Breast cancer is the most commonly diagnosed malignancy among women worldwide and represents the fifth leading cause of cancer-related mortality (1). Its heterogeneity is driven by a complex interplay of oncogenic signaling pathways and genetic alterations (2,3). Accordingly, critical components of effective breast cancer management include early detection, rigorous monitoring of treatment responses, and the discovery of novel molecular targets that may serve as biomarkers or be utilized in targeted therapies (3). Achieving these objectives necessitates a comprehensive elucidation of the molecular mechanisms that underlie tumor initiation, progression, and metastasis, as well as an in-depth analysis of their interactions within key cellular pathways (2).

Many studies have reported that the role of the unfolded protein response (UPR) is involved in a wide spectrum of diseases, such as metabolic syndrome, inflammatory autoimmune diseases, neurodegenerative disorders as well as carcinogenesis (4-6). Endoplasmic reticulum stress (ER stress) presents a double face role either participating in physiological processes ensuring proper regulation of mechanisms such as differentiation of cells, metabolism, proteostasis, protein synthesis and quality folding or by playing a crucial role in pathological conditions such as inflammation, hypoxia, oxidative stress, infection, and nutrient deprivation, which provoke ER stress and induce accumulation of unfolded proteins (7). This results in the activation of the UPR, a protective mechanism aimed at alleviating ER stress and restoring cellular homeostasis or, if unresolved, triggering apoptosis (8).

Three sensor proteins mediate the activation of the UPR signaling pathway during ER stress: 1) Inositol-requiring kinase 1 (IRE1), which is activated through autophosphorylation and mediates the splicing of X-box binding protein-1 (XBP1) mRNA, leading to the production of a potent transcription factor that upregulates genes involved in the UPR and ER-associated degradation and, also, generates a potent transcription factor that enhances cellular adaptation, proliferation, and VEGF-driven angiogenesis, yet its kinase arm can also trigger JNK-mediated apoptosis under irremediable stress such as in hypoxic microenvironment (9,10), 2) Protein kinase RNA-like ER kinase (PERK), which phosphorylates the a-subunit of translation initiation factor 2 (eIF2a), leading to the inhibition of global protein synthesis while selectively allowing translation of specific stress-related mRNAs (11). PERK selectively upregulates ATF4 and NRF2, which mitigate oxidative DNA damage, support vascularization, and sustain tumor cell viability in low-oxygen niches; paradoxically, sustained PERK activation can enforce G₁/S arrest through cyclin D1 suppression, fostering tumor dormancy (10).

3) Activating transcriptional factor 6 (ATF6), which translocates to the Golgi apparatus upon ER stress, where it is cleaved and activated. The resulting fragment acts as a transcription factor that enhances the expression of genes involved in protein folding and maturation, including XBP1 mRNA (11). Generally, the ER is considered perturbed in cancer cells due to uncontrolled proliferation, an altered tumor microenvironment, and exposure to chemotherapy and radiation therapy. These factors can lead to deprivation of glucose and oxygen, as well as insufficient vascularization, resulting in a hypoxic environment (12,13). To survive in such hostile conditions, cancer cells appropriately trigger the UPR response (14). Recently, numerous evidence has emerged supporting the crucial role of ER stress and UPR in cancer development and progression. These processes contribute to cancer hallmarks such as tumor survival, evasion of cell death, metastasis, and invasion through interactions with multiple oncogenic pathways across several cancer types including breast, endometrial, ovarian, colorectal, pancreatic, hepatocellular, skin, glioblastoma, prostate, and hematological malignancies (14-18). Together, these findings underscore the therapeutic promise of finely tuned IRE1-XBP1 or PERK-eIF2a pathway inhibitors agents that must balance the suppression of malignancy-promoting adaptive responses against the risk of inadvertently abrogating stress-induced tumor-suppressive programs.

ER stress and the UPR play a critical role in breast cancer development and progression by interacting with multiple oncogenic pathways (12). Emerging evidence suggests that ER stress exerts a paradoxical function in cancer cells, either promoting survival or inducing apoptosis, depending on the balance between protein folding demand and the ER’s capacity to manage that load (19). Moderate ER stress enhances cancer cell survival and contributes to chemotherapeutic resistance, whereas acute severe ER stress triggers apoptosis (19,20). Moreover, ER stress and autophagy have been implicated in cisplatin-induced apoptosis in lung cancer cells (21). Multiple studies have reported that activation of the UPR is associated with resistance to endocrine treatment and disease recurrence (20). Moreover, several reports have shown considerable crosstalk between ER signaling and the UPR (22,23). As a result, inhibiting the pathways leading to activation of the UPR is considered a novel potential target to restore drug sensitivity (22).

This study aimed to investigate the expression levels of IRE1 and PERK proteins in breast cancer tissues and their potential clinical significance. Utilizing immuno-histochemical (IHC) analysis and the Immunoreactive Score (IRS), we systematically evaluated their expression patterns and assessed their associations with clinico-pathological characteristics and patient survival. By examining these key components of the UPR, we sought to elucidate their role in breast cancer pathophysiology, assess their prognostic relevance, and explore their potential as novel biomarkers.

We retrospectively searched the electronic database and medical records of the 2^nd University Department of Obstetrics and Gynecology at the General Hospital of Thessaloniki “Hippokration”, Aristotle University of Thessaloniki, Greece. A total of 72 patients with breast cancer who underwent surgical treatment at the same institution between 2017 and 2024 were included in the study. All cases were invasive breast carcinomas, primarily of the ductal carcinoma of no special type (NST), with a few cases of lobular carcinoma. Several clinicopathological parameters were analyzed, including patient age, tumor size (diameter), stage, histological type and grade, estrogen receptor (ER), progesterone receptor (PR), Ki-67 (MIB1) proliferation index, human epidermal growth factor receptor 2 (HER-2) expression, neoadjuvant therapy, sentinel lymph node biopsy, axillary lymph-adenectomy, and type of surgical procedure (lumpectomy or mastectomy). Due to the limited number of cases, further stratification of invasive breast carcinoma (IBC) into molecular subtypes was not feasible. Histo-pathological diagnosis and immunohistochemical assessments were performed by specialized pathologists. Tumor classification followed the TNM system, and histological grading was based on the Elston and Ellis method (24). The study was approved by the Scientific Committee of the University Hospital of Ioannina (approval number: 589/9-8-2022).

Immunohistochemistry. IRE-1 and PERK expression was analyzed in 2-2.5 μm-thick tissue sections. Slides were incubated at 70˚C for 30 min, then left overnight in an oven with a gradual temperature decrease from 60˚C to 23˚C to facilitate the removal of excess paraffin. IHC staining was performed using the automated BOND-MAX system (Leica Biosystems, Nussloch, Germany). After deparaffinization and antigen retrieval (EDTA, 98˚C, 20 min), peroxidase blocking was applied for 10 min. Tissue sections were incubated with primary antibodies rabbit polyclonal IRE-1 (E-AB-93217, Elabscience, Houston, TX, USA) and mouse monoclonal PERK (B-5: sc-377400, Santa Cruz, Dallas, TX, USA), both at 1:50 dilution, followed by polymer incubation (10 min) and DAB chromogen application (10 min). Sections were counterstained with Mayer’s hematoxylin, dehydrated, cleared in xylene, and examined under a Nikon Eclipse E200 microscope.

Staining was evaluated using the IRS ranging from 0 to 12, derived by multiplying the staining intensity (0-3) by the percentage of positively stained cells (0-4). Two independent examiners assessed the slides, resolving discrepancies by consensus. Routine IHC markers (ER, PR, Ki-67, HER2) were previously recorded and included in the study.

Ethical considerations. This study was conducted in accordance with the Declaration of Helsinki and its later amendments or comparable ethical standards. The study protocol was approved (Approval number: 589/9-8-22) by the Institutional Review Board of University Hospital of Ioannina, Greece.

Statistical analysis. All analyses were conducted using Stata 16.0 (StataCorp, College Station, TX, USA). Categorical variables were compared using the chi-squared or Wilcoxon signed-rank test, while continuous variables were analyzed using Spearman’s rho, Mann-Whitney U, or Kruskal-Wallis tests. ROC curve analysis determined optimal cut-off values for categorizing IRE1 and PERK IRS scores into low (0-4) and high (5-12) expression groups. A p-value <0.05 was considered significant. Baseline comparisons between cases and controls were performed using the Wilcoxon rank-sum test. Descriptive statistics were reported as medians with interquartile ranges (IQR) for continuous variables and as frequencies with percentages for categorical variables. Univariable analyses compared IRS groups using Wilcoxon rank-sum for continuous and chi-squared for categorical variables. Survival was analyzed using Kaplan-Meier curves with log-rank tests. Cox proportional hazards models assessed associations between IRS groups and overall survival, adjusting for age, tumor stage, and histological grade. Follow-up continued until death, study end (December 2024), or loss to follow-up.

The results of immunohistochemical intensity expression of IRE1 and PERK are depicted in Figure 1 and Figure 2, respectively.

Patient and tumor characteristics. For the entire cohort the median age of the patients was 62 years (IQR=51-67.8). A total of 72 patients and 16 controls were analyzed. The follow-up duration for all patients was five years. The basic tumor characteristics are shown in Table I. Our data analysis provides intriguing findings regarding the expression of IRE1 and PERK and their clinico-pathological correlations and patients' survival in breast cancer.

Immunohistochemical analysis of IRE1 and PERK protein expression. Age was not associated with IRE1 expression (p=0.778); however, patients with high PERK expression were significantly older (mean: 62 vs. 58 years, p=0.038). IRE1 expression was significantly higher in patients (median: 9, IQR=8-12) than in controls (median: 4, IQR=3-8, p<0.001). Similarly, PERK expression was elevated in patients (median: 4, IQR=3-8) compared to controls (median: 2, IQR=0-4, p<0.001). A strong positive correlation was observed between IRE1 and PERK expression (Figure 3; Spearman’s ρ=0.55, p<0.001).

Associations of IRE1 and PERK with clinicopathological findings. Tumor size was not significantly associated with IRE1 (p=0.362) or PERK (p=0.778) expression. High IRE1 expression was more frequent in grade 3 tumors (29.5% vs. 0%), though not statistically significant (p=0.112). In contrast, high PERK expression was significantly associated with grade 3 tumors (34.9% vs. 10.3%, p=0.042).

Patients with high IRE1 expression had a significantly higher frequency of stage 3-4 tumors (5.6% vs. 0%, p<0.001), while PERK expression was not associated with TNM stage (p=0.325). IRE1 and PERK expression showed no significant association with tumor histological subtype (p=0.148, p=0.159) or with the presence of lobular carcinoma (p=0.203, p=0.35). Multifocal breast cancer was not linked to IRE1 expression (p=0.382), while PERK showed a non-significant trend (p=0.091).

Lymphovascular space invasion (LVSI) was more frequent in the high IRE1 group (54.5% vs. 26.2%), with a borderline non-significant trend (p=0.061). No association was found between PERK expression and LVSI (p=0.265). No significant differences were observed in neoadjuvant chemotherapy status for either IRE1 (p=0.884) or PERK (p=0.591).

Associations with lymph node involvement and surgical management. Axillary lymph node involvement was more frequent in the low IRE1 expression group (54.5% vs. 23.0%, p=0.031), whereas no association was found with PERK expression (p=0.612). High IRE1 expression was significantly linked to sentinel lymph node positivity (88.5% vs. 45.5%, p=0.001), whereas PERK expression showed no association (p=0.271). Axillary lymphadenectomy was not influenced by IRE1 (p=0.464) or PERK (p=0.957). Surgical treatment (lumpectomy or mastectomy) did not significantly differ between IRE1 (p=0.247) or PERK (p=0.891) expression groups, indicating that protein expression levels do not impact surgical management decisions.

Associations with hormone receptors and tumor markers. High IRE1 expression was significantly associated with ER positivity (100% vs. 68.9%, p=0.031) and PR positivity (90.9% vs. 55.7%, p=0.028), while PERK expression was linked to ER positivity (93.1%, p=0.002) but not PR (p=0.529). IRE1 expression was also significantly associated with HER2 positivity (44.3% vs. 9.1%, p=0.028) and higher Ki-67 levels (60% vs. 20%, p=0.003), suggesting a more aggressive tumor phenotype. PERK expression showed no significant associations with HER2 (p=0.891) or Ki-67 (p=0.415).

Impact of IRE1 and PERK expression on overall survival. Kaplan-Meier analysis showed a marginally non-significant difference in overall survival between low and high IRE1 expression groups (p=0.08, Figure 4), while no difference was observed for PERK expression (p=0.8, Figure 5). However, in the multivariate Cox regression analysis, high IRE1 expression (IRS 5-12) was associated with an increased mortality risk (HR=12.19, 95%CI=0.99-150.35, p=0.05). Similarly, high PERK expression was significantly linked to worse survival outcomes (IRS 4-12) (HR=12.10, 95%CI=1.16-126.30, p=0.04). Age was also a significant mortality risk factor in both models (HR=1.14 per year, p=0.05). Tumor grade was not associated with survival, while tumor stage was the strongest predictor of mortality (IRE1 model: HR=79.89, p<0.01; PERK model: HR=63.10, p<0.01) (Table II).

In this study, we analyzed the expression of ER stress-related proteins IRE1 and PERK in breast cancer specimens and their associations with tumor characteristics and survival. Our findings highlight the role of the UPR in breast cancer progression, linking high IRE1 and PERK expression to aggressive tumor features.

Elevated IRE1 expression was significantly associated with advanced TNM stage, ER/PR positivity, HER-2 positivity, and increased Ki-67 index. Prior studies have suggested that UPR activation enhances tumor aggressiveness by promoting survival, invasiveness, and therapy resistance (25). IRE1 activation drives XBP1 splicing, up-regulating UPR target genes and reinforcing the IRE1-XBP1 oncogenic axis in breast cancer (13,22,26). XBP1s increases ER expression, fostering estrogen-independent growth and resistance to antiestrogen therapy by preventing cell cycle arrest and suppressing apoptosis (23). This suggests that targeting the IRE1-XBP1 pathway could improve treatment outcomes in ER-positive breast cancer.

Similarly, high PERK expression was associated with aggressive disease markers, supporting evidence that PERK regulates protein synthesis and survival under ER stress (27-29). The PERK/eIF2α/ATF4 axis modulates androgen receptor expression in triple-negative breast cancer (TNBC) and prostate cancer, highlighting UPR’s role in hormone signaling (24,30). Additionally, progesterone influences ER-associated degradation and UPR activation in androgen-sensitive prostate cancer cells, indicating a potential interplay between steroid hormones and UPR pathways (27,31). Our findings suggest that PERK signaling similarly contributes to breast cancer progression by facilitating cellular adaptation to stress and enhancing tumor cell survival.

Beyond breast cancer, our findings align with reports showing that IRE1 activation drives epithelial-to-mesenchymal transition (EMT) and enhances metastatic potential in other malignancies (32). In colorectal cancer, the IRE1α-XBP1 pathway promotes EMT, leading to increased invasion and poorer survival (33). Similarly, high IRE1 expression in lung adenocarcinoma correlates with decreased recurrence-free survival, reinforcing the role of IRE1-driven UPR activation in tumor progression (21).

Our results also align with studies on HER-2-mutant breast cancer, where deregulation of the ERK, AKT, and mTOR pathways enhances UPR activation (34,35). Under these conditions, excessive PERK-ATF4-CHOP signaling triggers apoptosis via TRAIL-R2 and caspase-8 activation, suggesting a potential therapeutic vulnerability (8,13). Additionally, meta-analyses have shown that low XBP1 mRNA expression predicts better prognosis in serous ovarian cancer (36), while high XBP1 correlates with poor outcomes in glioma, TNBC, lymphoblastic leukemia, and ovarian cancer (20,32,37,38).

ER stress-related genes have also been implicated in vascular invasion, angiogenesis, and extracellular matrix production (37,39). In lung adenocarcinoma, their high expression correlates with improved recurrence-free survival; however, their role in breast and gastric cancers remains unclear (21,40,41). Our study adds to this understanding by showing that IRE1 and PERK may regulate GRP78 transcription, while PDI, a redox-sensitive activator, influences PERK regulation (32,38).

We further demonstrated that IRE1 expression correlates with sentinel lymph node positivity, reinforcing its role in promoting breast cancer invasion and lymph node metastasis. High PERK expression was associated with increased tumor grade, in line with previous reports linking IRE1’s downstream effector, XBP1, to oncogenic pathways through HIF1α in breast cancer, MMP9 in esophageal squamous cell carcinoma, β-catenin in bladder cancer, and the PI3K/mTOR pathway in osteosarcoma (28,42-45). These findings highlight the pivotal role of IRE1-driven signaling in cancer proliferation, invasion, and survival.

In contrast, activation of the PERK-eIF2α pathway has been linked to tumor-infiltrating lymphocytes in HER2-positive breast cancer, suggesting an immune-modulatory role for PERK in the tumor microenvironment (11,46). Anchored to the ER, PERK can initiate downstream cascades via ATF4, CHOP, and GADD34, influencing both apoptosis and survival through mechanisms such as autophagy, enhanced protein folding, and regulated protein degradation (37,47). PERK-dependent phosphorylation of Nrf2 facilitates redox homeostasis, contributing to chemoresistance (48). Additionally, PERK-mediated adaptive responses enable cancer cells to withstand oncogenic stressors like c-Myc and BRAF mutations, fostering angiogenesis, metastasis, and increased chemoresistance (47).

Despite these mechanistic insights, neither IRE1 nor PERK expression in our study significantly correlated with tumor size, histology, axillary lymph node involvement, axillary lymphadenectomy, neoadjuvant chemotherapy, or surgical approach. The near-significant association between IRE1 expression and LVSI (p=0.061) suggests a subtle role in metastasis. Beyond breast cancer, UPR signaling exhibits a dualistic nature (49). IRE1 modulates cell adhesion and migration in glioblastoma, while in TNBC, XBP1s-expressing cells show an increased propensity for lung metastasis (18,20). However, IRE1 signaling may also suppress migration via RIDD-mediated SPARC mRNA degradation, leading to attenuated RhoA signaling and enhanced sensitivity to doxorubicin, suggesting potential chemosensitization strategies (50). Similarly, PERK has been implicated in drug resistance via ABC transporter up-regulation and autophagy activation through the PERK-eIF2α-ATF4-CHOP pathway (as seen in chemoresistant colorectal adenocarcinoma HT29 cells) (51,52). However, its genetic inactivation may also induce genomic instability and oxidative DNA damage in mammary carcinoma, underscoring its context-dependent role in tumorigenesis (19,51-53).

Of paramount importance, Kaplan-Meier analysis in our study revealed a marginal survival difference between low and high IRE1-IRS groups (p=0.08), whereas no significant difference was observed for PERK-IRS groups (p=0.8). However, multivariate Cox regression analysis revealed that high IRE1-IRS expression was associated with a 12.19-fold increased mortality risk (95%CI=0.99-150.35, p=0.05), and high PERK-IRS expression corresponded to a 12.10-fold increased risk (95%CI=1.16-126.30, p=0.04), after adjusting for age, tumor stage, and grade. Tumor stage emerged as the strongest predictor of mortality, while age had a marginal effect, and tumor grade was not significantly associated with survival. These findings underscore the prognostic relevance of elevated IRE1-IRS and PERK-IRS expression in our cohort. Similarly, in OSCC, high XBP1 expression in tumors and lymph node metastases is linked to advanced disease and poor survival, while its inhibition suppresses AXL receptor tyrosine kinase expression, reducing invasiveness (54).

Recent studies underscore innovative strategies to overcome therapeutic resistance and induce apoptosis in breast cancer by targeting ER stress pathways. The retinoid N-(4-hydroxyphenyl)retinamide (4-HPR) and its non-hydrolysable analog 4-hydroxybenzylretinone (4-HBR) provoke ER stress and concomitantly activate both intrinsic and extrinsic apoptotic cascades; notably, 4-HPR also sensitizes TRAIL/Apo2L-resistant MCF-7 cells via GADD153-dependent upregulation of death receptor 5 (DR5), even though GADD153 itself is not strictly required for 4-HPR-induced apoptosis (55). Similarly, Suwannalert et al. demonstrated that the HSP90 inhibitor 17-AAG induces apoptosis in breast cancer cells by disrupting ER protein homeostasis through PERK activation and increased eIF2α phosphorylation (56).

In summary, our findings strongly associate high IRE1 and PERK expression with aggressive tumor features, emphasizing the multifaceted role of ER stress signaling in breast cancer. Rather than isolated observations, our data align with growing evidence that the UPR exerts a complex, context-dependent influence on tumor progression, affecting proliferation, invasion, survival, and chemoresistance. The interplay between IRE1 and PERK underscores their potential as both prognostic biomarkers and therapeutic targets. However, the intricate network of UPR pathways necessitates further investigation through integrated genomic, proteomic, and functional analyses. Such studies will be crucial for elucidating precise molecular mechanisms and translating these insights into targeted therapies that overcome treatment resistance and improve patient outcomes in breast cancer.

The Authors declare that they have no conflicts of interest in relation to this study.

Conceptualization and methodology: SF, GM, KD. Data curation: SF, GM, IN, NT, GMS, EAA. Writing original draft: SF, GM. Review and editing: IN, KD, SP, CMS. Investigation, formal analysis: SF, GM, CC, ES, KF, EAA, GMS. Technical support: ES, NT. Validation: All Authors have read and agreed to the published version of the manuscript.

This research received no external funding.

No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.