Acquired immune deficiency syndrome (AIDS) mortality has sharply decreased since 1996 because of highly active antiretroviral therapies (HAART). AIDS-related deaths, thus, represented 47% of all deaths observed in the HIV-positive population in 2000, then decreasing to 25% in 2010 . This decrease was, however, accompanied by an increase in mortality because of non-AIDS defined malignancies (11% of all deaths in 2000 and 22% in 2010) . Several studies have cited lung cancer as the most common non-AIDS-defined malignancy . The risk of lung cancer for HIV-positive patients is two-fold to six-fold higher than in HIV-undetermined patients, primarily because of this population's more prevalent tobacco use. Yet, smoking does not explain all the excess risk in the HIV-positive population. Other cofactors are probably involved, such as cannabis use, the direct oncogenic role of the HIV virus, chronic pulmonary infections, and immunosuppression, which seem to represent an independent risk . In most published studies, lung cancer in HIV-positive patients has a poorer prognosis than in HIV-undetermined patients, a prognosis, which was not improved by the emergence of HAART . This difference may be accounted for by a more advanced stage at diagnosis, more aggressive tumors, immunosuppression , as well as major healthcare disparities . Finally, in at least one study, HIV-positive patients treated with standard of care showed similar stage-for-stage outcomes as would be predicted by the Surveillance, Epidemiology, and End Results (SEER) program database .
Programmed cell death-ligand 1 (PD-L1) is overexpressed in numerous cancers, and interaction between PD-L1 on tumor cells and PD-1 (programmed cell death 1) on T cells leads to inhibition of the antitumor immune response. In lung cancer, PD-L1 is frequently overexpressed, affecting 19–100% of patients . This wide range could be explained by the absence of standardized antibodies and analysis techniques. Recent data revealed that blockade of PD-1/PD-L1 interaction can restore proliferation and interferon gamma production of tumor-infiltrating PD1+ CD8+ T cells . In clinical trials, inhibition of this checkpoint has appeared to be one of the most promising approaches for treating several cancer types. Thus, inhibiting the PD-1 pathway leads to a long-term benefit with sustainable response , especially for tumors with more than 50% PD-L1 staining . Because of its great proven benefits to overall survival and progression-free survival, antibodies targeting PD-1 or PD-L1 have been approved in USA and Europe in first-line (Pembrolizumab; Merck, Kenilworth, New Jersey, USA) or second-line [nivolumab, Bristol-Myers Squibbb (Milan, Illinois, USA) and atezolizumab, Roche (Basle, Switzerland)] of patients with advanced or metastatic nonsmall cell lung cancer (NSCLC).
PD-L1 status and the benefits of immunotherapies targeting the PD-1/PD-L1 pathway have not yet been assessed in HIV-positive patients because of the exclusion of this population from clinical trials. Nevertheless, immunotherapies blocking the PD-1 pathway could offer a second benefit in this population. PD-1 expression on T cells is associated with poorer HIV disease control, namely increased viral load, CD8+ T-cell dysfunction, and decreased CD4+ T-cell load [12,13]. In mouse and monkey models, inhibiting the PD-1 pathway led to the restoration of CD4+ and CD8+ T-cell loads and functions. In this population, immunotherapies could, therefore, ensure better control of both tumor disease and HIV infection .
The present study sought to assess PD-L1 status and tumor immune-cell infiltration in NSCLC from HIV-positive patients and to compare the results to those of HIV-undetermined patients.
Patients and methods
HIV patients and controls
All consecutive HIV adults treated for NSCLC at Tenon hospital (Paris, France) between 1996 and 2014 were enrolled. Histology samples were centrally reviewed by a pathologist specialized in thoracic oncology (M.A.). Clinical characteristics were collected from patient medical charts (P.Q. and A.L.), with special focus on age, sex, smoking history, Eastern Cooperative Oncology Group (ECOG) performance status, tumor node metastasis (TNM) stage, type of HIV contamination, blood CD4+ lymphocyte count, plasma HIV-viral load, delay between HIV diagnosis and cancer diagnosis, antiretroviral treatment, as well as HIV comorbidities. Histological type was determined according to the 2015 World Health Organization (WHO) classification (M.A.), according to morphological features, and p40, CK5/6, TTF1, CK7, and CK20 expression, analyzed by immunohistochemistry. Epidermal growth factor receptor (EGFR) and Kirsten-Ras (KRAS) mutational status, as well as anaplastic lymphoma kinase (ALK) expression, were routinely determined for adenocarcinomas and large cell carcinomas.
Consecutive surgical samples from NSCLC tumors (20 adenocarcinomas, 19 squamous cell carcinomas, and 15 large cell lung cancers) of HIV-undetermined patients were used as controls. Their histology was centrally reviewed using the same criteria than those applied for HIV tumors (M.A.). Patient characteristics have been provided in Table 1. These surgical samples were used to perform tissue microarray (TMA) analysis, with 1.0-mm diameter cores. Three cores were selected from each tumor, notably two from the center and one from the periphery, using a Minicore tissue arrayer (Excilone, Plaisir, France; C.H.).
Each patient signed a consent form, as required by national guidelines. The samples were collected in accordance to French legislation and ethical codes. A nonopposition for use of tumor samples was obtained from each patient.
Analysis of programmed cell death-ligand 1 expression
PD-L1 expression was analyzed by means of immunohistochemistry (C.D. and M.A.). Briefly, 3-μm sections were deparaffinized, rehydrated, and washed with phosphate-buffered saline (PBS). The PD-L1 antigen was retrieved by means of a tris-EDTA buffer (pH 9; Dako, Courtaboeuf, France), applied for 40 min in a steamer. The samples were blocked for peroxidase endogenous activity using the Dako Dual Endogenous and Serum-free blocks, prior to incubation with the primary antibody (1/500 dilution) overnight at +4 °C. Amplification was performed using the Dako EnVision antimouse immunoperoxidase method, with diaminobenzidine as the chromogen, for detection purposes. Irrelevant mouse IgG1 for mouse antibody (Dako) was used as negative control. Placenta tissue was used as positive control. Two PD-L1 antibodies were employed, provided by Dr Lieping Chen (clone 5H1; Yale University, New Haven, Connecticut, USA) and the Cell Signaling laboratory (clone E1L3N; Danvers, Massachusetts, USA). These antibodies have been validated in previously published studies [15,16]. The score used was calculated as the product of intensity (measured as 0, 1, 2, or 3) by positive cell percentage. We have employed a positivity threshold of 5% to assess the percentage of tumor cells [17–19]. The PD-L1 score was considered positive if greater than 5 (minimum positive cell percentage of 5 × minimum intensity of 1).
Analysis of immune infiltrate
In order to characterize the immune-cell populations, immunohistochemistry studies were performed by means of a Benchmark System (Ventana medical system, Tucson, Arizona, USA), using 3-μm formalin-fixed paraffin-embedded (FFPE) slides in order to characterize T lymphocytes as CD3 (clone SP7; dilution: 1 : 200; incubation: 20 min; Labvision, Fremont, California, USA), CD4 (clone 1F6; dilution: 1 : 25; incubation: 60 min; Novocastra, Nanterre, France), CD8 (clone C8/144b; dilution: 1 : 50; incubation: 32 min, Dako), or CD20 (clone L26, dilution 1 : 400, 40 min, Dako) cells, along with macrophages by means of CD163 (clone 10D6; dilution: 1 : 100; incubation: 32 min; Novocastra) and neutrophils by myeloperoxidase (MPO; clone 59A5; dilution: 1 : 100; incubation: 32 min; Novocastra) activity. The percentage of immune-positive cells was evaluated for each tumor sample based on the percentage of positive cells, the threshold of positivity being at least 5%.
Categorical variables were compared by mean values, and nonnormal continuous variables were expressed as medians (range). The Mann–Whitney U-test was used to compare the quantitative variables between HIV-positive and HIV-undetermined patients, such as intensity of PD-L1 expression and percentage of cells expressing markers for immune infiltrate. The agreement between both antibodies was estimated by means of the Kappa coefficient (high concordance between 0.61 and 0.80). Fisher's exact test was employed to compare PD-L1 score (positive versus negative score, with a cut-off of 5) and percentage of PD-L1 expression between the two groups (positive versus negative percentage, with a cut-off of 5). A Cox model was applied for survival analysis. Results were considered significant if the P-value was less than 0.05. Statistical tests were performed using Graphpad Version 23.0 (Graphpad Inc., San Diego, California, USA).
A total of 34 HIV-positive patients were included in the analysis; with patient characteristics provided in Table 1. Forty-seven percentage of samples were obtained by surgery, and 53% by a biopsy procedure. The most common type of HIV transmission risk was homosexual contact (35.3%), followed by intravenous drug use (29.4%). In 11.8%, cancer diagnosis led to the HIV-positive diagnosis. The interval between HIV infection diagnosis and lung cancer was 11.5 years (0–27.4 years). At the time of lung cancer diagnosis, 83% of patients exhibited a blood CD4+ lymphocyte count greater than 200/μl. Eighty-five percentage of patients had a performance status of 0 or 1. The average tobacco consummation was 38.5 packs of cigarettes per year (0–110). Approximately three-quarters of the patients received at least one antiretroviral therapy, whereas half were undergoing tritherapy. The remaining 23.5% patients were not receiving any antiretroviral therapy at the time of lung cancer diagnosis. Of the 14 patients with HIV-related comorbidities, the primary infections were hepatitis C infection (28.6%), pneumocystosis (21.4%), toxoplasmosis (14.3%), and hepatitis B infection (14.3%).
Programmed cell death-ligand 1 expression
Tumor samples that still contained tumor cells following routine diagnosis were evaluated for PD-L1 expression using two antibodies: clone E1L3N (n = 32) and clone 5H1 (n = 30; Supplementary Figure 1, http://links.lww.com/QAD/B201). PD-L1 staining was present at the membrane site of tumor cells (Fig. 1). Four tumors exhibited PD-L1 coexpression on inflammatory cells, such as macrophages. With a 5% threshold of positivity, 6 out of 32 (18.7%) and 3 out of 30 (10%) HIV patients were positive for PD-L1 with E1L3N and 5H1 clones, respectively, and 9.3% (E1L3N) and 5.6% (5H1) for HIV undetermined cohort. For both antibodies, the expression frequency did not differ between HIV-positive and HIV-undetermined patients (Fisher's test; data not shown). Concordance test between these two antibodies revealed substantial agreement, with a kappa of 0.72. Staining intensity (ranging from 0 to 3) was strong in the HIV-positive patients, yet always weak in the HIV-undetermined patients, although this difference was not statistically significant (Mann–Whitney U-test; data not shown). The PD-L1 score, considering both staining frequency and intensity, was higher in HIV-positive patients for E1L3N antibody [median (range) 0 (0–150) versus 0 (0–26.7), P = 0.047), whereas there was no statistically significant difference for 5H1 antibody [0 (0–120) versus 0 (0–26.7), P = 0.07] (Fisher's test; Fig. 2). The scores were widely distributed in the HIV-positive patients, with some very high, whereas all the HIV-undetermined scores were low (data not shown).
Immune and inflammatory infiltrates
Tumor samples that still contained tumor cells were evaluated for immune and inflammatory cell infiltration (n = 22; Supplementary Figure 1, http://links.lww.com/QAD/B201). The HIV tumors consisted primarily of macrophages and CD8+ T-infiltrating cells (Fig. 3 and Table 2). Compared with the HIV-undetermined tumors, the HIV tumors exhibited more CD8+ T cells (P < 0.0001), more CD20 lymphocytes (P = 0.005), and more macrophages (CD163) (P < 0.0001; Mann–Whitney U-test). There was no statistical difference for CD3 and CD4 lymphocytes and MPO neutrophils (P = 0.74, P = 0.76, and P = 0.13, respectively).
Correlation between clinical characteristics and programmed cell death-ligand 1 score
There was no association between clinical characteristics and PD-L1 score, regardless of the antibody used (Supplementary Table 1, http://links.lww.com/QAD/B201). There was no correlation between viral load or blood CD4+ count and PD-L1 score, nor between blood CD4+ count and tumor CD4 staining (data not shown). Furthermore, there was no impact of PD-L1 score on overall survival (Supplementary Figure 2, http://links.lww.com/QAD/B201). For the E1L3N antibody, the median overall survival was 13.5 months for patients with positive scores [95% confidence interval (95% CI) 1.7–25.3) and 16.5 months for those with negative scores (95% CI 10.1–22.9; P = 0.89). For the 5H1 antibody, the median overall survival was 13.5 months for patients with positive scores (95% CI, not applicable) and 16.5 months for those with negative scores (95% CI 8.4–24.6; P = 0.52).
We demonstrated that NSCLC in HIV-positive patients exhibited higher PD-L1 expression as that of HIV-undetermined patients, for E1L3N antibody. PD-L1 expression was not associated with any clinical characteristics, neither with blood CD4+ count nor HIV viral load. Significantly, immune-cell infiltration was higher in HIV tumors, with more CD8+ T cells, more B cells, and more macrophages than in HIV-undetermined patients.
Immunotherapies targeting the PD-1 pathway are emerging treatments, especially in lung cancer, in which 19–100% of HIV-undetermined tumors have been reported overexpressing PD-1/PD-L1 [8,20]. We have selected two antibodies in our study: L. Chen, clone 5H1, given that it was the first to be evaluated in a Phase I study, and Cell Signaling clone E1L3N, as the latter is commercially available and has shown good correlations with the other clones [21,22]. We reported that PD-L1 expression was higher for HIV-positive patients with E1L3N clone. Furthermore, high staining intensity was observed only in some HIV-positive patients, yet in no undetermined ones. Differences of PD-L1 expression between 5H1 and E1L3N clones are not surprising. Indeed, prevalence of PD-L1 expression could vary dramatically, probably because of differences in binding domain on PD-L1 protein . We also cannot exclude the possibility that the lack of significant difference with 5H1 antibody was caused by our study's lack of power because of its small number of included patients, even if all consecutive patients with lung cancers and HIV infection were included based on our department's prospective database. Another important bias was the control cohort constituting surgical samples from localized disease. Although PD-L1 expression does not seem to be dependent on disease stage [23,24], recent data has revealed that lung biopsies underestimated PD-L1 status compared with the corresponding resected surgical specimens . Lastly, the differences between both cohorts in terms of tobacco exposure and histology could be another bias. Indeed, smoking history has been associated with increased PD-L1 expression in lung cancer, whereas adenocarcinoma, the most significant histologic subtype in our HIV cohort, has been associated with lower PD-L1 expression than squamous cell carcinoma. Even if we did not observe any difference between both antibodies according to the type of sample (data not shown), this could explain also the low rates of PD-L1 expressions in both cohorts.
Stronger PD-L1 expression and immune infiltration in HIV patients would not be surprising for several reasons. Indeed, these patients have chronic immune activation and accumulation of CD8+ T cells, leading to lymphocytic alveolitis [25,26]. There is also a direct induction of PD-L1 expression by the virus. Furthermore, increased immune infiltration has been found in lymphoma from HIV-positive patients . Lastly, HIV patients have been shown to display important smoking exposure, possibly resulting in high mutational loads and high neoantigen burden, which were reported to be associated with high immune checkpoint inhibitors (ICI) responses.
In our study, we have used a score combining the proportion of positive tumor cells and the staining intensity. The proportion of positive cells is usually used in clinical trials and articles evaluating the predicting value of PD-L1 to ICI, and as other studies, we used a threshold of 5%. The PD-L1 positivity cut-off has not yet been defined, with variations observed depending on the clone, treatment line, and molecule used. Yet, the 5% cut-off has been validated in several studies [17–19]. However, only very few data are available in the scientific literature regarding the relevance of assessing PD-L1-staining intensity. Indeed, staining intensity has been reported to be relevant to better select patients that could benefit or not from PD-L1 immunotherapies, as in colorectal mismatch-repair deficient cancer or in castration-resistant prostate cancer [28,29].
The immune infiltration of tumors from HIV-positive patients exhibited higher rates of CD8, CD20, and macrophages, perhaps related to frequent infections and inflammatory processes. We did not observe any correlation between this infiltration and smoking history, CD4+ cell counts, and HIV viral load, likely accounted for by the small patient number in our cohort. The presence of this intense inflammatory and immune infiltration suggests that PD-L1 activation does mostly result from a preexisting immune response rather than intrinsic oncogenic events in these tumors. The presence of tumor-infiltrating lymphocytes, as a marker of a preexisting immune response, associated with positive PD-L1 expression, is currently the best known predictive factor of efficacy with respect to these therapies .
Furthermore, immunotherapies targeting tumors in HIV patients potentially display an impact on HIV-disease control. HIV stimulates PD-L1 expression, leading to T lymphocyte exhaustion and viral proliferation. PD-1/PD-L1 blockade can enable immune restoration and better viral control . All these findings provide real justifications for using PD-L1 or PD-1 immunotherapy in this specific population.
To our knowledge, this is the first report of PD-L1 expression in lung tumors pertaining to HIV-positive patients. We have demonstrated that HIV-infected tumors expressed higher PD-L1 than HIV-undetermined ones. Immune infiltration was more significant in HIV-positive patients, with more CD8+ T cells, more B cells, and more macrophages. These results must now be confirmed using a larger patient cohort, thereby supporting the argumentation for clinical trials to be conducted on immunotherapies in the HIV-positive population.
This work was supported by the Fondation du Souffle (Lauréat 2015 AO automne) and by the Fonds de Dotation ‘Recherche en Santé Respiratoire.’ The funding source had no influence over the study conduct, data collection or analysis.
Conflicts of interest
There are no conflicts of interest.
1. Morlat P, Roussillon C, Henard S, Salmon D, Bonnet F, Cacoub P, et al. Causes of death among HIV-infected patients in France in 2010 (national survey): trends since 2000
2. Lavolé A, Toper C, Belmont L, Ruppert A-M, Wislez M, Cadranel J. Cancer du poumon et infection par le Virus de l’Immunodéficience Humaine
. Rev Mal Respir
3. Mani D, Haigentz M, Aboulafia DM. Lung cancer in HIV infection
. Clin Lung Cancer
4. Pakkala S, Ramalingam SS. Lung cancer in HIV-positive patients
. J Thorac Oncol
5. Brock MV, Hooker CM, Engels EA, Moore RD, Gillison ML, Alberg AJ, et al. Delayed diagnosis and elevated mortality in an urban population with HIV and lung cancer: implications for patient care
. J Acquir Immune Defic Syndr
6. Suneja G, Lin CC, Simard EP, Han X, Engels EA, Jemal A. Disparities in cancer treatment among patients infected with the human immunodeficiency virus
7. D’Jaen GA, Pantanowitz L, Bower M, Buskin S, Neil N, Greco EM, et al. Human immunodeficiency virus-associated primary lung cancer in the era of highly active antiretroviral therapy: a multiinstitutional collaboration
. Clin Lung Cancer
8. Ji M, Liu Y, Li Q, Li X-D, Zhao W-Q, Zhang H, et al. PD-1/PD-L1 pathway in nonsmall-cell lung cancer and its relation with EGFR mutation
. J Transl Med
9. Zhang Y, Huang S, Gong D, Qin Y, Shen Q. Programmed death-1 upregulation is correlated with dysfunction of tumor-infiltrating CD8+ T lymphocytes in human nonsmall cell lung cancer
. Cell Mol Immunol
10. Brahmer JR, Tykodi SS, Chow LQM, Hwu W-J, Topalian SL, Hwu P, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer
. N Engl J Med
11. Herbst RS, Baas P, Kim D-W, Felip E, Pérez-Gracia JL, Han J-Y, et al. Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced nonsmall-cell lung cancer (KEYNOTE-010): a randomised controlled trial
12. Day CL, Kaufmann DE, Kiepiela P, Brown JA, Moodley ES, Reddy S, et al. PD-1 expression on HIV-specific T cells is associated with T-cell exhaustion and disease progression
13. Trautmann L, Janbazian L, Chomont N, Said EA, Gimmig S, Bessette B, et al. Upregulation of PD-1 expression on HIV-specific CD8+ T cells leads to reversible immune dysfunction
. Nat Med
14. Velu V, Titanji K, Zhu B, Husain S, Pladevega A, Lai L, et al. Enhancing SIV-specific immunity in vivo by PD-1 blockade
15. Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, et al. Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion
. Nat Med
16. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer
. N Engl J Med
17. Larkin J, Chiarion-Sileni V, Gonzalez R, Grob JJ, Cowey CL, Lao CD, et al. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma
. N Engl J Med
18. Borghaei H, Paz-Ares L, Horn L, Spigel DR, Steins M, Ready NE, et al. Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer
. N Engl J Med
19. Brahmer JR, Drake CG, Wollner I, Powderly JD, Picus J, Sharfman WH, et al. Phase I study of single-agent anti-programmed death-1 (MDX-1106) in refractory solid tumors: safety, clinical activity, pharmacodynamics, and immunologic correlates
. J Clin Oncol
20. Garon EB. Current perspectives in immunotherapy for non-small cell lung cancer
. Semin Oncol
2015; 42 (suppl 2):S11–S18.
21. Adam J, Rouquette I, Damotte D, Badoual C, Danel C, Damiola F, et al. PL04a.04: Multicentric French Harmonization Study for PD-L1 IHC testing in NSCLC
. J Thorac Oncol
22. Rimm DL, Han G, Taube JM, Yi ES, Bridge JA, Flieder DB, et al. A prospective, multiinstitutional, pathologist-based assessment of 4 immunohistochemistry assays for PD-L1 expression in non-small cell lung cancer
. JAMA Oncol
23. Pan Z-K, Ye F, Wu X, An H-X, Wu J-X. Clinicopathological and prognostic significance of programmed cell death ligand1 (PD-L1) expression in patients with nonsmall cell lung cancer: a meta-analysis
. J Thorac Dis
24. Ilie M, Long-Mira E, Bence C, Butori C, Lassalle S, Bouhlel L, et al. Comparative study of the PD-L1 status between surgically resected specimens and matched biopsies of NSCLC patients reveal major discordances: a potential issue for anti-PD-L1 therapeutic strategies
. Ann Oncol
25. Twigg HL, Knox KS. Impact of antiretroviral therapy on lung immunology and inflammation
. Clin Chest Med
26. Twigg HL, Soliman DM, Day RB, Knox KS, Anderson RJ, Wilkes DS, et al. Lymphocytic alveolitis, bronchoalveolar lavage viral load, and outcome in human immunodeficiency virus infection
. Am J Respir Crit Care Med
1999; 159 (5 Pt 1):1439–1444.
27. Chao C, Xu L, Silverberg MJ, Martínez-Maza O, Chen L-H, Castor B, et al. Stromal immune infiltration in HIV-related diffuse large B-cell lymphoma is associated with HIV disease history and patient survival
28. Massari F, Ciccarese C, Caliò A, Munari E, Cima L, Porcaro AB, et al. Magnitude of PD-1, PD-L1 and T lymphocyte expression on tissue from castration-resistant prostate adenocarcinoma: an exploratory analysis
. Target Oncol
29. Lee LH, Cavalcanti MS, Segal NH, Hechtman JF, Weiser MR, Smith JJ, et al. Patterns and prognostic relevance of PD-1 and PD-L1 expression in colorectal carcinoma
. Mod Pathol
30. Herbst RS, Soria J-C, Kowanetz M, Fine GD, Hamid O, Gordon MS, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients
human immunodeficiency virus; immunohistochemistry; lung neoplasms; programmed cell death protein ligand 1 expression; tumor-associated immune cells
Supplemental Digital Content
Copyright © 2018 Wolters Kluwer Health, Inc.