Pembrolizumab in an HIV-infected patient with glioblastoma

Abstract

Persons living with human immunodeficiency virus (PLWH) carry increased risk for developing malignancies, including glioblastoma. Despite extensive investigations, both human immunodeficiency virus (HIV) and glioblastoma are incurable. Treatment for a patient with combined glioblastoma and HIV remains an unexplored need. Preliminary evidence suggests that immunotherapy may be effective for the simultaneous treatment of both HIV and cancer by reversing HIV latency and T cell exhaustion. We present a case of glioblastoma in a PLWH who was treated with pembrolizumab. Treatment was well tolerated and safe with a mixed response. Our patient did not develop any opportunistic infections, immune-related adverse events, or worsening of his immunodeficiency. To our knowledge, this is the first reported case of a PLWH and glioblastoma treated with immunotherapy.

Plain Language Summary

Persons living with human immunodeficiency virus (PLWH) are at increased risk for cancers, including glioblastoma. Despite extensive research, both human immunodeficiency virus (HIV) and glioblastoma are incurable. The optimal treatment for concurrent HIV and glioblastoma is unknown. Early evidence suggests that immunotherapy can deplete residual HIV and restore immune function. We present a case of glioblastoma in a PLWH treated with immunotherapy. Treatment was well tolerated and safe. To our knowledge, this is the first reported case of a PLWH and glioblastoma treated with immunotherapy.

Plain language summary

Article highlights.

Human immunodeficiency virus cannot be fully eradicated

• Full eradication of the virus is hampered by reservoirs of latent HIV residing within memory CD4⁺ T cells.

Persons living with human immunodeficiency virus carry an increased risk for cancer and treatment for these patients is an unmet need

• Immunotherapy appears safe and effective in persons living with HIV with cancer.

• Immunotherapy for glioblastoma is under investigation.

• Preliminary evidence suggests that immunotherapy may also reverse HIV latency.

1. Background

Despite the advent of highly active antiretroviral therapy (HAART), human immunodeficiency virus (HIV) remains an incurable disease [1]. Though HIV-1 RNA viral load levels (VL) can be undetectable in persons living with HIV (PLWH), PLWH can only achieve functional cure [2,3]. Full eradication of the virus cannot be achieved due to reservoirs of latent HIV residing within memory CD4⁺ T cells [2,3]. This intracellular location provides a selective advantage for the virus to persist because it is inaccessible by HAART [3,4,5,6,7]. The mechanism by which this occurs is driven by a positive tropism between latent HIV and programmed cell death protein (PD)-1 [6,8]. In PLWH on HAART, PD-1 is overexpressed on the surface of memory CD4⁺ T cells and draws latent HIV into the cells [6,8]. Accordingly, PD-1 expression correlates with the extent of these reservoirs [9,10]. Permanent suppression of this pool of transcriptionally silent HIV is one treatment approach under exploration [3]. However, there remains a potential for latent HIV to reactivate, as evidenced by the remarkable case of the ‘Mississippi Child’ wherein an HIV-infected child in remission subsequently relapsed from reactivation of dormant HIV [3,5,11]. Depletion of HIV from these reservoirs may represent a better approach to eliminate the risk for reactivation [3,4,12]. Histone deacetylase and PD-1 inhibitors are tumor-directed therapies repurposed as HIV latent reversal agents to force HIV out of intracellular hiding where it can be eliminated by HAART [4,12,13,14].

Beyond the challenges of disease control with HIV, PLWH also carry an increased risk of developing malignancies compared with the general population [15,16,17]. This increased risk is driven by impaired immune surveillance [16,18]. Understanding the underlying mechanism of this immunodeficiency is essential to understanding how cancer treatment in PLWH can be addressed. First, HIV preferentially destroys CD4⁺ helper T cells, whose function is to activate the CD8⁺ T effector cells that clear virally infected cells and tumor cells [19]. Second, the function of CD8⁺ T cells is further impaired by the chronic immune activation from HIV [16,18]. Chronic immune activation upregulates PD-1 expression in previously primed CD8⁺ T cells, which leads to T cell exhaustion [16,18]. The loss of CD8⁺ T cell function from both these mechanisms then permit neoplastic development in PLWH [16,18]. Last, irrespective of HIV, cancers are marked by immunosuppression and T cell exhaustion [20]. On the basis of this understanding, the optimal therapy would inhibit the interaction between PD-1 and its ligands [6,8,9,17]. This blockade would concurrently address HIV latency and T cell exhaustion [17,21,22,23]. For these reasons, PD-1 inhibitors, such as pembrolizumab, have been the focus of studies for PLWH and cancer.

Early limited results are promising, but none of these studies included glioblastoma, a non-AIDS defining malignancy [13,16,17,21,25]. Malignancies, including glioblastoma, have been traditionally excluded from HIV clinical trials [13,17,27]. Correspondingly, PLWH have historically been excluded from glioblastoma trials [13,14,28,29,30,31,32,33,34]. Cancer care disparities in PLWH are now being recognized with efforts underway to bridge this knowledge gap [32,33,34,35]. We present a case of a well-controlled PLWH with progressive glioblastoma who was treated with pembrolizumab. We hypothesized that PD-1 inhibition could potentially reverse both HIV latency and T cell exhaustion to treat HIV and glioblastoma concurrently.

2. Case presentation

A 39-year-old male PLWH presented with headache. His neurologic exam was unremarkable. His outside lab lymphocyte subset panel 5 revealed that the plasma VL was 82,900 copies/ml and the absolute CD4⁺ count was 299 cells/ml³ (Figure 1). HAART was initiated. Brain MRI with and without contrast showed left parietal ring enhancing intra-axial mass (Figure 2A), and he underwent subtotal resection of the tumor.

Figure 1.

Timeline of patient's viral load level and CD4⁺ counts pre and post-pembrolizumab therapy. The viral load was undetectable prior to his first and second pembrolizumab infusions. Following these infusions, low-level viremia was detected, suggestive of HIV latency reversal and the purging of quiescent HIV out of the CD4⁺ T cells into the peripheral blood.

Figure 2.

Contrast-enhanced T1-weighted brain MRI axial sections, show masses in the left parietal, left temporal and left frontal lobes pre-pembrolizumab (A–C) and post three infusions of pembrolizumab (D–F). (A) 4.2 cm left parietal enhancing mass. (B) 3.2 cm left temporal enhancing mass. (C) 0.8 cm left frontal enhancing mass. (D) Interval regression (3.2 cm) of the left parietal mass. (E) Interval growth (6.0 cm) of the left temporal mass. (F) Interval enlargement (1.7 cm) of the left frontal heterogeneously enhancing mass.

Histologic sections showed a diffusely infiltrating glial neoplasm with extensive necrosis, composed of highly pleomorphic cells with bizarre, irregular nuclei, smudgy chromatin and variable amounts of eosinophilic cytoplasm intermixed with poorly differentiated cells with polygonal hyperchromatic nuclei and scant cytoplasm in the background of abundant geographic necrosis and numerous regions with microvascular proliferation (Figures 3A–C). On immunohistochemistry (IHC), tumor cells were patchy positive with GFAP, showed increased expression of p53 (suggestive of TP53 mutation) and had a variable Ki-67 proliferative index ranging from 20 to 80% in some areas (Figures 3D–F). IHC for CD3 and CD20 highlighted scattered T- and B-cells, respectively, but were negative in tumor cells (Figures 3G-H). In situ hybridization for EBER (Epstein-Barr virus [EBV]-encoded small RNA) was also negative. Next-generation sequencing (NGS) performed at a commercial laboratory showed a hot spot mutation in the promoter region of the TERT gene and a damaging missense mutation in the TP53 tumor suppressor gene. There were no mutations in ATRX, IDH1, IDH2, H3-3A and H3C3 genes (by NGS), and EGFRvIII was absent (by real-time PCR). Fluorescence in situ hybridization showed deletion of PTEN gene with no evidence of MET, MYCN, PDGFRA amplifications, BRAF rearrangement or 1p/19q codeletion. RNA sequencing was negative for RAF1, FGFR1/2/3 and NTRK1/2/3 fusions. PD-L1 LDT expression was detected by IHC in 60% Tumor Cells/Total Viable Tumor Cells. While significant atypia in a relatively young patient raised the possibility of a hypermutated tumor, tumor mutational burden (TMB) was low (5 mutations/Mb), and there was no microsatellite instability. MGMT promoter was unmethylated. A final integrated diagnosis of Glioblastoma, IDH-wildtype was determined based on both morphologic features and molecular findings [36,37].

Figure 3.

Histologic sections show a high-grade glioma (A) in the background of extensive necrosis (arrow). (B) Some tumor cells show abundant eosinophilic cytoplasm and bizarre nuclei. Microvascular proliferation (arrows) is present. (C) In other areas, immature-appearing tumor cells with scant cytoplasm and polygonal, hyperchromatic nuclei are arranged around vessels. (D) GFAP immunohistochemical stain is patchy and positive supporting the diagnosis of a glioma. (E) P53 stain is positive in more than 90% of the tumor cells, suggestive of TP53 gene mutation. (F) Ki-67 proliferative index is increased, reaching 80% in some areas. (G) CD3 stain highlights scattered small T-lymphocytes and (H) CD20 stain highlights scattered small B-lymphocytes. All scale bars are 100 microns.

Temozolomide was withheld for neutropenia and his tumor was treated with radiation (60 Gy in 30 fractions) alone. On Day 208, surveillance brain MRI with and without contrast showed 0.8 cm dural-based left frontal and 3.2 cm left temporal enhancing masses (Figures 2B–C). His neutropenia resolved, but he was not amenable to chemotherapy. Additional treatment options were offered, including pembrolizumab and lenvatinib [38]. He expressed a preference for intravenous route of administration, and he decided to proceed with pembrolizumab (200 mg every 3 weeks) and bevacizumab (10 mg/kg every 2 weeks). On Day 230, he received his first pembrolizumab infusion. His pre-pembrolizumab plasma VL was undetectable and his CD4⁺ count was 408 cells/ml³. On Day 240, his VL increased to 52 copies/ml, which subsequently dropped to undetectable levels on Day 246. At the time of his second pembrolizumab infusion on Day 252, his VL remained undetectable and his CD4⁺ count rose to 808 cells/ml³. At the time of his last bevacizumab infusion as well as his third and last pembrolizumab infusion on Day 276, he developed recurrent low-level viremia (VL: 94 copies/ ml) and his CD4⁺count dropped to 173 cells/ml³. He reported compliance with HAART. No opportunistic infections developed from the CD4⁺count decrease below 200 cells/ml³. On Day 290, his brain MRI demonstrated continued regression of the left parietal tumor, but interval enlargement of both left frontal and left temporal masses (Figures 2D–F). On Day 315, his VL was again undetectable and his CD4⁺ count increased to 661 cells/ml³. He was initiated on dexamethasone 4 mg twice daily with improvement in his headaches and further pembrolizumab was held. He experienced no immune-related adverse events from the pembrolizumab and no adverse events from bevacizumab. The decision was made to offer neurosurgical intervention for the 6 cm left temporal tumor for symptomatic relief and tissue diagnosis. The patient decided to proceed with resection and further bevacizumab was held in anticipation for surgery. Unfortunately, on Day 323, 11 months from his initial diagnosis, he passed away from causes unrelated to his HIV and glioblastoma. Ethical guidelines set out by the declaration of Helsinki were followed in the preparation of this report, and the patient provided written consent.

3. Discussion

Immunotherapy, such as PD-1 inhibitors, has revolutionized the landscape of treatment for many cancers. It is widely accepted that the inhibitory action of PD-1 augments the immune response and restores the function of CD8⁺ T cells to induce apoptosis in both virally infected cells and tumor cells [6,8,9,17,20]. However, PD-1 inhibitors have shown limited success for the treatment of glioblastoma due to its immunosuppressive tumor microenvironment (TME), with tumor infiltrating lymphocytes representing only <1%–2% of the TME [39]. Though high TMB may be useful as a predictive marker for response to immunotherapy in specific solid cancers, evidence suggests that high TMB is not associated with immunotherapy response for glioblastoma [40]. Conversely, TMB may be inversely associated with immunotherapy efficacy [41]. Our patient's TMB was low, which has been associated with prolonged survival in recurrent glioblastoma patients treated with immunotherapy [41]. Limited evidence exists for the use of PD-1 inhibitors in HIV and comorbid cancer. Even less evidence exists for its use in HIV and glioblastoma. When comorbid HIV and glioblastoma occur in the same patient, there are added complexities that must be taken into consideration. Immunodeficiency from the virus coupled with immunodeficiency from the tumor may impair the ability to mount a response to therapy [13,14,16,28,29,30,31,32,33,34,42,43]. Furthermore, drug–drug interactions between HAART and chemotherapy can exacerbate depressed CD4⁺ counts in a PLWH [44,45]. To mitigate these barriers, PD-1 inhibitors are being explored for PLWH and cancer. PD-1 inhibitors offer immune reconstitution without the conventional myelosuppression that is associated with chemotherapy and can potentially dually treat HIV and glioblastoma with no added risk of toxicity, opportunistic infections, or worsening immunosuppression [16,42,43,44,45,46].

Evidence suggests that PD-1 inhibitors do not cause detrimental effect on the VL in PLWH and malignancy [17,47]. Low level viremia bearing no clinical consequence has been reported, but a recent study suggests that low level viremia may be due to latency reversal [12,13,17]. In our patient with undetectable VL prior to pembrolizumab therapy, we observed two occurrences of low level viremia during pembrolizumab therapy. On both occasions, his VL subsequently returned to undetectable levels. This raises the possibility that low level viremia is a surrogate marker for HIV latency reversal and can be explained by quiescent HIV being purged out of the CD4⁺ T cells into peripheral blood giving rise to viremia then the subsequent elimination by HAART returning the VL back to down to undetectable levels [5,6,13].

PD-1 inhibitors have also not been shown to detrimentally affect CD4⁺ counts in PLWH and malignancy and do not increase the risk for opportunistic infections [17,21,47,48]. Our case corroborates these findings. CD4⁺ counts can steadily increase or decrease, but can also fluctuate [47,49,50,51], as was observed in our patient. Baseline to best response CD4⁺ count showed a single fold increase from 408 cells/ml³ to 808 cells/ml³. Though his CD4⁺ count fell below 200 cells/ml³ on one occasion, no opportunistic infections were unmasked.

PD-1 inhibitors are considered to be overall safe and well tolerated in PLWH and malignancy [13,17,21,25]. There are no apparent added risks for the development of immune-related adverse events (irAEs) in this patient population [17]. Reported risk factors include a higher number of cycles (median = 4 cycles) and higher cumulative dose (median = 600 mg) and [52]. In our patient, there were no irAEs following 3 administered cycles of pembrolizumab with a cumulative dose of 600 mg.

In addition to safety, preliminary evidence shows that pembrolizumab may be effective in PLWH with durable response [17,21,22,47,49,53,54,55,56,57]. Moreover, the combination of pembrolizumab and bevacizumab may be synergistic [58]. Anti-angiogenic therapy upregulates the expression of adhesion molecules, thereby enhancing the interaction of CD8⁺ effector T cells allowing CD8⁺ T cells to cross the blood-brain barrier [58]. However, this combinatorial therapy produced a prolonged response in a case of metastatic glioblastoma, but it remains to be determined if this will translate to improved overall survival [58,59].

In our patient, tumor regression (1/fourfold decrease) was noted in the intraparenchymal left parietal tumor, the original site of disease. Though it is possible this was a result of bevacizumab-induced pseudoresponse [60], there was no rebound growth 1.5 months after bevacizumab was discontinued. However, conflicting responses to pembrolizumab have been reported [17]. Our case also substantiates these findings. In our patient, we observed a mixed tumor response with tumor enlargement in the new dural-based left frontal and left temporal tumors with 1.1-fold and 3.6-fold increases compared with baseline, respectively. The paradoxical regression of the intraparenchymal tumor and the enlargement of the dural-based tumors was unexpected. It is widely accepted that glioblastomas are immunologically cold tumors and trafficking of CD8⁺ T cells beyond the blood–brain barrier is a known obstacle of PD-1 inhibitors [48,61,62]. The dural-based location is expected to provide an advantage by ameliorating the limitation of the blood-brain barrier [63,64]. We reason that these dural-based lesions were not previously treated as was the previously radiated left parietal lesion, which may have effected synergistic therapeutic benefit [65]. Further, it remains unknown if the enlargement of the left frontal and left temporal tumors were a result of true progression or pseudoprogression. Immunotherapy-induced inflammation can mimic true progression and can occur in MGMT unmethylated glioblastoma, though the probability is lower compared with methylated glioblastoma [21,47,66,67]. In the absence of tissue diagnosis due to our patient's untimely death from unrelated causes, a final diagnosis could not be made.

Our case adds to the mounting evidence supporting the use of immunotherapy in PLWH and cancer. In our patient, pembrolizumab was well tolerated, safe and at least partially effective with no apparent increased risk for toxicity, opportunistic infections, or worsening immunosuppression. To our knowledge, this is the first reported case of the use of pembrolizumab in a PLWH and glioblastoma.

There are limitations to this study. A single case is not generalizable to the entire patient population. Given our patient's untimely death, the duration for follow-up was short and the true efficacy of pembrolizumab alone could not be further assessed. The retrospective aspect of this study allowed for trending HIV-1 RNA, but not CD4⁺ T cell-associated HIV RNA [13,68]. There was a possibility that the observed low-level viremia was a result of HAART non-compliance as opposed to reversal of HIV latency. However, this scenario is less likely given the subsequent return of VL to undetectable levels and subsequent rise in CD4⁺ count.

4. Conclusion

The long-term survival of persons living with HIV is altering the clinical spectrum for the development of non-AIDS defining malignancies, including glioblastoma. Immunotherapy may potentially treat HIV and glioblastoma concurrently. Pembrolizumab was well tolerated, safe and effective in our patient. Further studies are needed to support the safety and efficacy of immunotherapy in persons living with HIV and cancer.

Author contributions

C Yuen – study concept, data collection, analysis, interpretation, manuscript drafting, revision and final approval. S Bao - data collection, analysis and final approval. M Pekmezci – data collection, analysis, interpretation, manuscript drafting, revision and final approval. F Mo – data collection and final approval. X-T Kong – manuscript revision and final approval.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

X-T Kong: Received honorarium from Zai Lab for invited speeches for symposiums prior to July of 2021. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The Community Regional Medical Center Institutional Review Board approved this study. Consent for publication: Written consent was obtained from the patient and written consent for the final publication was obtained from the patient's next of kin.

Data availability statement

Availability of data and material: The datasets from this study are available from the corresponding author upon reasonable request.

Funding: No funding was received for this manuscript.