Adipocyte-derived kynurenine stimulates malignant transformation of mammary epithelial cells through the aryl hydrocarbon receptor

Abstract

Anti-hormone therapies are not efficacious for reducing the incidence of triple negative breast cancer (TNBC), which lacks both estrogen and progesterone receptors. While the etiology of this aggressive breast cancer subtype is unclear, visceral obesity is a strong risk factor for both pre- and post-menopausal cases. The mechanism by which excessive deposition of visceral adipose tissue (VAT) promotes the malignant transformation of hormone receptor-negative mammary epithelial cells is currently unknown. We developed a novel in vitro system of malignant transformation in which non-tumorigenic human breast epithelial cells (MCF-10A) grow in soft agar when cultured with factors released from VAT. These cells, which acquire the capacity for 3D growth, show elevated aryl hydrocarbon receptor (AhR) protein and AhR target genes, suggesting that AhR activity may drive malignant transformation by VAT. AhR is a ligand-dependent transcription factor that generates biological responses to exogenous carcinogens and to the endogenous tryptophan pathway metabolite, kynurenine. The serum kynurenine to tryptophan ratio has been shown to be elevated in patients with obesity. Herein, we demonstrate that AhR inhibitors or knockdown of AhR in MCF-10A cells prevents VAT-induced malignant transformation. Specifically, VAT-induced transformation is inhibited by Kyn-101, an inhibitor for the endogenous ligand binding site of AhR. Mass spectrometry analysis demonstrates that adipocytes metabolize tryptophan and release kynurenine, which is taken up by MCF-10A cells and activates the AhR to induce CYP1B1 and promote malignant transformation. This novel hormone receptor-independent mechanism of malignant transformation suggests targeting AhR for TNBC prevention in the context of visceral adiposity.

1. Introduction

Currently, there are no means of pharmacological prevention for triple-negative breast cancer (TNBC) — an aggressive breast cancer subtype encompassing 10–15% of all breast cancers that tests negative for the expression of estrogen receptor, progesterone receptor, and lacks the amplification of the human epidermal growth factor receptor 2. TNBC is more prevalent in younger women and has an unclear etiology. A better understanding of its causes would lead to novel precision strategies for prevention.

Visceral obesity is a strong risk factor for both pre- and post-menopausal TNBC but how excessive deposition of visceral adipose tissue (VAT) promotes the malignant transformation of hormone receptor-negative mammary epithelial cells is unknown. We previously published that VAT-derived fibroblast growth factor-2 (FGF2) stimulates malignant transformation by activating its primary receptor, fibroblast growth factor receptor-1 (FGFR1), in human estrogen receptor negative breast epithelial cells (MCF-10A cells) [1]. One limitation of these studies was that the levels of recombinant FGF2 that significantly induced malignant transformation in MCF-10A cells (10–30 ng/mL) [1,2], were higher than the levels found in our visceral fat tissue filtrates (FTF) (~5 ng/mL) which stimulated MCF-10A growth in soft agar [2]. Growth in soft agar is a type of anchorage-independent, 3D cell growth that is a surrogate marker for malignant transformation and associates with tumorigenicity in mice [3,4]. Moreover, a FGFR1 blocking antibody only moderately attenuated FTF-induced growth in soft agar [1]. Collectively, these findings suggest additional factors in VAT contribute to malignant transformation. The objective of the study was to investigate other adipokines that increase TNBC risk.

The aryl hydrocarbon receptor (AhR) is a transcription factor that is activated by both exogenous and endogenous ligands to induce Phase I and Phase II metabolizing genes. While Phase II genes add hydrophilic groups to make xenobiotics water soluble for excretion, Phase I genes add reactive or polar groups. For an example of the latter, upon the binding of the pro-carcinogen benzo(a)pyrene, the AhR translocates to the nucleus, binds to xenobiotic response elements of DNA, and induces the transcription of CYP1A1, CYP1B1 and epoxide hydrolase [5-7]. The products of these genes are enzymes that metabolize benzo(a)pyrene into the epoxide intermediate which can initiate tumor formation by reacting with DNA, creating adducts and DNA damage, a hallmark of tumor initiation [8-10].

In addition to its role in tumor initiation, the AhR has been implicated in tumor progression by driving proliferation, suppressing apoptosis, and increasing migration and invasion properties implicated in a variety of different cancers [11-17]. This corroborates with our analysis of patient data showing worse overall survival in breast cancer patients with high expression of AhR with the magnitude of change in survival being greatest in TNBC tumors compared to all breast cancer specimens. Moreover, investigation of genes differentially expressed in VAT-transformed MCF-10A cells led to the hypothesis that malignant transformation by VAT is dependent on AhR activation, as both CYP1A1 and CYP1B1 gene expression were elevated in MCF-10A cells that acquired the ability to grow in soft agar. In this manuscript, we describe a novel mechanism for malignant transformation of MCF-10A cells and suggest the AhR as a potential target for TNBC prevention in the context of visceral adiposity.

2. Materials and methods

2.1. Cell culture and reagents

MCF-10A (human mammary epithelial cells) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). Cells were cultured in DMEM/Ham’s F12 (Invitrogen, Carlsbad, CA, USA) supplemented with 5% horse serum (HS), 100 ng/mL cholera toxin (Sigma,), 20 ng/mL epidermal growth factor (EGF; Peprotech, Rocky Hill, NJ, USA), 10 μg/mL insulin (Sigma, St. Louis, MO, USA), 0.5 mg/mL hydrocortisone (Sigma, St. Louis, MO, USA), 7.5% sodium bicarbonate, 15 mM HEPES, 2 mM L-Glutamine, and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA). MCF-10A cells were singularly used for transformation studies since there are no other suitable human non-tumorigenic cell lines that can be stimulated to grow in soft agar. Primary mouse bone marrow stromal cells (BMSCs) were a generous gift from Dr. Izabela Podgorski (Wayne State University, Detroit, MI, USA). BMSCs were cultured in DMEM and supplemented with 10% FBS and 1% penicillin/streptomycin. To induce differentiation of BMSCs to mature adipocytes, BMSCs were plated in a 3D collagen I matrix and treated with an adipogenic cocktail (30% StemXVivo Adipogenic Supplement, 1 μM insulin, 10 μM Rosiglitazone in complete medium). Medium was replenished every 48–72 h for a total of three or four treatments as previously described [18-20]. Differentiated adipocytes were washed three times with PBS before being used in experiments. Cells routinely test negative for mycoplasma contamination using the MycoAlert Mycoplasm Detection Kit (Lonza, Basel, Switzerland). MCF-10A cells were authenticated using LabCorps STR profiling services and were a 100% match.

2.2. Generation of mouse fat tissue filtrate (mFTF)

Six- to eight-week-old, female, SKH-1 mice were kept on low fat diet (LFD control diet, F4031 Bio-Serve Flemington, NJ, USA) or high fat diet (HFD, F3282 Bio-Serve Flemington, NJ, USA) for a total of four weeks. Visceral adipose tissue was then excised and homogenized in serum-free MEM on ice using the Tissue Ruptor (Qiagen, Hilden, Germany) on medium speed and filtered using hanging 15-mm wide 0.4 μm filter inserts (Millipore, Burlington, MA, USA) in to a six-well plate on a rocker at room temperature for one hour as previously described [1,2]. Filtrates were then quantified using a BCA assay and treatments were normalized to total protein. A concentration of 200 μg/mL of mFTF was used for soft agar experiments. Animal studies were approved by Michigan State University IACUC PROTO202200123. Euthanasia was performed at the termination of all experiments or in any case where discomfort was not adequately relieved by carprofen as described above. Mice were euthanized by giving a high dose of FatalPlus^® to ensure a quick and painless death, in accord with recommendations of the Panel of Euthanasia of the AVMA.

2.3. Soft agar transformation assay

MCF-10A cells were seeded in an agar layer at a density of 1000 cells per well in a 24-well plate in 200 μL of complete media with 0.3% agar on top of a 0.5% layer of agar in complete MCF-10A medium. Serum free DMEM (mFTF control), 200ug/ml of mFTF, 0.5 N HCl (L-kynurenine vehicle control), L-kynurenine, DMSO (antagonists vehicle control) or the AhR antagonists, CH223191, α-naphthoflavone, and Kyn101, were added to the agar layer containing the cells. Soft agar plates were then left at room temperature for 30 min before adding 200 μL of complete MCF-10A medium and incubated at 37 °C. Medium was replaced every 3–4 days for 14 days. After two weeks, the colonies in the wells were fixed in 70% ethanol and stained with 200 μL of 0.01% crystal violet. Colonies were counted using the BioTek Cytation 3 imaging plate reader using Gen5 3.04 software (BioTek Instruments, Inc., Winooski, VT, USA) and only colonies greater than 25 μm were recorded.

2.4. Generation of mouse fat tissue filtrate-transformed MCF-10A cells

MCF-10A cells that were treated with 200 μg/ml mFTF in soft agar were removed via pipette tip from agar and expanded in a 96-well plate in complete MCF-10A media. Media was replenished every 2–3 days and cells were grown out to obtain mFTF-transformed MCF-10A cells.

2.5. Taqman quantitative PCR analysis of aryl hydrocarbon receptor target genes

MCF-10A cells were lysed and RNA was isolated using the RNeasy MiniKit (Qiagen, Hilden, Germany). 300 to 1000 ng of RNA was used to generated cDNA for gene expression analysis using the High-Capacity cDNA Reverse Transcription Kit (ThermoFisher, Pittsburgh, PA, USA) The analysis of AhR and AhR target genes CYP1A1 and CYP1B1 was performed using TaqMan Individual Gene Expression assays for Human AhR (Hs00169233), CYP1A1 (Hs01054797), and CYP1B1 (Hs00164383). Assays were done in biological and technical triplicates using TaqMan Universal Mastermix II, no UNG (Thermofisher; Pittsburgh, PA, USA) with 50 ng of cDNA per well in a 96 well plate was run using the QuantStudio6 Flex Real-Time PCR (Applied Biosystems). DataAssist software was used for comparative analyses.

2.6. Generation of MCF-10A empty vector (EV) and aryl hydrocarbon Receptor^low cells

Parental MCF-10A cells were spinfected (2,000 RPMs for 2 h at 37 °C) with lentiviral particles encapsulating Cas9-producing plasmids (pLentiCRISPR_v2 backbone, GenScript, Piscataway, NJ, USA) containing either an empty vector (EV) or sgRNA targeting the AhR (AhR^Low; sgRNA: TTGCTGCTCTACAGTTATCC). Lentiviral particles were generated by collecting cell supernatant from HEK293T-LentiX (Takara, Kusatsu, Japan) cells transfected with PAX2, MD2, and pLentiCRISPR plasmids (EV or AhR sgRNA) using Lipofectamine LTX (ThermoFisher, Pittsburgh, PA, USA). MD2 and PAX2 plasmids were purchased through Addgene. Confirmation of lentivirus in the cell supernatant was performed using Lenti-X GoStix (Takara,Kusatsu, Japan). MCF-10A cells spinfected with lentiviral particles containing the EV or AhR sgRNA plasmid were selected for using 0.5 μg of puromycin. AhR levels were tested using immunoblotting with the AhR antibody (Cell Signaling Technologies, Danvers, MA, USA) and Taqman qPCR for AhR as previously described.

2.7. In vitro models of Adipocyte-MCF-10A interactions

2.7.1. Transwell co-culture

To allow for direct crosstalk between adipocytes and MCF-10A cells, 80,000 MCF-10A cells were seeded on the top layer of a Transwell filter (0.2 μm pore size; Corning, Corning, NY, USA) separating the cells from differentiated adipocytes embedded in a Collagen I (Advanced Biomatrix, Carlsbad, CA, USA) layer below. Cells were cultured together for 48 h and then washed with PBS, trypsinized and harvested for RNA to assess changes in gene expression upon direct exposure to adipocytes.

2.7.2. Adipocyte conditioned media

BMSCs were differentiated to mature adipocytes and washed three times with PBS. Media was replaced with serum-free DMEM and allowed to incubate for 48-hours, where adipocyte-secreted factors were collected and stored as conditioned media at −80 °C. 50,000 MCF-10A cells were plated in a 6-well plate and incubated overnight at 37 °C. The following day, the media was removed and replaced with a 50:50 mixture of complete MCF-10A media with serum-free DMEM (Control) or adipocyte conditioned media (AdipoCM) for 24-hours. After 24 h, cells were washed with PBS and collected for RNA as previously described.

2.8. Metabolomic profiling and stable isotope labeling

To quantify metabolites, the adipocytes and MCF10A cells were seeded in triplicate (n = 3) in 6-well plates with collagen + media (adipocytes) or only media (cells) as described in the cell culture methodology section until achieving approximately 80% confluency. For stable isotope labeling, media was refreshed on the plates and incubated for 2 h. Plates were then washed with PBS then switched to the respective media containing either 0.016 g/L ¹³C₁₀-Tryptophan (Cambridge Isotope Laboratories, Inc, Tewksbury, MA, CLM-4290-H-0.1) or × 0.025 g/L ¹³ C₁₀-kynurenine (Cambridge Isotope Laboratories, Inc, Tewksbury, MA, CLM-9884-H). Samples were collected at 0 min as an unlabeled control) and a range of timepoints as indicated by the figures. Metabolite extraction was performed as described previously [21]. Briefly, each well was washed with 0.9% saline (VWR, Radnor, Pennsylvania, 16005–092), then 500 μL of HPLC grade methanol was added followed by 300 μL of HPLC-grade water containing 0.5 μM PIPES as an internal control. Adipocytes were incubated with 0.1% collagenase in PBS per well for 15–30 min, washed with PBS and centrifugated 3 times (130xg at 4 °C for 5 min). On last spin, all PBS was removed and 500 μL of HPLC grade methanol was added to the pellet followed by 300 μL of HPLC-grade water containing 0.5 μM transferred to a 1.5 mL Eppendorf tube containing 500 μL of HPLC-grade chloroform. Samples were vortexed for 10 min, then centrifuged at 16,000g for 15 min. The polar layer was then removed and dried by lyophilization. In addition to intracellular metabolites, 100 μL media from adipocytes or cell culture and from the fat tissue filtrate (mFTF)-treated samples were also taken from each well at the same time points and extracted using the same procedure outline above. Protein extracted from the cells was dissolved in 0.2 M potassium hydroxide aqueous solution overnight and quantified using Pierce BCA Protein Assay Kit (Fisher Scientific, Hampton, NH, PI23225). Extracted metabolites were then resuspended in HPLC-grade water containing 0.6% of formic acid.

LC-MS/MS analysis was performed with ion-pairing reverse phase chromatography using an ACQUITY Premier HSS T3 Column (1.8 μm, 2.1 × 150 mm, Waters Corporation, Milford, MA, USA) for separation and a Waters Xevo TQ-XS triple quadrupole mass spectrometer. Metabolite peak processing was performed in MAVEN [22]. For unlabeled samples, the data was scaled by the internal standard (D,L-Tryptophan-d3 / T947202 - TRC, 2-Picolinic-d4 Acid / D-5289 - CDN Isotopes and L-Kynurenine-d4 Trifluoroacetic Acid Salt / K661007 - TRC) and protein content. The entire dataset was then normalized by Probabilistic Quotient Normalization [23]. For isotope labeled samples, the data was corrected for the natural ¹³C abundance using IsoCor [24] and reported as the fraction of labeled metabolite compared to the total amount of metabolite.

3. Results

3.1. AhR activity and breast cancer progression

Previous studies from our laboratory demonstrate that visceral adipose tissue (VAT) from mice fed a high-fat diet for 4 weeks or from human visceral adipose tissue promotes the malignant transformation of MCF-10A cells [1]. To better understand if the effects of VAT on MCF-10A cell transformation are permanent or temporal, a 3D-growing colony of MCF-10A cells transformed with mouse fat tissue filtrate (mFTF) was isolated, expanded in 2D, and re-seeded in 3D. Colony growth was measured compared to parental MCF-10A cells. mFTF-transformed cells exhibit ~60% more colony forming potential, suggesting lasting reprogramming of MCF-10A cells during transformation (Fig. 1A). Transformed cells were then analyzed to assess differential gene expression compared to the MCF-10A parental line to better understand the mechanisms driving these phenotypic changes. Interestingly, two direct target genes of the aryl hydrocarbon receptor (AhR), CYP1A1 and CYP1B1, were elevated in the transformed cells compared to the parental lines (Fig. 1B). The transformed cells also demonstrated elevated AhR mRNA and protein (Fig. 1B).

Fig. 1.

Aryl hydrocarbon receptor (AhR) target genes are induced in mammary epithelial cells transformed by mouse fat tissue filtrate (mFTF) and high AhR mRNA expression predicts poor survival in hormone-positive and hormone-negative breast cancers. A. 3D colony formation in MCF-10A cells cultured in the presence of mFTF, previously transformed by mFTF, and untreated parental cells. mFTF-transformed MCF-10A cells were generated by culturing MCF-10A cells in the presence of mFTF in the 3D soft agar assay, then isolating clones. Data are presented as a mean ± S.D. of values from biological triplicates, with each point representing an individual sample. Values **, p < 0.01 are considered statistically significant. B. Taqman qPCR gene expression analysis of AhR and AhR target genes, CYP1A1 and CYP1B1 and immunoblot analysis of AhR protein levels in mFTF-transformed or parental MCF-10A cells. TBP was used as an internal loading control for qPCR and Actin was used as a loading control for protein levels. All breast cancer patients (C) and TNBC breast cancer patients (D) with high AhR mRNA expression have decreased overall survival. Data was generated by combining mRNA data from GEO, TCGA, and EGA databases using the Kaplan Meier Plotter. The top 25% (highest quartile) of AhR expression was compared with the bottom 75% (lowest 3-quartiles). The log-rank test was utilized to test statistical significance.

To determine the translational relevance of this pre-clinical finding, AhR gene expression was analyzed from publicly available bulk RNA-sequencing data from clinical breast cancer patient samples that were either hormone receptor positive (estrogen receptor, progesterone receptor, and HER2, n = 2976) or triple negative breast cancer (TNBC, n = 126) subtypes from the Km Plotter database [25]. Patients that showed AhR gene expression within the highest quartile had significantly worse overall survival compared with those patients falling within the lower three quartiles in both patient groups (hormone positive and negative) (Fig. 1C) and when filtered for just TNBC subtypes (Fig. 1D). These data suggest a role for AhR in breast tumor progression and outcome; however, whether AhR activity is a driver of malignant transformation or a bystander is unknown.

3.2. Visceral adipose tissue-stimulated 3D growth is dependent on AhR

To test the hypothesis that transformation by VAT is dependent on AhR activity, MCF-10A cells were incubated in soft agar with 200 μg/ml mFTF in the presence of AhR modulators. As previously reported [1], MCF-10A cells in the presence of mFTF transformed as shown by increased colony numbers. MCF-10A cells were treated with: 1) Kyn-101, a recently identified AhR inhibitor, 2) CH223191, a well-characterized antagonist of AhR activity, or 3) α-naphthoflavone, an AhR modulator. Both Kyn-101 and CH223191 significantly attenuated mFTF-stimulated colony formation at 1 μM (Fig. 2A); however, α-naphthoflavone failed to significantly inhibit the effect of mFTF (Fig 2A). CH223191 inhibits AhR activation by selective agonists through the prevention of its translocation to the nucleus. Due to its agonist-specific inhibition of AhR, it is presumed that it either binds to the ligand binding domain to prevent ligand binding or binds to the AhR, modifying its conformation and availability of the ligand binding domain to certain agonists [26-29]. Kyn-101 similarly inhibits AhR nuclear localization and transcriptional activity by competing for the endogenous ligand binding site with an IC₅₀ of 22–25 nM [30-32]. Interestingly, Kyn-101 decreased the effect of mFTF on colony formation in a dose-dependent manner and did so significantly at the IC₅₀ (Fig. 2B). Importantly, Kyn-101 concentrations that inhibited colony formation were not cytotoxic in 2D growth assays, suggesting a critical role of AhR in transformation but not proliferation (Fig. 2C). These data also suggest that there are different mechanisms governing 2D versus 3D growth.

Fig. 2.

AhR activity mediates the transforming effects of mouse fat tissue filtrate (mFTF) on mammary epithelial cells (MCF-10A). A. Soft agar colony formation assay of MCF-10A cells in the presence of mFTF or AhR antagonists (CH223191 and Kyn-101) or an AhR modulator (alpha-naphthoflavone (α-NF). B. Soft agar colony formation assay in the presence of mFTF and Kyn-101 within IC₅₀ ranges. C. MCF-10A cell viability measured by CellTiter-Fluor after 24 h of Kyn-101 treatment. Viability is recorded as a percent compared to the vehicle treated control cells. D. Taqman qPCR analysis measuring AhR mRNA transcript levels in MCF-10A cells expressing empty vector (EV) or a sgRNA targeting AhR (AhR KD) and immunoblot analysis of AhR protein levels in the EV and AhR KD MCF-10A cells. Gene expression was normalized to the internal control, TBP. Actin and Cas9 were used as loading controls for AhR immunoblots. E. EV and AhR KD MCF-10A cells grown in 3D in soft agar in the presence of mFTF. For all soft agar experiments data are presented as a mean ± S.D. of values from biological triplicates, with each point representing an individual sample. A one-way ANOVA was performed to assess statistical significance between treatments groups. Values **p < 0.01, *** p < 0.001, and **** p < 0.0001 are considered statistically significant.

To further characterize the role of AhR in malignant transformation and rule out off-target effects of the antagonists, lentiviral delivery of a plasmid containing both Cas9 and a sgRNA targeting the AhR gene was utilized to knockout gene expression of AhR in MCF-10A cells. Analysis of the AhR KO cells compared to the EV Cas9 expressing cells showed a ~ 50% reduction in gene expression and protein levels (Fig. 2D). Importantly, mimicking the chemical antagonists, AhR^Low MCF-10A cells failed to form colonies in the presence of mFTF, whereas mFTF significantly stimulated colony formation in Empty Vector (EV) MCF-10A cells (Fig. 2E), further supporting the role of the AhR in VAT-driven transformation.

3.3. Adipocytes as a source of endogenous kynurenine through tryptophan metabolism

The ability of Kyn-101 to prevent mFTF-driven transformation of MCF-10A cells suggests the presence of an endogenous agonist secreted from VAT that has the propensity to activate the AhR and induce 3D growth. Kynurenine, an endogenous AhR ligand formed through the catabolism of tryptophan by indoleamine 2,3-dioxygenase (IDO), has received increasing interest for its recently identified function in obesity and diabetes [33-37]. Additionally, limited studies have shown that obese patients and mice fed high fat diets have elevated serum kynurenine/tryptophan ratios compared to lean controls [36,38]. To determine if kynurenine was present in the mFTF, liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis was performed to test the relative levels of kynurenine in serum free DMEM (Media Control), mFTF from female SKH-1 mice fed low fat diet (LFD mFTF) or a high fat diet (HFD mFTF).

Transformation-inducing HFD mFTF contained a significantly higher kynurenine/tryptophan ratio, indicative of elevated IDO activity, compared to LFD mFTF (Fig. 3A), moreover, the concentration of HFD mFTF (~0.3 μM) is approximately 6-fold higher when compared to LFD mFTF (~0.05 μM; Fig. 3B), supporting the role of visceral adiposity in TNBC risk.

Fig. 3.

Adipocytes with increased IDO activity excrete endogenous AhR ligand kynurenine, which is internalized and metabolized by mammary cells. A. Following treatment with HFD (high fat diet), LFD (low fat diet) or cell culture media (Control), mFTF (mouse fat tissue filtrate) was collected for liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. The kynurenine to tryptophan ratio (Kyn/Trp) was determined using signal intensity to verify indoleamine 2,3-dioxygenase (IDO) activity. B. Quantification of kynurenine concentrations in mFTF in mice fed LFD or HFD using LC-MS/MS. C. Simplified diagram of the tryptophan metabolism pathway. D. Adipocytes were cultured with ¹³C₁₀-tryptophan at 1, 6 and 12 h. Intracellular metabolites from cells and extracellular metabolites from cell media were extracted for analysis by LC-MS/MS. The graph indicates the percentage of the metabolite labeled fraction (M10) compared to the unlabeled fraction (M0) from adipocytes (blue) and cultured media (red). Unlabeled media and cells cultured in unlabeled media were used as the control (Ctrl). E. MCF-10A cells were incubated with ¹³C₁₀-kynurenine enriched media ranging from 1 min to 72 h to perform isotope labeling experiments. Intracellular metabolites from cells and extracellular metabolites from cell media were extracted and analyzed by LC-MS/MS section to evaluate the cellular uptake of kynurenine and further metabolism by MCF-10A cells. Graphs represent the intracellular labeled fraction (M10) of kynurenine, picolinic acid, and kynurenic acid compared to the unlabeled fraction in MCF-10A cells (yellow) and cultured media (red). Unlabeled media and cells cultured in unlabeled media were used as the control (Ctrl). Data are displayed as means ± S.E.M., *Statistical comparison between the treatment and control groups for unlabeled studies, and between ¹³C-labeled control fraction and ¹³C-labeled treatment fraction for labeled studies. n = 3. (**p-value ≤ 0.01, ****p-value ≤ 0.0001).

mFTF contains factors from all cell types in adipose tissue. To determine whether adipocytes specifically generate and excrete kynurenine to the cellular milieu and consequently other metabolites in the downstream pathway, such as picolinic acid (Fig. 3C), we performed isotope labeling studies using ¹³C₁₀–tryptophan. Adipocytes were generated from bone marrow stromal cells as previously described [18-20] and incubated with media containing ¹³C₁₀–tryptophan. Metabolites were extracted from cells and analyzed using LC-MS/MS. Adipocytes incorporate ¹³C₁₀-tryptophan within 1 h and convert it to ¹³C₁₀-kynurenine (Fig. 3D). To test whether adipocytes secrete kynurenine derived from tryptophan, we also extracted metabolites from ¹³C₁₀-tryptophan containing media cultured with adipocytes. The presence of labeled ¹³C₁₀-kynurenine in the culture media confirms that adipocytes secrete kynurenine metabolized from tryptophan (Fig. 3D), evidence that adipocytes are a source for endogenous secreted kynurenine.

Further, we sought to verify that MCF-10A cells can uptake kynurenine from the environment. MCF-10A cells were incubated with ¹³C₁₀-kynurenine, and metabolites were extracted over a range of time points. Indeed, MCF-10A cells rapidly uptake ¹³C₁₀-kynurenine and metabolize it to downstream metabolites including kynurenic acid and picolinic acid (Fig. 3E). Together, these data demonstrate that adipocytes excrete kynurenine that can be taken up by MCF-10A cells.

3.4. AhR signaling is activated in MCF-10A cells exposed to adipocytes

To determine if paracrine activity between adipocytes and MCF-10A cells leads to activation of the Kyn/AhR signaling axis in MCF-10A cells, MCF-10A cells were then either 1) co-cultured with differentiated adipocytes in transwell (AdipoTW), where there is continuous crosstalk between the two cell types or 2) treated with adipocyte-conditioned medium (AdipoCM), where only adipocyte-secreted factors are present. RNA was isolated and AhR target gene CYP1B1 was measured as a direct readout of AhR activity. CYP1B1 was significantly elevated in MCF-10A cells in both models, suggesting adipocytes secrete factors that have the propensity to activate the AhR (Fig. 4A). Importantly, this effect is mediated through the AhR, where inhibition of the AhR using Kyn-101 abrogates CYP1B1 induction in MCF-10A cells both in AdipoTW or treated with AdipoCM (Fig. 4A). Due to continuous crosstalk between adipocytes and MCF-10A cells in the transwell co-culture, there is still a slight increase in expression of CYP1B1 likely due to the adipocytes continuing to secrete kynurenine, whereas in the AdipoCM there is finite amount of kynurenine to effectively inhibit the Kyn/AhR signaling axis. In both the transwell and conditioned medium models, AhR mRNA was slightly elevated (p < 0.05), complementing the findings with the mFTF-transformed cells (Fig. 1B). This induction was not dependent on AhR activity, as Kyn-101 failed to block the induction (Fig. 4B). The benefit of using these models is the ability to specifically interrogate the adipocyte component and allow continuous crosstalk between the adipocytes and MCF-10A cells, representing the biological microenvironment. Collectively, these data suggest that adipocytes induce AhR activation in MCF-10A, leading to CYP1B1 induction.

Fig. 4.

Factors secreted from adipocytes induce CYP1B1 through AhR activation. A. Gene expression analysis of CYP1B1 (Left) and AhR (Right) in MCF-10A cells in transwell co-culture with bone marrow-derived adipocytes (Adipo-TW) with the addition of Kyn-101 or vehicle control using Taqman qPCR. B. Taqman qPCR gene expression analysis of CYP1B1 (Left) and AhR (Right) in MCF-10A cells treated with conditioned media from bone marrow-derived adipocytes in the presence or absence of Kyn-101. Data obtained from TaqMan qPCR analysis were normalized to human TBP and represent a mean of three biological replicates ± S.D., where each point represents an individual replicate. A one-way ANOVA was performed to determine statistical significance and values ** p < 0.01, **** p < 0.0001 are considered statistically significant. NS represents values that are Not Significant.

3.5. Kynurenine is sufficient to induce 3D growth

Our data suggests that 1) VAT-stimulated AhR activity is elevated and retained during malignant transformation of MCF-10A cells and 2) AhR activation in MCF-10A cells is dependent on the uptake of kynurenine from adipocytes. Tto determine whether kynurenine itself is sufficient to stimulate transformation or if it requires cooperation from other adipokines, MCF-10A cells were treated with either 2 μM or 5 μM kynurenine in soft agar (Fig. 5A). These doses were based upon concentrations of kynurenine measured in the serum of people with obesity [33,34]. Five μM L-kynurenine significantly induced 3D growth compared to the vehicle control in MCF-10A cells, suggesting a clear role of AhR signaling in driving MCF-10A cell transformation. Importantly, 5 μM L-kynurenine effectively activated the AhR in MCF-10A cells as measured through the induction of AhR target gene CYP1B1, a process inhibited by the addition of 0.025 μM Kyn-101 or 1 μM CH223191, which affirms that AhR signaling is modulated at these concentrations. Collectively, we have shown a novel role of visceral adipose tissuederived kynurenine in MCF-10A cell transformation through the kynurenine/AhR signaling axis, unveiling a potential therapeutic target for obesity-associated TNBC prevention.

Fig. 5.

The transforming effect of kynurenine (Kyn) is prevented in the presence of Kyn-101. A. Soft agar transformation assay of MCF-10A cells in the presence of mFTF, kynurenine, or vehicle control. B. 3D colony growth assay of MCF-10A cells treated with kynurenine, Kyn-101, a combination of kynurenine and Kyn-101, and vehicle control. C. Taqman qPCR analysis of CYP1B1 in MCF-10A cells treated for 6 h with vehicle control, Kyn-101, or CH223191 in the presence or absence of kynurenine. Soft agar data are represented as a colony count of colonies over 25 μm and presented as the mean of biological triplicates ± S.D. where each individual point represents an individual replicate. Data obtained from TaqMan qPCR analysis were normalized to human TBP and represent a mean of three biological replicates ± S.D., where each point represents an individual replicate. A one-way ANOVA was performed to determine statistical significance and values * p < 0.05, ** p < 0.01 **** p < 0.0001 are considered statistically significant.

4. Discussion

The increase in frequency of both obesity and cancer highlights the need to better understand pathophysiology and how these two diseases may be connected. Our previous studies demonstrated that high-fat diet-induced visceral adiposity promotes ultraviolet B-induced skin tumor formation in vivo and that FGF2 from adipose tissue stimulates malignant transformation by activating the FGFR1 receptor on skin and mammary epithelial cells in vitro [1,2]. These studies provided pre-clinical support that adipose tissue distribution plays a role in tumor promotion. We applied our novel model of visceral FTF-promoted malignant transformation to investigate hormone-independent tumor promoters for TNBC, a breast cancer subtype in which visceral adiposity is a strong risk factor for both premenopausal and postmenopausal onset. Our data suggest that adipocyte-released kynurenine activates the AhR in mammary epithelial cells to promote malignant transformation. This key finding implicates 1) a hormone-independent mechanism of tumor promotion and 2) the AhR as a novel therapeutic target of TNBC progression in the context of visceral adiposity.

A role for the AhR at every stage of breast carcinogenesis has been supported [12,39-45]. The canonical role of the AhR is to act as a sensor of environmental mediators by inducing expression of a battery of canonical Phase I and Phase II metabolism genes [46,47]. As an example, epidemiological evidence points to poly aromatic hydrocarbons (PAHs) increasing breast cancer risk [48-51]. PAHs act largely through the AhR-mediated transcriptional upregulation of cytochrome P450s, which cause the biotransformation of the parent compounds into reactive intermediates. Moreover, the activity of AhR in oncogenic transformation and progression, independent of exogenous ligand, has been widely investigated in the context of breast cancer. In fact, the overexpression of AhR alone is sufficient to transform mammary epithelial cells [52]. Our findings corroborate this data showing that knocking down or antagonizing the AhR prevents transformation by mFTF (Fig. 2E). While we recognize that the fold-change in response to MFTF appears similar between the EV control cells and the AhR KD cells, there was no significant difference between AhR KD treated with mFTF and AhR KD without mFTF. The trend of a higher colony count with mFTF in AhR KD cells may be explained by the presence of a small amount transcription factor.

The impact of AhR is particularly evident in ER⁻ tumor progression, as studies suggest that constitutive expression of AhR is a marker of ER⁻ breast cancer [53]. Our patient data analysis corroborates these findings, demonstrating worse outcomes in ER⁻ breast cabcer patients with high AhR gene expression (Fig. 1C). While this is true for all breast cancer patients, the magnitude of difference becomes greater when analyzed for TNBC patients (Fig. 1C). Moreover, it has been suggested that constitutively active AhR mediates CYP1B1 expression, whereas ligand activation mediates both CYP1A1 and CYP1B1 [54]. This supports context-dependent molecular outcomes. We demonstrate that VAT-transformed MCF-10A cells have elevated levels of both CYP1A1 and CYP1B1 (Fig. 1B), whereas adipocyte co-culture was only able to induce CYP1B1. This suggests that CYP1A1 is a bystander of transformation while CYP1B1 induction is a contributor to the driving mechanism. Additionally, we and others have shown that non-transformed mammary epithelial cells express little to no CYP1A1 at baseline levels [54], making detection of expression changes using transient assays difficult. The downstream mechanism by which CYP1B1 promotes anchorage-independent growth (malignant transformation) in the absence of an exogenous AhR ligand is unclear. Literature shows that CYP1B1 may drive ER⁻ breast cancer by degrading estrogen and producing ROS to form DNA adducts [55], stimulating the epithelial to mesenchymal transition through WNT/β-catenin signaling [56] or regulating levels of claudin [57], an important component of tight junctions.

AhR activity [58] and adiposity [59-63] have been each shown to independently have oncogenic effects in breast cancer. We now demonstrate a novel link between AhR activation by adipocyte-derived kynurenine and malignant transformation. Kyn-101, which preferentially competes for the endogenous ligand binding site for the AhR, significantly inhibited the transforming effects of VAT (Fig. 2B). This finding sparked our investigation of mFTF-derived kynurenine and its tumor promoting activity. While culturing MCF-10A cells with 5 μM of kynurenine was required to significantly stimulate growth in soft agar, LC-MS/MS analysis demonstrated presence of ~0.3 μM of kynurenine in transforming VAT based on extrapolation from standards (Fig. 3B). Moreover, our studies reveal that adipocytes can metabolize kynurenine into kynurenic acid, an activator of AhR with greater potency than kynurenine [64]. Therefore, kynurenine may cooperate with kynurenic acid or other growth factors to promote malignant transformation.

The kynurenine pathway has received increasing interest as its roles in obesity, diabetes and neurological conditions have become apparent. Relevant to our study, Huang et al., recently demonstrated that mice fed a high fat diet have increased kynurenine and IDO in visceral adipose tissue depots [36]. Our HFD-fed mice also demonstrated a higher kynurenine/tryptophan ratio in the VAT filtrates (Fig. 3A), a characteristic that correlates with the functional activity of the VAT. The transforming activity of VAT was significantly and concentration-dependently blocked with Kyn-101 (Fig. 2B) and kynurenine alone stimulated growth in soft agar (Fig. 5A), but at a much higher concentration than was present in mFTF. This suggests that while mFTF-derived kynurenine is necessary for growth in soft agar, it is not independently sufficient. Kynurenine is likely working in concert with other growth factors and tumor promoters (and kynurenic acid). However, there is still the possibility that a single administration of kynurenine might induce malignant transformation. It has been hypothesized that a single activation of the AhR may lead to constitutive activation and transformation by upregulating IDO1 to produce more AhR ligands within the cell [11].

While our original interpretation of this study was that excess VAT promotes tumorigenesis, we now speculate that VAT secretions, by inducing CYP1B1, may enhance the likelihood that human mammary epithelial cells will be transformed upon exposure to carcinogens. This hypothesis will be explored in future studies. Overall, these studies demonstrate a mechanism of VAT-promoted triple negative breast cancer, identify a kynurenine as a potential biomarker of risk, and add rigor to existing data that the AhR is a druggable target for novel preventive therapies.

Abbreviations:

AhR

aryl hydrocarbon receptor

BMSC

bone-marrow stromal cells

CYP1A1

cytochrome P450, family 1, subfamily A, polypeptide 1

CYP1B1

cytochrome P450, family 1, subfamily B, polypeptide 1

DMEM

Dulbecco’s modified essential medium

EGF

epidermal growth factor

FBS

fetal bovine serum

FGF2

fibroblast growth factor 2

FGFR1

fibroblast growth factor receptor-1

HEPES

N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid

HS

horse serum

IDO

indoleamine 2,3-dioxygenase

Kyn

L-Kynurenine

MEM

modified essential medium

mFTF

mouse fat tissue filtrate

PBS

phosphate buffered saline

TNBC

triple-negative breast cancer

VAT

visceral adipose tissue

Data availability

Data will be made available on request.