitor therapy. Unlike our patient who had a tumor with low TMB (2 Mut/Mb), the patients with basal cell cancer all had high TMB (median = 50 Mut/Mb).25 Regardless, in our case, combination checkpoint inhibition therapy led to a complete PET-negative ongoing response that has been durable for over 12 months despite our patient pre- senting with clinically near terminal, pathologic low TMB and microsatellite stable, disease. The lack of significant adverse immune-related events was striking given the higher starting dose of ipilimumab (3 vs. 1 mg/kg) and his rate and extent of disease response. Further study will need to determine the role and sequencing of targeted therapy and pazopanib, and checkpoint inhibitor dosing in the salvage setting for advanced epithelioid sarcoma and possibly other INI1-deficient sarcomas.

CONCLUSION

In conclusion, combined checkpoint inhibition therapy warrants further study in the salvage setting in patients with INI1-deficient epithelioid and other INI1-deficient sarcomas, seemingly regardless of prior therapy, TMB, and micro- satellite status and extent of disease and performance status.