Deep Learning-based Modeling for Preclinical Drug Safety Assessment.

In drug development, assessing the toxicity of candidate compounds is crucial for successfully transitioning from preclinical research to early-stage clinical trials. Drug safety is typically assessed using animal models with a manual histopathological examination of tissue sections to characterize the dose-response relationship of the compound - a time-intensive process prone to inter-observer variability and predominantly involving tedious review of cases without abnormalities. Artificial intelligence (AI) methods in pathology hold promise to accelerate this assessment and enhance reproducibility and objectivity. Here, we introduce TRACE, a model designed for toxicologic liver histopathology assessment capable of tackling a range of diagnostic tasks across multiple scales, including situations where labeled data is limited. TRACE was trained on 15 million histopathology images extracted from 46,734 digitized tissue sections from 157 preclinical studies conducted on Rattus norvegicus. We show that TRACE can perform various downstream toxicology tasks spanning histopathological response assessment, lesion severity scoring, morphological retrieval, and automatic dose-response characterization. In an independent reader study, TRACE was evaluated alongside ten board-certified veterinary pathologists and achieved higher concordance with the consensus opinion than the average of the pathologists. Our study represents a substantial leap over existing computational models in toxicology by offering the first framework for accelerating and automating toxicological pathology assessment, promoting significant progress with faster, more consistent, and reliable diagnostic processes.

PARPi, BRCA, and gaps: controversies and future research.

In recent years, various poly(ADP-ribose) polymerase (PARP) inhibitors (PARPis) have been approved for the treatment of several cancers to target the vulnerability of homologous recombination (HR) deficiency (e.g., due to BRCA1/2 dysfunction). In this review we analyze the ongoing debates and recent breakthroughs in the use of PARPis for BRCA1/2-deficient cancers, juxtaposing the 'double-strand break (DSB)' and 'single-stranded DNA (ssDNA) gap' models of synthetic lethality induced by PARPis. We spotlight the complexity of this interaction, highlighting emerging research on the role of DNA polymerase theta (POLθ) and ssDNA gaps in shaping therapy responses. We scrutinize the clinical ramifications of these findings, especially concerning PARPi efficacy and resistance mechanisms, underscoring the heterogeneity of BRCA-mutated tumors and the urgent need for advanced research to bridge the gap between laboratory models and patient outcomes.

Establishment of Primary Cell Cultures from Canine Oral Melanomas via Fine-Needle Aspiration: A Novel Tool for Tumorigenesis and Cancer Progression Studies.

Oral melanomas are the most common oral malignancies in dogs and are characterized by an aggressive nature, invasiveness, and poor prognosis. With biological and genetic similarities to human oral melanomas, they serve as a valuable spontaneous comparative model. Primary cell cultures are widely used in human medicine and, more recently, in veterinary medicine to study tumorigenesis, cancer progression, and innovative therapeutic approaches. This study aims to establish two- and three-dimensional primary cell lines from oral canine melanomas using fine-needle aspiration as a minimally invasive sampling method. For this study, samples were collected from six dogs, represented by four primary oral melanomas and five lymph nodal metastases. The cells were digested to obtain single-cell suspensions, seeded in flasks, or processed with Matrigel® to form organoids. The cell cultures were characterized through flow cytometry using antibodies against Melan-A, PNL2, and Sox-10. This technique offers a minimally invasive means to obtain cell samples, particularly beneficial for patients that are ineligible for surgical procedures, and enables the establishment of in vitro models crucial for comparative studies in mucosal melanoma oncology. To the best of our knowledge, this is the first work establishing neoplastic primary cell cultures via fine-needle aspiration in dogs.

Host cell CRISPR genomics and modelling reveal shared metabolic vulnerabilities in the intracellular development of Plasmodium falciparum and related hemoparasites.

Parasitic diseases, particularly malaria (caused by Plasmodium falciparum) and theileriosis (caused by Theileria spp.), profoundly impact global health and the socioeconomic well-being of lower-income countries. Despite recent advances, identifying host metabolic proteins essential for these auxotrophic pathogens remains challenging. Here, we generate a novel metabolic model of human hepatocytes infected with P. falciparum and integrate it with a genome-wide CRISPR knockout screen targeting Theileria-infected cells to pinpoint shared vulnerabilities. We identify key host metabolic enzymes critical for the intracellular survival of both of these lethal hemoparasites. Remarkably, among the metabolic proteins identified by our synergistic approach, we find that host purine and heme biosynthetic enzymes are essential for the intracellular survival of P. falciparum and Theileria, while other host enzymes are only essential under certain metabolic conditions, highlighting P. falciparum's adaptability and ability to scavenge nutrients selectively. Unexpectedly, host porphyrins emerge as being essential for both parasites. The shared vulnerabilities open new avenues for developing more effective therapies against these debilitating diseases, with the potential for broader applicability in combating apicomplexan infections.

PatchSorter: a high throughput deep learning digital pathology tool for object labeling.

The discovery of patterns associated with diagnosis, prognosis, and therapy response in digital pathology images often requires intractable labeling of large quantities of histological objects. Here we release an open-source labeling tool, PatchSorter, which integrates deep learning with an intuitive web interface. Using >100,000 objects, we demonstrate a >7x improvement in labels per second over unaided labeling, with minimal impact on labeling accuracy, thus enabling high-throughput labeling of large datasets.

Tumour microenvironment characterisation to stratify patients for hyperthermic intraperitoneal chemotherapy in high-grade serous ovarian cancer (OVHIPEC-1).

BACKGROUND: Hyperthermic intraperitoneal chemotherapy (HIPEC) improves survival in patients with Stage III ovarian cancer following interval cytoreductive surgery (CRS). Optimising patient selection is essential to maximise treatment efficacy and avoid overtreatment. This study aimed to identify biomarkers that predict HIPEC benefit by analysing gene signatures and cellular composition of tumours from participants in the OVHIPEC-1 trial. METHODS: Whole-transcriptome RNA sequencing data were retrieved from high-grade serous ovarian cancer (HGSOC) samples from 147 patients obtained during interval CRS. We performed differential gene expression analysis and applied deconvolution methods to estimate cell-type proportions in bulk mRNA data, validated by histological assessment. We tested the interaction between treatment and potential predictors on progression-free survival using Cox proportional hazards models. RESULTS: While differential gene expression analysis did not yield any predictive biomarkers, the cellular composition, as characterised by deconvolution, indicated that the absence of macrophages and the presence of B cells in the tumour microenvironment are potential predictors of HIPEC benefit. The histological assessment confirmed the predictive value of macrophage absence. CONCLUSION: Immune cell composition, in particular macrophages absence, may predict response to HIPEC in HGSOC and these hypothesis-generating findings warrant further investigation. CLINICAL TRIAL REGISTRATION: NCT00426257.

H2AX promotes replication fork degradation and chemosensitivity in BRCA-deficient tumours.

Histone H2AX plays a key role in DNA damage signalling in the surrounding regions of DNA double-strand breaks (DSBs). In response to DNA damage, H2AX becomes phosphorylated on serine residue 139 (known as γH2AX), resulting in the recruitment of the DNA repair effectors 53BP1 and BRCA1. Here, by studying resistance to poly(ADP-ribose) polymerase (PARP) inhibitors in BRCA1/2-deficient mammary tumours, we identify a function for γH2AX in orchestrating drug-induced replication fork degradation. Mechanistically, γH2AX-driven replication fork degradation is elicited by suppressing CtIP-mediated fork protection. As a result, H2AX loss restores replication fork stability and increases chemoresistance in BRCA1/2-deficient tumour cells without restoring homology-directed DNA repair, as highlighted by the lack of DNA damage-induced RAD51 foci. Furthermore, in the attempt to discover acquired genetic vulnerabilities, we find that ATM but not ATR inhibition overcomes PARP inhibitor (PARPi) resistance in H2AX-deficient tumours by interfering with CtIP-mediated fork protection. In summary, our results demonstrate a role for H2AX in replication fork biology in BRCA-deficient tumours and establish a function of H2AX separable from its classical role in DNA damage signalling and DSB repair.

TurboID mapping reveals the exportome of secreted intrinsically disordered proteins in the transforming parasite Theileria annulata.

Theileria annulata is a tick-transmitted apicomplexan parasite that gained the unique ability among parasitic eukaryotes to transform its host cell, inducing a fatal cancer-like disease in cattle. Understanding the mechanistic interplay between the host cell and malignant Theileria species that drives this transformation requires the identification of responsible parasite effector proteins. In this study, we used TurboID-based proximity labeling, which unbiasedly identified secreted parasite proteins within host cell compartments. By fusing TurboID to nuclear export or localization signals, we biotinylated proteins in the vicinity of the ligase enzyme in the nucleus or cytoplasm of infected macrophages, followed by mass spectrometry analysis. Our approach revealed with high confidence nine nuclear and four cytosolic candidate parasite proteins within the host cell compartments, eight of which had no orthologs in non-transforming T. orientalis. Strikingly, all eight of these proteins are predicted to be highly intrinsically disordered proteins. We discovered a novel tandem arrayed protein family, nuclear intrinsically disordered proteins (NIDP) 1-4, featuring diverse functions predicted by conserved protein domains. Particularly, NIDP2 exhibited a biphasic host cell-cycle-dependent localization, interacting with the EB1/CD2AP/CLASP1 parasite membrane complex at the schizont surface and the tumor suppressor stromal antigen 2 (STAG2), a cohesion complex subunit, in the host nucleus. In addition to STAG2, numerous NIDP2-associated host nuclear proteins implicated in various cancers were identified, shedding light on the potential role of the T. annulata exported protein family NIDP in host cell transformation and cancer-related pathways.IMPORTANCETurboID proximity labeling was used to identify secreted proteins of Theileria annulata, an apicomplexan parasite responsible for a fatal, proliferative disorder in cattle that represents a significant socio-economic burden in North Africa, central Asia, and India. Our investigation has provided important insights into the unique host-parasite interaction, revealing secreted parasite proteins characterized by intrinsically disordered protein structures. Remarkably, these proteins are conspicuously absent in non-transforming Theileria species, strongly suggesting their central role in the transformative processes within host cells. Our study identified a novel tandem arrayed protein family, with nuclear intrinsically disordered protein 2 emerging as a central player interacting with established tumor genes. Significantly, this work represents the first unbiased screening for exported proteins in Theileria and contributes essential insights into the molecular intricacies behind the malignant transformation of immune cells.

Effective detection of BRAFV595E mutation in canine urothelial and prostate carcinomas using immunohistochemistry.

Canine urothelial carcinoma (UC) and prostate carcinoma (PC) frequently exhibit the BRAFV595E mutation, akin to the BRAFV600E mutation common in various human cancers. Since the initial discovery of the BRAF mutation in canine cancers in 2015, PCR has been the standard method for its detection in both liquid and tissue biopsies. Considering the similarity between the canine BRAFV595E and human BRAFV600E mutations, we hypothesized that immunohistochemistry (IHC) using a BRAFV600E-specific antibody could effectively identify the canine mutant BRAFV595E protein. We tested 122 canine UC (bladder n = 108, urethra n = 14), 21 PC, and benign tissue using IHC and performed digital droplet PCR (ddPCR) on all 122 UC and on 14 IHC positive PC cases. The results from ddPCR and IHC were concordant in 99% (135/136) of the tumours. Using IHC, BRAFV595E was detected in 72/122 (59%) UC and 14/21 (65%) PC. Staining of all benign bladder and prostate tissues was negative. If present, mutant BRAF staining was homogenous, with rare intratumour heterogeneity in three (4%) cases of UC. Additionally, the BRAFV595E mutation was more prevalent in tumours with urothelial morphology, and less common in glandular PC or UC with divergent differentiation. This study establishes that BRAFV600-specific IHC is a reliable and accurate method for detecting the mutant BRAFV595E protein in canine UC and PC. Moreover, the use of IHC, especially with tissue microarrays, provides a cost-efficient test for large-scale screening of canine cancers for the presence of BRAF mutations. This advancement paves the way for further research to define the prognostic and predictive role of this tumour marker in dogs and use IHC to stratify dogs for the treatment with BRAF inhibitors.

Persisting cancer cells are different from bacterial persisters.

The persistence of drug-sensitive tumors poses a significant challenge in cancer treatment. The concept of bacterial persisters, which are a subpopulation of bacteria that survive lethal antibiotic doses, is frequently used to compare to residual disease in cancer. Here, we explore drug tolerance of cancer cells and bacteria. We highlight the fact that bacteria, in contrast to cancer cells, have been selected for survival at the population level and may therefore possess contingency mechanisms that cancer cells lack. The precise mechanisms of drug-tolerant cancer cells and bacterial persisters are still being investigated. Undoubtedly, by understanding common features as well as differences, we, in the cancer field, can learn from microbiology to find strategies to eradicate persisting cancer cells.

PARG-deficient tumor cells have an increased dependence on EXO1/FEN1-mediated DNA repair.

Targeting poly(ADP-ribose) glycohydrolase (PARG) is currently explored as a therapeutic approach to treat various cancer types, but we have a poor understanding of the specific genetic vulnerabilities that would make cancer cells susceptible to such a tailored therapy. Moreover, the identification of such vulnerabilities is of interest for targeting BRCA2;p53-deficient tumors that have acquired resistance to poly(ADP-ribose) polymerase inhibitors (PARPi) through loss of PARG expression. Here, by performing whole-genome CRISPR/Cas9 drop-out screens, we identify various genes involved in DNA repair to be essential for the survival of PARG;BRCA2;p53-deficient cells. In particular, our findings reveal EXO1 and FEN1 as major synthetic lethal interactors of PARG loss. We provide evidence for compromised replication fork progression, DNA single-strand break repair, and Okazaki fragment processing in PARG;BRCA2;p53-deficient cells, alterations that exacerbate the effects of EXO1/FEN1 inhibition and become lethal in this context. Since this sensitivity is dependent on BRCA2 defects, we propose to target EXO1/FEN1 in PARPi-resistant tumors that have lost PARG activity. Moreover, EXO1/FEN1 targeting may be a useful strategy for enhancing the effect of PARG inhibitors in homologous recombination-deficient tumors.

DNA repair, gap suppression, or fork protection: BRCA2 needs a break!

In this issue of Molecular Cell, Lim et al.1 reveal new insights into the distinct roles of BRCA2 in coping with DNA breaks, highlighting homologous recombination as the pivotal function that affects tumorigenesis and therapy response.

Profiling the heterogeneity of colorectal cancer consensus molecular subtypes using spatial transcriptomics.

The consensus molecular subtypes (CMS) of colorectal cancer (CRC) is the most widely-used gene expression-based classification and has contributed to a better understanding of disease heterogeneity and prognosis. Nevertheless, CMS intratumoral heterogeneity restricts its clinical application, stressing the necessity of further characterizing the composition and architecture of CRC. Here, we used Spatial Transcriptomics (ST) in combination with single-cell RNA sequencing (scRNA-seq) to decipher the spatially resolved cellular and molecular composition of CRC. In addition to mapping the intratumoral heterogeneity of CMS and their microenvironment, we identified cell communication events in the tumor-stroma interface of CMS2 carcinomas. This includes tumor growth-inhibiting as well as -activating signals, such as the potential regulation of the ETV4 transcriptional activity by DCN or the PLAU-PLAUR ligand-receptor interaction. Our study illustrates the potential of ST to resolve CRC molecular heterogeneity and thereby help advance personalized therapy.

DNA Damage Response and Mismatch Repair Gene Defects in Advanced and Metastatic Prostate Cancer.

Alterations in DNA damage response (DDR) and related genes are present in up to 25% of advanced prostate cancers (PCa). Most frequently altered genes are involved in the homologous recombination repair, the Fanconi anemia, and the mismatch repair pathways, and their deficiencies lead to a highly heterogeneous spectrum of DDR-deficient phenotypes. More than half of these alterations concern non- BRCA DDR genes. From a therapeutic perspective, poly-ADP-ribose polymerase inhibitors have demonstrated robust clinical efficacy in tumors with BRCA2 and BRCA1 alterations. Mismatch repair-deficient PCa, and a subset of CDK12-deficient PCa, are vulnerable to immune checkpoint inhibitors. Emerging data point to the efficacy of ATR inhibitors in PCa with ATM deficiencies. Still, therapeutic implications are insufficiently clarified for most of the non- BRCA DDR alterations, and no successful targeted treatment options have been established.

Canine follicular cell and medullary thyroid carcinomas: Immunohistochemical characterization.

Research on modulation of iodine uptake by thyroid cells could help improve radioiodine treatment of dogs with thyroid tumors. The aim of this study was to characterize the immunohistochemical expression of thyroid transcription factor-1 (TTF-1), thyroglobulin, thyrotropin receptor (TSHR), sodium iodide symporter (NIS), pendrin, thyroid peroxidase (TPO), vimentin, and Ki-67 in follicular cell thyroid carcinomas (FTCs) and medullary thyroid carcinomas (MTCs), and to compare protein expression between FTC causing hyperthyroidism and FTC of euthyroid dogs. Immunohistochemistry was performed in 25 FTCs (9 follicular, 8 follicular-compact, and 8 compact) and 8 MTCs. FTCs and MTCs were positive for TTF-1, and expression was higher in FTCs of euthyroid dogs compared with FTCs of hyperthyroid dogs (P= .041). Immunolabeling for thyroglobulin was higher in follicular and follicular-compact FTCs compared with compact FTCs (P = .001), while vimentin expression was higher in follicular-compact FTCs compared with follicular FTCs (P = .011). The expression of TSHR, NIS, pendrin, and TPO was not significantly different among the different subtypes of FTCs or between FTCs causing hyperthyroidism and FTCs in euthyroid dogs. TSHR, NIS, pendrin, and TPO were also expressed in MTCs. Ki-67 labeling index was comparable between FTCs and MTCs, and between FTCs causing hyperthyroidism and FTCs in euthyroid dogs. Proteins of iodine transport were also expressed in canine MTCs, which could have implications for diagnosis and treatment. The different expression of thyroglobulin and vimentin between FTC histological subtypes could reflect variations in tumor differentiation.

The prognostic value of tumor-stroma ratio and a newly developed computer-aided quantitative analysis of routine H&E slides in high-grade serous ovarian cancer.

Introduction: Tumor-stroma ratio (TSR) is prognostic in multiple cancers, while its role in high-grade serous ovarian cancer (HGSOC) remains unclear. Despite the prognostic insight gained from genetic profiles and tumor-infiltrating lymphocytes (TILs), the prognostic use of histology slides remains limited, while it enables the identification of tumor characteristics via computational pathology reducing scoring time and costs. To address this, this study aimed to assess TSR's prognostic role in HGSOC and its association with TILs. We additionally developed an algorithm, Ovarian-TSR (OTSR), using deep learning for TSR scoring, comparing it to manual scoring. Methods: 340 patients with advanced-stage who underwent primary debulking surgery (PDS) or neo-adjuvant chemotherapy (NACT) with interval debulking (IDS). TSR was assessed in both the most invasive (MI) and whole tumor (WT) regions through manual scoring by pathologists and quantification using OTSR. Patients were categorized as stroma-rich (≥ 50% stroma) or stroma-poor (< 50%). TILs were evaluated via immunohistochemical staining. Results: In PDS, stroma-rich tumors were significantly associated with a more frequent papillary growth pattern (60% vs 34%), while In NACT stroma-rich tumors had a lower Tumor Regression Grading (TRG 4&5, 21% vs 57%) and increased pleural metastasis (25% vs 16%). Stroma-rich patients had significantly shorter overall and progression-free survival compared to stroma-poor (31 versus 45 months; P < 0.0001, and 15 versus 17 months; P = 0.0008, respectively). Combining stromal percentage and TILs led to three distinct survival groups with good (stroma-poor, high TIL), medium (stroma-rich, high TIL, or; stroma-poor, Low TIL), and poor(stroma-rich, low TIL) survival. These survival groups remained significant in CD8 and CD103 in multivariable analysis (Hazard ratio (HR) = 1.42, 95% Confidence-interval (CI) = 1.02-1.99; HR = 1.49, 95% CI = 1.01-2.18, and HR = 1.48, 95% CI = 1.05-2.08; HR = 2.24, 95% CI = 1.55-3.23, respectively). OTSR was able to recapitulate these results and demonstrated high concordance with expert pathologists (correlation = 0.83). Conclusions: TSR is an independent prognostic factor for survival assessment in HGSOC. Stroma-rich tumors have a worse prognosis and, in the case of NACT, a higher likelihood of pleural metastasis. OTSR provides a cost and time-efficient way of determining TSR with high reproducibility and reduced inter-observer variability.

CRISPR/Cas9-Mediated Targeting of BPV-1-Transformed Primary Equine Sarcoid Fibroblasts.

Equine sarcoids (EqS) are fibroblast-derived skin tumors associated with bovine papillomavirus 1 and 2 (BPV-1 and -2). Based on Southern blotting, the BPV-1 genome was not found to be integrated in the host cell genome, suggesting that EqS pathogenesis does not result from insertional mutagenesis. Hence, CRISPR/Cas9 implies an interesting tool for selectively targeting BPV-1 episomes or genetically anchored suspected host factors. To address this in a proof-of-concept study, we confirmed the exclusive episomal persistence of BPV-1 in EqS using targeted locus amplification (TLA). To investigate the CRISPR/Cas9-mediated editing of BPV-1 episomes, primary equine fibroblast cultures were established and characterized. In the EqS fibroblast cultures, CRISPR-mediated targeting of the episomal E5 and E6 oncogenes as well as the BPV-1 long control region was successful and resulted in a pronounced reduction of the BPV-1 load. Moreover, the deletion of the equine Vimentin (VIM), which is highly expressed in EqS, considerably decreased the number of BPV-1 episomes. Our results suggest CRISPR/Cas9-based gene targeting may serve as a tool to help further unravel the biology of EqS pathogenesis.

Artificial Intelligence to Predict the BRAF V595E Mutation in Canine Urinary Bladder Urothelial Carcinomas.

In dogs, the BRAF mutation (V595E) is common in bladder and prostate cancer and represents a specific diagnostic marker. Recent advantages in artificial intelligence (AI) offer new opportunities in the field of tumour marker detection. While AI histology studies have been conducted in humans to detect BRAF mutation in cancer, comparable studies in animals are lacking. In this study, we used commercially available AI histology software to predict BRAF mutation in whole slide images (WSI) of bladder urothelial carcinomas (UC) stained with haematoxylin and eosin (HE), based on a training (n = 81) and a validation set (n = 96). Among 96 WSI, 57 showed identical PCR and AI-based BRAF predictions, resulting in a sensitivity of 58% and a specificity of 63%. The sensitivity increased substantially to 89% when excluding small or poor-quality tissue sections. Test reliability depended on tumour differentiation (p < 0.01), presence of inflammation (p < 0.01), slide quality (p < 0.02) and sample size (p < 0.02). Based on a small subset of cases with available adjacent non-neoplastic urothelium, AI was able to distinguish malignant from benign epithelium. This is the first study to demonstrate the use of AI histology to predict BRAF mutation status in canine UC. Despite certain limitations, the results highlight the potential of AI in predicting molecular alterations in routine tissue sections.

MND1 and PSMC3IP control PARP inhibitor sensitivity in mitotic cells.

The PSMC3IP-MND1 heterodimer promotes meiotic D loop formation before DNA strand exchange. In genome-scale CRISPR-Cas9 mutagenesis and interference screens in mitotic cells, depletion of PSMC3IP or MND1 causes sensitivity to poly (ADP-Ribose) polymerase inhibitors (PARPi) used in cancer treatment. PSMC3IP or MND1 depletion also causes ionizing radiation sensitivity. These effects are independent of PSMC3IP/MND1's role in mitotic alternative lengthening of telomeres. PSMC3IP- or MND1-depleted cells accumulate toxic RAD51 foci in response to DNA damage, show impaired homology-directed DNA repair, and become PARPi sensitive, even in cells lacking both BRCA1 and TP53BP1. Epistasis between PSMC3IP-MND1 and BRCA1/BRCA2 defects suggest that abrogated D loop formation is the cause of PARPi sensitivity. Wild-type PSMC3IP reverses PARPi sensitivity, whereas a PSMC3IP p.Glu201del mutant associated with D loop defects and ovarian dysgenesis does not. These observations suggest that meiotic proteins such as MND1 and PSMC3IP have a greater role in mitotic DNA repair.

Multi-omics analysis reveals distinct non-reversion mechanisms of PARPi resistance in BRCA1- versus BRCA2-deficient mammary tumors.

BRCA1 and BRCA2 both function in DNA double-strand break repair by homologous recombination (HR). Due to their HR defect, BRCA1/2-deficient cancers are sensitive to poly(ADP-ribose) polymerase inhibitors (PARPis), but they eventually acquire resistance. Preclinical studies yielded several PARPi resistance mechanisms that do not involve BRCA1/2 reactivation, but their relevance in the clinic remains elusive. To investigate which BRCA1/2-independent mechanisms drive spontaneous resistance in vivo, we combine molecular profiling with functional analysis of HR of matched PARPi-naive and PARPi-resistant mouse mammary tumors harboring large intragenic deletions that prevent reactivation of BRCA1/2. We observe restoration of HR in 62% of PARPi-resistant BRCA1-deficient tumors but none in the PARPi-resistant BRCA2-deficient tumors. Moreover, we find that 53BP1 loss is the prevalent resistance mechanism in HR-proficient BRCA1-deficient tumors, whereas resistance in BRCA2-deficient tumors is mainly induced by PARG loss. Furthermore, combined multi-omics analysis identifies additional genes and pathways potentially involved in modulating PARPi response.