PARP-inhibition reprograms macrophages toward an anti-tumor phenotype

Abstract

Poly(ADP)ribosylation inhibitors (PARPis) are toxic to cancer cells with homologous recombination (HR) deficiency but not to HR-proficient cells in the tumor microenvironment (TME), including tumor-associated macrophages (TAMs). As TAMs can promote or inhibit tumor growth, we set out to examine the effects of PARP inhibition on TAMs in BRCA1-related breast cancer (BC). The PARPi olaparib causes reprogramming of TAMs toward higher cytotoxicity and phagocytosis. A PARPi-related surge in NAD+ increases glycolysis, blunts oxidative phosphorylation, and induces reverse mitochondrial electron transport (RET) with an increase in reactive oxygen species (ROS) and transcriptional reprogramming. This reprogramming occurs in the absence or presence of PARP1 or PARP2 and is partially recapitulated by addition of NAD derivative methyl-nicotinamide (MNA). In vivo and ex vivo, the effect of olaparib on TAMs contributes to the anti-tumor efficacy of the PARPi. In vivo blockade of the "don't-eat-me signal" with CD47 antibodies in combination with olaparib improves outcomes in a BRCA1-related BC model.

Trial registration: ClinicalTrials.gov NCT02624973.

Keywords: CP: Cancer; NAD+; PARP-inhibitor; macrophages; tumor immunology; tumor metabolism.

Conflict of interest statement

Declaration of interests N.Y.R.A. is key opinion leader for Bruker Daltonics, scientific advisor to Invicro, and receives support from Thermo Finnegan and EMD Serono. H.P.E. reports personal fees (honoraria, consulting, or advisory role) from Amgen, AstraZeneca, Abbvie, Bristol-Myers-Squibb, Daiichi Sankyo, Dagens Medisin, Eli Lilly, HAI Interaktiv AS, MSD, Novartis, Pfizer, Pierre Fabre, Roche, Sanofi, and Seagen. G.M.W. reports grants from Merck & Co. and institutional support from Glaxo Smith Kline outside the submitted work. In addition, G.M.W. has US patent 20090258352 A1 Pin1 as a marker for abnormal cell growth licensed to Cell Signaling, R&D Systems. O.S. is a 2seventy bio employee and obtains compensation and has equity in the company. H.P.E. received institutional funding from AstraZeneca (ESR-14-10077) and Pfizer (GMGS 51752519).

Copyright © 2022 The Authors. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1.. Olaparib augments anti-tumor macrophage activity

(A and B) Macrophage ablation decreases the efficacy of olaparib. (A) Experimental design. K14-Cre BRCA1f/f p53 f/f tumors (hereafter referred to as K14 tumors) were implanted in syngeneic mice, and mice were randomized when the tumors reached 4–7 mm in diameter. Clodronate liposomes or PBS control liposomes were administered twice a week for the first week (100 μL/mouse), followed by once a week; olaparib treatment was daily. (B) Survival statistics, mice stratified according to treatment, log rank test. The preset endpoint was a tumor diameter of 20 mm. In brackets is shown the median survival for each cohort in days from the start of treatment. (C–E) Macrophages in olaparib-treated tumors are large and highly phagocytic. (C) Reanalysis of forward scatter signal (FSC) as a readout for cell volume for F4/80+ cells as published (Pantelidou et al., 2019; n = 5 mice for olaparib and control each). Tumors harvested from mice treated with olaparib and corresponding controls were dissociated, stained with F4/80, and subjected to flow cytometry. Data are presented as mean ± SD. Significance was determined by an unpaired t test. (D and E) K14 tumors were implanted in syngeneic mice, randomized when the tumors reached 4–7 mm in diameter, and euthanized after 10 days, and olaparib-treated (n = 8) and control (n = 6) tumors were subjected to dual-stain immunohistochemistry with γH2AX (red) and MAC2 (brown), scale bar = 25μm (D), and in each tumor, dual-positive cells were counted in six random fields (E). Data are presented as mean ± SD. Significance was determined by the Mann Whitney test, ***p

Figure 2.. Olaparib directly reprograms alternatively activated macrophages toward an anti-tumor activity

(A and B) Mouse bone marrow cells (6 donor mice) were cultured on culture slides in the presence of IL-4, IL-10, and M-CSF and olaparib or DMSO. Cells were stained with hematoxylin and eosin on day 6, scale bar = 25μm (A). Cell diameter measured by Cellometer, 4 replicates per group, 3 independent experiments; data are presented as mean ± SD, unpaired t test (B). (C) Increased phagocytosis upon olaparib treatment. Mouse bone marrow cells were cultured as in (A) but in 10 cm plates, harvested, and co-cultured with GFP-expressing K14-BRCA1f/f p53f/f cancer cells (hereafter referred to as K14-GFP cells) for 4 h at a ratio of 1:1 and stained with F4/80 antibodies. The percentage of dual-positive cells was determined by flow cytometry (4 donor mice, 4 replicates per group, data presented as mean ± SD). (D) Mouse bone marrow cells were treated as in (A) and then co-cultured with 5,000 K14-GFP cells for 24 h at ratios as indicated (5 replicates per group, data presented as mean ± SD). Bioluminescence assay (hereafter referred to as BLI assay) was performed to detect the fraction of surviving cancer cells. (E and F) Adoptive transfer of macrophages precultured ex vivo into NSG mice (10 tumor-bearing mice/arm). Bone marrow from FVB donor mice (5 donor mice per time point) was cultured as in (A) in the absence or presence of olaparib or vehicle for 6 days. Bone marrow-derived macrophages (BMDMs) were intratumorally injected in recipient NSG mice weekly 9 times (5 × 105/mouse), and olaparib was administered daily starting at day 15 after the first macrophage injection (E). Survival according to Kaplan Meier after adoptive transfer of macrophages pretreated with olaparib or DMSO ex vivo, and median survival in days is shown in parentheses (F). Cell culture data are presented as mean ± SD. Significance was determined by two-way ANOVA, ***p < 0.001. In vivo data were subjected to the log rank test.

Figure 3.. PARP inhibition induces a pro-inflammatory macrophage phenotype

(A and D–G) Mouse bone marrow cells were cultured as indicated in the absence or presence of olaparib or DMSO for 6 days (hereafter referred to as BMDMs), 3 donor mice for each experiment. (A) Total mRNAs were extracted and subjected to quantitative real-time PCR. Mean ± SD of three replicates, representative of three separate experiments. (B and C) K14 tumors from mice treated for 10 days with olaparib were subjected to dual-stain IHC with iNOS (red) and MAC2 (blue), scale bar = 25μm (B), and 8 fields of each slide from 6 control tumor-bearing mice and 8 olaparib treated tumor-bearing mice were counted for double positive cells. Mean ± SD, Mann Whitney test (C). (D) RNA from macrophages from 4 donor mice per experiment was differentiated ex vivo in the absence or presence of olaparib (quadruplicates for each group) and subjected to RNA-seq analysis. Pathway analysis of 500 top up-regulated genes. Green bars represent up-regulated genes in Hallmark pathways, and blue bars represent up-regulated genes in Reactome pathways. (E) Polar metabolites were extracted from BMDMs from 4 donor mice per experiment, cultured as indicated, and analyzed by mass spectrometry. Data presented as mean ± SD of 4 replicates, and comparisons were made using an unpaired t test. (F and G) Glycolytic flux is up-regulated and oxygen consumption down-regulated in olaparib-treated macrophages. BMDMs were harvested and 1 × 105/well seeded in a Seahorse assay 24-well plate and cultured overnight. Extracellular acidification (F) and oxygen consumption rates (G) were determined using a Seahorse analyzer (each time point represents the mean ± SD of 5–6 cell culture replicates of bone marrow from 3 donor mice per experiment, representative of two individual experiments).

Figure 4.. PARP inhibition facilitates ROS production via RET and induces macrophage reprogramming independent of PARP1/2

(A) Acute increase of macrophage NAD levels in response to olaparib. BMDMs were treated with olaparib for 6 h prior to lysis for NAD determination (3 donor mice per experiment, data presented as mean ± SD of duplicate cultures, representative of two independent experiments). (B) ROS production is rotenone sensitive. BMDMs were harvested and suspended in ROS detection buffer in the presence of rotenone as indicated (20 min). Incubation with DCFCA for another 30 min, followed by flow cytometry. (C and D) BMDMs were treated with olaparib alone or together with trolox. (C) Cells were stained with DCFDA, and ROS levels were analyzed by flow cytometry. (D) Cells were co-cultured with K14-GFP cells for 24 h, and BLI assay was performed to evaluate phagocytosis of the BMDMs. (E–I) BMDMs from PARP1/2-null mice and wild-type controls were cultured as in Figure 3A. (E) NAD determination and (F and G) ROS determination. 3 donor mice per experiment in (A)–(G). Data presented as mean ± SD of triplicate cultures. (H and I) Phagocytosis assay (co-culture with K14-GFP cells followed by BLI assay). Data presented as mean ± SD of 5 (H) or 6 (I) replicates obtained from 3 donor mice per experiment, representative of a duplicate experiment. (J) BMDMs were cultured as in Figure 3A with DMSO, olaparib in the absence or presence of rotenone, or MitoQ prior to co-culture with K14-GFP cells as indicated, and analysis of phagocytosis with the BLI assay. Data represent 6 replicates obtained from 3 donor mice, presented as mean ± SD, representative of a duplicate experiment. For all data, significance was determined by the unpaired Student’s t test. *p

Figure 5.. PARP inhibition leads to a shift in nicotinamide metabolism including accumulation of methyl-nicotinamide (MNA), which can phenocopy olaparib’s effect on BMDMs

(A) K14-Cre BRCA1f/fp53f/f tumor-bearing mice were treated with olaparib for 10 days, with 4 tumor-bearing mice per group. Euthanasia 2 h after the last olaparib dose, tumors were immediately snap frozen, and frozen sections were prepared for in situ mass spectrometry and immunofluorescence staining. Hematoxylin and eosin (H&E), olaparib levels, and NAD precursors nicotinamide mononucleotide, nicotinic acid, and 1-MNA were analyzed on serial frozen sections by use of matrix-assisted laser desorption ionization mass spectrometry imaging (MALDI-MSI) on 5 μm frozen sections. Brightening color represents the increasing abundance of the measured parameter. F4/80 antibodies (red fluorescence) were used to stain macrophages (6th and 7th row). Red (F4/80) and white (MNA) outlined squares correspond to each other in these serial sections. Scale bar = 5mm. (B) Biosynthesis of MNA. NAM, nicotinamide; NMN, nicotinamide mononucleotide; NMNAT, nicotinamide mononucleotide adenylyl transferase; NAMPT, nicotinamide phosphoribosyltransferase; NNMT, nicotinamide N-methyltransferase. (C) MNA induces a prophagocytic phenotype. Mouse BMDMs were cultured as in Figure 3A with DMSO, olaparib, or MNA and co-cultured with K14-GFP cells, followed by phagocytosis via BLI assay (6 donor mice per experiment, 6 culture replicates per experiment, data represent the mean ± SD of a representative duplicate experiment). For all data, significance was determined by unpaired Student’s t test. *p

Figure 6.. Olaparib-treated macrophages increase macrophage recruitment and enhance CD8 T cell cytotoxicity in the TME

(A–G) Autologous, ex vivo-induced and olaparib-treated BMDMs were injected intratumorally on days 0 and 7. Mice (5 tumor-bearing mice/arm) were euthanized at day 10, and tumors were subjected to CyTOF. Data are presented as mean ± SD for the respective population. (A) CD45+ cell populations identified by CyTOF. (B–G) Quantitative analysis of subpopulations. (H and I) Ex vivo, PARP-inhibitor treatment of macrophages promotes the expansion of CD8+ T cells. Immortalized BMDMs (iBMDMs) were co-cultured with T cells from syngeneic wild-type (H; 5 donor mice) or OTI mice (I; 5 donor mice) in the presence or absence of olaparib for 72 h, and CellTrace was added during culture. Cells were stained with anti-CD3 (BV510) and subjected to flow cytometry (5 donor mice for WT or OT1 mice; data are presented as mean ± SD for the respective population). Significance was determined by unpaired Student’s t test. *p < 0.05, **p < 0.01, and ****p < 0.0001. (J and K) Syngeneic implantation of K14 tumors as in Figure 1A. Randomization to isotype control (n = 10) or anti-CD47 (n = 10) antibody as single agents or in combination with olaparib or MNA (50 mg/kg per day orally). Kaplan-Meier curves for MNA treatments with (n = 13) or without CD47 (n = 12) antibodies (J) or for olaparib (n = 22) with or without CD47 (n = 17) antibodies (K). Generation of Kaplan Meier curves and survival analysis with Gehan-Wilcoxon test. Median survival is shown in parentheses.