Inhibition of autotaxin activity with IOA‐289 decreases fibrosis in mouse E0771 breast tumors

Abstract

Tumor‐associated fibrosis contributes to an immunosuppressive microenvironment that hinders effective anti‐tumor immune responses. This study investigates the potential of IOA‐289, a novel autotaxin (ATX) inhibitor, which blocks lysophosphatidate (LPA) production and signaling, in modulating fibrosis in breast tumors. Bioinformatic analysis of human breast tumors revealed a strong correlation between levels of LPA_1,‐4 receptors and extracellular matrix (ECM) genes. Interaction of ECM molecules and integrin β1/CD44 between myofibroblasts and other cell types had the highest contribution to cell–cell communication. We showed that LPA induced α‐smooth muscle actin mRNA in mouse mammary fibroblasts and increased expressions of collagen type‐I α1 chain (COL1A1) and lamininγ1. IOA‐289 decreased the expressions of COL1A1, fibronectin‐1, and transforming growth factor β1 (TGFβ1) in E0771 breast tumors in mice. Masson's trichrome staining revealed a marked decrease in collagen deposition within breast tumors of IOA‐289‐treated mice. Decreased tumor fibrosis aligns with previous findings that IOA‐289 enhanced the infiltration of CD8⁺ cytotoxic T cells and decreased fibrotic factors including leukemia inhibitory factor and transforming growth factor‐beta1 in tumors. We also demonstrated that E0771 cells express negligible ATX and LPA receptors. Therefore, ATX inhibition did not affect cancer cells directly in our model. These results underscore the potential of ATX inhibitors in reprogramming the tumor microenvironment to favor anti‐tumor immunity and attenuate fibrosis. ATX inhibitors are in clinical trials for treating idiopathic pulmonary fibrosis and pancreatic cancer. Our results support the development of ATX inhibitors as a strategy for improving the treatment of breast cancer and other diseases involving fibrosis.

What's new?

Lysophosphatidate is one of the critical mediators of fibrosis formation in the tumor microenvironment, which hinders immune surveillance and therapy. Building on their previous finding that blocking lysophosphatidate production with the autotaxin inhibitor, IOA‐289 decreased tumor growth by approximately 60%, here the authors show that mouse breast cancer cells cannot respond directly to lysophosphatidate signaling due to the lack of receptors. Instead, lysophosphatidate signaling plays a crucial role in the activation of tumor‐associated fibroblasts and production of extracellular matrix molecules. The findings underscore the potential of autotaxin inhibitors to reprogram the tumor microenvironment to attenuate fibrosis and favor anti‐tumor immunity.

Abbreviations

ATX

autotaxin

ECM

extracellular matrix

EMT

epithelial‐mesenchymal transition

FBSC

Charcoal‐treated fetal bovine serum

GVSA

gene set variation analysis

LIF

leukemia inhibitory factor

LPA

lysophosphatidate, lysophosphatidic acid

LPC

lysophosphatidylcholine

MF

mammary fibroblast

MMP

matrix metalloproteinase

MMT

macrophage‐myofibroblast transition

PCA

principal component analysis

TGFβ1

transforming growth factorβ1

TME

tumor microenvironment

TNF

tumor necrosis factor

tSNE

t‐distributed stochastic neighbor embedding

αSMA

α smooth muscle actin

1. INTRODUCTION

Breast cancer is one of the most prevalent cancers worldwide, characterized by its diverse histopathological features and complex interactions within the tumor microenvironment (TME). ¹ , ² The TME, which comprises various stromal cells, extracellular matrix (ECM) components, and immune cells, plays a pivotal role in tumor progression, metastasis, and therapeutic resistance. ³ A key aspect of the TME in breast cancer is the presence of tumor fibrosis, driven by the excessive deposition of ECM proteins such as collagen and fibronectin. ⁴ Tumor fibrosis not only enhances the physical barriers that protect cancer cells from immune surveillance and therapy, but it also facilitates a pro‐tumorigenic environment. ⁵

Among the critical mediators of fibrosis in the TME is lysophosphatidate (lysophosphatidic acid [LPA]), which signals through a family of six G‐protein‐coupled LPA receptors (LPARs). ⁶ Extracellular LPA is generated primarily by autotaxin (ATX), a secreted enzyme that converts lysophosphatidylcholine (LPC) into LPA. ⁷ The LPA–LPAR signaling axis has been implicated in various processes associated with several diseases, including cancer progression involving cell proliferation, survival, migration, inflammation, and fibrosis. ⁸ , ⁹ , ¹⁰ , ¹¹ It is significant that in breast tumors that ATX is produced mainly by fibroblasts and endothelial cells, with negligible production by the breast cancer cells. ¹² , ¹³ Importantly, LPA signaling downstream of ATX secretion also promotes the activation of fibroblasts into myofibroblasts, cells that are major contributors to ECM remodeling and fibrosis in tumors. ¹⁴ , ¹⁵ , ¹⁶

In recent years, the therapeutic targeting of ATX and LPA signaling has gained attention as a potential strategy to improve tumor fibrosis ¹⁷ as well as in improving the efficiency of chemo‐ and radiotherapy in breast cancer. ¹⁸ , ¹⁹ , ²⁰ IOA‐289, cambritaxestat, a potent ATX inhibitor, has demonstrated promising anti‐tumor effects in preclinical models. ²¹ , ²² Our previous study showed that IOA‐289 treatment not only decreased the growth of E0771 breast tumors, but it also led to reductions in the protein levels of leukemia inhibitory factor (LIF) and transforming growth factor beta 1 (TGFβ1), two cytokines known to induce fibroblast activation. ¹³ , ²³ , ²⁴ , ²⁵ Additionally, IOA‐289 treatment increased the infiltration of CD8⁺ cytotoxic T cells in tumors, suggesting a more immunogenic TME following ATX inhibition. ¹³ , ²²

Despite these encouraging findings, the precise mechanisms by which IOA‐289 exerts its anti‐tumor effects remain to be fully elucidated, particularly concerning its impact on tumor fibrosis and ECM remodeling. LPA signaling occupies a central role in fibrosis ²⁶ and the remodeling of the TME. ²⁷ It is, therefore, critical to understand how IOA‐289 influences these processes and whether its effects in decreasing tumor growth are mediated through direct inhibition of LPA signaling in cancer cells or through modulation of the stromal compartment.

In this study, we investigated the effects of IOA‐289 on fibrosis within E0771 breast tumors. We focused on the functional status of LPA signaling in breast cancer cells and mammary fibroblasts (MF), as well as its impact on ECM of the TME. By analyzing the cell–cell communication profile in breast tumors, we aimed to elucidate how IOA‐289 modulates the TME and contributes to decreased tumor growth. Our present results show that inhibiting ATX activity in fibroblasts with IOA‐289 decreased fibrosis in E0771 breast tumors. This finding provides additional insights into the therapeutic potential of ATX inhibition in decreasing breast tumor growth and enhancing anti‐tumor immunity.

2. MATERIALS AND METHODS

2.1. Bio‐informatic analysis

Gene expression analysis was conducted using the Gene Expression Profiling Interactive Analysis (GEPIA) web tool, which leverages data from The Cancer Genome Atlas (TCGA) and Genotype‐Tissue Expression (GTEx, RRID: SCR_018294) projects. Expression levels of specific genes were compared between normal breast tissues and breast tumors to assess differential expression patterns. In particular, the analysis focused on key genes encoding ECM components, such as COL1A1 (collagen type‐I α1 chain), COL1A2 (collagen type‐I α2 chain), COL2A1 (collagen type‐II α1 chain), COL3A1 (collagen type‐III α1 chain), and FN1 (fibronectin 1). Additionally, GEPIA was employed to investigate the correlation between LPA receptors and the expression of these ECM genes. Spearman correlation coefficients were used to evaluate the strength and direction of these associations.

Single‐cell RNA sequencing (scRNA‐seq) data analysis and cell–cell communication analysis were conducted using the publicly available dataset GSE199219 and R packages Seurat (RRID: SCR_016341) and CellChat (RRID: SCR_021946). Cells with fewer than 200 or more than 5000 detected genes, as well as cells with a mitochondrial gene expression exceeding 25%, were excluded from the analysis. Principal component analysis (PCA) was conducted to reduce dimensionality after log normalization and scaling of the data. We chose 12 as the number of principal components based on the elbow plot and jackstraw method. Cells were then clustered using the Louvain algorithm implemented in the Seurat package, with a resolution parameter of 0.5 to capture biologically relevant clusters. The data were visualized by t‐distributed stochastic neighbor embedding (tSNE). Clusters were annotated based on known marker genes: PTPRC, CD3D, CD3E, and CD8A for cytotoxic T cells; PTPRC, CD3D, CD3E, and CD4 for T helper cells (Th); PTPRC, CD3D, CD3E, CD4, and FOXP3 for regulatory T cells; PTPRC and CD14 for macrophages; PDGFRB, COL3A1, FN1, and ACTA2 for myofibroblasts; PECAM1 and VWF for endothelial cells; EPCAM, KRT8, and KRT18 for epithelial (cancer) cells; AIF1 and CD79A for plasma cells; CD79A and MS4A1 for B‐cells; EPCAM, KRT8, KRT18, and VIM for epithelial cells undergoing epithelial–mesenchymal transition (EMT); PTPRC, CD14, COL3A1, and ACTA2 for macrophages undergoing macrophage–myofibroblast transition.

The whole CellChatDB database was used for analyzing cell–cell communication at single‐cell level. The data for each signaling pathway, which defined as all communication probability among all the pairs of cell groups in inferred network, were calculated and visualized.

2.2. Isolation of mouse mammary fibroblasts

Mouse MF were isolated as described previously ²⁸ with modification. Briefly, mammary fat pads were harvested from 8‐ to 12‐week‐old female C57BL/6J mice and immediately placed in ice‐cold PBS. The tissues were minced into small pieces and subjected to enzymatic digestion using a cocktail of three enzymes: 1 mg/mL collagenase type XI (C7657, Millipore Sigma, Oakville, ON, Canada), 0.1 mg/mL hyaluronidase type V (C5138, Millipore Sigma), and 20 U/mL DNase I (E091, Applied Biological Materials Inc., Richmond, BC, Canada) in serum‐free Dulbecco's minimum essential medium (DMEM). The tissue‐enzyme mixture was incubated at 37°C in a shaker for 3 h. The cell suspension was collected carefully, avoiding the top fat layer, filtered through a 70 μm cell strainer to remove undigested tissue clumps, and then centrifuged at 400g for 5 min. The cell pellet was washed twice with phosphate‐buffered saline (PBS) and then resuspended in complete DMEM medium supplemented with 10% fetal bovine serum (FBS). Cells were placed in a 10‐cm dish and incubated at 37°C with an atmosphere of 5% CO₂. After 24 h, non‐adherent cells were removed by washing with PBS, and the medium was replaced. Cells remaining on the dish were primary fibroblasts and used for experiments between Passages 2 and 5.

2.3. Cell culture and stimulation

E0771 (CRL‐3461, ATCC, Manassas, VA, USA, RRID: CVCL_GR23) and 4T1 (CRL‐2539, ATCC, RRID: CVCL_0125) mouse breast cancer cells, NIH3T3 fibroblasts (CRL‐1658, ATCC, RRID: CVCL_0594), and primary mouse MF were cultured in DMEM (supplemented with 10% FBS. Hs578Bst human breast fibroblasts (HTB‐125, ATCC, RRID: CVCL_0807) and matched Hs578T human breast cancer cells (HTB‐126, ATCC, RRID: CVCL_0332) were cultured in DMEM medium supplemented with 10% FBS, 0.01 mg/mL insulin (I2643, Millipore Sigma), and 30 ng/mL EGF (E4127, Millipore Sigma). Cells were maintained at 37°C in a humidified atmosphere with 5% CO₂. All cell lines were authenticated using short tandem repeat (STR) profiling within the last 3 years. All experiments were performed with mycoplasma‐free cells.

For stimulation with LPA or TGFβ1, cells were seeded in 12‐well plates at 2 × 10⁵ cells per well. Cells were serum‐starved for 4 h when they reached 70%–80% confluence and treated with 5–20 μM LPA (857128P, Millipore Sigma) or 10 ng/mL TGFβ1 (5231LF, New England Biolabs, Whitby, Ontario, Canada) for different times. Some of the cells were pretreated with 0.5 μM Ki16425, an inhibitor of LPA₁ and ₋₃ receptors (10012659, Cayman Chemical, Ann Arbor, Michigan, USA) or 5 μM parthenolide, an inhibitor of NFκB (70080, Cayman Chemical) for 30 min before stimulation.

2.4. Cell co‐culture

Co‐culture experiments were performed using Hs578Bst and Hs578T cells. Cells were co‐cultured in six‐well plates using inserts with 0.4 μm pore size (12‐565‐012, Fisher Scientific, Ottawa, Ontario, Canada). To initiate the co‐culture, Hs578T breast cancer cells were seeded into the lower compartment of the six‐well plate at a density of 1 × 10⁵ cells/well. Hs578Bst fibroblasts were seeded into the upper compartment of inserts at a density of 5 × 10⁴ cells. Both types of cells were co‐cultured in the same wells with DMEM medium supplemented with 10% charcoal‐treated FBS (FBSC) where lipids including LPA were removed, ²⁹ 0.01 mg/mL insulin, 30 ng/mL EGF, and 50 μΜ LPC (855575P, Millipore Sigma). Some of the cells were treated with 100 nM IOA‐289 (iOnctura, Genève, Switzerland). The co‐cultures were maintained at 37°C in a humidified incubator with 5% CO₂ for 48 h.

2.5. Syngeneic orthotopic breast tumor model in mice

A syngeneic orthotopic mouse breast cancer model was established as previous. ¹³ Briefly, 1 × 10⁶ E0771 breast cancer cells suspended in PBS with 50% Matrigel (354230, Corning, Glendale, Arizona, USA) were injected into the fourth left mammary fat pad of female C57BL/6J mice (strain#: 000664, The Jackson Laboratory, Bar Harbor, ME, USA). Mice were randomly assigned to the test and control groups and housed under 12 h light/dark cycles with free access to water and standard mouse chow containing 4% fat. IOA‐289 was ground into a fine powder in a mortar and suspended at 20 mg/mL in 0.5% methyl cellulose (182312500, Acros Organics, Morris Plains, NJ, USA). IOA‐289 treatment began 2 days post‐injection, with mice receiving oral gavage of 100 mg/kg IOA‐289 or vehicle twice daily until sacrifice on Day 20. This twice per day dosage of IOA‐289 was established previously because ATX activity recovered by ~88% over the 12 h between doses. ¹³ The IC50 for IOA‐289 is about 36 nM. ²²

2.6. Quantitative PCR (qPCR)

Extraction of mRNA was performed using the EZ‐10 DNAaway RNA miniprep kit (BS88136, Bio Basic Inc., Markham, ON, Canada) followed by reverse transcription using the All‐In‐One 5× RT Master Mix (G490, Applied Biological Materials Inc.). qPCR was performed using BlasTaq 2× qPCR Master Mix (G892, Applied Biological Materials Inc.). The primer sequences for human ATX (ENPP2) forward: 5′ACAACGAGGAGAGCTGCAAT3′ and reverse: 5′AGAAGTCCAGGCTGGTGAGA′. Human αSMA (ACTA2) forward: 5′AATGCAGAAGGAGATCACGG3′ and reverse: 5′TCCTGTTTGCTGATCCACATC3′. Human TGFβ1 (TGFB1) forward: 5′GCCTTTCCTGCTTCTCATGG3′ and reverse: 5′TCCTTGCGGAAGTCAATGTAC3′. Mouse αSMA (ACTA2) forward: 5′GTGAAGAGGAAGACAGCACAG3′ and reverse: 5′GCCCATTCCAACCATTACTCC3′. Mouse collagen Iα1 (COL1A1) forward: 5′CATAAAGGGTCATCGTGGCT3′ and reverse: 5′TTGAGTCCGTCTTTGCCAG3′. Mouse collagen IIIα1 (COL3A1) forward: 5′GAAGTCTCTGAAGCTGATGGG3′ and reverse: 5′TTGCCTTGCGTGTTTGATATTC3′. Mouse lamininα1 (LAMA1) forward: 5′AAAGGAAAGTGTCAGTACCAGG3′ and reverse: 5′TTCTCTAAGCATCGCAAGGG3′. Mouse lamininγ1 (LAMC1) forward: 5′CCCAACTCCATCAACCTCAC3′ and reverse: 5′GTACTGATAAGGAATCCAGGGC3′. Mouse fibronectin (FN1) forward: 5′CTTTGGCAGTGGTCATTTCAG3′ and reverse: 5′ATTCTCCCTTTCCATTCCCG3′. Mouse TGFβ1 (TGFB1) forward: 5′CCTGAGTGGCTGTCTTTTGA3′ and reverse: 5′GTGGAGTTTGTTATCTTTGCTG3′. Mouse LIF forward: 5′GAGTCCAGCCCATAATGAAGG3′ and reverse: 5′ACAGGTGGCATTTACAGGG3′. Mouse LPAR1 forward: 5′CTATGTTCGCCAGAGGACTATG3′ and reverse: 5′GCAATAACAAGACCAATCCCG3′. Mouse LPAR2 forward: 5′CACACTCAGCCTAGTCAAGAC3′ and reverse: 5′GTACTTCTCCACAGCCAGAAC3′. Mouse LPAR3 forward: 5′GCCCGGTGTGCAATAAAA3′ and reverse: 5′CTTAAAAGCCCCAGAAGTGATG3′. Mouse LPAR4 forward: 5′ACAGGCATGAGCACATTCTC3′ and reverse: 5′TGGAGGCAGACGATCAGAG3′. Mouse LPAR5 forward: 5′CTAGTGTGCAGAAGGGAACAA3′ and reverse: 5′CTACATACTGGAGCAGGGTTTC3′. Mouse LPAR6 forward: 5′CACATCTGAATAGCAAAGGCG3′ and reverse: 5′TGAACATGCACCCGTACAG3′. GAPDH was used as an internal control. The primers for human and mouse GAPDH are forward: 5′ACTTTGTCAAGCTCATTTCC3′ and reverse: 5′TCTTACTCCTTGGAGGCCAT3′.

2.7. Western blotting

The levels of specific proteins were measured by western blotting as described previously. ³⁰ Rabbit anti‐collagen Iα1 (72026, Cell Signaling Technology, Danvers, MA, USA, RRID: AB_2904565), rabbit anti‐lamininγ1 (45788, Cell Signaling Technology), rabbit anti‐αSMA (19245, Cell Signaling Technology, RRID: AB_2734735), mouse anti‐phospho‐AKT (4051, Cell Signaling Technology, RRID: AB_331158), rabbit anti‐AKT (4691, Cell Signaling Technology, RRID: AB_915783), mouse anti‐phospho‐ERK (9106, Cell Signaling Technology, RRID: AB_331768), rabbit anti‐ERK (9102, Cell Signaling Technology, RRID: AB_330744) antibodies were used. Immunoblots were analyzed by the Odyssey infrared imaging system (LI‐COR Biosciences, Lincoln, Nebraska, USA).

2.8. Multiplex protein measurement

Matrix metalloproteases (MMPs) in tumor tissue were measured using a multiplexing laser bead assay by Eve Technologies (Calgary, Alberta, Canada) as reported previously. ¹³ Sample preparation was performed following the posted instructions (https://www.evetechnologies.com/sample-preparation-guide/).

2.9. Masson's trichrome staining

Formalin‐fixed, paraffin‐embedded tissue samples were sectioned at 5 μm thickness and mounted on glass slides. Sections were treated with xylene and rehydrated through a graded series of ethanol. ³¹ After rehydration, slides were stained using a Masson's Trichrome Staining Kit (87019, Thermo Scientific, Waltham, MA, USA) according to the manufacturer's protocol. Five representative fields from each tumor were analyzed to provide a mean result.

2.10. Statistical analysis

All statistical analyses were performed using GraphPad Prism (RRID: SCR_002798). An unpaired two‐tailed Student's t‐test was used for comparisons between two groups. For comparisons involving more than two groups, one‐way or two‐way ANOVA was conducted, followed by Holm–Šídák post hoc test for multiple comparisons. A p‐value of less than .05 was considered statistically significant.

3. RESULTS

3.1. Expression of fibrosis‐related genes is upregulated in breast tumors and this correlates with that of LPA receptors

We used the GEPIA platform to assess changes in fibrosis‐related genes and the expression levels of key ECM genes, including COL1A1, COL1A2, COL3A1, and FN1. These genes are significantly upregulated in breast tumor tissues compared to normal breast tissues (Figure 1A). We also explored the correlation between these ECM genes and LPA receptors. The results revealed strong correlations between COL1A1, COL1A2, COL3A1, and FN1 with LPAR1 and LPAR4, with correlation coefficients exceeding 0.4 (Figure 1B). Additionally, these genes also showed positive correlations with LPAR5, LPAR6, and ENPP2 (ATX), though with lower correlation coefficients (Figure S1).

FIGURE 1.

(A) Expression levels of collagen type‐I α1 chain (COL1A1), collagen type‐I α2 chain (COL1A2), collagen type‐II α1 chain (COL2A1), collagen type‐III α1 chain (COL3A1), and fibronectin 1 (FN1) in human breast tumors compared to normal breast tissues. The Y‐axis represents log₂ transcripts per million (TPM + 1). *p < .01. (B) Spearman's correlation analysis between the expression levels of lysophosphatidate, lysophosphatidic acid (LPA) receptors (LPAR1 and LPAR4) and fibrosis‐related genes (COL1A1, COL1A2, COL3A1, and FN1). (C) t‐distributed stochastic neighbor embedding (tSNE) clustering of cells from six human breast tumor samples, identifying 11 cell types: CTL (cytotoxic T cells), Th (T helper cells), Treg (regulatory T cells), Macro (macrophages), myoFb (myofibroblasts), Endo (endothelial cells), Epi (epithelial cells), Plasma (plasma cells), B‐cells, Epi‐EMT (epithelial cells in epithelial–mesenchymal transition), and Macro‐MMT (macrophages in macrophage‐myofibroblast transition). (D) Heat map showing the average expression of extracellular matrix genes and LPA receptors.

3.2. Single‐cell RNA sequencing reveals cell type‐specific gene expression

We identified 18 distinct cell clusters and classified them into 11 cell types using scRNA‐seq data from dataset GSE199219, which includes six human breast tumor samples (Figure 1C). These cell types include cytotoxic T cells (CTL, expressing PTPRC, CD3D, CD3E, and CD8A), T helper cells (Th, expressing PTPRC, CD3D, CD3E, and CD4), regulatory T cells (Treg, expressing PTPRC, CD3D, CD3E, CD4, and FOXP3), macrophages (Macro, expressing PTPRC and CD14), myofibroblasts (myoFb, expressing PDGFRB, COL3A1, FN1, and ACTA2), endothelial cells (Endo, expressing PECAM1 and VWF), epithelial (cancer) cells (Epi, expressing EPCAM, KRT8, and KRT18), plasma cells (Plasma, expressing AIF1 and CD79A), B‐cells (expressing CD79A and MS4A1), epithelial cells undergoing epithelial–mesenchymal transition (Epi‐EMT, expressing EPCAM, KRT8, KRT18, and VIM), and macrophages undergoing macrophage‐myofibroblast transition (Macro‐MMT, expressing PTPRC, CD14, COL3A1, and ACTA2). Expressions of these markers in 18 clusters are shown in Figure S2.

The expressions of COL1A1, COL1A2, COL3A1, FN1, and ACTA2 were localized predominantly in myofibroblasts, Epi‐EMT, and Macro‐MMT cells (Figure 1D). In contrast, LPA receptors were broadly expressed across multiple cell types, with LPAR4 being primarily expressed in myofibroblasts and endothelial cells. This suggests that the correlation between ECM genes and LPAR _1/4 is likely driven by LPA receptor‐mediated signaling rather than co‐expression in specific cell types.

3.3. Tumor myofibroblasts predominantly signal through the extracellular matrix

Our CellChat analysis revealed that collagen‐ and fibronectin‐mediated signaling constitutes half of the top 50 ligand‐receptor pairs involved in cell communication within the TME (Figure S3). Heatmaps of communication probabilities showed that myofibroblasts and macrophage‐MMT are more likely to be the primary sources of collagen and fibronectin signaling (Figure 2A,B), compared to other networks such as TGFβ, tumor necrosis factor (TNF), C‐C‐motif ligand (CCL), chemokine (C‐X‐C motif) ligand (CXCL), interleukin‐6 (IL6), leukemia inhibitory factor receptor (LIFR), and colony stimulating factor (CSF) (Figure 2C–I). These signals predominantly interact with CD44 and integrinβ1, which are broadly expressed by other cell types within the TME (Figure 2J,K).

FIGURE 2.

Cell–cell communication probability heatmaps of (A) collagen, (B) fibronectin, (C) TGFβ, (D) TNF, (E) CCL, (F) CXCL, (G) IL6, (H) LIFR, and (I) CSF. Violin plot of expression levels of (J) collagen and interactors of collagen, and (K) fibronectin and interactors of fibronectin in breast tumors. CTL, cytotoxic T cells; Endo, endothelial cells; Epi, epithelial cells; Epi‐EMT, epithelial cells in epithelial–mesenchymal transition; Macro, macrophages; Macro‐MMT, macrophages in macrophage–myofibroblast transition; myoFb, myofibroblasts; Plasma, plasma cells; Th, T helper cells; Treg, regulatory T cells.

Gene set variation analysis (GSVA) also demonstrated that myofibroblasts and other ACTA2‐positive cells, such as epithelial cells in EMT and macrophages in MMT, showed higher GSVA enrichment scores in pathways that regulate ECM‐related signaling pathways in the breast tumor (Figure S4).

3.4. LPA stimulated expressions of αSMA, collagen Iα1, and lamininγ1 in mouse primary mammary fibroblasts

To investigate the effects of LPA on myofibroblast activation, primary mouse MF were isolated from the mammary tissue of C57BL/6J mice. Fibroblast markers fibroblast activation protein (FAP), α smooth muscle actin (αSMA, ACTA2), and S100A4 were highly expressed in isolated fibroblasts relative to E0771 breast cancer cells (Figure S5A). Treatment with 5 μM LPA significantly increased the mRNA expression of ACTA2, a marker of myofibroblast differentiation, by ~2.5‐fold at 12 h and ~2.6‐fold at 24 h (Figure 3A). ACTA2 expression increased correspondingly when fibroblasts were treated with 5, 10, and 20 μM LPA for 24 h, demonstrating a dose–response relationship. Notably, this LPA‐induced increase in ACTA2 expression was inhibited by 0.5 μΜ Ki16425, an LPAR_1/3 antagonist, and 5 μΜ parthenolide, an NF‐κB inhibitor (Figure 3B).

FIGURE 3.

(A) Time course analysis of mRNA for ACTA2 expression in mouse mammary fibroblasts stimulated with 5 μM lysophosphatidate, lysophosphatidic acid (LPA). (B) ACTA2 mRNA expression in mouse mammary fibroblasts treated with different doses of LPA for 24 h, with or without the presence of 0.5 μM Ki16425 (an LPA receptor 1/3 inhibitor) and 5 μM parthenolide (an NF‐κB inhibitor). *p < .05, compared with LPA [0]; ^$ p < .05, ^$$$ p < .001, compared with LPA alone (the same concentration). (C) ACTA2 mRNA expression in mouse mammary fibroblasts treated with 5 μM LPA, 10 ng/mL transforming growth factorβ1 (TGF‐β1) or 5 μM LPA + 10 ng/mL TGF‐β1 for 24 h. (D) Protein levels of αSMA, collagen Iα1, and lamininγ1 in mouse mammary fibroblasts treated with 5 μM LPA, 10 ng/mL TGF‐β1 or 5 μM LPA + 10 ng/mL TGF‐β1 for 48 h. *p < .05, **p < .01.

To compare the effects of TGFβ1, fibroblasts were treated with 10 ng/mL TGFβ1 for 24 h, which also resulted in a similar increase in ACTA2 expression. However, the combination of 5 μM LPA and 10 ng/mL TGFβ1 did not produce a significant interactive effect (Figure 3C).

Both 5 μM LPA and 10 ng/mL TGFβ1 treatments for 48 h increased the levels of αSMA protein (Figure 3D). Five μM LPA also upregulated the expression of collagen Iα1 and lamininγ1, key components of the ECM. Interestingly, TGFβ1 did not significantly affect the expression of collagen Iα1 or lamininγ1, indicating that LPA could have a more specific role in regulating ECM protein production in MF.

3.5. Co‐culture with breast cancer cells induces ACTA2 expression in fibroblasts

To assess the impact of LPA on myofibroblast differentiation further, Hs578Bst cells, a human MF line, were cultured alone or co‐cultured with its matched breast cancer cell line Hs578T. The serum‐containing medium was treated with charcoal to remove endogenous lipids and then supplemented with 50 μM LPC to provide the substrate for ATX. The conversion of LPC to LPA by ATX secreted from the cells was expected to drive cellular responses.

Co‐culture with Hs578T cells did not significantly affect the expression levels of ATX (ENPP2) (Figure 4A) or TGFB1 (Figure 4B) in Hs578Bst cells. However, it led to a significant increase in ACTA2 expression in Hs578Bst cells, with a ~6.5‐fold elevation compared to cells cultured alone (Figure 4C). This marked increase in ACTA2 expression was blocked completely by inhibiting ATX with 100 nM IOA‐289 (Figure 4C), indicating that the LPA produced by ATX was responsible for the observed myofibroblast activation. We also measured mRNA levels of LPA receptors in Hs578Bst cells. Only LPAR1 (Figure 4D) and LPAR2 (Figure 4E) were detectable and were not affected by co‐culture with HS578T cells.

FIGURE 4.

mRNA levels of (A) ENPP2, (B) TGFB1, (C) ACTA2, (D) LPAR1, and (E) LPAR2 in Hs578Bst cells (fibroblasts) cultured alone or co‐cultured with Hs578T cells (matched breast cancer cells), with or without the presence of 100 nM IOA‐289, an autotaxin inhibitor. *p < .05, **p < .01. LPA, lysophosphatidate, lysophosphatidic acid; TGFβ1, transforming growth factorβ1.

While IOA‐289 blocked the increase in ACTA2, it simultaneously upregulated ENPP2 expression in Hs578Bst cells co‐cultured with Hs578T cells (Figure 4A). This indicates a compensatory mechanism in response to ATX inhibition resulting from decreased feedback inhibition by LPA on ATX transcription and secretion. ³² Although there was increased ATX expression, it should be recognized that this ATX is not active because it is inhibited by IOA‐289.

3.6. IOA‐289 reduces tumor fibrosis in a syngeneic breast tumor model

We established a syngeneic orthotopic breast tumor model in C57BL/6J mice using E0771 breast cancer cells. Mice were treated with IOA‐289 at 100 mg/kg twice daily. Treatment with IOA‐289 significantly decreased mRNA levels of COL1A1 and FN1, as well as the fibrotic marker TGFB1 (TGFβ1) in the tumors (Figure 5A). Additionally, IOA‐289 treatment decreased the protein levels of collagen Iα1, while the levels of lamininγ1 and αSMA remained unchanged (Figure 5B). Protein levels of MMP2, MMP3, MMP8, proMMP9, and MMP12 in the tumors were not affected by IOA‐289 treatment (Figure 5C). Masson's trichrome staining of tumor tissues demonstrated a decrease in fibrotic tissue in the IOA‐289 treated group compared to the control group (Figure 5D). These findings indicate that IOA‐289 effectively decreases the collagen content within the TME, establishing its potential role in attenuating tumor fibrosis.

FIGURE 5.

(A) mRNA levels of collagen type‐I α1 chain (COL1A1), collagen type‐III α1 chain (COL3A1), mouse lamininα1 (LAMA1), mouse lamininγ1 (LAMC1), fibronectin 1 (FN1), and human TGFβ1 (TGFB1) in breast tumors derived from E0771 cells in C57BL/6J mice, with or without IOA‐289 treatment at 100 mg/kg twice daily. (B) Protein levels of α smooth muscle actin (αSMA), collagen Iα1, and lamininγ1 in tumors. (C) Protein levels of matrix metalloproteinase 2 (MMP2), MMP3, MMP8, proMMP9, and MMP12 in tumors. (D) Representative images of hemotoxylin and eosin (H&E) and Masson's trichrome staining of control and IOA‐289 treated tumors. Right panel is quantification of fibrotic tissue from Masson's trichrome stained slides. *p < .05, **p < .01.

3.7. E0771 breast cancer cells lack functional LPA signaling

Quantitative PCR analysis was performed to compare the mRNA expression levels of LPAR1, LPAR2, LPAR3, LPAR4, LPAR5, and LPAR6 in NIH3T3, mouse MF, E0771, and 4 T1 cells. E0771 cells exhibited very low expression levels of LPAR2, LPAR5, and LPAR6, while LPAR1, LPAR3, and LPAR4 were undetectable. By contrast, NIH3T3 cells, mouse MF, and 4T1 breast cancer cells express LPAR1 and LPAR2 at varying levels, with NIH3T3 cells showing the highest expression (Figure 6A).

FIGURE 6.

(A) mRNA levels of LPAR1, LPAR2, LPAR3, LPAR4, LPAR5, and LPAR6 in NIH3T3, mouse mammary fibroblasts (MF), E0771, and 4T1 cells. (B) Time course stimulation by lysophosphatidate, lysophosphatidic acid (LPA). LPA increased LIF expression in NIH3T3, mouse MF, 4T1 breast cancer cells, Hs578Bst fibroblasts, and Hs578T cancer cells, but not in E0771 cells. (C) Time course stimulation by LPA. LPA increased phosphorylation of AKT and ERK in NIH3T3 cells, but not in E0771 cells. (D) LPA increased phosphorylation of AKT and ERK in Hs578Bst fibroblasts and Hs578T cancer cells. *p < .05, **p < .01, ***p < .001, compared with 0 h (no LPA treatment).

Functional assays revealed that E0771 cells did not respond to LPA stimulation with LIF expression, whereas this was observed in NIH3T3, mouse MF, 4T1, Hs578Bst, and Hs578T cells (Figure 6B). Additionally, LPA‐induced phosphorylation of AKT and ERK was detected in NIH3T3, Hs578Bst, and Hs578T cells, but was absent in E0771 cells (Figure 6C,D). These results show that E0771 cells lack functional LPA signaling through LPARs, which indicates an indirect response to LPA in the TME.

LIF‐induced STAT3 phosphorylation was also lacking in E0771 cells. TGFβ1‐induced Smad2 phosphorylation was also much lower in E0771 cells than in NIH3T3 cells (Figure S5B). This indicates both LIF and TGFβ1 signaling in the E0771 tumor model are more active in fibroblasts than in the cancer cells.

4. DISCUSSION

Excessive LPA signaling is one of the reasons for increased tumor growth in many types of cancers, including breast and hepatocellular cancer. ³³ , ³⁴ Previous work demonstrated that the ATX inhibitor, IOA‐289 (cambritaxestat), decreased breast tumor growth in a syngeneic mouse model using E0771 breast cancer cells. ¹³ , ²² The present study demonstrates a significant impact of IOA‐289 in decreasing tumor fibrosis in this mouse model. The results highlight the potential of targeting the LPA signaling pathway to modulate the TME by influencing the stromal compartment.

The E0771 breast cancer cell line lacks the capacity to respond to LPA stimulation, as evidenced by the low or undetectable expression levels of LPA receptors. This is confirmed by the absence of LPA‐induced signaling events such as through AKT and ERK phosphorylation. Therefore, the decrease in tumor growth observed with IOA‐289 treatment is more likely to be caused through effects on the TME through other cell types within the tumor, rather than direct inhibition of LPA signaling in E0771 cancer cells. IOA‐289 could influence stromal cells, such as fibroblasts, which are more responsive to LPA signaling, contributing to the overall reduction in tumor growth. In many cancer models, LPA signaling contributes directly to cancer cell proliferation, survival, and migration. ³⁵ , ³⁶ However, the lack of LPA responsiveness in E0771 cells shifts the focus to the stromal compartment, where LPA appears to be critical for activating fibroblasts and for promoting the differentiation of these cells into αSMA⁺ myofibroblasts. Myofibroblasts are key drivers of fibrosis, contributing to ECM remodeling and the creation of a dense, fibrotic stroma that can hinder immune cell infiltration and facilitate tumor progression. ³⁷ , ³⁸ Myofibroblasts also directly suppress immune cell activity. ³⁹ By using the E0771 model, we were able to isolate the effects of LPA signaling in stromal cells, providing clearer evidence that ATX inhibitors such as IOA‐289 can modulate the TME by targeting stromal rather than cancer cells, which is sufficient to inhibit tumor growth.

One of the most significant findings from our previous research and that of others is that IOA‐289 treatment increased the infiltration of CD8⁺ cytotoxic T cells into the tumors and did not affect the numbers of CD45⁺, F4/80⁺, CD206⁺, and FoxP3⁺ cells in TME. ¹³ , ²² This observation aligns with the current study, which shows that IOA‐289 decreases collagen and fibronectin deposition, key components of the fibrotic ECM. Fibrosis in the TME can act as a physical and biochemical barrier, restricting the infiltration and movement of immune cells such as CD8⁺ T cells, ⁴⁰ which are critical for anti‐tumor immunity. By reducing fibrosis, IOA‐289 likely alters the ECM in a way that facilitates better immune cell access to cancer cells, thereby enhancing the effectiveness of immune surveillance and attack. The decrease in collagen fibers observed in our Masson's trichrome staining assays supports this hypothesis, suggesting that IOA‐289's anti‐fibrotic effects could contribute directly to the improved immune cell infiltration seen in treated tumors. In addition, LPA signaling through LPAR₅ suppresses T cell function via multiple mechanisms including disrupting T cell receptor signaling, ⁴¹ inhibiting immune synapse formation ⁴² and metabolic reprogramming. ⁴³ Our data base analysis of human breast tumors also showed positive correlations with LPAR5 and ENPP2 (ATX). Thus, the action of IOA‐289 in blocking ATX activity and thus LPA signaling in the TME will attenuate the immune‐suppressive effects of LPA.

Increasing evidence indicates that ECM molecules in TME play an important role in tumor development. Interaction of type I collagen and CD44 between fibroblasts and malignant cells facilitated cancer progression as reported in head and neck cancer. ⁴⁴ Integrinβ1 also interacted with type I collagen to promote gastric cancer progression ⁴⁵ and promotes chemo‐resistance in breast cancer cells. ⁴⁶ We demonstrated that CD44 and integrinβ1 are two major communicators of ECM molecules in E0771 breast tumors, and this could have similar effects in other cancers.

The therapeutic potential of ATX inhibitors extends beyond preclinical models, as evidenced by ongoing clinical trials exploring the use of these inhibitors in various diseases, including idiopathic pulmonary fibrosis and cancer. ¹⁷ , ⁴⁷ These trials underscore the relevance of targeting the LPA signaling pathway in fibrotic diseases. In the context of cancer, the ability of ATX inhibitors to modulate the TME by reducing fibrosis and enhancing CD8⁺ T cell infiltration could make them valuable components of combination therapies. Combining ATX inhibitors with immune checkpoint inhibitors could boost the efficacy of immunotherapy by making tumors more accessible to immune cells. Additionally, as our study suggests, the stromal effects of ATX inhibition could complement existing therapies that primarily target cancer cells, leading to more comprehensive anti‐tumor responses.

AUTHOR CONTRIBUTIONS

Xiaoyun Tang: Conceptualization; methodology; formal analysis; investigation; writing – review and editing; data curation; writing – original draft. Humayara Khan: Methodology; investigation; writing – review and editing. Karolina Niewola‐Staszkowska: Writing – review and editing; funding acquisition. Frank Wuest: Writing – review and editing; funding acquisition. David N. Brindley: Conceptualization; methodology; funding acquisition; formal analysis; supervision; writing – review and editing; project administration; resources.

FUNDING INFORMATION

This study was financed through an operating grant from the Canadian Institutes of Health Research (PJT‐169140) and a MITACS award for a postdoctoral fellowship to Humayara Khan that was partly financed by iOnctura.

CONFLICT OF INTEREST STATEMENT

David N. Brindley has served as a member of the Scientific and Clinical Advisory Boards for iOnctura and received partial funding for a postdoctoral fellow from iOnctura. Karolina Niewola‐Staszkowska is an employee and shareholder of iOnctura.

ETHICS STATEMENT

All procedures were performed in accordance with the Canadian Council of Animal Care as approved by the University of Alberta Animal Welfare Committee.

Supporting information

Data S1. Supporting Information.

Tang X, Khan H, Niewola‐Staszkowska K, Wuest F, Brindley DN. Inhibition of autotaxin activity with IOA‐289 decreases fibrosis in mouse E0771 breast tumors. Int J Cancer. 2025;157(6):1205‐1217. doi: 10.1002/ijc.35471

DATA AVAILABILITY STATEMENT

Data sources and handling of the publicly available datasets used in this study are described in Section 2. Further details and other data that support the findings of this study are available from the corresponding author upon request.