A, Adaptive PD-L1 expression demonstrated with PD-L1 IHC, whereby PD-L1 is expressed by both tumor cells and immune cells (ICs) at the host-tumor interface (blue arrowheads). B, Tumor cells express PD-L1 constitutively at a low level (pink asterisk), which is accentuated at the interface with ICs (blue asterisks), consistent with a mixed expression pattern. C, Constitutive tumor cell PD-L1 expression independent of an IC infiltrate, suggesting an oncogene-driven mechanism of display. IHC indicates immunohistochemistry; SCC, squamous cell carcinoma.
The orange dots represent genes that are upregulated in specimens from HIV-positive compared with HIV-negative patients. The HIV-positive group overexpressed some cytokines or cytokine receptors (IL1A, IL17RC, IL10RB) and some markers related to T-cell activation (ICOSLG, P2RX1), and marginally overexpressed LAG3. Interleukin 18, a proinflammatory cytokine previously reported to be present at elevated levels in sera from HIV-positive individuals, was the most significantly overexpressed gene in the HIV-associated tumors (fold-change, 12.69, P < .001). IFNG, an inflammatory cytokine associated with adaptive PD-L1 expression, was expressed equivalently between the 2 groups. The horizontal dotted line indicates P = .05 and vertical dotted lines indicate 2.0-fold difference in expression levels. Data are normalized to PTPRC (CD45, pan immune cell marker). There was no significant difference in PTPRC expression among the samples. SCC indicates squamous cell carcinoma.
eMethods and Results
eTable 1. Baseline patient characteristics by HIV status
eTable 2. Correlation of peripheral CD4 T cell counts with tumor infiltrating immune cell (IC) densities among 19 HIV(+) patients with anal SCC and documented peripheral blood CD4 counts
eTable 3. Genes selected for multiplex qRT-PCR array
eTable 4. Expression of candidate immune-related genes in anal SCC specimens derived from 3 HIV(+) versus 4 HIV(-) patients
eTable 5. Comparison of features of the tumor immune microenvironment captured in tissue microarray cores taken from the central versus the peripheral areas of the tumor
eFigure 1. Adaptive pattern of PD-L1 expression in an anal SCC from a HIV(+) patient
eFigure 2. Correlation of immune cell (IC) PD-L1 expression with density of T cell infiltrates
eFigure 3. The prevalence of PD-L1 expression and IC infiltration are similar in anal SCCs from HIV(+) and HIV(-) patients
eFigure 4. The HALO image analysis platform was used to digitally quantify the density of immune cell subsets identified using immunohistochemistry
Customize your JAMA Network experience by selecting one or more topics from the list below.
Yanik EL, Kaunitz GJ, Cottrell TR, et al. Association of HIV Status With Local Immune Response to Anal Squamous Cell Carcinoma: Implications for Immunotherapy. JAMA Oncol. 2017;3(7):974–978. doi:10.1001/jamaoncol.2017.0115
Does the tumor immune microenvironment differ between HIV-positive and HIV-negative patients with anal squamous cell carcinoma (SCC)?
In this study of 40 tumor specimens from patients with anal SCC, there were no significant differences in programmed death-ligand 1 (PD-L1) expression or programmed cell death protein 1 (PD-1), CD3, CD4, CD8, CD68, or LAG-3 immune cell subsets by immunohistochemical analysis. Immune-related gene expression profiles were also similar between HIV-positive and HIV- negative patients.
HIV status alone does not affect the degree of immune cell infiltration or checkpoint molecule expression in tumors from patients with anal SCC, supporting the ongoing clinical investigation of anti–PD-1 and PD-L1 therapies in this population.
The programmed cell death protein 1 (PD-1) and programmed death-ligand 1 (PD-L1) pathway play an important immunosuppressive role in cancer and chronic viral infection, and have been effectively targeted in cancer therapy. Anal squamous cell carcinoma (SCC) is associated with both human papillomavirus and HIV infection. To date, patients with HIV have been excluded from most trials of immune checkpoint blocking agents, such as anti–PD-1 and anti–PD-L1, because it was assumed that their antitumor immunity was compromised compared with immunocompetent patients.
To compare the local tumor immune microenvironment (TME) in anal SCCs from HIV-positive and HIV-negative patients.
Design, Setting, and Participants
Anal SCC tumor specimens derived from the AIDS and Cancer Specimen Resource (National Cancer Institute) and Johns Hopkins Hospital included specimens. Tumors were subjected to immunohistochemical analysis for immune checkpoints (PD-L1, PD-1, LAG-3) and immune cell (IC) subsets (CD3, CD4, CD8, CD68). Expression profiling for immune-related genes was performed on select HIV-positive and HIV-negative cases in PD-L1+ tumor areas associated with ICs.
Main Outcomes and Measures
Programmed death-ligand 1 expression on tumor cells and ICs, PD-L1 patterns (adaptive vs constitutive), degree of IC infiltration, quantified densities of IC subsets, and gene expression profiles in anal SCCs from HIV-positive vs HIV-negative patients.
Approximately half of 40 tumor specimens from 23 HIV-positive and 17 HIV-negative patients (29 men and 11 women; mean [SD] age, 51 [9.9] years) demonstrated tumor cell PD-L1 expression, regardless of HIV status. Median IC densities were not significantly decreased in HIV-associated tumors for any cellular subset studied. Both adaptive (IC-associated) and constitutive PD-L1 expression patterns were observed. Immune cell PD-L1 expression correlated with increasing intensity of IC infiltration (r = 0.52; 95% CI, 0.26-0.78; P < .001) and with CD8+ T-cell density (r = 0.35; 95% CI, 0.11-0.59; P = .03). Gene expression profiling revealed comparable levels of IFNG in the TME of both HIV-positive and HIV-negative patients. A significant increase in IL18 expression levels was observed in HIV-associated anal SCCs (fold change, 12.69; P < .001).
Conclusions and Relevance
HIV status does not correlate with the degree or composition of IC infiltration or PD-L1 expression in anal SCC. These findings demonstrate an immune-reactive TME in anal SCCs from HIV-positive patients and support clinical investigations of PD-1/PD-L1 checkpoint blockade in anal SCC, irrespective of patient HIV status.
Tumors may co-opt the programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) immune checkpoint pathway to avoid host immune surveillance. Therapeutic blockade of this pathway unleashes the host immune response, often leading to durable tumor regressions in patients with advanced cancers.1 Monoclonal antibodies targeting PD-1 or PD-L1 have been approved by the US Food and Drug Administration to treat melanoma, non–small-cell lung carcinoma (NSCLC), Hodgkin lymphoma, and carcinomas of the kidney, urothelial tract, and squamous cell carcinoma of the head and neck (SCCHN).
Immunosuppressed patients with cancer, including those with HIV, have historically been excluded from immunotherapy trials and it is unknown if immune checkpoint pathways are active in the tumors of these patients. However, many HIV-associated cancers are linked to oncogenic viruses, such as human papillomavirus (HPV) in anal squamous cell carcinoma (SCC).2 In immunocompetent patients, virus-associated cancers contain immune-active microenvironments and some have responded well to anti–PD-1 therapies.3,4 In patients with HIV, the presence of strong viral antigens in the tumor immune microenvironment (TME) could potentially offset the effects of HIV-associated systemic immune dysfunction. The current study compared the TME in HIV-positive vs HIV-negative patients with anal SCC, a virus-associated cancer that has not previously been characterized for PD-1/PD-L1 expression and where new treatment options would be of great value.
Forty-eight anal SCC specimens were identified from 29 HIV-positive and 19 HIV-negative patients; 8 specimens were excluded for not meeting pathologic criteria (eMethods in the Supplement). Tumor tissue microarrays (TMAs) using central tumor cores were stained with immunohistochemistry (IHC) for immune checkpoint molecules (PD-1, PD-L1, LAG-3) and immune cell (IC) subsets (CD3, CD4, CD8, CD68). Membranous (cell surface) PD-L1 expression on tumor or infiltrating ICs, intensity of IC infiltration, and patterns of PD-L1 expression associated with IC infiltrates were scored as previously described.5 Immune cell subsets were quantified using HALO image analysis (Indica Laboratories). Tumor cells from areas of adaptive PD-L1 expression adjacent to ICs were captured by laser capture microdissection and used to compare gene expression profiles between tumors from HIV-positive and HIV-negative patients. Analyses were blinded to patient HIV status. Statistical significance was determined with a 2-sided α level of .05 unless otherwise stated. Full methods are provided in the Supplement.
All specimens were collected in compliance with the institutional review boards of the AIDS and Cancer Specimen Resource and Johns Hopkins Hospital; all patients had signed consent or waiver of consent forms for tissue use.
Forty anal SCCs from 23 HIV-positive and 17 HIV-negative unique patients were studied (eTable 1 in the Supplement). Peripheral blood CD4 counts were available for 19 HIV-positive patients, with 10 (53%) below 200 cells/mm3 (range, 60-597 cells/mm3). Nearly half (49%) of 35 patients with treatment data had received chemotherapy and/or radiation prior to specimen procurement. Details regarding treatment effect on immune parameters are provided in the eMethods and eResults in the Supplement.
Immune cell densities were similar between the 2 groups for all cellular subsets analyzed (Table 1) (eFigure 1 in the Supplement), although there was a trend toward decreased CD4 to CD8 ratio in HIV-associated anal SCC. Peripheral CD4 counts were not associated with the density of any intratumoral IC subset or checkpoint expression (eTable 2 in the Supplement). Varied patterns of tumor cell PD-L1 expression, including adaptive, constitutive, and mixed,1,5 were observed in tumors from both HIV-positive and HIV-negative patients (Figure 1). When all 40 cases were assessed, IC PD-L1 expression was correlated with the degree of IC infiltration (r = 0.52; 95% CI, 0.26-0.78; P < .001) (Table 2) and with intratumoral CD8+ T-cell density (r = 0.35; 95% CI, 0.11-0.59; P = .03) (eFigure 2 in the Supplement). In contrast, tumor cell PD-L1 expression was not significantly correlated with the degree of IC infiltration or intratumoral CD8+ T-cell density.
No significant differences were observed in tumor cell PD-L1 expression, adaptive PD-L1 expression, or presence of moderate to severe IC infiltrates in anal SCC from HIV-positive vs HIV-negative patients, although there was a trend toward increased IC PD-L1 expression in HIV-associated tumors (eFigure 3 in the Supplement). Although tobacco use was common among patients with anal SCC, smoking history was not associated with expression of the studied parameters. Whereas immune-related gene expression (including IFNG) in tumors from HIV-positive and HIV-negative patients was generally similar (Figure 2) (eTable 3 and eTable 4 in the Supplement), interleukin-18 (IL-18) expression was significantly elevated in the HIV-positive group.
Little is known about whether HIV-related systemic immune dysfunction alters local antitumor immunity. Our findings demonstrate an immune-reactive microenvironment in anal SCCs from both HIV-positive and HIV-negative patients, irrespective of peripheral CD4 counts in HIV-positive patients. The association of IC PD-L1 expression with the infiltrating IC intensity and CD8+ T-cell density, and IFNG gene expression in these tumors, suggests that active cytokine secretion by tumor infiltrating lymphocytes drives adaptive PD-L1 expression in some anal SCCs.
Programmed death-ligand 1 may also be expressed on tumor cells constitutively, in the absence of immune infiltration, as seen in some of these cases. However, in other cases, we observed an adaptive superimposed on constitutive PD-L1 expression pattern, as previously described for other tumors with squamous differentiation, including SCCHN, NSCLC, and urothelial carcinomas.1 Anti–PD-1 and anti–PD-L1 therapies mediate tumor regression in approximately 20% to 30% of patients with these advanced cancers.4,6-10 Programmed cell death-ligand protein 1 expression may enrich for response to anti–PD-1 and anti–PD-L1, but the role of PD-L1 as a biomarker in anal SCC has not yet been established.
The immunologic similarity of anal SCCs from HIV-positive and HIV-negative patients may have several explanations. HIV infection increases the risk of HPV infection, persistence, and early progression11 but may have little effect on later steps in tumorigenesis or on signaling pathways driving constitutive PD-L1 expression. In addition, virus-driven cancers may elicit strong immune responses because of the presence of viral proteins,12 suggesting that anal SCC may be a particularly good candidate for anti–PD-1 and anti–PD-L1 immunotherapy. T cells expressing PD-1, LAG-3, and Tim3 have been described in anal SCC.13 The expression of LAG-3 was also observed in this study by IHC, supporting a potential second therapeutic immune checkpoint target in this cancer type.
Encoding an inflammatory cytokine, IL18 was the most highly upregulated gene in HIV-associated anal SCCs. Interleukin 18 (IL-18) is increased in the blood of patients with HIV,14 and the dysregulated cytokine milieu that characterizes late-stage HIV infection is thought to alter the balance of IL-18 activity to favor HIV replication and disease progression.15 In this study, elevated levels of IL18 thus may reflect advanced HIV infection, consistent with very low peripheral CD4 counts in 2 of the 3 patients studied (60 and 67 cells/mm3).
This study has several limitations. Almost half of the patients were pretreated and 31% of the specimens were from recurrent tumors, and thus our findings may not be generalizable to all patients with anal SCC. In addition, assessment of central cores from TMAs may underestimate immune activity at the tumor-stromal interface or in draining lymph nodes. The limited number of specimens might also have affected our ability to detect small differences in IC subsets, checkpoint molecules, and immune-related gene expression between HIV-positive and HIV-negative patients, or between patients with CD4 counts above and below 200 cells/mm3. However, for the 2 HIV-positive patients with very low CD4 counts, we did not identify a commensurate decrease in intratumoral CD4 density, and IFNG expression in these tumors appeared comparable to expression levels in tumors from HIV-negative patients.
This study suggests a similar degree of IC infiltration and expression of the clinically targetable immune checkpoints PD-L1 and PD-1 in anal SCCs from HIV-positive and HIV-negative patients. To our knowledge, this is the first study of its kind in anal SCC, a relatively rare tumor for which new treatment options are needed. Our findings, as well as early clinical data,13 support ongoing clinical investigations of anti–PD-1, alone or in combination with anti–CTLA-4, in HIV-positive patients (NCT02314169 and NCT02408861). Furthermore, our findings provide a rationale for targeting LAG-3 in patients with advanced anal SCC.
Corresponding Author: Janis M. Taube, MD, Division of Dermatopathology, Johns Hopkins University School of Medicine, Blalock 907, 600 N Wolfe St, Baltimore, MD 21287 (firstname.lastname@example.org).
Accepted for Publication: December 21, 2016.
Published Online: March 23, 2017. doi:10.1001/jamaoncol.2017.0115
Author Contributions: Drs Taube and Engels had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Yanik and Kaunitz contributed equally.
Concept and design: Yanik, Xu, Ogurtsova, Cornish, Topalian, Engels, Taube.
Acquisition, analysis, or interpretation of data: Yanik, Kaunitz, Cottrell, McMiller, Ascierto, Esandrio, Cornish, Topalian, Engels, Taube.
Drafting of the manuscript: Yanik, Kaunitz, McMiller, Ascierto, Xu, Ogurtsova, Taube.
Critical revision of the manuscript for important intellectual content: Yanik, Kaunitz, Cottrell, McMiller, Esandrio, Cornish, Topalian, Engels, Taube.
Statistical analysis: Yanik, Ascierto.
Obtained funding: Topalian, Taube.
Administrative, technical, or material support: Kaunitz, Cottrell, McMiller, Esandrio, Xu, Ogurtsova, Cornish, Taube.
Conflict of Interest Disclosures: Dr Topalian has received research grants from Bristol-Myers Squibb and is a consultant for Five Prime Therapeutics. Dr Taube has received research support from Bristol-Myers Squibb and is a consultant for Bristol-Myers Squibb, Merck, and AstraZeneca. Dr Lipson has received research support from Bristol-Myers Squibb and Merck, and is a consultant for Bristol-Myers Squibb, EMD Serono, Merck, and Novartis. No other disclosures are reported.
Funding/Support: This work was supported by the Dermatology Foundation (Taube), WW. Smith Foundation (Taube), NIH R01 CA142779 (Taube, Topalian), NIH T32 CA193145 (Cottrell), the Bloomberg~Kimmel Institute for Cancer Immunotherapy at Johns Hopkins, and the Intramural Research Program of the National Cancer Institute (Yanik, Engels). Drs Taube and Topalian were also supported by a Stand Up To Cancer—Cancer Research Institute Cancer Immunology Translational Research Grant (SU2C-AACR-DT 1012). Stand Up To Cancer is a program of the Entertainment Industry Foundation administered by the American Association for Cancer Research. Specimens were provided by the AIDS and Cancer Specimen Resource, funded by the National Cancer Institute (UM1CA181255).
Role of the Funder/Sponsor: The funding institutions and corporations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We would like to thank Elizabeth Montgomery, MD and Robert Anders, MD, PhD, Johns Hopkins University School of Medicine, for helpful discussions. We thank Alan Berger, PhD, Johns Hopkins University School of Medicine, for statistical analysis of gene expression data. These contributors were not compensated for this work.