Multidrug resistance (MDR) is a common and leading cause of treatment failure and mortality in cancer patient. While numerous biological processes contribute to drug resistance, recent studies have revealed that most anticancer drugs rapidly bind to and damage newly synthesized proteins upon entering cells. This process, termed acute drug protein damage (ADPD), occurs before the drugs act on their canonical targets. To evade the lethal effects of ADPD, cancer cells rapidly initiate a series of protective responses collectively termed the protein damage response (PDR). This cascade includes damage recognition via protein ubiquitination, damage clearance through the proteasome system, and subsequent mitophagy to remove damaged mitochondria caused by the co-import of drugs and damaged proteins. Here, we review the current understanding of multiple biological processes underlying drug resistance, with a focus on the mechanisms of ADPD induced by anticancer drugs, the pivotal role of PDR in driving MDR, and its potential applications in predicting and overcoming drug resistance.

Rationale:The preservation of the tumor immune microenvironment (TIME) ex vivo is essential for investigating tumor-immune interactions and developing effective immunotherapies. However, current culture models often fail to maintain autologous immune cells or support high-throughput testing. To overcome these limitations, we establish and validate a novel miniaturized-tumor culture (MTC) platform. Methods:To build the MTC platform, we took tumor tissues from both mouse models and human patients. We processed these tissues into 100-500 μm fragments. Supplementation with IL-2 and IL-7 in culture maintained long-term intra-tumoral T cell survival. We then evaluated whether the platform could maintain autologous lymphoid populations and respond to immune checkpoint blockade (ICB), through co-culture assays with peripheral blood mononuclear cells (PBMCs) or splenocytes. The system accurately recapitulates drug response and resistance in both immunocompetent and immunodeficient models. Finally, we screened a high-throughput drug library to identify agents that re-sensitize tumors to anti-PD-L1 therapy. To figure out the underlying mechanisms, we used bulk RNA-sequencing, flow cytometry, and targeted CXCL13 antibody neutralization. Results:The MTC platform demonstrated better preservation of autologous lymphoid populations and remained responsive to ICB. Co-culture assays revealed enhanced immune cell infiltration upon ICB treatment. Furthermore, the MTC models recapitulated drug response and resistance phenotypes. Our screen identified axitinib (AXI) as the most potent agent for re-sensitizing tumors to anti-PD-L1 therapy across different cancer types. Mechanistically, AXI potentiates antigen presentation in tumor and dendritic cells. It also enhances cytotoxic T-cell function via the upregulation of CXCL13. Finally, CXCL13 blockade effectively abrogated AXI-induced T cell recruitment and tumor regression. Conclusions:The MTC platform serves as a high-fidelity, high-throughput tool for modeling the TIME. It provides a valuable framework for developing novel anti-cancer strategies and elucidating their underlying mechanisms of action.

ABSTRACT Drugs induce coordinated phenotypic changes across multiple modalities, including transcriptional reprogramming and cellular morphological remodeling. Predicting these drug-induced modality changes is central to drug discovery, mechanism-of-action studies and precision therapeutics, however, prediction performance depends critically on how both drug compounds and cellular states are represented. Despite rapid advances in drug molecular and gene representation methods, a systematic evaluation of these methods remains lacking. Herein, we introduce MVCBench, a comprehensive benchmarking framework for evaluating drug molecular and gene representation methods in predicting drug-induced multimodal virtual cell (MVC) phenotypes. MVCBench leverages large-scale transcriptomic and high-content imaging data and systematically evaluates 24 representation methods (12 drug molecular and 12 gene representation methods) across nearly 1.1 million drug-induced profiles, under both in-distribution and out-of-distribution settings spanning unseen compounds, cell lines, assay plates and datasets. Our benchmarking reveals a pronounced modality-dependent asymmetry: advanced drug molecular representations substantially improve the prediction of drug-induced morphological phenotypes but provide only limited gains for gene expression prediction relative to classical fingerprints, whereas task-specific gene representations outperform general-purpose foundation models in predicting drug-induced transcriptomic responses. Predictive performance also deteriorates sharply under distribution shift, highlighting persistent challenges in cross-dataset and cross-platform generalization. We further show that integrating transcriptomic and morphological modalities consistently improves prediction accuracy, and derive practical design principles for MVC architectures, including modality-aware loss calibration and fusion strategies. Together, MVCBench provides a systematic foundation for evaluating representation methods and offers guidance for developing robust MVC models of drug-induced cellular responses.

ABSTRACT Triple‐negative breast cancer (TNBC) is the most aggressive breast cancer subtype with high metastatic potential and limited treatment options. Natural killer (NK) cells represent a promising immunotherapy strategy due to their innate tumor‐killing capacity, but their efficacy against TNBC remains limited. We found that TNBC cells, particularly the mesenchymal‐like subtype, exhibited greater resistance to NK cells compared to non‐TNBC cells. Mechanistic studies indicate that TNBC cells’ survival in response to NK cells occurs in three phases. First, within 1 h of NK cell co‐culture, TNBC cells accumulate reactive oxygen species (ROS), which upregulate C‐X‐C motif chemokine ligand 1 (CXCL1) and interleukin 8 (IL8) expression in an NF‐κB− and ERK/JNK−AP‐1‐dependent manner. Second, secreted CXCL1/IL8 binds to C‐X‐C motif chemokine receptor 1/2 (CXCR1/2), activating the AKT−BCL‐2 pathway to enhance cancer cell survival and suppress NK cell function by downregulating NKG2D, TRAIL, and IFN‐γ expression. Third, CXCL1/IL8−CXCR1/2 autocrine loop further amplifies their own synthesis and induces programmed cell death 1 ligand 1 (PD‐L1) expression via NF‐κB and ERK/JNK−AP‐1 pathways. High CXCL1/IL8 expression correlates with reduced NK cell infiltration and shorter distant‐metastasis‐free survival in breast cancer patients. Combinatorial application of CXCR1/2 inhibitor with anti‐PD‐L1 antibody can overcome NK cell dysfunction and reduce TNBC metastasis.

BACKGROUND:Insufficient infiltration of CD8+ T cells in the tumor microenvironment (TME) critically restricts antitumor immunity and cancer immunotherapy efficacy. The purpose of this study was to identify novel tumor cell-intrinsic regulators of T-cell infiltration and to elucidate their mechanisms of action. METHODS:We performed a genome-wide Sleeping Beauty transposon mutagenesis screen in murine breast cancer models. Protein-protein interactions were identified by mass spectrometry and validated by co-immunoprecipitation. Gene and protein expression levels were assessed by reverse transcription and quantitative PCR and western blotting. T-cell infiltration and function were evaluated using flow cytometry, immunohistochemistry (IHC), multiplex IHC, and by analyzing bulk and single-cell RNA sequencing data complemented by bioinformatic analysis. The specific dephosphorylation sites on LGALS1 were confirmed through phosphomimetic mutant experiments. T-cell infiltration was further validated using an in vitro T-cell transendothelial migration assay and in vivo mouse models. RESULTS:Our screening identified 39 candidate genes, with tumor cell-intrinsic dual-specificity phosphatase 22 (DUSP22) expression correlating with enhanced CD8+ T-cell accumulation and suppressed tumor progression. Overexpression of DUSP22 resulted in increased CD8+ T-cell infiltration and enhanced T-cell function. Mechanistically, DUSP22 binds to LGALS1 and dephosphorylates it at the Ser8 and Thr58 residues, leading to LGALS1 degradation and subsequent alleviation of LGALS1-mediated immunosuppression. In human breast cancer samples, LGALS1 expression was negatively correlated with both DUSP22 levels and CD8+ T-cell infiltration. Therapeutic targeting of the DUSP22-LGALS1 axis significantly enhanced CD8+ T-cell infiltration and synergized with anti-programmed cell death protein-1 therapy to boost antitumor responses. CONCLUSIONS:Our findings unveil a novel phosphorylation-dependent DUSP22-LGALS1 axis that reprograms the immunosuppressive TME. This work thus proposes a promising therapeutic strategy to overcome immune checkpoint blockade resistance in breast cancer.

Proteasome inhibitors show limited efficacy against solid tumors. We previously show the combination of Bortezomib with ammonium tetrathiomolybdate (TM) or AMD3100 could inhibit breast cancer growth rely on intact immune system. However, it remains to be explored the broad application of these drug combinations in cancer treatment. We show drug combinations inhibit the growth of multiple tumor models but fails against B16F10. We identify GSDME-mediated pyroptosis triggered by drugs as the critical switch: GSDME mediated pyroptosis drives CCL5 release, recruiting dendritic cells (DCs) and CD8⁺ T cells to initiate adaptive immunity. B16F10 cells express minimal GSDME and therefore undergo apoptosis to retain CCL5 intracellular and abrogating antitumor immunity. Overexpression of GSDME in B16F10 restores pyroptosis, CCL5 secretion, and treatment sensitivity, while GSDME loss abolishes efficacy in responsive models. Notably, pyroptosis in GSDME-proficient subsets propagates systemic immunity, controlling distant GSDME-deficient lesions. In human cancers, GSDME expression correlates with CD8⁺ T cell and DC infiltration. High GSDME predicts superior survival in BTZ-treated multiple myeloma patients. These findings establish GSDME mediated pyroptosis as the primary route for CCL5 secretion, explaining differential efficacy across tumor models and suggesting patients with high GSDME expression may benefits more from proteasome inhibitor-based therapies.

Radiotherapy (RT) is a cornerstone of cancer treatment; however, its efficacy is frequently hampered by its adverse effects on normal tissues. By studying the effects of high-dose radiotherapy (HDRT) and low-dose radiotherapy (LDRT), we found that cancer cells adapt distinct responses to these doses to reduce cytotoxicity. Upon HDRT, cancThese authors contributed equally to this worker cells initiate a strong DNA damage response (DDR) to gain resistance through rapid production and/or activation of proteins for cell cycle arrest and DNA damage repair. In contrast, LDRT has a milder effect on the DDR and promotes resistance by triggering the synthesis of new proteins, including those essential for DNA repair and protein damage clearance. We showed that the inhibition of proteasome activity using a proteasome inhibitor (PI) result in the accumulation of damage to both proteins and DNA, leading to the profound death of cancer cells. Mechanistically, LDRT enhances protein synthesis through both increased mTOR signaling and 80S ribosome assembly. On the basis of these findings, we designed a chemoradiotherapy strategy that combines LDTR with PI to treat cancer while minimizing non-targeted toxicity.

Poly(ADP-ribose) polymerase 1 (PARP1) inhibition represents a promising targeted therapy for BRCA-deficient cancer patients based on the synthetic lethality theory. Recent evidence shows that the efficacy of DNA damage drugs depends on two aspects: DNA repair signaling and immune response. Applying a functional proteomics approach, we find that the function of the spliceosome is perturbed by PARP inhibitors via enhancing interaction between PARP1 and SF3B1, a key factor of the spliceosome. We demonstrate that differential alternative spliced mRNA and accumulation of double-stranded RNA (dsRNA) are induced by perturbation of the spliceosome upon PARP inhibitor treatment, resulting in triggering dsRNA antiviral mimicry innate immune response. Moreover, we identify a novel function of BRCA1, through which BRCA1 regulates innate immune response, leading to compromising of the innate immune signaling by downregulation of IRF3 in BRCA1-deficient breast cancer cells, which reduces the sensitivity to PARP inhibitors and causes intrinsic resistance. Polyinosinic-polycytidylic acid (poly(I:C)) is a dsRNA synthetic analog sensitizing PARP inhibitors through further triggering dsRNA signaling. Finally, we show that the combination of PARP inhibitors and poly(I:C) enhances anti-tumor efficiency in vivo. Overall, our study reveals that BRCA1 deficiency impedes tumor cell intrinsic innate immune response, inducing intrinsic resistance to PARP inhibitors that can be overcome when poly(I:C) is combined.

Chemotherapy-induced drug resistance is a very complex process. About 20% of cancer patients have received effective therapy with platinum (Pt) medicines, and drug resistance has since emerged through several mechanisms. To investigate the complicated heterotypic interactions between cancer cells and immune cells during the resistant tumor outgrowth processes under Pt treatment, we dissected the crosstalk signals with single-cell RNA sequencing and performed functional studies on Pt-induced drug resistance using overexpression, knockout, and DOX-inducible technologies. In this study, we have identified that: 1) S100a6-Lars2 oncogenic signaling could promote TGF-β production from tumor and immune cells and induce expressions of PD1, Tim3, and Selplg on T cells, which could contribute to resistant tumor outgrowth under Pt treatment; 2) Pt drugs could promote the survival of cancer stem cells (CSCs) with elevated hyaluronan-mediated motility receptor (Hmmr) and mTOR oncogenic pathway; 3) we validated Lars2 oncogenic action in resistant tumor outgrowth in FVB immunocompetent mice and the dynamic changes of the immune cell population; 4) combinatory treatment involving the inhibition of TGF-β signaling, the mTOR pathway, and Hmmr action, along with PD1 antibody, shows a significant effect in inhibiting tumor growth. This study advances the understanding of the mechanism of Pt drug-induced resistant tumor outgrowth, which may allow us to better manage the tumor immune microenvironment (TIME) and postpone the onset of resistance. Yuqing Wang, Chuxia Deng. S100a6-Lars2 oncogenic signaling shaping immunosuppressive environment contributes to platinum-based resistance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2170.

Supplementary Table S2. PCR primers of the 10 genes in the subnetwork for mouse model.

Breast cancer initiation and progression are driven by various oncogenic factors and their effects on the surrounding microenvironments. Through integrative analysis of ChIP-sequencing and RNA-sequencing with fast proliferating mammary epithelial cells from pregnant Brca1MKO and wild type (WT) mice, we found that elevated Smyd3-Shcbp1 signaling is featured with activation of the Ras-MAPK pathway and increased transcription activity in both premalignant mammary epithelium and tumor cells. Smyd3-Shcbp1 signaling shapes the tumor immunosuppressive microenvironment (TIME) and is associated with immune therapy resistance to PD1 antibody treatment. Trametinib, a potent inhibitor of MEK/MAPK, could reverse the expression of Smyd3 and Shcbp1 in both Brca1 mutant and WT tumor bearing mice. We further demonstrated that the combinatory treatment of trametinib together with PD1 antibody enhances the function of effector T cells, sensitizing tumors with elevated Smyd3 and Shcbp1 signaling to αPD1 treatment. This study advances the understanding of breast tumor progression and provides a new selective strategy for breast cancer patients.

Supplementary Table S12. MREs between mRNAs and miRNAs in the survival-related ceRNA network.

Supplementary Figure S6. Construction of subnetwork of prognosis-related ceRNA network.

Supplementary Table S8. Annotation for the biotypes of DE transcriptome between low and high TIILs subtypes.

BRCA1‑deficient triple‑negative breast cancer (TNBC) presents significant treatment challenges owing to the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) targets, exhibits marked molecular heterogeneity that precludes the application of effective targeted therapies, and harbors a highly immunosuppressive tumor microenvironment. Here, we used the Brca1co/co MMTV‑Cre mouse model that recapitulates human BRCA1‑mutant TNBC, characterized by early dominance of CD11b⁺Gr‑1⁻F4/80Low blood‑derived macrophages and subsequent enrichment of F4/80High tissue macrophages within adipose‑rich mammary glands. PD‑1 blockade with anti‑mPD‑1 monoclonal antibodies (mAb) significantly delayed primary tumor progression, reduced proliferation marker levels (PCNA, Ki‑67), enhanced apoptosis (as indicated by increased cleaved PARP levels), and selectively impaired PI3K/AKT signaling. In a post‑resection setting, anti-mPD-1 treatment extended recurrence‑free survival rates, with elevated CD4, CD8α, and cleaved PARP levels observed in recurrent tumors. Mice with the longest relapse‑free intervals exhibited the strongest T cell marker expression. A combination of focal 20 Gy irradiation and PD-1 blockade exerted a potent synergistic effect. Specifically, irradiation reduced extracellular matrix deposition and enhanced tumor cell apoptosis (evidenced by increased cleaved caspase-3 and cytosolic PCNA) while PD-1 blockade stimulated robust inflammatory responses, in particular, expansion of CD8α⁺ T cell infiltration. These mechanistic insights align with clinical strategies for TNBC that integrate DNA damaging agents and immunotherapy and validate this model as an optimal in vivo platform for preclinical evaluation of novel treatment modalities for BRCA1‑associated breast cancer.

Inherited mutation in breast cancer susceptibility gene 1 (BRCA1) is strongly associated with mammary tumors that exhibit triple-negative characteristics, are insensitive to endocrine-targeted therapies, and show basal-like properties, including aggressive phenotypes [1, 2]. It has been reported that the average cumulative risk of breast cancer for BRCA1 mutation carriers by age 70 years is 57% (95% confidence interval [CI]: 47%-66%) [3]. Despite the high incidence and aggressive characteristics of BRCA1-associated breast cancer, few substantial improvements in preventing or treating this cancer have been made, largely due to the challenges of clinic-based cohort studies. During malignant transformation, cancer progression is facilitated by metabolic reprogramming–one of the hallmark characteristics of cancer. Previously, we found that inhibition of AKT is a potential strategy for the prevention and therapeutic management of Brca1-mutant mammary tumors. However, pharmacological inhibition proved less effective and less safe compared to genetic perturbation, limiting its potential for clinical application [4]. Meanwhile, mTOR, a key regulator of metabolism and a downstream target of the PI3K/AKT signaling pathway, has emerged as a promising therapeutic target for several diseases, including treatment of cancer [5]. In addition to identifying the contribution of mTOR signaling to BRCA1-deficient cells (Supplementary Figure S1), we provide genetic and pharmacological evidence using multi-orthogonal preclinical models [6-8] that mTOR is closely involved in the development and growth of Brca1-mutated mammary tumors (Figure 1A). To investigate the role of mTOR in the absence of BRCA1, we assessed the development of mammary glands in post-pubertal Brca1/Mtor-mutant mice by examining ductal and lobular development of the fourth mammary gland. Measurements of mammary gland density using the Branch software (ver. 1.1 [9]) showed that ductal length and branching were significantly diminished in the mammary glands of Brca1co/coMtorco/coMMTV-Cre mice (Figure 1B,C, Supplementary Figure S2). To determine whether mTOR contributes to BRCA1-deficient mammary tumor formation, we examined tumor formation in cohorts of Brca1co/co (n = 28), Brca1co/coMtorco/co (n = 30), Brca1co/coMMTV-Cre (n = 24), and Brca1co/coMtorco/coMMTV-Cre (n = 29) mice (Top left of Figure 1A). Brca1co/co and Brca1co/coMtorco/co mice showed no signs of mammary abnormalities, including tumors, up to 24 months of age. In contrast, Brca1co/coMMTV-Cre mutant mice developed breast cancer, reaching a high incidence (37.5%; 9/24) by 24 months of age. During the same period, Brca1co/coMtorco/coMMTV-Cre mice exhibited a lower incidence of breast cancer (6.9%; 2/29) and significantly better tumor-free survival compared to Brca1co/coMMTV-Cre mice (P = 0.008, log-rank test) (Figure 1D). Next, we examined whether mTOR inhibition using a clinically applicable pharmacological approach would produce similar effects as genetic ablation. To test pharmacological inhibition of mTOR, we administered everolimus (20 mg/kg, oral, 5 times/week) or vehicle to 4-month-old Brca1co/coMMTV-Cre mice for 11 months (lower left of Figure 1A, Supplementary Figure S3). During this period, Brca1co/coMMTV-Cre mice in both groups spontaneously developed palpable mammary tumors. At the end of the study period (15 months of age), vehicle-treated Brca1co/coMMTV-Cre mice showed a high incidence of mammary tumors (93%; 13 of 14). During the same period, everolimus-treated Brca1co/coMMTV-Cre mice exhibited a breast cancer incidence of 46% (5 of 11) and significantly longer tumor-free survival compared to their vehicle-treated counterparts (P = 0.0117, log-rank test) (Figure 1E). Moreover, while multiple tumors were found in 2 of 14 (14%) vehicle-treated mice, no cases of multiple tumors were detected in everolimus-treated mice (Figure 1F). In addition to tumor formation, whole-mount analysis of non-tumor–bearing mammary glands revealed that everolimus treatment reduced total ductal length and branch number by approximately 30% compared to vehicle treatment (Supplementary Figure S3). Notably, everolimus treatment significantly reduced the formation of abnormal hyperplastic foci (0.6 vs. 5.5 foci/mammary gland; P < 0.01) in non-tumor–bearing mammary glands of Brca1-mutant mice (Figure 1G,H). Taken together, these results suggest that genetic ablation and pharmacological inhibition of mTOR signaling prevents the proliferation of mammary epithelial cells and reduce tumor formation in Brca1-mutant mice. Contribution of mTOR in the initiation and progression of BRCA1-associated mammary tumors. (A) Overview of the experimental designs to investigate the attenuation of tumor initiation and progression by mTOR inhibition in Brca1-mutant mice. (B) Representative whole-mount staining of mammary glands from 2-month-old mice with the indicated genotypes. Arrows indicate the presence of sprouting ducts in the mammary glands. Scale bar, 5 mm. (C) Total length (light gray) and branch numbers (dark gray) of ducts between the lymph node and the end tip in the 4th mammary glands of 2-month-old Brca1co/co (n = 9), Brca1co/coMtorco/co (n = 16), Brca1co/coMMTV-Cre (n = 18), and Brca1co/coMtorco/coMMTV-Cre (n = 25) mice, estimated using Branch software (**P < 0.01, one-way ANOVA with Tukey's post hoc correction). (D) Kaplan-Meier curves of tumor-free survival showing normal survival (100%) of Brca1co/co (n = 28) and Brca1co/coMtorco/co (n = 30) mice and significantly diminished tumor-free survival in Brca1co/coMMTV-Cre mice (n = 24) compared with Brca1co/coMtorco/coMMTV-Cre mice (n = 29) (*P = 0.008, log-rank test). (E-F) Kaplan-Meier curves (E) and Summary data (F) showing tumor-free survival of Brca1co/coMMTV-Cre mice treated with vehicle (n = 14) or everolimus (EVE, n = 11, 20 mg/kg, oral, 5 times/week). (G) Representative whole-mount staining of non-tumor-baring mammary glands from 15-month-old Brca1co/coMMTV-Cre mice treated with vehicle or everolimus for 11 months. Small arrowheads indicate abnormal foci. The panels on the right are magnifications of the boxed areas in adjacent panels. Scale bar, 5 mm. (H) Number of foci per mammary gland in vehicle- (n = 6) and everolimus-treated (n = 11) non-tumor-baring mammary glands (**P < 0.01). (I) Upon spontaneous tumor appearance, tumor-bearing mice were randomized into vehicle (n = 11) or everolimus (n = 15, 20 mg/kg, oral, 5 times/week) treatment. Tumor growth progression was monitored weekly by MRI. Representative MRI scans of tumor-bearing mice at baseline and following the indicated treatments. (J) Graphs showing RTVs (left panel) between post-treatment and baseline (start of treatment) and analysis of weekly progression (right panel). Tumor growth was assessed by the RTV = tumor volume at a given time (cm3)/tumor volume at the initiation of treatment (cm3). Weekly progression in the everolimus-treated group was significantly lower than that in the vehicle-treated group (P = 0.0002, chi-square test). (K) Summary data showing tumor progression and mouse survival following vehicle or everolimus treatment. (L) Responsiveness of spontaneously developed mammary tumors from Brca1co/coMMTV-Cre mice to everolimus, segregated based on mouse survival: non-responder (NR), survival ≤ 9 weeks; responder (R), survival > 9 weeks. (M) Representative MRI scans of non-responder and responder tumor-bearing mice at baseline and the indicated times. (N) Summary data showing everolimus-responsiveness of tumor progression (RTV) and mouse survival. (O) Graph showing calculated RTVs for engrafted tumors treated with vehicle (black lines) or everolimus (red lines). Spontaneously developed mammary tumors (n = 22) were collected from Brca1co/coMMTV-Cre mice and transplanted into nude mice. Growth of the corresponding tumors in sham-treated mice versus mice treated with everolimus (5 mg/kg, oral, 5 times/week) was tested. All mice were sacrificed when any tumors from a shared origin reached ∼3 cm3. (P) Comparison of RTVs and tumor weights at the end of the study between vehicle-treated (n = 22) and everolimus-treated (n = 22) mice. (Q) Heat maps showing protein (prot1 and prot2) and phosphopeptide (phos1 and phos2) signatures (rows) defining Sub1 and Sub2. The numbers of proteins and phosphopeptides are indicated in parentheses. (R) Box plots showing ratios of RTVs (left panel) and tumor weights (right panel) at endpoints in everolimus-treated samples compared with their paired vehicle-treated samples in Sub1 and Sub2. **, P < 0.01; *, P < 0.05 from two-sample t-test. (S) Cellular pathways significantly enriched by proteins (prot1 and 2) and phosphoproteins of selected phosphopeptides (phos1 and phos2) defining Sub1 and Sub2. The heat map shows pathway enrichment significance, represented as z-scores computed as -N−1(P-value), where P-value is the enrichment P-value from DAVID or ConsensusPathDB, and N−1(P-value) is the inverse normal distribution. (T) Network model showing interactions between proteins and phosphorylated proteins involved in leukotriene metabolic process (top), actin cytoskeleton regulation (bottom left), and neutrophil extracellular trap formation (NETosis, bottom right). Pink nodes indicate proteins (prot2) defining Sub2, while circled P on a node indicates phosphoproteins containing the phosphopeptides (phos2) defining Sub2. Solid arrows indicate direct activation; dotted arrows indicate indirect activation; arrows with "+p" denote phosphorylation. (U) Representative immunohistochemistry images of ALOX5 for the indicated responder and non-responder. Scale bar, 50 µm. Abbreviations: AA, arachidonic acid; DAG, diacylglycerol; PI, phosphatidylinositol. To determine whether mTOR inhibition also suppresses the progression of BRCA1-associated breast cancer, we tested the efficacy of everolimus on spontaneously developed mammary tumors in Brca1co/coMMTV-Cre mice through periodic observation and palpation. Tumor-bearing mice (size < 0.5 cm3) were then randomized to receive either vehicle or everolimus via oral gavage (Top right of Figure 1A, Supplementary Figure S4). Tumor volumes at baseline and during progression were measured weekly using magnetic resonance imaging (MRI) until the tumors reached a volume of ∼3 cm3 (Figure 1I). Tumors in vehicle-treated mice grew more rapidly than those in everolimus-treated mice (Figure 1J, left panel). An analysis of weekly progression showed that 73% (28 of 38) of tumors in vehicle-treated mice exhibited greater than 50% progression, compared to only 33% (46 of 141) in everolimus-treated mice (Figure 1J, right panel; P = 0.0002, chi-square test). Additionally, the weekly increase in tumor volume in the everolimus-treated group (51.0%, 95% CI: 40.3%-61.8%) was significantly lower (P < 0.001) than that in the vehicle-treated group (89.1%, 95% CI: 70.4%-107.7%). Moreover, everolimus-treated mice showed significantly longer survival (2.8-fold on average) compared to vehicle-treated mice (Figure 1K; P < 0.001). Importantly, while everolimus treatment significantly improved therapeutic outcomes in Brca1-mutant tumors, responses to everolimus showed heterogeneity among individual mice. Specifically, 8 of the 15 mice, designated as responder mice, exhibited a significant reduction in the ratio of tumor volume (RTV) in response to everolimus. In contrast, the remaining 7 mice, designated as non-responders, displayed a higher RTV than responders and vehicle-treated mice (Figure 1L,M). Additionally, the survival of responders (13.0 weeks) was nearly double that of non-responders (6.6 weeks) (Figure 1N). To further examine the therapeutic efficacy of everolimus, we employed an engraft model for preclinical evaluation. Tumor tissues were collected from 22 individual spontaneously developed mammary tumors in Brca1co/coMMTV-Cre mice, orthotopically transplanted into nude female mice, amplified, re-transplanted, and subsequently treated with either vehicle or everolimus. Tumor progression was monitored (bottom right of Figure 1A, Supplementary Figure S5), and all mice were sacrificed when any tumors in either vehicle- or everolimus-treated group reached ∼3 cm3 (Figure 1O, Supplementary Figure S6). Tumors from everolimus-treated mice showed significant reductions in RTV (43%) and weight (38%) compared to tumors from vehicle-treated mice (Figure 1P, Supplementary Figure S5). These findings suggest effective management of BRCA1-associated breast cancer by pharmacological mTOR inhibition. To explore this heterogeneity, we conducted global proteome and phosphoproteome profiling of vehicle- and everolimus-treated allograft tumors (Supplementary Figure S7A). Two distinct sample clusters (Sub1 and Sub2) were identified using both protein and phosphopeptide data (Supplementary Figure S7B,C). We identified 304 and 323 proteins that were upregulated, and 251 and 291 phosphopeptides that were upregulated, in Sub1 and Sub2, respectively (Figure 1Q). Sub2, characterized by higher RTVs and weights, represented the non-responders, whereas Sub1 corresponded to the responders (Figure 1R). The upregulated proteins and phosphoproteins in Sub2 were associated with neutrophil extracellular trap formation (NETosis) and leukotriene metabolism (Figure 1S). Enzymes involved in phosphatidylinositol and arachidonic acid formation/metabolism were upregulated in non-responders, leading to the release of leukotrienes (Figure 1T). Upon leukotriene binding, (1) proteins and phosphorylations mediating actin polymerization required for neutrophil migration, and (2) proteins involved in NETosis, were upregulated in non-responders (Figure 1T). Western blotting and immunohistochemistry confirmed the upregulation of representative markers of the leukotriene and NETosis pathways (Figure 1U, Supplementary Figure S7D,E). Therefore, our findings provide preclinical evidence that targeting mTOR inhibition is a potential strategy for the prevention and therapeutic management of BRCA1-associated breast cancer. Additionally, activation of the leukotriene-neutrophil activation axis can serve as a predictive marker of resistance to targeted mTOR inhibition. We further discussed the leukotriene signaling as a predictive biomarker and potential clinical translational value of this study in supplementary information. Chang-il Hwang, Daehee Hwang, and Sang Soo Kim conceived the study, designed, and supervised the experiments. Tae Hyun Kim, Chu-Xia Deng, Sung Chul Lim, Chang-il Hwang, Daehee Hwang, and Sang Soo Kim wrote and revised the manuscript. Eun Jung Park, Tae Hyun Kim, Dong Hoon Shin, Heesun Cheong, Chu-Xia Deng, Sung Chul Lim, Chang-il Hwang, Daehee Hwang, and Sang Soo Kim contributed to the data analysis and interpretation. Hye Jung Baek, Jihao Xu, and Heesun Cheong contributed to the in vitro experiments. Eun Joo Cho, Min Kyung Ki, and Dong Hoon Shin conducted the in vivo animal experiments. Geun Hee Han performed bioinformatical analysis. Tae Hyun Kim executed the statiscal analysis. Sung Chul Lim performed the pathological analysis. Eun Jung Park performed immunological analysis. All authors read and approved the final manuscript. We would like to thank Core facilities at National Cancer Center Korea for supporting analysis. The authors declare no conflict of interest. This work was supported by the National Cancer Center of Korea (NCC-2210680/2410880) and the National Research Foundation of Korea (2023R1A2C1004000). All procedures involving animals and their care were approved by the Institutional Animal Care and Use Committee of the National Cancer Center of Korea (NCC-15-295). All generated global and phosphoproteomic data have been deposited in the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository, with PXD054301 as the identifier [10]. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

All supplementary tables in one excel file

Supplementary Table S3. DE genes in C2 (High) compared with C1 (Low) subtype.

Multidrug resistance is a significant barrier in cancer therapy largely due to poorly understood regulatory mechanisms. Here we reveal that certain anticancer drugs can bind to newly synthesized proteins prior to reaching their canonical targets, resulting in various forms of protein damage. This binding disrupts protein functions, particularly those of mitochondrial proteins, resulting in substantial cytotoxicity. The protein damage is further exacerbated by mitochondrial reactive oxygen species generated as a consequence of the initial damage, creating a positive feedback loop. In response, cancer cells rapidly initiate a chain of events, which we term the Protein Damage Response (PDR). This includes damage recognition primarily mediated by protein ubiquitination and subsequent damage clearance via the proteasome system. Notably, patients with advanced, drug-resistant metastatic breast or colon cancers exhibit elevated proteasome activity. In an effort to predict drug resistance, we developed a sensitive kit for detecting proteasome levels, enabling the identification and subtyping of patients with high proteasome activity to support tailored therapeutic strategies. Using a three-dimensional tumor slice culture-based drug sensitivity assay and an investigator-initiated clinical trial, we demonstrate that three clinically approved proteasome inhibitors effectively overcome multidrug resistance in colon and breast cancer patients with elevated proteasome activity.

Supplementary Table S17. CD4+/CD8+ T cells infiltration levels in tumor tissues of mice bearing with S180.