Breast Cancer Reveals Latent -Related Susceptibility to Pulmonary Hypertension.

Methods
All supporting data are included in this article. Additional details or inquiries can be directed to the corresponding author. Methods are described in the Supplemental Material. Experimental procedures involving human samples conformed to the principles of the Declaration of Helsinki and were approved by the institutional ethics committee (CER-20735). All animal experiments were performed and analyzed in a blinded manner, following the guidelines outlined previously,20 in accordance with the ethics guidelines of Université Laval (protocol 20219-060) and in compliance with institutional biosafety and Canadian Council on Animal Care standards.

Bioinformatic and Data Mining
Gene–disease associations were analyzed using the DisGeNET2r package,21 while meta-analysis was performed using the Breast Cancer Gene Expression Miner v4.9 “Expression Module.”22 BMPR2 genetic variants were extracted from cBioPortal using datasets published after 2020 and filtered with the keyword “breast.”23–25 Copy number alteration analysis was limited to cBioPortal datasets with available Genomic Identification of Significant Targets in Cancer data.26

Animal Model
Heterozygous Bmpr2+/Δ71 Sprague-Dawley rats were generated as described previously19 and kindly provided by Dr F. Perros. Animals were housed in pairs per cage under identical environmental conditions (12-hour light/dark cycle, controlled temperature and humidity), regardless of genotype, with ad libitum access to food and water. Animal genotyping was performed using polymerase chain reaction on genomic DNA extracted from tail tips following a standard proteinase K digestion protocol (20 µL of proteinase K per 100 µL of double distilled water). polymerase chain reaction was performed using Taq DNA polymerase (5 U/µL; VWR, CA95057-686), dNTP nucleotides (10 nM; Amresco, VWR, CA97063-232), and primers flanking the mutant sequence: forward primer 5′-AAGCTAGGTCCTTCGCATCTG-3′ (IDT, 285630186, 100 µM) and reverse primer 5′-TAGGGACGGGAAACTACACG-3′ (IDT, 285630185, 100 µM). Polymerase chain reaction products, along with a 100-bp DNA molecular weight ladder (Invitrogen, Carlsbad, CA; 1769808), were resolved on 1.7% agarose gels and stained with RedSafe DNA dye (Froggabio, 21141) for visualization. Sample size was based on prior exploratory studies using the Bmpr2+/Δ71 model and adjusted for ethical feasibility. A 5% failure rate in hemodynamic measurements was anticipated when determining group sizes. For the spontaneous tumor development follow-up, all animals from our colony were monitored between September 2020 and September 2023, corresponding to the active lifespan of the colony.

Experimental Design

Spontaneous Tumor Development
Spontaneous tumor development was assessed weekly in a cohort of 71 Bmpr2+/Δ71 (54% female) and 45 wild-type (WT) (46% female) Sprague-Dawley rats over a median follow-up period of 264 days.

7,12-Dimethylbenz[a]anthracene–Induced Tumor Development
Seven-week-old (180 g) WT and Bmpr2+/Δ71 female Sprague-Dawley rats were randomized into 4 groups. In group A, 27 WT rats received 7,12-dimethylbenz[a]anthracene (DMBA) (D3254-1G, lot SLCC3858, Sigma) to induce mammary tumors via oral administration (20 mg/mL in soybean oil at a dose of 1 mL per 183 g of body weight). In group B, 13 WT rats received vehicle only (1 mL of soybean oil per 183 g of body weight). In group C, 27 Bmpr2+/Δ71 rats received DMBA under the same conditions as group A. In group D, 13 Bmpr2+/Δ71 rats received vehicle only.
DMBA-treated animals were monitored weekly until the appearance of the first tumor, which occurred after an average of 6 weeks. Subsequently, tumor progression (volume and number) was monitored daily. The study end point was reached when the tumor size reached 10 cm3, as per ethics committee recommendations. Age-matched controls were also euthanized at this time point. At euthanasia, organs, including the lungs, right ventricle, tumors, mammary glands, and blood, were collected for histological and biochemical analyses. Hemodynamic measurement, histology, and biochemistry were performed as described previously27 and in the Supplemental Material. All successfully genotyped rats (Bmpr2+/+ and Bmpr2+/Δ71) completing the experimental protocol were included. Animals were excluded in cases of poor-quality hemodynamic recordings, technical issues, or early euthanasia recommended by veterinary staff.

Hemodynamic Measurement
Invasive closed-chest right heart catheterization was performed as described previously.27 Briefly, anesthesia was induced with 5% isoflurane and maintained with 3% isoflurane during the procedure. A high-fidelity pressure–volume catheter (Scisense; Transonic, London, ON, Canada) was advanced into the right ventricle via the jugular vein and right atrium in closed-chest animals. Right ventricular pressure and volume were continuously recorded using the Scisense ADV500 pressure–volume measurement system (Transonic) and LabScribe2 software (iWorx, Dover, NH). Hemodynamic parameters, including right ventricular systolic pressure, stroke volume, end-diastolic volume, and end-systolic volume, were directly obtained from pressure–volume traces. Total pulmonary resistance was calculated as total pulmonary resistance=mean pulmonary artery pressure (mPAP)/cardiac output, where cardiac output was derived as stroke volume×heart rate, and mPAP was estimated using the formula 0.61×right ventricular systolic pressure+2.28

Cell Isolation, Culture, and Treatment
Human and rat PASMCs were isolated from total lungs as described previously29 and in the Supplemental Material.

Microscopy, Immunofluorescence, Immunohistochemistry, and Western Blotting:
Antibody details are provided in Table S1. Additional methodological details are described in the Supplemental Material.

RNA Sequencing

RNA Extraction and Sequencing
Total RNA was extracted from DMBA-induced mammary tumor samples of Bmpr2+/Δ71 and WT female rats (n=3 per condition) using the Direct-zol RNA Miniprep Kit (R2051, Zymo Research). RNA libraries were prepared using the Illumina Stranded mRNA Library Preparation Kit and sequenced on an Illumina NovaSeq platform. Sequencing generated 150-bp paired-end reads targeting 50 million clusters per sample (100 million reads total). FASTQ files were aligned to the mRatBN7.2 rat reference genome with Ensembl gene annotations (build 98) using Spliced Transcripts Alignment to a Reference (v2.6.1a).30 Gene quantification was performed using RNA-Seq by Expectation–Maximization (v1.3.1), and quality control analysis was carried out with multiQC. Downstream analysis was conducted in R 4.0, with separate analysis for each tissue type. Initial filtering of genes was performed using the filterByExpr() function from the edgeR package. Differential expression analysis was then conducted using the DESeq2 protocol.31 Unsupervised clustering of samples was visualized via principal component analysis. Transcripts with adjusted P<0.05 and log2 fold change>1 were considered significant. Gene Ontology and pathway enrichment analyses were performed using the Kyoto Encyclopedia of Genes and Genomes, Gene Ontology Biological Process, Gene Ontology Cellular Component, and Gene Ontology Molecular Function databases via ShinyGO.32 The datasets generated and analyzed during this study will be available in the Gene Expression Omnibus repository (GSE314019).

The French National Healthcare Database Epidemiology Study
The French National Healthcare Database (Système National des Données de Santé [SNDS]) is a nationwide database of claims and hospital care with almost 20 years of history. Two cohorts were defined (PAH cohort and breast cancer cohort) to estimate the incidence of breast cancer in patients with PAH and the incidence of PAH in patients with breast cancer between January 1, 2009 and December 31, 2024. PAH was identified using a validated algorithm (ie, hospitalization with a specific PAH diagnosis code (I27.0), right heart catheterization procedure 1 year around this hospitalization, the dispensation of specific PAH treatment 6 months before or ever after hospitalization or right heart catheterization, and the exclusion of patients with riociguat as first-line treatment or pulmonary angioplasty or endarterectomy procedure).33 The date of diagnosis was defined as the date of the first hospitalization, the first right heart catheterization, or the first specific treatment dispensation, whichever came first. Patients with breast cancer were identified using specific diagnosis codes (C50 and D05), and date of cancer was defined as the first occurrence. In the PAH cohort, patients who first presented with PAH and subsequently developed breast cancer were described with their associated diseases and comorbidities at the time of entry into the cohort (PAH diagnosis) and compared with patients who did not develop breast cancer. In the breast cancer cohort, patients who subsequently developed PAH were described with associated diseases, comorbidities, and the types of other cancers at cohort entry (breast cancer diagnosis) and then compared with patients who did not develop PAH. Moreover, in this cohort, we extracted all antineoplastic exposures (Anatomical Therapeutic Chemical Classification System, group L) after cohort entry (up to PAH diagnosis in patients presenting with PAH and up to randomly assigned dates that follow the distribution of time to onset of PAH in the PAH group), and we estimated the proportion of patients exposed to each drug class (Anatomical Therapeutic Chemical level 3) during the follow-up time. Incidence rates and corresponding 95% CIs were calculated using the Poisson distribution, standardized by age categories (using the distribution of the whole French population), and stratified by sex. Standardized incidence rate ratios of PAH in the breast cancer cohort and of breast cancer in the PAH cohort, compared with the general population, were also estimated. The total number of person-years was determined from the SNDS data, from the time patients entered the database until their date of death or December 31, 2024.

Statistical Analysis
Values are expressed as fold change or mean±SEM. Statistical tests were selected based on data distribution. For normal distribution, unpaired t tests were used for comparisons between 2 groups, while 1-way ANOVA followed by Tukey-Kramer post hoc tests were applied for comparisons among more than 2 groups. For non-normal distribution, the Kruskal-Wallis test was used for comparisons involving more than 2 groups, followed by Dunn multiple comparisons test. Paired analyses were conducted when the same cells were exposed to different conditions. For BMPR2 expression analysis in human breast cancer and normal-like tissues, comparisons were performed using 1-way ANOVA followed by Dunnett-Tukey-Kramer post hoc tests. Correlations were assessed using Pearson or Spearman correlation tests, depending on data distribution. All statistical analyses were performed using GraphPad Prism (version 9, GraphPad Software, La Jolla, CA) and are detailed in the Supplemental Material. The statistical methodology was reviewed and validated by a biostatistician.

Results

Decreased BMPR2 Expression Is Associated With Breast Cancer
Using the DisGeNET database, we found that BMPR2 is also associated with breast cancer (Figure 1A; Table S2). Among 21 pathological features linked to BMPR2, the majority (67%) were related to PAH, as expected, but, notably, 24% were associated with breast cancer. To explore this relationship, we performed a meta-analysis of publicly available sequencing data to assess the presence of rare, predicted deleterious BMPR2 variants in patients with breast tumors. These were predominantly somatic mutations (variant allele frequency <0.5) and included 16 missense, 4 nonsense, and 1 frameshift deletion and 1 frameshift insertion (Figure 1B; Table S3). Importantly, 1 of the identified variants (G83R) affects the same amino acid position previously implicated in PAH pathogenesis.34 To contextualize these findings, we compared the burden of somatic BMPR2 mutations in breast cancer with that observed across other tumor types. As anticipated, colorectal cancer exhibited the highest mutation burden (10.9% of screened cases), consistent with prior studies implicating BMPR2 loss in colorectal tumorigenesis (Table S4).11 Increased frequencies were also observed in liver (2.1%), prostate (1.6%), and bladder (1.2%) cancers. Conversely, low or no BMPR2 mutations were detected in cohorts of brain, lung, uterine, pancreatic, and ovarian cancers. We next analyzed copy number alterations across 2599 samples. A total of 76 samples exhibited copy number alterations in the BMPR2 locus, of which 70 (92%) were classified as deep deletions (Figure 1C). In an independent publicly available cohort, we analyzed BMPR2 mRNA expression across 4 subtypes of human breast cancer and normal breast-like tissues. Compared with normal breast-like tissues (n=869), BMPR2 expression was significantly reduced in breast cancer tissues across all subtypes (n=3411) (Figure 1D; Table S5). Consistently, we also observed reduced BMPRII protein expression in mammary glands from patients with breast cancer compared with healthy mammary tissue from noncancer controls (Figure 1E). To assess tissue specificity, we compared BMPRII expression across various tumor types. Among these, breast carcinoma exhibited the lowest expression (Figure S1A and S1B).

Bmpr2 Depletion Is Associated With Increased Risk of Mammary Tumor Development in Rats
To further explore the relationship of BMPR2 dysregulation and breast cancer, we assessed the spontaneous development of mammary tumors in Bmpr2+/Δ71 rats. First, we confirmed the presence of a 71-bp deletion (Δ71) in exon 1 of the Bmpr2 gene (Figure S2A). This mutation resulted in an ~50% reduction in BMPRII protein levels in the lungs of Bmpr2+/Δ71 rats compared with WT controls (Figure S2B) and was associated with decreased phosphorylation of SMAD1/5/9, a key downstream effector of BMPR2 signaling (Figure S2C). Over a median follow-up of 264 days, 10 of 71 Bmpr2+/Δ71 rats (14%) developed spontaneous mammary tumors compared with only 1 of 45 WT rats (2%) (Figure 1F and 1G; Figure S2D). Histopathological analysis showed that 82% of mammary tumors were benign, encompassing lesions with predominantly acinar hyperplasia and adenomatous or fibroadenomatous architecture, while 18% exhibited malignant features consistent with adenocarcinoma (Figure 1H). Notably, reduced BMPRII protein expression correlated with decreased BRCA1 protein expression in rat mammary glands (Figure 1I and 1J).35 These findings align with previous reports17,18 and further support the hypothesis that reduced BMPR2 expression may increase susceptibility to breast tumorigenesis.

Mammary Tumors Exacerbate PH in Bmpr2+/Δ71 Rats
We next investigated whether breast tumors could exacerbate pulmonary vascular disease in the context of Bmpr2 deficiency. Mammary adenocarcinomas were induced in female WT and Bmpr2+/Δ71 rats using the carcinogen DMBA. Tumor onset occurred at similar times in both genotypes (48.7±3.4 days in WT versus 51.2±5.2 days in Bmpr2+/Δ71; Figure S3A and S3B), with comparable tumor burden (Figure S3C and S3D). All tumors were confirmed as adenocarcinomas (Figure S3E). We then assessed pulmonary hemodynamics in tumor-bearing animals and age-matched controls without tumors. Notably, Bmpr2+/Δ71+tumor rats exhibited significantly elevated right ventricular systolic pressure, mPAP, and total pulmonary resistance, alongside reduced stroke volume and cardiac output, compared with all other groups (Figure 2A through 2D; Figure S4A). Strikingly, 67% of Bmpr2+/Δ71+tumor rats had mPAP>25 mm Hg versus 13% of WT+tumor and 0% of nontumor animals (Figure S4B). Histological analysis revealed increased pulmonary vascular remodeling and PASMC proliferation as well as a trend toward reduced PASMC apoptosis in Bmpr2+/Δ71+tumor rats (Figure 2E through 2G; Figure S4C). In contrast, no significant changes in PASMC proliferation or endothelial layer thickness were observed across groups (Figure S4D and S4E), though Bmpr2 deficiency alone tended to reduce both metrics. Importantly, no significant hemodynamic/histologic differences were found between WT, Bmpr2+/Δ71, WT+tumor, or Bmpr2+/Δ71 rats without tumors (Figure 2A through 2F), suggesting that tumor burden, not genotype or DMBA exposure alone, drives PH exacerbation. Furthermore, DMBA-treated rats without tumors (WT–DMBA–no tumors and Bmpr2+/Δ71–DMBA–no tumors) showed no PH phenotype, confirming that DMBA itself does not induce PH (Figure S5A through S5C). Thus, taken together, these results suggest that mammary tumors exacerbate adverse pulmonary vascular remodeling and PH development in Bmpr2+/Δ71 animals.

Bmpr2+/Δ71 Rats With Mammary Tumors Exhibit a Proinflammatory Phenotype
The impaired pulmonary hemodynamics and adverse vascular remodeling observed in Bmpr2+/Δ71+tumor animals were associated with increased lung monocyte and leukocyte accumulation (Figure 2G through 2I). Additionally, compared to the WT, Bmpr2+/Δ71 (without tumors), and WT+tumor groups, lungs from Bmpr2+/Δ71+tumor animals exhibited significantly elevated expression of the proinflammatory cytokine IL-1β and increased protein levels of the transcription factor NF-κB (Figure 2J and 2K). Correlation analysis revealed that lung IL-1β levels were moderately but significantly associated with pulmonary hemodynamic and vascular remodeling parameters, including right ventricular systolic pressure, total pulmonary resistance, and vascular remodeling (Table S6). Interestingly, no significant differences were observed in lung IL-6 expression across all experimental conditions (Figure S6A). In addition, we evaluated the expression of factors known to be associated with BMPR2 depletion and PAH progression.36,37 PGC1α protein expression was elevated in Bmpr2+/Δ71 lungs regardless of tumor status (Figure S6B). OPG (osteoprotegerin) levels were increased in the lungs of Bmpr2+/Δ71, WT+tumor, and Bmpr2+/Δ71+tumor animals compared with WT controls (Figure S6C).

Bmpr2 Mutation Is Associated With Proinflammatory Mammary Tumors and Increased Circulating Inflammatory Mediators
Tumors extracted from Bmpr2+/Δ71 animals displayed a distinct transcriptomic profile compared with those from WT animals, with 273 upregulated (including IL-1β) and 43 downregulated genes (adjusted P<0.05, log₂ fold change (log2FC)>1) (Figure 3A; Figure S7A). Increased IL-1β expression in Bmpr2+/Δ71 tumors was validated by quantitative reverse transcription–polymerase chain reaction (qRT-PCR) (Figure S7B). Functional Gene Ontology analysis revealed that the biological functions associated with upregulated differentially expressed genes were predominantly related to inflammation (Figure 3B; Figure S7C). Consistent with the transcriptomic findings, histological analysis showed increased infiltration of inflammatory cells, including monocytes and leukocytes, in Bmpr2+/Δ71 tumors (Figure 3C and 3D). Supporting this observation, Bmpr2+/Δ71+tumor animals showed higher overall inflammation compared with WT+tumor animals, as evidenced by the measurement of 19 circulating proinflammatory cytokines (Figure 3E and 3F). At the individual cytokine level, IL-1α, IL-1β, IFN-γ, LIX, and IL-6 were significantly increased in the blood of Bmpr2+/Δ71+tumor animals (Figure 3E). Notably, IL-1β was the only cytokine consistently elevated across the tumor, blood, and lungs of Bmpr2+/Δ71+tumor animals compared with WT+tumor rats.

Tumor-Secreted IL-1β Induces Bmpr2+/Δ71 r-PASMC Proliferation
To further investigate the association between mammary tumors and PH, we conducted in vitro experiments using cells isolated from Bmpr2+/Δ71 and WT tumors as well as rat pulmonary arterial smooth muscle cells (r-PASMCs) isolated from the lungs of Bmpr2+/Δ71 and WT rats. Conditioned media (CM) experiments were performed, revealing that media from Bmpr2+/Δ71 tumor cells (CM-Bmpr2+/Δ71) induced proliferation of Bmpr2+/Δ71 r-PASMCs (Figure 4A; Figure S8A through S8C). In contrast, CM from WT tumor cells did not significantly induce Bmpr2+/Δ71 r-PASMC proliferation (Figure 4A; Figure S8A through S8C). Consistent with this observation, tumor cells isolated from Bmpr2+/Δ71 animals exhibited significantly higher IL-1β expression compared with those from WT animals (Figure 4B and 4C). Supernatants from Bmpr2+/Δ71 tumor cells contained elevated IL-1β concentrations and exhibited a higher overall inflammatory profile (Table S7). Notably, neither CM-Bmpr2+/Δ71 nor CM from WT tumor cells induced proliferation in WT r-PASMCs (Figure S8D through S8G). To investigate the role of IL-1β in Bmpr2+/Δ71 r-PASMC proliferation, we conducted IL-1β antibody blockade experiments. Neutralizing IL-1β partially reversed the proproliferative effect of CM-Bmpr2+/Δ71 on Bmpr2+/Δ71 r-PASMCs (Figure 4D; Figure S8H through S8J). These findings suggest that IL-1β secreted by Bmpr2+/Δ71 tumor cells promotes r-PASMC proliferation, providing a potential mechanistic link between tumor-derived inflammation and PH.

BMPR2 Mutation Exacerbates IL-1β–Induced Human PAH-PASMC Proliferation
Human PASMCs derived from patients with PAH with BMPR2 mutations and those without mutations showed no significant differences in baseline proliferation (Figure S9A through S9C). However, human PASMCs derived from patients with PAH with BMPR2 mutations exhibited a significant decrease in phospho-SMAD1/5/9 levels compared with nonmutated human PASMCs (Figure S9D). IL-1β treatment (2.5 and 5 µg/mL) increased PASMC proliferation in a dose-dependent manner across all conditions (Figure S9E through S9I). However, a significant increase in proliferation was only observed in human PASMCs derived from patients with PAH with BMPR2 mutations stimulated with 5 µg/mL of IL-1β (Figure S9E, S9H, and S9I). These findings suggest that, similar to observations in r-PASMCs, PASMCs from patients with PAH carrying BMPR2 mutations exhibit heightened sensitivity to IL-1β–induced proliferation, supporting a role of inflammation in BMPR2-related PAH progression.

Increased Breast Cancer Incidence in Patients with PAH
Regardless of genetic status, impaired BMPR2 signaling is a hallmark of PAH; for instance, lungs from patients with PAH show reduced BMPR2 expression compared with non-PAH controls (Figure S10A).4,5 To investigate the clinical association between PAH and breast cancer, we analyzed nationwide data from the SNDS. Overall, we identified 9964 patients with PAH, 1 300 397 patients with breast cancer, and 328 patients with both diseases between January 1, 2009 and December 31, 2024.
In the PAH cohort, 111 patients developed breast cancer after PAH diagnosis, of whom 98.2% were women, with a median age of 65.6 years (interquartile range, 57.2–71.7). The distribution of PAH-associated conditions and other comorbidities appeared to be similar between the 2 cohorts (Table 1). The overall standardized incidence rate of breast cancer in the PAH cohort was 198.25 per 100 000 per year, compared with 96,71 in the general population, yielding an incidence rate ratio of 2.05 (1.70–2.46) (Table 2). When restricting the analysis to women, the incidence rose to 315.59 among women with PAH versus 187.61 in the general female population, yielding an incidence rate ratio of 2.32 (1.93–2.80) (Table 2). In the PAH cohort, the incidence of breast cancer was higher at younger ages compared with patients without PAH (Figure 5A and 5B). A trend toward increased breast cancer incidence was also noted in the male PAH subgroup; however, this observation should be interpreted cautiously given the very low number of men with breast cancer (n=2) among patients with PAH (Figure S10B; Table 2).
In the Breast cancer cohort, 217 patients developed PAH after breast cancer diagnosis. Most (99.1%) were women, with a median age at breast cancer diagnosis of 63.0 years (interquartile range, 55–69). Patients with breast cancer developing PAH had more cardiovascular comorbidities compared with those without PAH (Table 1). The overall standardized incidence rate of PAH in patients with breast cancer was 20.98 per 1 000 000 person-year compared with 7.41 in the general population, yielding an incidence rate ratio of 2.83 (95% CI, 2.48–3.24) (Table 2). In women, PAH incidence was higher in patients with previous breast cancer (Table 2). In the breast cancer cohort, the incidence of PAH was higher at younger ages compared with patients without breast cancer (Figure 5C and 5D). While interpretation in men should be cautious due to the limited number of cases (n=2 among patients with PAH), a comparable trend was observed (Table 2; Figure S10C). Types of antineoplastic drugs used in patients developing PAH were similar to those used in patients not developing PAH (Table 1).

Discussion
From data mining to population-level epidemiology, combined with in vivo and in vitro experiments, our work suggests a link between BMPR2 and breast cancer development. Conversely, our findings indicate that breast cancer, likely mediated by a proinflammatory mammary tumor microenvironment, may exacerbate PAH in genetically/epigenetically predisposed populations (Figure 6).
Our findings align with previous reports first showing that Bmpr2+/Δ71 rats are prone to mammary tumors and that cancer with a proinflammatory microenvironment exacerbates the PH phenotype.12,16–18 In lung cancer, both patient samples and mouse models demonstrate perivascular inflammatory cell infiltration, vascular remodeling, and impaired pulmonary hemodynamics, with inflammation driving PASMC proliferation, migration, and resistance to apoptosis.16 Clinically, PH worsens survival of cancer and can result from pulmonary tumor microembolism or pulmonary tumor thrombotic microangiopathy, most often linked to gastric, breast, and lung cancers.38–41 While we could not assess pulmonary tumor embolism/pulmonary tumor thrombotic microangiopathy, we found that 67% of Bmpr2+/Δ71 tumor-bearing rats developed PH versus 13% in WT with tumors, associated with a proinflammatory phenotype and increased PASMC sensitivity to IL-1β, consistent with the role of BMPR2 in limiting inflammation.9,12,37
In a population-level study, we found a bidirectional association between breast cancer and PAH; patients with PAH had a higher incidence of breast cancer, and breast cancer incidence was also increased in patients with pre-PAH, with PAH occurring more often in those with breast cancer. This aligns with recent work identifying BRCA1 as a novel BMPR2 signaling target and BRCA1-associated protein as a potential PAH gene.35,42 In the mammary gland, BMP ligands are widely expressed and influence proliferation in three-dimensional collagen matrices, often synergizing with other growth factors.43–45 Loss of BMPR2 signaling, known for tumor-suppressive effects in other cancers, may similarly contribute to breast cancer development in patients with PAH.11,46 Although our study focused on the BMPR2/breast cancer/PAH axis, other PAH-related pathways may contribute to tumorigenesis, and the tumor-suppressive role of BMPR2 suggests that the PAH-cancer link may extend beyond breast malignancies. With an aging population with PAH, updated hemodynamic definitions (mPAP>20 mm Hg and pulmonary vascular resistance>2 Wood Units) and improved survival owing to targeted therapies, the emergence of cancer in at-risk patients with PAH might warrant increased clinical attention.47 PAH therapies may also influence this risk; for instance, prostacyclin and prostacyclin synthase overexpression in breast cancer correlates with poor prognosis, enhanced migration, and apoptosis resistance, and stable prostacyclin analogues promote migration in vitro.48–50 Whether long-term PAH therapy increases breast cancer risk in susceptible patients remains to be determined.
In line with our preclinical observations, we observed an increased incidence of PAH in the population with breast cancer. Similarly, the DELPHI-2 study (NCT01600898), which followed 55 asymptomatic BMPR2 mutation carriers with annual screening over a minimum of 2 years, reported that 1 of the 5 patients who developed PAH during follow-up had previously been diagnosed with breast cancer.51 Notably, this patient had been treated with the chemotherapeutic agent trastuzumab emtansine, which has recently been associated with the development of PH.52,53 In addition, other cancer-directed therapies commonly used in breast cancer, including alkylating agents such as cyclophosphamide and thoracic radiation, have been associated with pulmonary vascular injury and secondary PH.54–57 While our preclinical data indicate that breast cancer can promote PH development independent of chemotherapy in the setting of reduced Bmpr2 expression, the epidemiological findings, particularly the markedly higher incidence of PAH observed in the breast cancer cohort (per million person-years versus 7.41 per million in the general population), likely reflect a more complex interplay. This excess risk is plausibly driven by a combination of intrinsic cancer-related mechanisms, tumor-associated inflammation, and, in a subset of patients, treatment-related vascular toxicity from chemotherapy or radiation.
Finally, although BMPR2 mutations are the most common genetic risk factor for heritable PAH, only ~20% of carriers develop disease, implying a requirement for a “second hit.” We propose that tumor-associated inflammation could serve as this second hit. Supporting this hypothesis, Tian et al. demonstrated that exogenous lung inflammation induced by adenoviral delivery of 5-lipoxygenase triggers PH in Bmpr2-mutant rats but not in WT littermates.58 Similarly, in mice, chronic lipopolysaccharide exposure induced PH in Bmpr2+/− animals, whereas WT controls remained unaffected.59 In line with our findings, these models exhibited pulmonary vascular remodeling and inflammation-driven PH, reinforcing the concept that inflammatory stress can unmask latent susceptibility associated with impaired BMPR2 signaling.
In conclusion, our study reveals a potential bidirectional association between breast cancer, BMPR2 signaling, and PAH, with breast cancer potentially facilitating PAH development through a proinflammatory tumor microenvironment, and, conversely, impaired BMPR2 signaling and established PAH being linked to an increased risk of breast cancer. Beyond advancing our mechanistic understanding, these findings may help raise clinical awareness of a possible link between PAH and breast cancer.

Study Limitation
PAH and breast adenocarcinoma are both diseases with a strong female predominance.60,61 Therefore, our study focused exclusively on female rats, and male rats were not included. The potential association between breast cancer, PH, and BMPR2 mutations in men remains unexplored. While numerous BMPR2 mutations have been linked to PAH development,62 our study specifically investigated the impact of mammary tumors in Bmpr2+
Δ71 female rats. We did not extend our research to other Bmpr2-mutant rat models, such as Bmpr2+/Δ527 and Bmpr2+/Δ16, developed by Tian et al.58 Due to ethical and animal welfare considerations (namely, the advanced age of the colony and declining reproductive performance), the breeding colony was euthanized, limiting additional in vivo experiments. Finally, at the population level, our epidemiological analyses using the SNDS revealed a bidirectional association between breast cancer and PAH; however, several limitations inherent to claims-based databases must be acknowledged. First, the observational nature of these data precludes causal inference, and the population-level findings should be interpreted strictly as associative. In addition, the SNDS lacks granular clinical and treatment-related information, limiting adjustment for potential confounders. Notably, genetic data are not available in the SNDS, preventing determination of BMPR2 mutation status among patients with PAH in the population cohort. Furthermore, comprehensive information on cancer-directed therapies is incomplete, particularly for commonly used agents such as anthracyclines and cyclophosphamide, which may independently influence pulmonary vascular risk.

Article Information

Acknowledgments
The authors gratefully acknowledge Dr. Pierre Hélie (Department of Pathology and Microbiology, Faculty of Veterinary Medicine, Université de Montréal, Canada) and Dr. Andréanne Gagné (Department of Pathology, IUCPQ-Université Laval) for expertise in the histopathological characterization of rat mammary tumors as well as Michèle Orain (CRIUCPQ, Québec, Canada) for technical assistance with the tumor microarray. The authors also extend their gratitude to the Department of Pathology at IUCPQ-Université Laval for valuable support.

Sources of Funding
This work was supported by the Canadian Institutes of Health and Research (IC132974) and Fondation du Centre de Recherche de l’Institut de Cardiologie et de Pneumologie de Québec (2024-01).

Disclosures
None.

Supplemental Material
Figures S1–S10
Tables S1–S7
Uncropped blot
ARRIVE checklist

Supplementary Material