Chemopreventive Effect of Bitter Melon (Momordica charantia) Juice on Pancreatic Carcinogenesis in LSL‐Kras^{G12D /+}; p48^Cre/+ Mice

ABSTRACT

Herein, using the LSL‐Kras^G12D/+; p48^Cre/+ transgenic mouse model [that recapitulates the full spectrum of human disease, progressing from early pancreatic intraepithelial neoplasia lesions to invasive pancreatic ductal adenocarcinoma (PDAC)], we determined the chemopreventive effect of long‐term (chronic) administration of Bitter Melon Juice (BMJ‐standardized lyophilized juice processed from the fruits of Momordica charantia) against pancreatic cancer. Starting at 6 weeks of age, female LSL‐Kras^G12D/+; p48^Cre/+ mice were subjected to BMJ intervention (oral gavage, dose: 200 mg/kg body wt., 5 days/week) for 35 weeks (study end: 41 weeks of age), and pathological and molecular changes in the pancreas were assessed. Results indicated that BMJ shifted the pancreatic tumor profile from moderately differentiated adenocarcinoma in controls towards a more heterogeneous profile enriched for well‐differentiated lesions indicating that prolonged exposure to BMJ bioactives results in a less aggressive phenotype. BMJ significantly reduced proliferative index (~24%) and induced a ~ 4‐fold increase in apoptosis in the pancreas. The effects on AMPK activation (~2‐fold increase) and decrease in the expression of [glucose (GLUT‐1) and lactate (MCT‐4) transporter levels (∼49%–69%,), NF‐κB/p65 (~40%), EMT marker‐Vimentin (~46%), fibrotic microenvironment marker (α‐SMA, ~29%) and Mucin‐1 (MUC‐1, ~45%)] indicate that BMJ modulates several interconnected pathways essential for PDAC growth and progression. Plasma proteomic profiling indicated that Fractalkine (CX3CL1), MIP‐3β (CCL19), MIP‐3α (CCL20), and SCF were decreased by BMJ, while Axl and TNF‐alpha were overexpressed in BMJ‐treated animals compared to positive controls. These results encourage further research into bitter melon compounds as a potential chemopreventive option for pancreatic cancer, especially for high‐risk groups.

Abbreviations

ADM

Acinar‐to‐ductal metaplasia

AMPK

AMP‐activated protein kinase

AXL

Axl receptor tyrosine kinase

BMJ

Bitter Melon Juice

CCL19 (MIP‐3β)

C‐C motif chemokine ligand 19/macrophage inflammatory protein‐3 beta

CCL20 (MIP‐3α)

C‐C motif chemokine ligand 20/macrophage inflammatory protein‐3 alpha

CCL24 (Etotaxin‐2)

C‐C motif chemokine ligand 24

CCR3

C‐C motif chemokine receptor 3

cDC1

conventional type 1 dendritic cells

CD30 Ligand (TNFSF8)

Tumor necrosis factor (TNF) ligand super family member 8

CD30 (TNFRSF8)

TNF receptor super family receptor 8

CDKN2A

Cyclin‐dependent kinase inhibitor 2 A

CD‐31 (PECAM‐1)

Cluster of differentiation 31/platelet endothelial cell adhesion molecule‐1

CX3CL1 (Fractalkine)

C‐X3‐C motif chemokine ligand 1

CXCL13 (BLC)

C‐X‐C motif chemokine ligand 13/B lymphocyte chemoattractant

CXCL1 (KC)

C‐X‐C motif chemokine ligand l/keratinocyte chemoattractant

EMT

Epithelial‐to‐mesenchymal transition

GLUT‐1

Glucose transporter‐1

IGFBP5

insulin‐like growth factor‐binding protein 5

IHC

Immunohistochemistry

IL1A

Interleukin‐1 alpha

IL1β

Interleukin‐1 beta

KRAS

Kirsten rat sarcoma viral proto‐oncogene

MCT‐4

Monocarboxylate transporter‐4

MUC‐1

Mucin‐1

NF‐κB

Nuclear factor kappa B

pAMPK

Phosphorylated AMPK

PanIN

Pancreatic intraepithelial neoplasia

PCNA

Proliferating cell nuclear antigen

PDAC

Pancreatic ductal adenocarcinoma

PSC

Pancreatic stellate cells

SCF

stem cell factor

SMAD4

Suppressor of mothers against decapentaplegic 4

TNF alpha

tumor necrosis factor alpha

TP53

Tumor protein 53

XCL1

XC motif chemokine ligand 1

α‐SMA

Alpha‐smooth muscle actin

1. Introduction

Pancreatic cancer ranks among the deadliest malignancies worldwide and is the third leading cause of cancer‐related death in the United States. An estimated 67,530 new cases are expected to be diagnosed in the year 2026, with approximately 52,740 deaths projected over the same period [1]. Pancreatic ductal adenocarcinoma (PDAC) accounts for more than 90% of pancreatic cancer cases, and its prognosis remains devastatingly poor, largely because the disease is rarely diagnosed before it has already metastasized [2]. Despite decades of modest progress in oncological treatment, pancreatic cancer continues to resist nearly every therapeutic approach developed against it. The overall survival rate has seen only marginal improvement, and the scale of clinical effort has yet to translate into meaningful outcomes [3, 4]. More than 481 Phase 1 and 85 Phase 3 clinical trials for metastatic pancreatic cancer have been completed, and five new drugs have received approval since 2000, yet median survival continues to hover below 1 year [4].

A significant contributor to this bleak reality is the anatomical location of the pancreas, nestled deep within the abdomen, which makes early detection inherently difficult. Also, PDAC in its early stages is frequently asymptomatic or presents with vague, nonspecific complaints that do not raise immediate clinical suspicion [2]. PDAC develops through the sequential accumulation of mutations in canonical driver genes (KRAS, CDKN2A, TP53, and SMAD4), progressing from microscopic precursor lesions such as pancreatic intraepithelial neoplasia (PanIN) and acinar‐to‐ductal metaplasia (ADM) to invasive and ultimately metastatic PDAC [5]. The dense desmoplastic tumor microenvironment, constitutive activation of NF‐κB, reprogrammed cancer metabolism, and epithelial‐to‐mesenchymal transition (EMT) collectively drive therapeutic resistance and poor prognosis [2, 5, 6, 7, 8, 9, 10]. Thus, there remains a critical need for safe, effective chemopreventive and intervention strategies, particularly those that can be administered to intercept pancreatic cancer growth and progression early on.

In this regard, extensive research studies by us, including studies by other labs, focused on bitter melon using standardized lyophilized bitter melon juice (BMJ) and/or bitter melon extracts derived from Momordica charantia, have shown that bitter melon has an anti‐pancreatic cancer effect [11]. Specifically, prior work from our group has demonstrated BMJ's anti‐pancreatic potential (alone and/or in combination with gemcitabine) using in vitro and in vivo preclinical models [12, 13, 14, 15, 16, 17]. However, the in vivo chemopreventive efficacy of BMJ when administered chronically has not been established. The present study addresses this critical question using the LSL‐Kras^G12D/+; p48^Cre/+ transgenic mouse model, a preclinical system that accurately recapitulates the full spectrum of human pancreatic cancer, progressing from early PanIN lesions to invasive PDAC as a function of mice age in a reproducible manner [18, 19]. Accordingly, the objective of the study was to determine the anti‐cancer effect of long‐term (chronic) administration of oral BMJ on pancreatic carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ mice.

2. Materials and Methods

2.1. BMJ Preparation

Commercially available Chinese variety bitter melon (Momordica charantia) was utilized in this study. The bitter melons were washed with water, deseeded, cut into pieces, juiced, and centrifuged as described previously [14, 15]. The resulting supernatant was subjected to sterile filtration and lyophilization. The lyophilized BMJ powder was stored in Wheaton serum vials at 4°C. For administration to the mice, the lyophilized powder was suspended in water and administered 5 days/week as oral gavage (100 μL) at a dose of 200 mg/kg body weight [14, 15].

2.2. Animal Models and Breeding

Initial breeding stock (developed on C57BL/6 background) included male LSL‐Kras^G12D/+ and female p48^Cre/+ mice procured from Prof. C.V. Rao at the University of Oklahoma Health Sciences Center [18, 19]. Non‐transgenic C57BL/6 mice used for breeding or used as wild‐type controls were initially purchased from Jackson Laboratory, (Bar Harbor, ME) and later generated by maintaining a breeding colony. This mouse study was conducted at the University of Colorado Denver, Anschutz Medical Campus (UC Denver‐AMC). All animal procedures were performed following protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Colorado Denver, Anschutz Medical Campus.

LSL‐Kras^G12D/+; p48^Cre/+ transgenic mice were generated by separately crossbreeding LSL‐Kras^G12D/+ and p48^Cre/+ mice, both on a C57BL/6 background, with wild‐type C57BL/6 mice to establish sufficient parent colonies. LSL‐Kras^G12D/+ males were then bred with p48^Cre/+ females to generate the LSL‐Kras^G12D/+; p48^Cre/+ transgenic model. The routinely obtained transgenic mice were randomized into BMJ‐treatment and disease‐positive control groups (n = 8 mice/group). LSL‐Kras^G12D/+; p48^Cre/+ female mice were subjected to BMJ exposure (oral gavage) at 6 weeks of mice age and treatment continued for 35 weeks till study end (41 weeks of mice age) for chronic exposure assessment. Positive control groups (untreated LSL‐Kras^G12D/+; p48^Cre/+) received regular drinking water. Wild‐type C57Bl/6 mice (without BMJ‐exposure) served as universal negative controls. Throughout the study, animals had access (ad libitum) to drinking water and AIN‐76A diet (Envigo). Animals were monitored daily and their body weights were recorded on a biweekly basis. No significant difference in body weight was observed between BMJ‐exposed mice compared to the untreated groups (data not shown). At study end, mice were euthanized by CO2 asphyxiation followed by exsanguination; plasma was collected, and pancreases were harvested. Pancreas [whole tissue or dissected in half (each section with a head and tail part)] was either formalin‐fixed (flat‐sandwiched in filter paper) or snap‐frozen for subsequent pathological or molecular analyses, as relevant. Archived samples related to the study were shipped, and corresponding studies were carried out at South Dakota State University.

2.3. Histopathological Analysis

Formalin‐fixed paraffin embedded tissues were processed as 5‐μm thick sections for histopathological assessment. H&E‐stained tissue sections (pancreatic head and tail) were assessed for pathological features in blinded manner and scored as 1 unit. Sections were microscopically evaluated based on updated pathological classification [20] and classified as low grade PanIN, carcinoma in situ (high grade PanIN) and pancreatic ductal adenocarcinoma (PDAC). Low‐grade PanIN lesions appeared as single ducts or as clusters of 3–4 ducts. PDAC lesions were further classified as well‐differentiated (WD), moderately differentiated (MD), or poorly differentiated (PD). Acinar to ductal metaplasia (ADM) regions were observed as ductal/ADM clusters (≥ 5 high‐grade PanIN ducts or ADM ducts as clusters); ductal/ADM clusters were additionally sub‐classified as small (≥ 5–7 count), medium (≥ 8–15 count), or large (> 15 count) clusters. Furthermore, the % area of the pancreas covered by the ADM lesions was also calculated. Lesion incidence was determined based on the highest histopathological grade identified in each mouse, with each animal assigned to its most advanced lesion category. Percent lesion incidence therefore reflected the distribution of mice according to the highest lesion grade observed.

2.4. Immunohistochemistry and Statistical Analysis

For immunohistochemical (IHC) analysis, paraffin‐embedded sections (5‐µm thick) were deparaffinized and stained using specific primary antibodies followed by 3,3′‐diaminobenzidine (DAB) staining. Primary antibodies used were PCNA (Abcam #ab29, RRID: AB_303394); platelet/endothelial cell adhesion molecule 1 (PECAM‐1)/CD‐31 (Santa Cruz Biotec., #sc‐376764, RRID: AB_2801330); Vimentin (Santa Cruz Biotec., #sc‐6260, RRID: AB_628437); α smooth muscle actin (Abcam #ab15030, RRID: AB_11155097); total NFκB/p65 (Abcam #ab16502, RRID: AB_443394); pAMPK^Thr172 (Cell Signaling Tech. #50081S, RRID: AB_2799368); cleaved Caspase‐3 (Cell Signaling Tech., #9661S, RRID: AB_2341188); MCT4 (Santa Cruz Bio. #sc‐376140, RRID: AB_2943304); GLUT‐1 (Protein Technologies #66290‐1‐Ig, RRID: AB_2881673); MUC‐1 (Abcam #ab15481, RRID: AB_301891). Biotinylated secondary antibodies (goat anti‐rabbit IgG, and goat anti‐mouse IgG) were sourced from Dako/Agilent technologies.

For PCNA and cleaved Caspase‐3 staining, positive cells were quantified by counting brown‐stained (DAB) cells among the total number of cells in five randomly selected fields at ×400 magnification. Immunoreactivity score (arbitrary values based on brown stain intensity in five randomly selected fields at ×400 magnification in heterogeneous staining and overall value in homogeneously stained) was assigned for all other markers. Both head and tail sections of the pancreas were assessed by IHC, and the average of these sections was the score assigned to an animal. The final score of a treatment group was the mean (±SEM) of the assigned scores for all animals in the group. For assessment of pAMPK^Thr172 expression, cytoplasmic staining intensity was considered, while for GLUT‐1 and MCT‐4 expression overall cellular staining intensity was observed (which included plasma membrane + cytoplasmic staining); intensity of brown staining was scored as 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). For assessment of total NF‐κB (p65) expression, cytoplasmic and nuclear staining were considered, and intensity of brown staining was scored as 0 [only blue (no brown staining), +1 (light bluish brown cytoplasmic staining), +2 (moderate brown cytoplasmic staining), +3 (moderate brown cytoplasmic and nuclear staining), and +4 (strong nuclear and cytoplasmic staining). Vimentin expression was scored based on intensity of brown cytoplasmic staining: 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). Vimentin and NF‐κB (p65) expression were observed in both stromal and ductal regions.

MUC‐1 staining was observed in the luminal (apical) side of ducts, while α‐SMA staining was observed in the extracellular matrix; both were scored based on intensity of brown stain: 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). For α‐SMA expression, the percentage area of pancreas tissue positive for DAB color was also scored as 0 (no‐staining), +1 (≤ 10% area), +2 (11%–25% area), +3 (26%–50% area), and +4 (50%–100% area). H‐Score analysis for α‐SMA expression was calculated as [% proportion area of the pancreas (positive for α‐SMA expression) x immunoreactivity score (α‐SMA intensity)]. Analysis of histopathology and immunohistochemistry data outcomes were performed using GraphPad Prism 10, with two‐sided p ≤ 0.05 considered significant. Fishers Exact test for incidence values and unpaired two‐tailed student's t‐tests were used as relevant. Microscopic analyses were conducted with a Zeiss Axio Imager A1 microscope, and photomicrographs were captured with an AxioCam MrC5 camera.

2.5. Plasma Proteomics Profiling

Equal volumes of plasma (collected at necropsy) from each mouse within the same group (BMJ, positive control, and wild‐type C57Bl/6) were pooled (n = 4) to minimize inter‐individual variability. Proteomic profiling (cytokine/chemokine‐focused) using the sandwich‐based immunoassay format (Raybiotech Mouse Cytokine Array Kit C3) was performed with 150 µL of pooled plasma per membrane, following the manufacturer's instructions. Protein levels were visualized via chemiluminescence, and the signal was captured on X‐ray film. The densities of protein spots on the film/image were analyzed using ImageJ software, normalized to control reference spots with a correction factor. Age‐matched wild‐type C57BL/6 mice served as reference without gene upregulation to calculate log2 fold changes; Qiagen Ingenuity Pathway Analysis (IPA) software was used to identify enriched pathways, networks, and upstream regulators.

3. Results

3.1. Effect on Pathological Changes in Pancreas During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice After Long‐Term (Chronic) Oral Administration of BMJ

Histopathological assessments of pancreatic tissues (Figure 1A) indicated that the positive control cohort had a higher incidence of advanced‐grade lesions (∼100% MD adenocarcinoma), whereas the BMJ‐treated group displayed ∼50% MD and ∼50% WD adenocarcinoma, indicating tumor arrest at a lower grade (Figure 1B). While low‐grade PanIN lesion counts were absent in both positive control and BMJ‐treated mice, the carcinoma in situ lesions were slightly increased by BMJ (Figure 1C). Importantly, the medium‐sized ductal/ADM clusters were also increased in BMJ (Figure 1D), while the % area of pancreas covered by the ADM lesions was also slightly increased in the BMJ‐treated female mice (∼39.4% ± 4.22 in BMJ vs. ∼33.75% ± 9.04 in the positive control group) (Figure 1E); however, it was not statistically significant.

Figure 1.

Effect of chronic BMJ administration on pancreatic pathological lesion development in LSL‐Kras^G12D/+; p48^Cre/+ mice. (A) Representative H&E pictographs (×200) of pancreas sections from positive control and BMJ‐treated female mice depicting PanIN/PDAC‐ductal clusters (block arrows), ADM region (straight line arrows), CIS (plus sign); normal acini (a), and thickened stroma (ts) characteristic features. (B) % lesion incidence (distribution of mice according to the highest lesion grade observed). (C) average lesion counts for LG PanIN subtypes (single ducts and clusters), CIS, and total ductal/ADM clusters. (D) average counts of ductal/ADM clusters by size. (E) % of pancreatic area covered by acini to ductal metaplasia (ADM). Data in C, D, and E graphs are expressed as mean ± SEM, n = 8/group. low‐grade (LG) PanIN, carcinoma in situ [CIS], well‐differentiated [WD], moderately differentiated [MD], and poorly differentiated [PD] adenocarcinoma. $, p ≤ 0.05.

Thus, chronic BMJ treatment spanning 6 to 41 weeks produced a consistent and sustained improvement in tumor differentiation throughout the treatment period. The shift from predominantly moderately differentiated adenocarcinoma in positive controls toward a more heterogeneous profile enriched for well‐differentiated lesions in BMJ‐treated animals indicates that prolonged exposure to BMJ bioactives modulates the tumor cells toward a less aggressive phenotype. This effect, which promotes differentiation, is unique to long‐term treatment, as it indicates ongoing engagement with pathways that regulate differentiation rather than a temporary interruption of signaling. The LSL‐Kras^G12D/+; p48^Cre/+ model, which advances from early PanIN to invasive PDAC within 6 to 30 weeks, offers a suitable biological context for this study [18, 19, 21, 22]. The 35‐week treatment period in the present study covers both the preneoplastic and invasive stages of the disease. The fact that BMJ can still affect differentiation even during this extended period indicates that its influence is not limited to a crucial early phase but reflects a continuous ability to alter tumor biology throughout the course of the disease evolution.

The simultaneous increase in ADM with improved differentiation in mice chronically treated with BMJ warrants further mechanistic investigation. ADM is a reversible trans‐differentiation process where acinar cells gain ductal characteristics and is recognized as an early stage in PDAC development [23]. However, not all ADM lesions develop into PanIN or carcinoma. Data from chronic BMJ treatment show that even with increased ADM frequency, these cells undergo a more controlled, less dysplastic trans‐differentiation process. This process does not lead to high‐grade disease as rapidly as observed in controls. This indicates that long‐term BMJ treatment could separate ADM formation from malignant progression. Such a distinction has important implications for analyzing chemoprevention outcomes at the histological level.

3.2. Effect on Proliferation and Apoptosis During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of BMJ

Next, we investigated whether BMJ inhibits proliferation and whether it could induce apoptosis during carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas. IHC staining for the detection of PCNA‐positive cells was performed (Figure 2A); the results indicated that the % of PCNA‐positive cells was 73.2% ± 2.9 in the positive control versus ∼55.3% ± 3.6 in the BMJ‐treated group. Thus, chronic BMJ treatment led to a significant decrease (∼24%, p ≤ 0.01) in the proliferative index during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice.

Figure 2.

Effect of chronic BMJ administration on proliferation and apoptosis in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas. Immunohistochemical (IHC) analysis of (A) proliferation marker: proliferating cell nuclear antigen (PCNA) and (B) apoptosis marker: cleaved caspase‐3 in pancreas sections from positive control and BMJ‐treated mice (6–41 weeks chronic study). Representative pictographs are shown at ×400 magnification (with digitally magnified zoomed‐in inset images). Positive cells were quantified by counting brown‐stained cells among the total number of cells at ×400. Data are expressed as mean ± SEM; IHC, 8 samples/group (heads and their associated tail sections). #, p ≤ 0.01; Ψ, p ≤ 0.02.

The ∼24% decrease in PCNA expression after long‐term BMJ administration offers strong evidence of diminished proliferative activity over the 35‐week exposure period, implying that BMJ continues to inhibit DNA replication in tumor cells throughout the disease course. PCNA acts as a processivity factor for DNA polymerase δ and is crucial for DNA replication, repair, and chromatin remodeling [24]. Higher PCNA levels in pancreatic cancer correlate with increased tumor aggression and poorer patient prognosis [9, 25].

Simultaneously, we assessed cleaved caspase‐3 as a marker of apoptotic cell death [26] (Figure 2B). Cleaved caspase‐3 was evaluated in BMJ‐treated tissues to help determine programmed cell death. BMJ treatment from 6 to 41 weeks of age increased the % of cleaved‐caspase 3‐positive cells. Specifically, there were 3.55% ± 1.9 cleaved caspase‐3‐ positive cells in the positive control versus 14.2% ± 2.2 in the BMJ‐treated group. Thus, BMJ treatment for a chronic period led to a significant increase (∼4‐fold, p ≤ 0.02) in apoptotic cell death during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice (Figure 2B).

3.3. Effect on Glucose and Lactate Transporters and Cell Metabolism Regulators During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of BMJ

A key outcome of this study was BMJ's significant influence on the cancer cell metabolism regulator AMP‐activated protein kinase (AMPK). Specifically, the immunoreactivity scores for the expression of phosphorylated AMPK (pAMPK^Thr172) were 1.66 ± 0.15 in the positive control versus 2.56 ± 0.27 in the BMJ‐treated group (Figure 3A). Thus, BMJ treatment for a chronic period led to a ∼ 2‐fold increase (p ≤ 0.05) in pAMPK^Thr172 expression during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice (Figure 3A).

Figure 3.

Effect of chronic BMJ administration on glucose and lactate transporters and cell metabolism regulators in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas. IHC analysis of (A) phosphorylated AMP‐activated protein kinase (pAMPK^Thr172), (B) glucose transporter 1 (GLUT‐1), and (C) monocarboxylate transporter 4 (MCT‐4) in pancreas sections from positive control and BMJ‐treated mice (6‐41 weeks chronic study). Representative pictographs are shown at ×400 magnification (with digitally magnified zoomed‐in inset images). For assessment of pAMPK^Thr172 expression, cytoplasmic staining intensity was considered, while for GLUT‐1 and MCT‐4 expression overall cellular staining intensity was observed (which included plasma membrane + cytoplasmic staining); intensity of brown staining was scored as 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). Data are shown as mean ± SEM; IHC, 8 samples/group (heads and their associated tail sections). #, p ≤ 0.01; $, p ≤ 0.05.

Regarding the expression of Glucose transporter GLUT‐1, the immunoreactivity scores were 2.23 ± 0.17 in the positive control versus 1.14 ± 0.2 in the BMJ‐treated group. Thus, chronic BMJ treatment led to a significant decrease (∼ 49%, p ≤ 0.01) in GLUT‐1 levels during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice (Figure 3B). Furthermore, the immunoreactivity scores for lactate transporter MCT‐4 were 1.59 ± 0.12 in the positive control versus 0.50 ± 0.25 in the BMJ‐treated group (Figure 3C). Thus, BMJ treatment for chronic period led to ∼ 69% reduction (p ≤ 0.01) in the MCT‐4 levels during carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas (Figure 3C). Taken together, these results support our previous research outcomes showing BMJ's role in modulating glucose metabolism and lactate efflux in pancreatic cancer models [13].

The simultaneous downregulation of GLUT‐1 and MCT‐4 indicates a coordinated disruption of the glycolytic phenotype. GLUT‐1 overexpression promotes higher glucose uptake to support aerobic glycolysis, and its increased levels in PDAC are associated with poor prognosis and lower survival rates [27]. Similarly, MCT‐4 facilitates lactate export, preventing the buildup of acid inside cells and sustaining glycolysis during hypoxia. Elevated MCT‐4 levels characterize a glycolytic subtype of pancreatic cancer associated with worse prognosis and distinctive metabolic requirements [28]. By simultaneously activating AMPK and inhibiting the machinery for glucose uptake and lactate export, BMJ effectively disrupts multiple aspects of cancer metabolism reprogramming.

This metabolic remodeling could lead to several downstream effects. For instance, decreased lactate production would lessen acidification in the tumor microenvironment, which might reduce lactate‐driven promotion of angiogenesis, immune suppression, and metastatic spread [29]. Furthermore, activating AMPK can trigger autophagy and reduce protein synthesis by suppressing mTOR, leading to energetic stress that specifically impacts rapidly dividing cancer cells [30].

3.4. Effect on Fibrotic Tumor Microenvironment During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of BMJ

The dense desmoplastic stroma in pancreatic cancer significantly contributes to its aggressive nature and resistance to chemotherapy [31]. Pancreatic cancer is characterized by an extensive desmoplastic reaction that constitutes up to 90% of tumor volume and significantly contributes to therapeutic resistance and poor outcomes [32]. In PDAC, the fibrotic environment is largely driven by activated pancreatic stellate cells (PSCs) that express α‐smooth muscle actin (α‐SMA). As a prominent marker of fibrosis, α‐SMA indicates the activation of PSCs, which play a crucial role in the development of the tumor stroma [33, 34].

Accordingly, we next assessed α‐SMA expression (Figure 4). Our study results indicated that the intensity of α‐SMA expression was substantially different between the positive control and the BMJ‐treated groups (3.0 vs. 2.12 ± 0.2, respectively, p≤ 0.01) (Figure 4A). However, the % proportion area of pancreas tissue staining positive for α‐SMA (arbitrary immunoreactivity scores) was not significantly different between the positive control and the BMJ‐treated group (1.75 ± 0.48 in the positive control vs. 1.5 ± 0.2 in the BMJ‐treated group) (Figure 4B). Importantly, BMJ decreased the H‐score from 5.25 to 3.16 (p ≤ 0.01) (Figure 4C). Overall, the intensity assessments indicated that chronic BMJ treatment led to a significant decrease (∼ 29%, p ≤ 0.01) in α‐SMA expression during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice.

Figure 4.

Effect of chronic BMJ administration on fibrogenic extracellular matrix and the EMT marker Vimentin in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas. IHC analysis of (A) α‐smooth muscle actin (α‐SMA) intensity, (B) % positive area of pancreas for α‐SMA expression, (C) H‐Score analysis for α‐SMA expression, (D) epithelial‐to‐mesenchymal transition (EMT) marker Vimentin in pancreas sections from positive control and BMJ‐treated mice (6‐41 weeks chronic study. Representative pictographs are shown at ×400 magnification (with digitally magnified zoomed‐in inset images). Vimentin expression was observed in both stromal and ductal regions, while α‐SMA staining was observed in the extracellular matrix; scoring was based on intensity of brown staining: 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). For α‐SMA expression, the percentage area of pancreas tissue positive for brown color was also scored as 0 (no‐staining), +1 (≤ 10% area), +2 (11–25% area), +3 (26–50% area), and +4 (50–100% area). H‐Score analysis for α‐SMA expression was calculated as [% proportion area of the pancreas (positive for α‐SMA expression) x immunoreactivity score (α‐SMA intensity)]. Data are shown as mean ± SEM; IHC, 8 samples/group (heads and their associated tail sections). *, p ≤ 0.001; #, p ≤ 0.01.

Activated PSCs are responsible for depositing extracellular matrix in PDAC, producing collagen, fibronectin, and other stromal proteins that create a dense fibrotic barrier. The desmoplastic stroma has complex and sometimes conflicting functions in pancreatic cancer biology. Although it can act as a physical barrier that limits drug penetration and contains cancer stem cells, recent research indicates that fully depleting the stroma might paradoxically speed up tumor growth in some cases [35, 36]. However, modulating rather than entirely ablating stromal components offers a more subtle therapeutic strategy.

The decrease in α‐SMA expression after BMJ treatment indicates a possible disruption of PSC activation or viability. PSC activation occurs due to paracrine signals from tumor cells, inflammatory cytokines, and growth factors such as TGF‐β, PDGF, and FGF [37]. The observed reduction in fibrosis could enhance drug delivery, decrease mechanical barriers to immune cell infiltration, and interfere with cancer cell‐stromal interactions that support tumor growth.

3.5. Effect on Epithelial to Mesenchymal Transition (EMT) Marker During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of Oral BMJ

In pancreatic cancer, EMT significantly contributes to metastasis, therapy resistance, and the emergence of cancer stem cell traits [38]. Vimentin, a type of intermediate filament protein, is a standard mesenchymal marker for EMT. Its expression in tumor epithelial cells signals mesenchymal differentiation and a more aggressive phenotype [39]. As a crucial biomarker of EMT, Vimentin is often upregulated during this process, contributing to tumor progression [40, 41]. In the current study, we assessed Vimentin expression by IHC to assess BMJ's effects on EMT (Figure 4D). Results indicated that the overall immunoreactivity score for pancreatic tissue was 2.69 ± 0.06 in the positive control versus 1.45 ± 0.09 in the BMJ‐treated group. Thus, chronic BMJ treatment led to a significant decrease (∼ 45.68%, p ≤ 0.001) in the expression of the mesenchymal marker Vimentin during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice (Figure 4D).

The Vimentin protein functions both as a biomarker and actively contributes to cancer cell migration, invasion, and interactions with the microenvironment by regulating focal adhesions and cytoskeletal dynamics [39]. EMT is regulated by transcription factors such as SNAIL, SLUG, TWIST, and ZEB1/2. These are activated by various signals, including pathways like TGF‐β, Wnt, Notch, and Hedgehog [42]. The decrease in Vimentin expression after BMJ treatment may indicate a broader suppression of EMT transcriptional programs, through the modulation of signaling pathways like NF‐κB, which we also observed to be significantly reduced in follow‐up studies.

3.6. Effect on Inflammation and Mucilage‐Associated Signaling Molecules During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of BMJ

Chronic inflammation is a key factor driving pancreatic cancer development and progression. NF‐κB, a central regulator of inflammatory responses, is constantly active in most PDAC cases and supports tumor cell survival, growth, invasion, and resistance to therapy [8, 43].

Our results indicated that the immunoreactivity scores for the expression of total NF‐κB (p65 sub‐unit) were 1.81 ± 0.19 in the positive control versus 1.07 ± 0.1 in the BMJ‐treated group (Figure 5A). Thus, BMJ treatment for chronic period led to a significant decrease (∼ 40.47%, p ≤ 0.02) in the expression of NF‐κB levels during carcinogenesis in the pancreas of LSL‐Kras^G12D/+; p48^Cre/+ mice (Figure 5A).

Figure 5.

Effect of chronic BMJ administration on NF‐κB (p65) and MUC‐1 expression in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas. IHC analysis of (A) NF‐κB (p65) and (B) MUC‐1 in pancreas sections from positive control and BMJ‐treated mice (6‐41 weeks chronic study). Representative pictographs are shown at ×400 magnification (with digitally magnified zoomed‐in inset images). NF‐κB (p65) expression was observed in both stromal and ductal regions, and both cytoplasmic and nuclear staining were considered; intensity of brown staining was scored as 0 [only blue (no brown staining), +1 (light bluish brown cytoplasmic staining), +2 (moderate brown cytoplasmic staining), +3 (moderate brown cytoplasmic and nuclear staining), and +4 (strong nuclear and cytoplasmic staining). MUC‐1 staining was observed in the luminal (apical) side of ducts, and intensity of brown stain was scored as: 0 [only blue (no brown staining)], +1 (light bluish brown), +2 (light brown), +3 (moderate brown), and +4 (dark brown). Data are shown as mean ± SEM; IHC, 8 samples/group (heads and their associated tail sections). *, p ≤ 0.001; Ψ, p ≤ 0.02.

NF‐κB activation in PDAC occurs via several mechanisms, such as oncogenic Kras signaling, cytokine loops that are autocrine and paracrine, and interactions within the stroma. Once activated, NF‐κB triggers the transcription of genes that promote inflammation (IL‐6, IL‐8, TNF‐α), cell survival (BCL‐2, BCL‐XL, survivin), proliferation (cyclin D1, c‐MYC), angiogenesis (VEGF), and metastasis (MMP‐2, MMP‐9) [44]. NF‐κB is especially relevant to therapeutic resistance, as it has been linked to gemcitabine resistance by increasing the levels of anti‐apoptotic proteins and drug efflux pumps like P‐glycoprotein [45]. By NF‐κB suppression, BMJ could make cancer cells more responsive to apoptotic signals, decrease the secretion of cytokines that promote tumors, and interfere with the inflammatory microenvironment that fosters tumor development.

Furthermore, our results indicated that the immunoreactivity scores for the expression of mucin 1 (MUC‐1) were 3.63 ± 0.17 in the positive control versus 1.98 ± 0.16 in the BMJ‐treated group (Figure 5B). Thus, chronic BMJ treatment led to a significant decrease (∼ 45%, p ≤ 0.001) in the expression of MUC‐1 during carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ mice pancreas (Figure 5B).

MUC‐1 is a transmembrane mucin that is abnormally overexpressed and glycosylated in pancreatic cancer, contributing to tumor growth, invasion, and immune evasion [46]. MUC‐1 interacts with growth factor receptors such as EGFR and HER2, enhancing their signaling and supporting oncogenic transformation [47]. The protein's cytoplasmic domain can move to the nucleus and bind with transcription factors like NF‐κB, further connecting MUC‐1 to inflammatory signaling [48].

Importantly, MUC‐1 has been identified as a mediator of gemcitabine resistance in pancreatic cancer. Our previous work demonstrated that gemcitabine treatment increases MUC‐1 expression in patient‐derived xenograft models, while BMJ combination therapy reduces MUC‐1 levels and overcomes drug resistance [12]. The sustained MUC1 reduction observed after chronic BMJ treatment suggests that this compound might prevent or slow down the development of chemoresistance, which is especially important considering the widespread emergence of gemcitabine resistance in PDAC patients.

3.7. Effect on Angiogenesis During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice Pancreas After Long‐Term (Chronic) Oral Administration of BMJ

CD‐31, also called platelet endothelial cell adhesion molecule‐1 (PECAM‐1), is a transmembrane glycoprotein found on vascular endothelial cells and is a widely used IHC marker for tumor angiogenesis. In tissue sections, CD‐31 primarily identifies endothelial cells and helps assess tumor angiogenesis, which can indicate rapid tumor growth. Unlike many solid tumors, PDAC has sparse, only partially functional blood vessels because the fibrotic stroma hampers blood vessel formation and function, thereby reducing drug delivery and the effectiveness of systemic chemotherapy [49].

In the present study, we assessed CD‐31 expression in pancreas tissues/tumors via IHC to evaluate the chronic chemopreventive effects of BMJ on tumor vasculature in LSL‐Kras^G12D/+; p48^Cre/+ mice. Chronic BMJ treatment from 6 to 41 weeks resulted in a ∼ 10.8% decrease in CD‐31 expression; immunoreactivity scores were 2.87 ± 0.07 in the positive control versus 2.56 ± 0.28 in the BMJ‐treated mice. However, this difference was not statistically significant (p = 0.349) (data not shown). The absence of a statistically significant reduction in CD‐31 suggests the likelihood of a hypovascular nature of PDAC, in which large shifts in microvessel density are not typically observed. Nonetheless, the lower vessel count in the BMJ‐treated group suggests a trend toward reduced angiogenic activity, which, when considered alongside the significant suppression of NF‐κB observed in this study, may reflect an early or partial disruption of pro‐angiogenic signaling, given that NF‐κB directly regulates VEGF transcription and other mediators of endothelial cell recruitment. Future studies with larger cohorts or complementary markers such as VEGF or Nestin may better capture the anti‐angiogenic potential of long‐term BMJ treatment in this model.

3.8. Effect on Plasma Cytokine/Chemokine Profile During Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice After Long‐Term (Chronic) Oral Administration of BMJ

Cytokines/chemokines and associated molecules differentially expressed in plasma from LSL‐Kras^G12D/+; p48^Cre/+ mice after long‐term chronic oral BMJ administration are shown in Figure 6. Expression of Eotaxin‐2 (CCL24) increased in wild‐type and positive control mice; however, BMJ significantly reduced expression (positive control vs. BMJ‐treated: log2FC = −1.03). Ltn (XCL1) was consistently under expressed in both positive control and BMJ‐treated groups compared with wild‐type mice (positive control vs. BMJ‐treated: log2FC = −0.76). Various cytokines and chemokines, including Fractalkine (CX3CL1), MIP‐3 beta (CCL19), MIP‐3 alpha (CCL20), and SCF, showed overexpression in the positive control group compared with wild‐type mice. Notably, in the BMJ‐treated group, these cytokines were under expressed compared with the positive control group (log2FC: CX3CL1/Fractalkine = −0.81; CCL19/MIP‐3β = − 0.30; CCL20/MIP‐3α = − 1.00; SCF = −1.78). CD30 Ligand (TNFSF8) and CD30 (TNFRSF8) exhibited consistent overexpression in both positive control and BMJ‐treated groups compared with wild‐type mice (positive control vs. BMJ‐treated: CD30/TNFRSF8, log2FC = +1.70). BLC (CXCL13) and CD40 (TNFRSF5) exhibited overexpression in BMJ‐treated groups compared with both positive control and wild‐type mice (positive control vs. BMJ‐treated log2FC: BLC/CXCL13 = +1.14; CD40 (TNFRSF5) = +1.00). Interestingly, Axl and TNF‐alpha were significantly overexpressed with BMJ treatment compared with the positive control group (log2FC: Axl = +1.32; TNF‐alpha = +1.09).

Figure 6.

Effect of chronic BMJ administration on plasma cytokine and chemokine profile in LSL‐Kras^G12D/+; p48^Cre/+ mice. (A) Heatmap displaying integrated density values of cytokines/chemokines and associated molecules across wild‐ type (WT: C57Bl/6 mice), positive control (untreated mice), and BMJ (treated mice). (B) Bar chart showing log2 fold change (log2FC) of differentially expressed plasma proteins in positive control versus BMJ‐treated group. (C) Venn diagram illustrating the overlap of differentially expressed proteins (|log2FC | ≥ 1; total = 17) among the three pairwise comparisons: wild type versus positive control; positive control versus BMJ‐treated; BMJ‐treated versus wild‐type.

Next, using these log2 fold changes for cytokine‐ and chemokine‐associated molecules, we performed pathway prediction analysis using the Ingenuity Pathway Analysis (IPA) software (Supplemental Figure S1). The IPA network analysis identified TNF and IL1A as highly upregulated hub proteins within the chronic inflammatory network (Supplemental Figure S1), consistent with the established KrasG12D‐driven feedforward loop in which IL‐1 and TNF activate NF‐κB to amplify pro‐inflammatory cytokine secretion [50]. Bitter melon bioactives have previously been shown to suppress NF‐κB, TNF‐alpha, IL‐1β, and downstream chemokine signaling in inflammatory models, supporting the biological plausibility of BMJ's modulation of these chronic inflammatory networks [51, 52]. Additional networks revealed upregulation of the Axl receptor tyrosine kinase and downregulation of leptin and IGFBP5, along with convergence of the MAPK/PI3K/NF‐κB pathways and a fifth network featuring non‐coding RNAs and TGFB1, underscoring the multi‐pathway oncogenic and epigenetic signaling landscape modulated by BMJ over the chronic period.

4. Discussion

From a translational standpoint, the chronic treatment protocol used in this study (lasting 6–41 weeks of mice age) simulates a preventive scenario where intervention takes place during the formation and early growth of premalignant lesions. The LSL‐Kras^G12D/+; p48^Cre/+ model develops PanIN lesions starting at approximately 6–8 weeks of age, progressing to invasive PDAC by 20–30 weeks [18, 19, 21, 22]. Our treatment, starting at 6 weeks, covers both preventing early lesion formation and intervening in established PanIN development. Our earlier research also showed that BMJ can improve gemcitabine effectiveness and overcome resistance mechanisms [12, 16], highlighting its potential as a complementary therapy for established disease. The multifaceted effects of BMJ observed in this chronic administration study appear to converge on several core hallmarks of cancer [53].

In the chronic plasma cytokine study spanning 6 to 41 weeks, the key finding was that Eotaxin‐2 (CCL24) was detected in all three groups (wild type, positive control, BMJ‐treated). BMJ treatment significantly decreased Eotaxin‐2 (CCL24) levels compared to the positive control and wild type; however, because the wild‐type group also showed increased Eotaxin‐2 (CCL24), the BMJ‐induced decrease cannot be correlated with an anti‐cancer effect. (Note: Eotaxin‐2 increase could possibly be an age‐related effect but needs further investigation.) Eotaxin‐2 (CCL24) is a CC‐family chemokine that promotes tumorigenesis through M2 macrophage polarization, angiogenesis, and cancer cell invasion via its receptor CCR3. Its reduction by BMJ suggests attenuation of eosinophil‐recruiting chemotactic signaling within the pancreatic TME [54]. Ltn (XCL1) was consistently under expressed in both the positive control and BMJ‐treated groups relative to wild type, indicating that impaired XCL1‐mediated cDC1 recruitment and cytotoxic T cell priming are a stage‐independent feature of KrasG12D‐driven disease that BMJ does not restore [55]. Among the differentially expressed molecules, Fractalkine (CX3CL1), MIP‐3β (CCL19), MIP‐3α (CCL20), and SCF were overexpressed in the positive control group and suppressed by BMJ treatment, while CD30 (TNFRSF8) and CD30 Ligand (TNFSF8) remained elevated in both positive control and BMJ‐treated groups, and Axl and TNF‐ alpha were overexpressed in BMJ‐treated animals compared to positive controls. The IPA networks reveal a multi‐layered signaling landscape in which T‐cell activation and cytokine‐mediated immune communication, TNF/IL1A‐driven chronic inflammation with matrix remodeling, kinase‐centric dysregulation of growth factors and Axl activation, MAPK‐dominated survival and cell death signaling, and TNF/IL1β ‐driven inflammatory epigenetic regulation with inflammation‐metabolism crosstalk and TGFβ1 fibrogenic signaling collectively define the chronic PDAC microenvironment in this transgenic model. BMJ treatment modulates multiple nodes across this interconnected signaling architecture, providing a mechanistic basis for its multi‐target chemopreventive and intervention activity.

Thus, several molecular targets and pathways could potentially contribute to BMJ's chemopreventive effects. The activation of AMPK appears to be a crucial upstream event that may lead to metabolic reprogramming, inhibition of mTOR, and subsequent effects on cell growth and survival. Also, AMPK activation has been demonstrated to suppress NF‐κB signaling via various mechanisms, such as direct phosphorylation of IKK and regulation of SIRT1 [56], potentially linking the metabolic and inflammatory effects we observed. Additionally, the decrease in NF‐κB activity might lead to reduced EMT, since NF‐κB directly regulates genes for EMT‐inducing transcription factors and cytokines that facilitate mesenchymal transition, which was seen in breast cancer [57]. The decrease in stromal fibrosis and EMT marker implies that BMJ might interfere with the mutual interactions between cancer cells and activated stroma, which promote tumor progression. The simultaneous regulation of multiple pathways suggests that BMJ works via multiple mechanisms rather than targeting a single pathway, which could account for its broad‐ranging effectiveness across different facets of pancreatic cancer biology.

The bioactive compounds responsible for BMJ's effects potentially include cucurbitane‐type triterpenoids, phenolic compounds, and various glycosides identified in bitter melon that exhibit anticancer properties through diverse mechanisms including cell cycle arrest, apoptosis induction, and inhibition of oncogenic signaling pathways [11]. Future research should aim to pinpoint the specific bioactive components and evaluate how each contributes to the observed chemopreventive effects. Thus, the safety and palatability of dietary options like BMJ offer notable benefits over synthetic chemotherapeutic drugs for chemoprevention in high‐risk groups. Nevertheless, moving to clinical use calls for careful attention to dosing, bioavailability, standardization of active ingredients, and potential interactions with existing treatments.

Author Contributions

Concept and Design: Rajesh Agarwal, Komal Raina. Development of Methodology: Komal Raina, Rajesh Agarwal, Deepanshi Dhar, Lakshmi Sai Pratyusha Bugata, Robin Kumar. Performed Experiments: Deepanshi Dhar, Rama Kant, Lakshmi Sai Pratyusha Bugata, Komal Raina, Md. Imtiazul Kabir, Robin Kumar. Acquisition of Data: Deepanshi Dhar, Rama Kant, Lakshmi Sai Pratyusha Bugata, Komal Raina.Analysis and Interpretation of Data: Lakshmi Sai Pratyusha Bugata, Komal Raina. Write, revise/review of MS: Lakshmi Sai Pratyusha Bugata, Komal Raina, Deepanshi Dhar, Rama Kant, Md. Imtiazul Kabir, Robin Kumar, Chapla Agarwal, Rajesh Agarwal. Administrative, technical, and material support: Rajesh Agarwal, Chapla Agarwal, Komal Raina. Study Supervision: Rajesh Agarwal, Komal Raina.

Ethics Statement

All animal studies were performed under Institutional guidelines using approved Institutional Animal Care and Use Committee (IACUC) protocol [# B‐57916(03)1E (#00074); approval date (03/01/2016)] in the specific pathogen‐free facility at UC Denver‐AMC. No human/patient samples were part of the study.

Conflicts of Interest

The authors declare no conflicts of interest.

AI Tools Used

Grammarly was used to assist with proofreading, focusing on correcting grammar and rephrasing sentences as needed.

Supplemental Data

Supplemental Figures are available as an addendum.

Supporting information

Figure S1: Ingenuity Pathway Analysis (IPA) network analysis of plasma cytokines/chemokines after chronic BMJ administration in LSL‐Kras^G12D/+; p48^Cre/+ mice. IPA network showing (A) Ligand‐GPCR complexes and associated signaling interactions among differentially expressed cytokines/chemokines; (B) IL‐10‐downregulated genes for extracellular proteins and scavenger receptor class A activity; (C) NF‐κB and inflammatory cytokine signaling cascades; (D) Collagen, integrin, and extracellular matrix (ECM) interactions; (E) miRNA‐mediated and metabolite‐associated regulatory interactions. Color coding: red = increased measurement, green = decreased measurement; orange ellipses = predicted activation, blue ellipses = predicted inhibition; glow indicates activity opposite to measurement direction.

Bugata L. S. P., Dhar D., Kant R., et al., “Chemopreventive Effect of Bitter Melon (Momordica Charantia) Juice on Pancreatic Carcinogenesis in LSL‐Kras^G12D/+; p48^Cre/+ Mice,” Molecular Carcinogenesis 65 (2026): 1383‐1397. 10.1002/mc.70164.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.