Protective Activity of Programmed Cell Death Protein 1 Blockade and Synergy With Caspofungin in a Murine Invasive Pulmonary Aspergillosis Model

Abstract

Pharmacological immune checkpoint blockade has revolutionized oncological therapies, and its remarkable success has sparked interest in expanding checkpoint inhibitor therapy in infectious diseases. Herein, we evaluated the efficacy of programmed cell death protein 1 (PD-1) blockade in a murine invasive pulmonary aspergillosis model. We found that, compared with isotype-treated infected control mice, anti–PD-1–treated mice had improved survival, reduced fungal burden, increased lung concentrations of proinflammatory cytokines and neutrophil-attracting chemokines, and enhanced pulmonary leukocyte accumulation. Furthermore, combined treatment with anti–PD-1 and caspofungin resulted in a significant survival benefit compared with caspofungin or anti–PD-1 therapy alone, indicating a synergistic effect between PD-1 inhibitors and immunomodulatory antifungal agents.

We studied the efficacy of immune checkpoint blockade in mice with invasive pulmonary aspergillosis. Anti–programmed cell death protein 1 improved survival, reduced fungal burden, induced proinflammatory lung cytokines, and enhanced pulmonary leukocyte accumulation, and, with caspofungin, synergistically improved survival.

Opportunistic mold infections including invasive pulmonary aspergillosis (IPA) remain a major source of disease and death in immunocompromised patients. Because current antifungal pharmacotherapy is suboptimal in the setting of impaired host immunity, there is an unmet need for adjunct immune enhancement strategies to potentiate the efficacy of antifungals [1]. Although cellular immunotherapies such as adoptive T-cell transfer or chimeric antigen receptor T cells or natural killer cells may have game-changing potential, their translatability to the bedside is complicated by logistical, regulatory, and economic hurdles [1]. Immune checkpoint inhibitor (CPI) therapy, a widely available immunotherapeutic strategy that already has proven value in modern oncology, could overcome these obstacles and is gaining interest for applications in the field of infectious diseases [2].

Blockade of the cytotoxic T-lymphocyte antigen 4 (CTLA-4) and programmed cell death protein 1 (PD-1)/Programmed cell death protein 1 ligand 1 (PD-L1) checkpoint pathways can prevent immune exhaustion and harness immunity to various classes of antigens, including clinically important molds and yeasts [2–7]. Specifically, in vitro studies revealed a potential role of CPI in promoting protective immunity to Aspergillus fumigatus, the most common opportunistic mold pathogen in immunocompromised patients [7]. Therefore, we sought to evaluate the tolerability and therapeutic efficacy of CTLA-4 and PD-1 blockade in a murine IPA model. We found improved survival and fungal clearance in mice receiving low-dose anti–PD-1 therapy, associated with increased pulmonary leukocyte accumulation and secretion of proinflammatory cytokines. In addition, we show potential synergism of anti–PD-1 with caspofungin, an antifungal with known immunomodulatory activity.

MATERIALS AND METHODS

Murine Infection Model

Female 8–12-week-old BALB/cAnNCrl inbred mice (Charles River Laboratories) with a weight of 20–25 g were immunosuppressed with cyclophosphamide (Sigma-Aldrich, 150 mg/kg body weight on days −4 and −1, 100 mg/kg on day +3) and cortisone acetate (Sigma-Aldrich, 300 mg/kg on day −1), as described elsewhere [8]. Mice were infected intranasally with 5 × 10⁴A. fumigatus Af-293 conidia, prepared as described in the Supplementary Methods section. Starting 6 hours after infection, mice were treated intraperitoneally with phosphate-buffered saline (PBS), immunoglobulin G2a antibody (isotype control; Leinco Technologies, I-1177), monoclonal PD-1 antibody (anti–PD-1; Leinco Technologies, P362), and/or monoclonal CTLA-4 antibody (anti–CTLA-4, Leinco Technologies, C1614). The dosing interval and drug concentrations for each experiment are detailed in the figure legends. To study combination therapy of checkpoint blockade and conventional antifungal therapy with immunomodulatory activity [9], daily intraperitoneal treatment with 1 mg/kg caspofungin (Sigma-Aldrich) or a mock injection with saline was administered. Survival was monitored daily until day 8 after infection. All experiments were approved by the MD Anderson Cancer Center Institutional Animal Care and Use Committee (protocol 00001734-RN00).

Determining Fungal Burden With Quantitative Polymerase Chain Reaction

Quantification of lung fungal burden with 18S quantitative polymerase chain reaction was performed as described elsewhere [10], with minor modifications as specified in the Supplementary Methods.

Cytokine Assays

Murine lung tissue was homogenized in PBS containing a broad-spectrum protease inhibitor (Roche) and 0.025% Tween20 (Bio-Rad), using a Mini-Beadbeater homogenizer (BioSpec Products) and 15 acid-washed glass beads (3 mm; Sigma-Aldrich). Homogenates were centrifuged at 3600 rpm for 5 minutes. Supernatants were cleaned from debris by another centrifugation step at 13 200 rpm. Cytokine and chemokine concentrations per gram of lung tissue were determined using a Luminex assay, as described elsewhere [11].

Histopathology

Lungs were fixed in 10% formaldehyde and embedded in paraffin wax. Representative sections were stained with Grocott-Gomori methenamine–silver nitrate (GMS) or hematoxylin-eosin (HE), imaged with a ScanScope slide scanner (Aperio; Leica Biosystems), and analyzed using ImageScope x64 software (Aperio; Leica Biosystems).

Statistical Analysis

Statistical analyses were performed using GraphPad Prism software, version 8. The Mantel-Cox log-rank test was used to determine significant differences in survival curves. Fungal burden was compared using the Kruskal-Wallis test with the Dunn multiple comparison test. Cytokine and chemokine concentrations were analyzed using the 2-sided Mann-Whitney test, with Benjamini-Hochberg correction for multiple testing. Significance levels were stratified at P < .05, P < .01, and P < .001.

RESULTS

In initial experiments, we compared the tolerability and efficacy of anti–PD-1 and anti–CTLA-4, as well as their combined administration in immunosuppressed A. fumigatus–infected mice (Supplementary Figure 1). To limit toxicity, we split the weekly dose of CPI described in murine oncological studies [12] into daily injections of 50 µg of anti–PD-1 (2–2.5 mg/kg) and/or 20 µg of anti–CTLA-4 (0.8–1 mg/kg) per mouse. In uninfected immunosuppressed mice, neither CPI elicited apparent toxic effects (data not shown). Although anti–PD-1 treatment led to improved survival rates among A. fumigatus–infected animals (Supplementary Figure 1), mice became lethargic, and 10 of 15 showed signs of stress, such as limping, hunched posture, dull or sluggish movements, frequent uncontrollable urination, and labored breathing. Groups receiving anti–CTLA-4 treatment or the combined anti–PD-1 plus anti–CTLA-4 regimen did not show significantly improved survival outcomes (Supplementary Figure 1). Furthermore, dual checkpoint blockade was even more toxic than individual CPI treatment in A. fumigatus–infected mice, with an earlier onset of apparent signs of stress and toxicity in 14 of 15 mice. Therefore, we focused on anti–PD1 monotherapy at an attenuated schedule in subsequent experiments.

Specifically, we reduced the anti–PD-1 dose to 0.25 mg/kg and increased the anti–PD-1 dosing interval to every other day. We further included an additional control group treated with an immunoglobulin G isotype antibody (Figure 1A). Infected mice receiving mock treatment with either PBS or the isotype antibody exhibited 8-day survival rates of 33% and 36%, respectively. In contrast, 68% of the mice in the PD-1 antibody treatment group survived (P < .001) (Figure 1A). Accordingly, pulmonary fungal burden was strongly reduced in anti–PD-1–treated mice (median spore equivalents, 0.39 × 10⁹) compared with both PBS treatment (2.71 × 10⁹; P = .01) and the isotype control (2.06 × 10⁹; P = .08) (Figure 1B). Histopathological analysis (Grocott-Gomori methenamine–silver nitrate staining) confirmed fewer and smaller foci of hyphal invasion in the lungs of anti–PD-1–treated compared with isotype-treated mice (Figure 1C).

Figure 1.

Low-dose anti–programmed cell death protein 1 (PD-1) monotherapy improves survival and enhances proinflammatory cytokine production in murine invasive pulmonary aspergillosis (IPA). A, BALB/c mice were immunosuppressed with cyclophosphamide and cortisone acetate and infected intranasally with 50 000 Aspergillus fumigatus conidia. Mice then received 250 µg/kg monoclonal PD-1 antibody, an immunoglobulin G (IgG) isotype control antibody, or phosphate-buffered saline (PBS) (mock treatment), administered intraperitoneally every other day. Survival curves were compiled based on the assessment of 24–28 mice per treatment in 3 independent experiments. †P < .01 (log-rank test). B, Fungal burden was determined on day +8 (filled diamonds) or on death (open diamonds), using an 18S ribosomal quantitative polymerase chain reaction (qPCR) assay (n = 10 mice per treatment). Individual and median values are shown. *P < .05 (Kruskal-Wallis test with Dunn multiple comparisons test). C, Additional A. fumigatus–infected mice receiving either isotype or anti–PD-1 treatment were euthanized on day 4 after infection. Lung sections were stained with Grocott-Gomori methenamine–silver nitrate (GMS) or hematoxylin-eosin (HE). The rectangular insets denote the borders of the corresponding 20x HE panel. Representative images are shown (red scale bars represent 500 µm; black scale bar, 100 µm). D, Cytokine and chemokine concentrations in lung tissue homogenates from A. fumigatus–infected mice receiving either isotype or anti–PD-1 treatment were quantified on day +4 using a 24-plex Luminex assay. Individual (n = 6 mice per treatment and time point) and median concentrations of selected cytokines and chemokines per gram of lung tissue are shown. *P < .05 (2-sided Mann-Whitney test with Benjamini-Hochberg adjustment for a false discovery rate of 0.05). Raw data including an additional sampling time point (day +1) are provided in Supplementary Figure 3. Abbreviations: IL-1β, interleukin 1β; TNF-α, tumor necrosis factor α.

Although enhanced fungal clearance contributed to improved survival in the PD-1–treated cohort, interindividual survival outcomes of PD-1–treated mice were not significantly correlated with fungal burden (Supplementary Figure 2A), suggesting a role of additional immunomodulatory effects of anti–PD-1. Therefore, we compared the pulmonary cytokine and chemokine milieu of mice with IPA treated with anti–PD-1 or isotype. No significant impact of anti–PD-1 treatment on early cytokine and chemokine release was seen, but as the infection progressed (day +4), the anti–PD-1–treated cohort showed significant up-regulation of the proinflammatory mediators tumor necrosis factor α, interleukin 1β, interleukin 6, and granulocyte-macrophage colony-stimulating factor as well as the neutrophil-targeted chemokines CXCL1, CXCL2, and CCL3 [13], compared with isotype-treated mice (Supplementary Figure 3; selected examples in Figure 1D). Correspondingly, HE-stained lung sections of anti–PD-1–treated mice revealed that hyphal invasion foci were surrounded by heavier leukocyte infiltrates, which displayed an increased ratio of neutrophils to lymphoid cells, compared with the isotype control (Figure 1C).

Finally, we sought to study combined application of anti–PD-1 and conventional antifungal therapy (Figure 2A). Therefore, we selected caspofungin, a drug that has modest efficacy against Aspergillus but has been associated with protective immunomodulatory activity [9]. Baseline mortality rates in infected and isotype-treated mice were higher than in the previous experimental series (Figure 2B), possibly owing to increased stress caused by daily manipulations. However, anti–PD-1 monotherapy prolonged the median survival time from 3 to 5 days and increased 8-day survival from 0% to 33% (P < .001 vs isotype) (Figure 2B), thus resulting in a similar survival advantage as in the preceding experiment (Figure 1A). While caspofungin in combination with the isotype control provided protection comparable to that achieved with anti–PD-1 treatment (median survival time, 4 days; 8-day survival rate, 33%; P = .002 vs isotype only), 83% of the mice receiving combined caspofungin and anti–PD-1 treatment were alive on day 8 after infection (P < .001 vs all other treatment arms) (Figure 2B).

Figure 2.

Anti–programmed cell death protein 1 (PD-1) and caspofungin (CAS) provide synergistic therapeutic activity in murine invasive pulmonary aspergillosis (IPA). A, Timeline of experimental procedures. BALB/c mice were immunosuppressed with cyclophosphamide and cortisone acetate and infected intranasally with 50 000 Aspergillus fumigatus conidia. Mice then received 250 µg/kg PD-1 antibody (anti–PD-1) or an immunoglobulin G (IgG) isotype control every other day, administered intraperitoneally. In addition, mice received daily intraperitoneal gavage of 1 mg/kg CAS or a mock injection with saline. B, Survival curves were compiled based on the assessment of 24 mice per treatment in 2 independent experiments. †P < .01; ‡P < .001 (log-rank test). C, Fungal burden was determined on day +8 or on death, using an 18S ribosomal quantitative polymerase chain reaction (qPCR) assay (n = 10 mice per treatment). Individual and median values are shown. ‡P < .001 (Kruskal-Wallis test with Dunn multiple comparisons test). D, Two additional mice from each treatment arm were euthanized on day 4 after infection. Lung sections were stained with Grocott-Gomori methenamine–silver nitrate (GMS) or hematoxylin-eosin (HE). The rectangular inserts denote the borders of the corresponding 20x HE panel. Representative images are shown (red scale bar represents 500 µm; black scale bars, 100 µm).

Both anti–PD-1 monotherapy and its combination with caspofungin significantly lowered pulmonary fungal burden as detected with quantitative polymerase chain reaction (P < .001), whereas caspofungin alone had a moderate, nonsignificant effect on fungal clearance (Figure 2C). As in the preceding experiment, enhanced fungal clearance was significantly correlated with improved survival outcomes across all treatment arms but not with interindividual outcomes in infected mice receiving combination treatment (Supplementary Figure 2B). HE-stained lung sections again revealed markedly increased pulmonary neutrophil accumulation in both anti–PD-1–treated cohorts, whereas the addition of caspofungin to anti–PD-1 treatment did not change the magnitude or type of pulmonary leukocyte infiltrates (Figure 2D). However, considerably less tissue injury and necrosis were seen in mice receiving anti–PD-1 combined with caspofungin, underscoring potentially protective immunomodulatory properties of combination therapy (Figure 2D).

DISCUSSION

Here, we report the first in vivo study of CPI as immunotherapy in a neutropenic IPA model. Collectively, our results indicate that anti–PD-1 checkpoint blockade during A. fumigatus infection, without antifungals, confers a survival advantage and improves fungal clearance from infected lungs. Thereby, our proof-of-concept study adds to the emerging body of preclinical evidence indicating a protective role of CPI in various invasive mycoses [2] as seen in animal models of cryptococcosis [3], candidiasis [4], and disseminated histoplasmosis [5] and in an intravenously infected postsepsis aspergillosis model [6].

Multiplex cytokine profiling in infected lung homogenates revealed 2 major groups of mediators significantly up-regulated in the anti–PD-1–treated cohort. On one hand, key proinflammatory cytokines such tumor necrosis factor α, interleukin 1β, and interleukin 6, which are important mediators of protective anti-Aspergillus immunity [13], were more prominently induced in anti–PD-1–treated mice, suggesting that the immune microenvironment in the lungs created by PD-1 inhibition may be more permissive for Aspergillus control via innate immune cell activation [13]. In addition, neutrophil-attracting chemokines were enriched in the lungs of anti–PD-1–treated mice, consonant with increased pulmonary neutrophil accumulation, as corroborated by histopathological analysis.

Previous studies found rather limited early changes to pulmonary T-cell accumulation in murine cryptococcosis [3] and candidiasis [4]. Accordingly, pulmonary concentrations of T-cell-derived cytokines in our IPA model displayed very minor differences between anti–PD-1–treated and isotype-treated mice on days 1 and 4 after infection. This observation could contribute to the explanation of significantly greater survival advantage facilitated by anti–PD1 compared with anti–CTLA-4 in our rapidly lethal immunosuppressed IPA model. Unlike CTLA-4, which primarily modulates T-cell activation [2], PD-1 is expressed in a wide array of peripheral tissues, and the PD-1/PD-L1 pathway is involved in the activation of non–T-cell leukocyte populations, including natural killer cells and mononuclear phagocytes, which are known to play a critical role in early protective anti-Aspergillus immunity [2, 13]. Long-term observation studies after recovery from immunosuppression should complement the present analysis in order to characterize the comparative contributions of T cells and other leukocyte subsets to anti–PD1-mediated protection. Specifically, there may be delayed effects of anti–PD-1 on sustained activation and exhaustion of adaptive cellular immunity in IPA, as seen in previous studies of CPI therapy in invasive mycoses, including postsepsis aspergillosis [3, 4, 6].

While contributing to enhanced neutrophil recruitment and fungal clearance, the strongly induced release of proinflammatory cytokines in conjunction with dose-dependent, likely hyperinflammatory toxicity also underscores the possibility that CPI may be a double-edged sword in an acute, high-inoculum infection setting, with potential detrimental immune activation and tissue damage. Specifically, the marked toxicity encountered in mice with dual checkpoint blockade matches previous experience in several oncological immunotherapy trials [14]. As is the case for antitumor immunotherapy, thorough immune profiling studies will be necessary to establish feasible biomarkers predicting anti-Aspergillus immune augmentation and the risk for immune-related toxicity [15].

Documenting a synergistic survival advantage and reduced lung tissue necrosis in mice receiving anti–PD-1 plus caspofungin, our study also underscores the promise of combining immune therapy with traditional antifungals. While displaying limited anti-Aspergillus activity, echinocandins disturb the equilibrium of the fungal cell wall and lead to unmasking of immunogenic epitopes, thereby enhancing phagocytic activity and neutrophil-induced fungal damage [9]. Because fungal antigen exposure drives up-regulation of the PD-1/PD-L1 axis [2–4, 7], we hypothesize that concomitant PD-1 blockade may contribute to improved and sustained activation of cellular immunity. As a step toward clinical translation of combined CPI and antifungal therapy, it would be interesting to investigate anti–PD-1 treatment along with Aspergillus-active antifungals, such as voriconazole or liposomal amphotericin B. Future studies should further compare infection outcomes depending on the timing of antifungal therapy and CPI administration, define dose-response correlations, and include a thorough evaluation of adverse inflammatory responses and autoimmune toxicity [2]. Specifically, enhancement of conventional antifungal therapy or prophylaxis with a single dose of anti–PD-1 therapy may present an appealing approach to exploiting potential synergistic activity while limiting CPI-related inflammatory toxicity.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Presented in part: ID Week 2019, Washington, DC, 2–6 October 2019. Oral presentation 974.

Financial support. This study was supported by the National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (grant 5R03AI137754 to D. P. K. and support from the NIAID Division of Intramural Research to M. S. L.), the Texas 4000 for Cancer Distinguished Professorship for Cancer Research (D. P. K.), the National Cancer Institute, National Institutes of Health (Cancer Center CORE support grant 16672), and the MD Anderson Cancer Center’s Division of Internal Medicine (research and quality improvement award to S. W.).

Potential conflicts of interest. D. P. K. reports research support from Astellas Pharma and honoraria for lectures from Merck, Gilead, and United Medical, and he has served as a consultant for Astellas Pharma, Cidara, Amplyx, Pulmocide, and Mayne Pharma. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.