Ivermectin converts cold tumors hot and synergies with immune checkpoint blockade for treatment of breast cancer

We show that treatment with the FDA-approved anti-parasitic drug ivermectin induces immunogenic cancer cell death (ICD) and robust T cell infiltration into breast tumors. As an allosteric modulator of the ATP/P2×4/P2×7 axis which operates in both cancer and immune cells, ivermectin also selectively targets immunosuppressive populations including myeloid cells and Tregs, resulting in enhanced Teff/Tregs ratio. While neither agent alone showed efficacy in vivo, combination therapy with ivermectin and checkpoint inhibitor anti-PD1 antibody achieved synergy in limiting tumor growth (p=0.03) and promoted complete responses (p<0.01), also leading to immunity against contralateral re-challenge with demonstrated anti-tumor immune responses. Going beyond primary tumors, this combination achieved significant reduction in relapse after neoadjuvant (p=0.03) and adjuvant treatment (p<0.001), and potential cures in metastatic disease (p<0.001). Statistical modeling confirmed bona fide synergistic activity in both the adjuvant (p=0.007) and metastatic settings (p<0.001). Ivermectin has dual immunomodulatory and ICD-inducing effects in breast cancer, converting ‘cold’ tumors ‘hot’, thus represents a rational mechanistic partner with checkpoint blockade.

and Tregs, resulting in enhanced Teff/Tregs ratio. While neither agent alone showed efficacy in 23 vivo, combination therapy with ivermectin and checkpoint inhibitor anti-PD1 antibody achieved 24 synergy in limiting tumor growth (p=0.03) and promoted complete responses (p<0.01), also 25 leading to immunity against contralateral re-challenge with demonstrated anti-tumor immune 26 responses. Going beyond primary tumors, this combination achieved significant reduction in 27 relapse after neoadjuvant (p=0.03) and adjuvant treatment (p<0.001), and potential cures in 28 metastatic disease (p<0.001). Statistical modeling confirmed bona fide synergistic activity in both 29 the adjuvant (p=0.007) and metastatic settings (p<0.001). Ivermectin has dual 30 immunomodulatory and ICD-inducing effects in breast cancer, converting 'cold' tumors 'hot', thus 31 represents a rational mechanistic partner with checkpoint blockade. 32 Checkpoint blockade (1, 2) has emerged as a revolutionary approach that harnesses a patient's 34 own immune system to treat cancer. However, checkpoint inhibitors as single agents are only 35 effective in a subset of patients and cancer types (2). Recent studies suggest that efficacy of 36 checkpoint inhibitors is primarily limited to cancers already infiltrated by T cells -often termed 37 'hot' tumors. In contrast, 'cold' tumors have little to no T cell infiltration and generally do not 38 respond to checkpoint blockade. Early clinical studies with checkpoint blockade therapy in breast 39 cancer have focused on triple negative breast cancer (TNBC), because this subtype has a higher 40 mutational load and is thought to be more 'immunogenic' (3). Even so, anti-PD1/PDL1 antibodies 41 have produced clinical responses in only a small subset (15-20%) of TNBC patients (4). As such, 42 there is considerable interest in identifying drugs capable of priming breast tumors (turning 'cold' 43 tumors 'hot') to synergize with checkpoint blockade. 44 A recently described phenomenon, termed immunogenic cell death (ICD) (5, 6), is a form of cell 45 death that induces an immune response from the host. ICD is distinguished from classical 46 apoptosis and other non-immunogenic or tolerogenic forms of cell death by several hallmarks, 47 activity, defined as an effect that is significantly greater than the sum of the drugs' individual 123 effects (submodel p=0.008, false discovery rate/FDR 3%, Table 1A). Complete tumor resolution 124 was observed in 6/15 mice on the combination treatment, 1/20 on ivermectin alone, 1/10 on anti-125 PD1 antibody alone, and 0/25 on no treatment. Mice that resolved tumors on the ivermectin and 126 anti-PD1 combination therapy were re-challenged with 100,000 4T1 cells in the contralateral 127 mammary fat pads. All of these mice resisted development of new tumors (Fig. 3B), while control 128 naïve animals all developed tumors (data not shown). This suggests that combined treatment 129 with ivermectin and anti-PD1 induces protective anti-tumor immunity in complete responders. 130 To gain further insight into the mechanism underlying efficacy of the combined treatment, we 131 compared the magnitude to which ivermectin, anti-PD1, and their combination potentiated the 132 infiltration of T cells. As shown visually in Fig. 3C and quantitatively in Fig. 3D, infiltration of both 133 CD4 + and CD8 + T cells into 4T1 tumors (Day 21) was greatest after treatment with the 134 combination of ivermectin and anti-PD1. To measure anti-tumor T cells, splenocytes were 135 isolated from untreated, single agent treated or ivermectin plus anti-PD1 combination treated 136 mice, then co-cultured with 4T1 cells as targets to measure CD107 mobilization and IFN-γ 137 expression as markers for functional T cell responses (19). A functional tumor-specific immune 138 response was confirmed by the presence of a discrete population of CD8 + T cells positive for 139 CD107 and IFN-γ in mice treated with ivermectin plus anti-PD1, but not in mice treated with anti-140 PD1 alone or untreated controls (p<0.01; Fig. 3E, F). 141

Combination therapy effective across spectrum of clinically relevant settings 142
Moving beyond control of primary tumors, we sought to test this combination immunotherapy 143 across the major clinically relevant settings: neoadjuvant, adjuvant, and metastatic treatments.
Interleukin-2 (IL-2). IL-2 was the first cytokine to be successfully used in the treatment of cancer 146 to induce T cell activation(20). A major challenge in the development of IL-2 as a therapeutic 147 antitumor agent is that IL-2 can act on both T cells and regulatory T cells (Tregs). The contrasting 148 actions of IL-2 has led to inconsistent responses and limited the development of high-dose IL-2 149 for cancer immunotherapy. Increasing the half-life of IL-2 has been shown to be a promising 150 strategy for improving IL-2 based immunotherapy. This can greatly reduce the dose of IL-2 151 required for therapeutic activity, enhancing both safety and efficacy (21,22). We explored the 152 secondary hypothesis that addition of a recombinant albumin-IL-2 fusion with extended half-life 153 to the ivermectin and anti-PD1 regimen (anti-PD-1 + IL-2 therapy, termed "IP" for simplicity) can 154 further improve the efficacy of our combined treatment. 155 Neoadjuvant therapy has come to play an increasingly prominent role in the treatment of cancer. 156 We tested treatment of ivermectin combined with anti-PD-1 and IL-2 (IP) by monitoring survival 157 of animals receiving neoadjuvant combination therapy followed by surgical resection of the 158 primary tumor on day 16 following tumor inoculation (schema in Fig. S1A). Development of loco-159 regional recurrence and distant metastases were monitored by bioluminescent imaging, and 160 animals were euthanized upon decline in body condition score and signs of morbidity. All 161 untreated animals required euthanasia due to lethal diseases around day 20-25 following 162 surgical resection of primary tumor (Fig. 4A). Treatment with IP therapy alone provided some 163 survival benefit with approximately 40% of animals remaining free of lethal disease. The best 164 survival outcome was seen with the combination of IP and ivermectin therapy, with 165 approximately 75% of animals becoming long-term survivors following surgical resection 166 (p<0.05, Fig. 4A). Surviving treated mice were re-challenged with 100,000 4T1 cells in the 167 contralateral mammary fat pads. The majority of IVM + IP treated mice did not develop new 168 tumors (Fig. 4B), while IP treated and control naïve nice all developed tumors. Splenocytes from 169 these animals were reactive (via ELISPOT) against 4T1 cells, demonstrating evidence for anti-170 tumor T cell responses in the IVM + IP treated animals (Fig. 4C). These results suggest that the 171 IVM + IP combination treatment is effective in the neoadjuvant setting and induces protective 172 anti-tumor immunity in responders. 173 Surgery remains the primary treatment for breast cancer; however, relapse is common 174 necessitating adjuvant therapy in high-risk patients post-surgery. We assessed the efficacy of 175 ivermectin, anti-PD1, and recombinant IL-2 alone or in combination as adjuvant immunotherapy 176 after surgery. 4T1 cells expressing luciferase (0.5x10 6 , 4T1-Luc) were injected into the 177 mammary pad of female BALB/c mice and allowed to grow into palpable tumors over 10 days, 178 after which tumors were surgically resected. Treatment was initiated on day 2 following surgery 179 to mimic adjuvant therapy (schema in Fig. S1A). Development of recurrence and metastasis was 180 monitored at multiple time points via bioluminescence imaging (Day 17 shown, Fig. 4D), then 181 animals were monitored until they met euthanasia criteria based on decline in body condition 182 score and signs of morbidity. Treatment with anti-PD1 or IVM alone led to similar survival as 183 untreated controls (Fig. 4E). Animals treated with the combination of ivermectin and anti-PD1 184 (with or without IL-2) had significantly prolonged survival, with approximately 40% of animals 185 becoming long-term survivors (p<0.001, Fig. 4E). Through statistical modeling, the ivermectin 186 and anti-PD1 combination was found to be highly synergistic compared to IVM or anti-PD-1 187 alone (submodel p=0.007, FDR 2%, Table 1B). Interestingly, addition of IL-2 did not further 188 enhance the survival benefit from the ivermectin and anti-PD1 combination (submodel p=0.51, 189 FDR 67%, Table 1B). These data demonstrate that treatment with ivermectin and anti-PD1 (with or without IL-2) is effective in the adjuvant setting, without evidence for drug related or synergistic 191 toxicity based on parallel body weight observations (Fig. S1B). survivors. The combined effect of IVM and anti-PD-1 on survival in the metastatic setting was 206 again found to be highly synergistic compared to IVM or anti-PD-1 alone (submodel p<0.001, 207 FDR <1%, Table 1C). As in the adjuvant setting, addition of IL-2 did not further enhance the 208 survival benefit from the ivermectin and anti-PD1 combination (submodel p=0.64, FDR 73%, 209 Table 1C). These data demonstrate that treatment with ivermectin and anti-PD1 (with or without 210 IL-2) is also effective in the metastatic setting.

Discussion 212
Since its discovery in the mid-1970s, ivermectin has been used safely by over 700 million people 213 worldwide to treat river blindness and lymphatic filariasis (23); it is inexpensive and accessible. 214 Our results demonstrate that treatment with ivermectin induces robust T cell infiltration into 215 breast tumors via induction of ICD, thus turning 'cold' tumors 'hot'. Unlike conventional 216 chemotherapy drugs, this agent has the added benefit of not suppressing host immune function, 217 but rather has beneficial immunomodulatory effects -making it a promising and mechanistic 218 partner for immune checkpoint blockade. The release and accumulation of high levels of 219 extracellular ATP has emerged as a key characteristic feature of the tumor microenvironment 220 by ivermectin is consistent with previous reports demonstrating that mouse splenic Tregs 228 (CD4+CD25+) have higher sensitivity to increasing (>100 µM) doses of extracellular ATP 229 compared to CD8+ and CD4+CD25-T cells (26). This differential sensitivity to extracellular ATP 230 is P2X7-dependent and directly associated with levels of surface P2X7 receptor expression 231 (CD4+CD25+ > CD4+CD25-> CD8+ T cells). Recent reports showed that the ATP/P2X7 axis 232 also operates in MDSC and MDSC-mediated immunosuppression (27,28). This is consistent 233 with our finding that ivermectin can selectively target expanded myeloid cells isolated from tumor-bearing mice ex vivo in a P2X7-dependent fashion. Further research will be needed to 235 elucidate the relative sensitivities of different subsets of MDSC and tumor-associated 236 macrophages/neutrophils (TAMs/TANs) to ivermectin, as well as to validate the in vivo effects 237 of ivermectin on various myeloid subsets within the tumor microenvironment and systemically. 238 While differential ATP/P2X7-dependent cytotoxicity may be one possible explanation for the 239 immunomodulatory effects of ivermectin in vivo, recent reports also implicate ATP release and 240 P2X4-dependent signaling in the CXCL12/CXCR4-mediated migration and inflammation-driven 241 recruitment of T cells (29). The role of P2X4 in T cell activation, proliferation and migration was 242 particularly pronounced in CD4 T cells, which is consistent with our own data demonstrating 243 ivermectin to be particularly potent at increasing the CD4+ Teff/Treg ratios in ex vivo treated 244 splenocytes (Fig. 2D) and augmenting intra-tumoral infiltration with CD4+ T cells (Fig. 3D). Thus,  in the paper will be deposited in the Supplementary Materials. Access to those data is 372 unrestricted.

Mice and treatment 375
Female BALB/c mice were purchased from Charles River Laboratories at 5-8 weeks of age 376 and housed in City of Hope's animal care facilities under pathogen-free conditions. All 377 procedures were performed under approval from City of Hope's Animal Care and Use 378 Committee. Mice were inoculated with 100,000 4T1 breast cancer cells in the right mammary 379 fat pad, then palpated to check for tumor engraftment before commencing their assigned 380 treatment regimen. Treatments included: 5 mg/kg of ivermectin (Sigma Aldrich, St. Louis MO) 381 given via oral gavage daily for 6 days; 10 mg/kg of anti-PD1 (BioXCell, West Lebanon NH) 382 treatment given subcutaneously once weekly; MSA-IL2 administered at 1.5 mg/kg by 383 intraperitoneal injection in 50 µL sterile PBS once weekly; combination treatments; or no 384 treatment (Fig. S1). Ivermectin was solubilized in 45% (2-Hydroxypropyl)-β-cyclodextrin (Sigma 385 Aldrich, 332593-1KG), as previously described (34). Tumor growth was measured 2-3 times a 386 week with a digital caliper for up to 56 days. Mice were euthanized when tumor growth reached 387 1.5 cm in length or width. Tumor volume was calculated as (length*width 2 )/2. Metastasis 388 experiments were performed by injecting 0.5x10 6 luciferase expressing 4T1 tumor cells (4T1-389 Luc) subcutaneously in the mammary gland of female BALB/c mice, followed by surgically 390 resection of the primary tumor on day 14 after inoculation. In vivo bioluminescence imaging was 391 used to monitor metastatic outgrowth, which was carried out on a Lago X optical imaging system 392 (Spectral Instruments Imaging, Tucson, AZ). Overall tumor burden per mouse was assessed 393 weekly via bioluminescence imaging. Recurrence of primary tumor was recognized when the 394 animal's luciferase value exceeded 600,000 photons/sec/cm 2 /steradian, a threshold chosen 395 because it was well above the lower limit of reproducible detection (510,000) and because, in optimization experiments, 600,000 was the lowest threshold consistently followed by ever-To keep the trials' overall risk of error below 5%, p values for the primary hypothesis for synergy 488 from combination treatment were subjected to the step-up Bonferroni adjustment of Hochberg 489 (37). Separately, p values for the secondary hypotheses underwent the same adjustment.