Fetal alcohol exposure alters mammary epithelial cell subpopulations and promotes tumorigenesis

Abstract

Background

Fetal alcohol exposure (FAE) increases the risk of mammary tumorigenesis in adult offspring; however, the underlying mechanism remains unknown. This study tested the hypothesis that FAE shifts the mammary epithelial cell (MEC) composition towards one that promotes tumorigenesis.

Methods

Pregnant FVB/NJ dams bred to MMTV-Wnt1 male mice were given ad-libitum access to 5% alcohol in 0.2% saccharin solution from GD9–10 and 10% alcohol in 0.2% saccharin from GD11-GD19 or 0.2% saccharin solution from GD9-GD19. Thoracic and inguinal mammary glands from WT and Tg female offspring were harvested at 5 and 10 weeks of age and dissociated to yield a single cell suspension enriched for MECs for flow cytometry, mammosphere assay, and gene analysis. A subset of Tg offspring was followed for tumor formation.

Results

WT glands of FAE animals exhibited a decreased basal cell population and increased luminal:basal ratio at 10 weeks of age. qRT-PCR analysis of total MECs found that Hey1 mRNA expression was increased in the WT FAE group at 10 weeks of age. In Tg glands FAE increased the luminal progenitor cell population at 5 weeks of age but did not alter MEC composition at 10 weeks of age. Tertiary mammosphere forming efficiency was greater in the WT glands of FAE animals at 10 weeks of age. Tumor latency was decreased in the FAE group. Flow cytometry analysis indicated that FAE females developed tumors with an increased basal cell population.

Conclusions

These data indicate that FAE can shift MEC subpopulations, increasing the proportion of cells that are potentially vulnerable to transformation and affecting cancer risk.

Introduction

Breast cancer is one of the most common cancers in women with over 2 million new cases worldwide reported in 2018 (Bray et al., 2018). Many of the known risk factors such as family history, race, early menarche, and late menopause, cannot be controlled. However, other risk factors such as alcohol consumption can be managed through lifestyle changes (Feng et al., 2018). Much of breast cancer research has focused on adulthood but the risk for breast cancer may begin during fetal development (Soto et al., 2008; Hilakivi-Clarke and de Assis, 2006; Simmen and Simmen, 2011). For example, maternal exposure to diethylstilbestrol or a high fat diet in rats increases the risk for tumorigenesis in their offspring (Kawaguchi et al., 2009; Hilakivi-Clarke et al., 1997). In the context of alcohol, studies have indicated that fetal alcohol exposure (FAE) promotes tumorigenesis in rodent models (Hilakivi-Clarke et al., 2004; Polanco et al., 2010; Crismale-Gann et al., 2016; Ma et al., 2015). These statistics become particularly alarming when considering that half of women of childbearing age report drinking and 18% of these women binge drink; additionally, 11.5% of women admit to drinking during pregnancy (Denny et al., 2019; Tan et al., 2015). In fact, a recent study found that the rates of fetal alcohol spectrum disorder may actually be greater than what has been previously reported, with incidence rates that may be greater than that of autism (May et al., 2018).

The mammary epithelial cell (MEC) lineage originates from stem cells that divide asymmetrically to produce luminal and basal progenitor cells which maintain the mature luminal and myoepithelial (basal) cells of the gland (Visvader and Stingl, 2014). Breast cancer is a heterogeneous disease composed of diverse phenotypes that exhibit differing histopathology and molecular signatures (Lim et al., 2009; Prat and Perou, 2011). There is accumulating evidence that each subtype may arise from an individual cell subpopulation as both stem and progenitor cells can represent targets of oncogenic transformation (Liu et al., 2016; Visvader and Stingl, 2014).

While MEC subpopulations and subsequent tumor incidence have been found to be altered by exogenous factors in utero or in adulthood, the effect of FAE has not been examined (Chang et al., 2012; Rahal et al., 2013; Kim et al., 2013; Wang et al., 2014). Data from our lab using Sprague-Dawley rats indicates that FAE affects normal mammary gland growth early in development, alters tumor susceptibility in adulthood and promotes the development of tumors with a more aggressive phenotype (Polanco et al., 2010; Crismale-Gann et al., 2016; Polanco et al., 2011). However, the mechanism by which this occurs has not been determined. The objective of the present work was to determine if FAE can impact the MEC lineage and subsequent tumor composition using the MMTV-Wnt1 mouse model in which tumors arise from an expansion of the stem cell population (Li et al., 2003; Liu et al., 2004).

Materials and Methods

Animals and treatment

All animal protocols were approved by Rutgers University Institutional Animal Care and Use Committees (IACUC 15–023). Animals were housed in a controlled environment with a 12-hour light/dark cycle. The MMTV-Wnt1 line on an FVB background (FVB.Cg-Tg(Wnt1)1Hev/J) was obtained as a gift from Dr. Pamela Cowin with permission from The Jackson Laboratories. Female WT FVB/NJ mice (JAX stock #01800) were bred to either WT FVB/NJ males or Tg MMTV-Wnt1 male mice and checked for seminal plugs each morning. The day a plug was identified was considered gestational day (GD) 0. On GD9, pregnant dams were randomly assigned to the alcohol or control group. In lieu of water, dams in the alcohol group were given ad libitum access to 5% v/v alcohol in 0.2% saccharin solution for two days and then 10% v/v alcohol in 0.2% saccharin until birth on GD19. Control dams were given ad libitum access to 0.2% saccharin solution from GD9 through birth. The amount consumed by each dam was measured every 24 h. Treatment stopped when dams gave birth and pups remained with the dam until weaning on postnatal day (PND) 21. Litters were normalized to 5 to 9 pups. Pup weight was recorded every three days until weaning, following which body weights were recorded once a week until the end of the study. WT and Tg offspring (n = 10–13) were sacrificed at 5 and 10 weeks of age and mammary glands were harvested (Figure 1).

Figure 1. Experimental design.

Pregnant dams were given ad libitum access to ethanol in 0.2% saccharin (5% v/v from gestational day (GD) 9–10 and 10% v/v from GD11–19) or water with 0.2% saccharin from GD 9–19. Treatments were stopped at birth. Mammary glands from wild type (WT) and transgenic (Tg) female offspring were harvested at 5 and 10 weeks of age and an enriched population of mammary epithelial cells (MECs) was isolated. A separate group of Tg females were monitored for tumor formation.

A separate group of Tg female offspring (one offspring from each of 10–11 dams) were monitored for tumor development. Beginning at 6 weeks of age mice were palpated twice a week for tumors. Tumor growth was measured every 3–4 days using a caliper and mice were sacrificed when the tumor reached 1–1.5 cm in diameter (Figure 1). Primary tumors were harvested and divided in half, with one half fixed in 10% neutral buffered formalin (NBF; ThermoFisher; Waltham MA) and the other used for isolation of tumor epithelial cells. Lungs were drop fixed in 10% NBF to determine the occurrence of lung metastases by histological analysis.

Tissue histology

Tissues were fixed for at least 24 h then subsequently washed in PBS 3 times for 20 min. After washing, fixed tissues were dehydrated in increasing concentrations of ethanol (70%, 80%, 95%, 100%) and xylene twice for 10 min. Tissues were then placed in a 1:1 mix of xylene and paraffin for 1 h and then placed in fresh paraffin overnight and embedded the following morning. Tissues were sectioned at 5 μm thickness and placed on Superfrost® Plus microscope slides (VWR; Dublin, Ireland). Slides were baked for 15 min at 60°C and then deparaffinized using xylene and rehydrated through decreasing concentrations of ethanol. Tissue sections were then stained with hematoxylin followed by eosin, dehydrated with increasing concentrations of ethanol followed by xylene, then mounted using Permount (Fisher Scientific; Waltham, MA). Tumor pathology was evaluated by a toxicological pathologist at the Environmental Occupational Health Sciences Institute at Rutgers University.

Blood alcohol levels

A separate group of animals (n = 16) were given 10% v/v ethanol with 0.2% saccharin for 1 week. Trunk blood was collected 2 to 4 h after lights off and allowed to clot for 30 min at room temperature then spun at 1500 × g for 10 min at 4°C. Serum was stored at −20°C until analysis. Blood alcohol levels were measured using an Analox Alcohol Analyzer (Analox instruments).

Mammary gland whole mount analysis

Left or right inguinal mammary glands were harvested from both genotypes and treatments at 5 and 10 weeks of age and processed as previously described (Stires et al., 2016). The number of terminal end buds (TEB) was determined in whole mounts from animals at 5 weeks of age by counting the bulb-like structures at the ends of ducts at the periphery of the ductal tissue. Digital images of whole mounts were obtained using a Leica MDG41 stereomicroscope equipped with a camera and Leica Acquire software. Ductal elongation was analyzed in FIJI by measuring the length of ductal growth from the distal end of the lymph node to the edge of the epithelial structure in the fat pad. Mammary epithelial area was quantified using FIJI image threshold, highlighting the epithelial structure and calculating the area of that structure in the gland.

Mammary epithelial cell isolation

At sacrifice, left and right thoracic and inguinal mammary glands were harvested and lymph nodes were removed. For Tg offspring, glands from one female per dam were analyzed while for WT offspring, glands from 2 littermates per dam were pooled. MEC isolation followed procedures as previously described with minor modifications (Prater et al., 2013). All reagents were from Stem Cell Technologies unless otherwise noted. Glands were placed into 15 ml conical tubes containing a 1X collagenase solution [DMEM/F12 media (Hyclone; Pittsburgh, PA) with 50 μg/ml gentamicin (VWR Amresco; Dublin, Ireland) containing 1 mg/ml collagenase and 10 U/ml hyaluronidase] and dissociated on a rotating incubator at 37°C overnight for 15 h. After dissociation, red blood cells were lysed with a 1:4 mixture of HBSS/2% FBS (HF; Hyclone; Pittsburgh, PA and Atlanta Biologicals; Flowery Branch, GA, respectively) and 0.8% ammonium chloride, resulting in a suspension of mammary organoids. A single cell suspension was obtained by pipetting organoids in 0.25% pre-warmed trypsin-EDTA for 1 min, followed by the addition of 10 ml HF. Cells were spun for 10 min at 1400 × g at 4°C and resuspended in pre-warmed 5 mg/ml dispase plus 1 mg/ml DNAse I and pipetted for 1 min followed by addition of 10 ml of HF and filtered through a 40 μm mesh cell strainer (VWR; Dublin, Ireland). Freshly isolated cells were immediately stained for flow cytometry analysis, as well as plated for mammosphere culture. Remaining cells were stored at −80°C for qRT-PCR analysis.

Tumors were minced with scissors or blades into a paste-like consistency and dissociated in a rotating incubator at 37°C for 2 h in DMEM/F12 containing 3000 U/ ml collagenase with 1000 U/ml hyaluronidase and 50 μg/ml gentamycin (VWR Amresco). After dissociation, red blood cells were lysed in ammonium chloride as described above. At this point, tumor organoids were frozen in DMEM/F12 media containing 10% FBS and 10% DMSO (Sigma; St. Louis, MO) for a slow freeze and stored at −80°C until further use. Upon thawing, organoids were resuspended in 10 ml pre-warmed DMEM/F12 media and spun at 1200 × g for 10 min and further dissociated into a single cell suspension as described above.

Flow cytometry analysis

Antibodies used for flow cytometry were titrated to determine the optimal concentration. Approximately 0.5–1 × 10⁶ freshly isolated mammary or tumor epithelial cells were resuspended in HF. MECs were labeled with the following antibodies, all from BD Pharmingen (San Jose, CA) unless otherwise stated: biotinylated anti-CD31 (1:100, clone 390), anti-CD45 (1:100, clone 30-F11) and anti-TER119 (1:100, clone TER-119), anti-CD24-FITC (1:400, clone M1/69), anti-CD49f-PE-Cy7 (1:200, clone GoH3; BioLegend; San Diego, CA), anti-CD61-PE (1:1000, clone 2C9.G2; BioLegend; San Diego, CA) and streptavidin PerCP-Cy5.5 (1:100). Tumor epithelial cells were similarly labeled with the following exceptions: biotinylated anti-CD31, anti-CD45, and anti-TER119 were used at 1:800 dilutions, anti-CD29-PE (1:160, clone HM Β−1) was used instead of anti-CD49f, and anti-CD61-AlexaFluor 647 (1:800, clone 2C9.G2) and streptavidin APC-Cy7 (1:800) were used. Cells were incubated with biotinylated antibodies for 30 min in the dark on ice, then washed with HF and incubated with remaining antibodies and streptavidin for an additional 30 min. After washing, cells were resuspended at a concentration of 1 × 10⁶ cells/ml and at least 100,000 events were acquired for analysis. DAPI (Invitrogen; Waltham, MA) was added just prior to analysis to label dead cells. Single color controls were used to set up compensation at the beginning of each experiment. Data acquisition was performed using a Beckman Coulter Gallios flow cytometer. Gating was based upon single color staining and fluorescence minus one controls. The gating strategy is presented in Figure S1. Data were analyzed using FlowJo software (Tree Star, Inc.).

Mammosphere and tumorsphere formation

To assess mammosphere forming efficiency (MFE), freshly isolated MECs were plated in triplicate on 6-well ultra-low attachment plates (Corning) in serum-free DMEM/F12 medium containing B27 (without vitamin A; Gibco, Waltham, MA), 10 ng/ml epidermal growth factor (Stem Cell Technologies; Vancouver, BC), 10 ng/ml basic fibroblast growth factor (Stem Cell Technologies), 10 μg/ml heparin (Stem Cell Technologies) and 50 μg/ml gentamycin (VWR Amresco; Dublin, Ireland). MECs were plated at 10,000 cells/well in triplicate and cultured for 7 days to allow for primary sphere formation. Fresh media (1 ml) was added every 3 days. After 7 days, spheres were pooled from triplicate wells and dissociated with 0.05% trypsin-EDTA (Gibco) for 10 min at 37°C. Cells were replated at 5,000 cells/well in triplicate to allow for secondary sphere formation. After 7 days, spheres were pooled from triplicate wells, dissociated, and again replated at 5,000 cells/well to allow for tertiary sphere formation. Mammospheres larger than 50 μM were counted after 7 days and MFE was calculated by dividing the number of tertiary mammospheres by the number of single cells plated.

To assess tumorsphere forming efficiency, freshly isolated tumor cells were plated on 6-well ultra-low attachment plates without serum as described above with the following modification: primary tumor cells were plated at 100,000 cells/well in triplicate. Spheres were dissociated and plated an additional two times at 5,000 cells/well to monitor secondary and tertiary mammosphere formation as described above. For all assays, tertiary mammosphere formation was analyzed.

RNA isolation and analysis

Total RNA was purified from isolated MECs using the Qiagen RNeasy micro kit according to the manufacturer’s directions (Qiagen; Germantown, MD). RNA quantity and quality were assessed using the Nanodrop ND-100 (Thermo Scientific; Waltham, MA) and Agilent 2100 Bioanalyzer with the Agilent RNA 6000 Nano kit (Agilent Technologies; Santa Clara, CA), respectively. RNA (500 ng) was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Life Technologies; Waltham, MA) and used for RT-PCR analysis. Primers were validated as previously described (Agostini-Dreyer et al., 2015) using a pool of RNA obtained from isolated MECs.

qRT-PCR was performed as previously described (Agostini-Dreyer et al., 2015) with the following modification: cDNA samples were diluted 1:20 or 1:200 based on the relative standard curve established for each gene during primer validation. Samples were analyzed by determining the fold-change relative to the calibrator using the 2^-ΔΔCt method. The calibrator was a cDNA pool of 2–3 cDNA samples from each treatment group. Cyclophilin A was used as the housekeeping gene. Primer sequences are presented in Table 1.

Table 1.

Primer sequences for genes analyzed by qRTPCR.

Gene	Accession no.	Primer sequence (5’ – 3’)	Product length
K8	NM_031170.2	F: AGTTCGCCTCCTCATTGAC R: GTCGCAACAGGCTCCACT	80
K14	NM_001313956.1	F: AGCGGCAAGAGTGAGATTT R: AATAACCTGGAGGAGACCAAAG	115
Elf5	NM_001145813.1	F: GAGACCAAGACTGGCATCAA R: TCCAAAGTTCTCACCTGTGG	111
Ki67	NM_001081117	F: CTGCCTGTTTGGAAGGAGTAT R: AAGTCAAAGAGCAAGAGGCA	97
Gata3	NM_001355111.1	F: CGAGATGGTACCGGGCACTA R: GACAGTTCGCGCAGGATGT	136
Notch1	NM_008714.3	F: ACAACAACGAGTGTGAGTCC R: ACACGTGGCTCCTGTATATG	221
Dll4	NM_019454.3	F: GGAACCTTCTCACTCAACATCC R: CTCGTCTGTTCGCCAAATCT	140
Hey1	NM_010423.2	F: TGAGCTGAGAAGGCTGGTAC R: ACCCCAAACTCCGATAGTCC	170
Hes1	NM_008235.2	F: CCCCAGCCAGTGTCAACAC R: ACACGTGGCTCCTGTATATG	67
Cyclophilin	NM_008907.2	F: TGCTGGACCAAACACAAACGGTTC R: CAAAGACCACATGCTTGCCAT	63
Cyclin D1	NM_007631.2	F: CAGAGGCGGATGAGAACAAG R: GAGGGTGGGTTGGAAATGAA	97
Myc	NM_001177353.1	F: CACCAGCAGCGACTCTGAA R: CCCGACTCCGACCTCTTG	98

Statistical analysis

Differences between FAE and control groups were analyzed using an unpaired t-test or an unpaired t-test with Welch’s correction if variances were not equal using GraphPad Prism version 6.0 (La Jolla, CA). Body weights were analyzed using repeated measures one-way ANOVA. Significance of the Kaplan-Meier curve was determined using Log-rank Mantel-Cox test. P ≤ 0.05 was considered statistically significant. Tendencies were considered for 0.05< p ≤ 0.1.

Results

Dam and pup parameters

Dams fed alcohol consumed 5–6 ml of ethanol each day, which equaled approximately 12 g ethanol/kg body weight/day, while control dams consumed 7–9 ml saccharin solution per day (Figure S2A, B). Blood alcohol level was analyzed and found to average 82.l ± 17.1 mg/dl, which corresponds with the legal limit of intoxication in humans. Dam body weights did not differ between treatment groups throughout pregnancy. At birth, neither litter size nor average pup weight was affected by treatment (Figure S2C, D).

FAE does not affect overall gland morphology in either wild-type or transgenic offspring

Whole mount analysis of mammary glands from 5- and 10-week WT or transgenic offspring indicated that expression of the Wnt-1 oncogene caused an increase in lobular-alveolar hyperplasia relative to WT controls (Figure S3) as previously described (Tsukamoto et al. 1988). However, FAE did not affect overall gland morphology of either group as TEB number (5 weeks only), ductal elongation and ductal area were not different between the two groups.

FAE decreases the basal epithelial cell population and increases mammosphere forming activity in the normal mammary gland

To determine if FAE alters the distribution of MEC subpopulations in the WT gland, MECs were isolated and analyzed by flow cytometry at 5 and 10 weeks of age. Luminal and basal epithelial cell populations were distinguished based on the expression of CD24 and CD49f within the lineage negative (TER119, CD45 and CD31) population, where luminal cells were labeled as CD24⁺CD49f^lo and basal cells were labeled as CD24⁺CD49f^hi (Figure 2A, B). The luminal progenitor cells within the luminal population were identified based on CD61 expression (Vaillant et al., 2008, Visvader and Smith, 2011). When mammary glands were analyzed at 5 weeks of age, proportions of luminal, basal and luminal progenitor epithelial cells were comparable between treatment groups (Figure 2C). At 10 weeks of age FAE significantly reduced the basal cell population (p < 0.05; Figure 2D), leading to an increase in the luminal to basal ratio (p < 0.01; Figure 2D). There was no treatment effect on the luminal progenitor cell population.

Figure 2. Fetal alcohol exposure decreases the wild type (WT) basal cell population at 10 weeks of age.

Mammary glands of WT offspring exposed to alcohol or saccharin (control) in utero were harvested and total mammary epithelial cells (MECs) were analyzed by flow cytometry. Representative flow cytometry pseudocolor plots of WT MECs from mice at A) 5 and B) 10 weeks of age. Lineage negative (Lin-) cells were separated into CD24⁺CD49f^lo (luminal) and CD24⁺CD49f^hi (basal) expression. The luminal CD24⁺CD49f^lo population was further divided into CD61⁺ cells to determine the proportion of luminal progenitors within the luminal cell population (n = 7 at 5 weeks, n = 7–8 at 10 weeks). Proportion of MECs in luminal and basal cell populations as well as the luminal to basal cell ratio at C) 5 (n = 10) and D) 10 weeks of age (n = 13). Bars represent mean ± SEM. Student’s t test, *p < 0.05, **p<0.01.

To determine if these changes in MEC populations corresponded with changes in MFE, MECs were plated in anchorage independent conditions to promote the growth of stem-like/progenitor cells (Pastrana et al., 2011; Shaw et al., 2012). As expected, mammosphere numbers visibly decreased from primary to tertiary passage (Figure 3A). While FAE did not affect tertiary MFE at 5 weeks of age (Figure 3B), MFE was increased from 0.22% in the control group to 0.37% in the FAE group at 10 weeks of age (p < 0.05; Figure 3C). These results indicate that FAE shifts the MEC composition in the normal WT gland at 10 weeks of age and expands a cell population with increased sphere forming capabilities.

Figure 3. Mammosphere forming efficiency (MFE) is increased in wild type (WT) fetal alcohol exposed (FAE) offspring at 10 weeks of age.

Mammary glands of WT offspring exposed to alcohol or saccharin (control) in utero were harvested and total mammary epithelial cells (MECs) were isolated and plated for mammosphere culture. A) Representative images of primary through tertiary spheres (independent of phenotype or treatment). Tertiary MFE was determined at B) 5 and C) 10 weeks of age by dividing the number of spheres formed over the total number of cells plated (n = 9–10). Bars represent mean ± SEM. Student’s t test, *p < 0.05.

FAE alters MEC subpopulations in pubertal transgenic offspring

MECs were isolated from 5- and 10-week old Tg glands and analyzed by flow cytometry to determine if FAE shifts the MEC subpopulations (Figure 4A, B). At 5 weeks of age the luminal and basal populations were not different between the two groups; however, the luminal progenitor population was significantly higher in the FAE group (p < 0.01; Figure 4C). At 10 weeks of age the luminal and basal cell populations as well as the luminal progenitor populations were similar between the groups (Figure 4D). Despite the increase in luminal progenitor populations at 5 weeks of age, MFE was not affected at either age (Figure 4E, F).

Figure 4. Fetal alcohol exposure alters transgenic (Tg) mammary epithelial cell (MEC) composition at puberty.

Mammary glands of Tg offspring exposed to alcohol or saccharin (control) in utero were harvested and total MECs were isolated and analyzed by flow cytometry. Representative images of CD61⁺ luminal progenitor populations in MECs from Tg offspring at A) 5 and B) 10 weeks of age. Proportion of MECs in luminal (CD24⁺CD49f^lo), basal (CD24⁺CD49f^hi) and luminal progenitor (CD24⁺CD49f^loCD61⁺) cell populations C) at 5 and D) 10 weeks of age. Tertiary MFE of E) 5 week and F) 10-week Tg MECs. Bars represent mean ± SEM, n = 10–12; Student’s t test, **p < 0.01.

FAE alters Notch signaling in WT glands, but not in the presence of the Wnt1 oncogene

To further explore the shift in MEC subpopulations in both the normal gland at 10 weeks of age and the Tg gland at 5 weeks of age, several markers of interest were measured in total MECs by qRT-PCR analysis. In agreement with the 10-week WT flow data, expression of cytokeratin 14 (K14), a marker of basal MEC, was significantly lower in the FAE group (p < 0.05; Figure 5A) while expression of cytokeratin 8 (K8), a marker of luminal cells, was not altered (data not shown). Interestingly, Ki67 expression tended to be lower in the alcohol group (p = 0.0844; Figure 5A). To further explore the mechanism underlying the increased luminal to basal ratio in the 10-week MEC data, expression of genes that play a role in luminal cell fate and commitment was analyzed. The Notch pathway affects self-renewal and lineage-specific differentiation of stem and progenitor cells in the adult mammary gland (Bouras et al., 2008; Dontu et al., 2004). FAE significantly increased the expression of Hey1 (p < 0.01; Figure 5B), a target gene of the Notch pathway, while other members of the Notch pathway (Notch1, Hes1 and Dll4) were not affected. Expression of Elf5 and Gata3, genes linked to luminal cell differentiation (Asselin-Labat et al., 2007; Chakrabarti et al., 2012), was not altered by FAE (data not shown). Expression of the same set of genes was not different between the two groups at 5 weeks of age (data not shown). These results suggest that FAE may shift WT MECs towards a luminal cell fate by altering Notch signaling. In the Tg gland, no differences in Notch gene expression were observed at 5 or 10 weeks of age (Figure 6A and 6B). Interestingly, Ki67 again tended to be lower at 5 weeks of age (p = 0.0575) (Figure 6A).

Figure 5. Hey1 expression is increased in 10-week wild type (WT) MECs.

Mammary glands of WT offspring exposed to alcohol or saccharin (control) in utero were harvested and total mammary epithelial cells (MECs) were isolated and used for RNA isolation. Relative mRNA expression was determined by qRT-PCR in 10-week WT MECs for markers of A) basal cells and proliferation and B) Notch pathway. Bars represent mean ± SEM, n = 7 – 9 individual samples consisting of 2 pooled littermate MECs per sample. Student’s t test *p < 0.05, **p < 0.01.

Figure 6. Fetal alcohol exposure does not affect Notch gene expression in transgenic (Tg) mammary glands.

Mammary epithelial cells (MECs) isolated from glands of Tg offspring exposed to alcohol or saccharin (control) in utero were analyzed by qRT-PCR. Relative mRNA expression of Notch genes and Ki67 in A) 5-week Tg MECs (n = 6–8) and B) 10-week Tg MECs (n = 10). Bars represent mean ± SEM.

To determine if FAE could alter MEC subpopulations by augmenting the Wnt pathway, Myc and cyclin D1, two genes activated by Wnt, were examined (Figure 7). Neither gene was significantly increased by FAE in either WT or Tg offspring at either 5 or 10 weeks of age, although Myc tended to be increased at 5 weeks of age in FAE WT animals (p = 0.077).

Figure 7. Fetal alcohol exposure does not affect markers of Wnt signaling.

RNA from mammary epithelial cells (MECs) isolated from glands of offspring exposed to alcohol or saccharin (control) in utero was analyzed by qRT-PCR. Relative mRNA expression of cyclin D1 and myc is shown. Bars represent mean± SEM. A) 5 wk WT (n = 5); B) 5 wk Tg (n = 8); C) 10 wk WT (n = 11–14); D) 10 wk Tg (n = 10).

FAE decreases tumor latency and alters tumor epithelial cell subpopulations

Previous studies have shown that FAE increases the risk for tumorigenesis in rat and mice offspring (Polanco et al., 2010; Hilakivi-Clarke et al., 2004; Crismale-Gann et al., 2016; Ma et al., 2015). To determine if FAE also affected tumorigenesis in the Wnt1 Tg model, offspring were palpated for tumor development starting at 6 weeks of age. As shown in Figure 8A, FAE significantly decreased tumor latency (p < 0.05). By 30 weeks of age, 50% of the FAE group had developed tumors compared to less than 30% of animals in the control group. Growth rate of tumors was not affected by FAE (Figure 8B). No lung metastases were found in either group at the time of sacrifice (data not shown). Histological analysis indicated that tumors from both treatment groups contained adenomatous glandular elements and foci of unequivocal adenocarcinoma characterized by high mitotic rate, nuclear pleomorphism, and cytologic atypia. Rare areas of squamous metaplasia/adenosquamous carcinoma were observed. There was no evidence of increased aggressiveness of the carcinomas in the alcohol group; in fact, there was some suggestion that the alcohol group exhibited tumors that were progressing less rapidly. This observation requires further study for confirmation.

Figure 8. Fetal alcohol exposure decreases tumor latency in transgenic (Tg) offspring and increases basal epithelial cell population in tumors.

Tumor formation was monitored in Tg offspring exposed to alcohol or saccharin (control) in utero. Mice were palpated for tumors twice a week starting at 6 weeks of age. A) Kaplan-Meier curve showing tumor latency (n = 10 per group). Log-rank Mantel-Cox test, *p < 0.05. B) Scatter plot of tumor growth rate. Initial and final tumor volumes were measured. Tumor growth rate was calculated as change in tumor volume in mm³/days of tumor growth. C) Bar graph analysis of proportion of luminal (CD24⁺CD29^lo), basal (CD24⁺CD29^hi) and luminal progenitor (CD24⁺CD29^loCD61⁺) tumor cells from alcohol and control groups (n = 8–10). D) Bar graph analysis of tertiary sphere forming efficiencies (TFE) of tumor cells (n = 5–6). Bars represent mean ± SEM. Student’s t test, *p < 0.05.

To determine if the decreased latency was associated with a shift in the proportions of epithelial cell subpopulations, tumors were harvested when they reached 1 – 1.5 cm in diameter and analyzed by flow cytometry. Overall, FAE increased the basal cell population in the mammary tumors while the luminal and luminal progenitor populations were comparable between the two groups (p < 0.05; Figure 8C). When cells were plated for tumorsphere culture, tertiary tumorsphere forming efficiency was not significantly different between the groups (Figure 8D).

Discussion

Alcohol is a known teratogen; however, the connection between FAE and the mammary gland is minimally explored compared to other fetal tissues. Numerous studies, including several reports from our group, have shown that alcohol acts as a fetal insult on the mammary gland, promoting increased tumorigenesis and tumor multiplicity in rodents exposed to alcohol in utero (Hilakivi-Clarke et al., 2004; Polanco et al., 2010; Crismale-Gann et al., 2016; Ma et al., 2015). In the current study, the hypothesis that alcohol may affect mammary gland tumorigenesis by targeting MEC subpopulations was explored. To test this hypothesis, we utilized the transgenic MMTV-Wnt1 mouse model that spontaneously develops mammary tumors due to an increased stem/progenitor cell pool, as opposed to other mouse models where tumor formation results from different cells of origin (Shackleton et al., 2006; Mikaelian et al., 2004; Taneja et al., 2009). This allowed for the investigation of FAE on the mammary stem/progenitor cell pool and its role in tumorigenesis. The results show that FAE decreases tumor latency in MMTV-Wnt1 offspring with an expansion in the basal cell population of the tumors. Additionally, the use of WT FVB/NJ littermates found that FAE targets the MEC subpopulations in the normal mammary gland in the absence of additional insults.

Interest in the concept that cancer risk can begin during fetal development has brought attention to the fact that tissue development and cell fate can be altered by the fetal environment (Soto et al., 2008). The primary role of the MEC lineage is to develop and maintain the epithelial structure of the mammary gland. However, studies have shown that both pre- and post-natal insults can shift the lineage and influence the risk for mammary tumorigenesis (Chang et al., 2012; Lambertz et al., 2017; Wang et al., 2014; Rahal et al., 2013). In the present work, analysis of normal WT glands found a decreased basal cell population and an increased luminal:basal ratio in the FAE group at 10 weeks of age. These data corresponded with a greater tertiary MFE in the 10-week FAE glands. The mammosphere forming assay promotes the growth of spherical colonies enriched in stem-like cells and serves as a readout for stem/progenitor cell activity (Pastrana et al., 2011). As the mammary hierarchy continues to be defined, both basal and luminal populations have been found to contain repopulating cells, a characteristic of true stem/progenitor-like cells (Keller et al., 2012). In future studies, it will be important to directly determine if the mammospheres from FAE animals contain an increase in stem/progenitor cells and whether these are found in the luminal or basal subpopulations of the spheres. MECs also exhibit cell plasticity, as differentiated cells are capable of acquiring stem-like properties (Chaffer et al., 2011). Similar to our findings in the WT glands, loss of the IGF1 receptor (IGFR) decreased the basal cell population and increased luminal progenitors in the developing murine mammary gland. When the loss of IGF1R was investigated in Wnt1-induced tumorigenesis, tumor latency was reduced in bigenic MMTV-dominant-negative igf1r (dnigf1r)/Wnt1 animals (Rota et al., 2014). In the present study, while the luminal progenitor population did not increase in the WT gland based on CD61 expression, the increased luminal:basal ratio indicates a shift in the MEC composition.

Based on the shift towards an increased luminal population at 10 weeks of age with alcohol in the WT group and the increase in luminal progenitor cells with alcohol in the TG animals at 5 weeks of age, we hypothesized that signaling pathways that control luminal cell fate would be altered. Of the genes examined, expression of Hey1 was significantly greater in the FAE WT group. Hey1 is a main downstream effector of Notch signaling and increased expression is indicative of Notch signaling activation (Bouras et al., 2008; Hassan et al., 2013). Activated Notch signaling promotes MaSC differentiation into cells of the luminal lineage (Bouras et al., 2008; Dontu et al., 2004). For example, Elf5-null mice exhibit increases in mammary Notch signaling and CD 61+ luminal progenitor cells (Chakrabarti et al., 2012). Alternatively, knockout of the Notch related transcription factor Rbpj in mammary progenitor cells alters cell fate specification and leads to an increase of proliferating basal cells (Buono et al., 2006). In vitro, treatment with a Notch activating peptide promotes an increase in mammosphere formation and spheres with increased Notch signaling express a greater number of multi-potent cells (Dontu et al., 2004). Activation of Notch signaling also promotes tumorigenesis in mice (Boyle et al. 2017; Callahan and Egan, 2004, Jhappan et al., 1992; Ling et al., 2010). The increased luminal:basal ratio in conjunction with increased MFE may have resulted from the increased Notch signaling, promoting cells towards a luminal cell fate. Despite no difference in the Notch 1 receptor and Dll4 ligand, there are several other receptors and ligands that play a role in the Notch pathway; therefore, it is possible that FAE may have affected other receptors and ligands not measured in this study. Additionally, measuring protein levels of Notch pathway members may also give a clearer indication of the activation of Notch signaling in the WT mammary gland.

In the Tg mammary gland, a change in MEC composition was evident earlier, around the time of puberty, with an expansion of the luminal progenitor population in the FAE group at 5 weeks of age. By 10 weeks of age no differences were observed in MEC composition. Similarly, results from our previous work showed advanced mammary development in FAE females, based on increases in MEC proliferation and TEB numbers, at PND20 that was lost by PND40. Despite this return to a morphologically normal state, FAE animals exhibited an increased susceptibility to carcinogen-induced tumorigenesis (Polanco et al., 2011). In 2014, Wang et al. reported that pubertal bisphenol A (BPA) exposure increased the basal cell population in Balb/c mammary glands at 6 weeks of age but not later at 2 or 4 months of age; however, increased neoplastic lesions were observed in regenerated mammary glands from cells of animals exposed to BPA (Wang et al., 2014). These results suggest that early alterations in the mammary gland can affect risk later in life despite the evidence of a morphologically normal gland.

The decrease in tumor latency in FAE females with tumors exhibiting an expansion of the basal cell population suggests that the luminal progenitor expansion found at 5 weeks of age may contribute to the increased tumorigenesis. The increase in tumor basal cells is interesting given that stem cells are generally thought to reside in the basal compartment; however, this subpopulation also contains myoepithelial cells as well as basal progenitor cells. Direct alcohol exposure increases the cancer stem cell population in many cancers (Xu and Luo, 2017). In MCF7 breast cancer cells overexpressing ErbB2, alcohol treatment increases the cancer stem cell population and mammosphere formation (Xu et al., 2016). Further studies will be needed to determine if FAE increases the cancer stem cell population of the Wnt-1 tumors.

Interestingly, the cell proliferation marker Ki67 tended to be lower in FAE offspring at 10 weeks of age for WT animals and at 5 weeks of age for Tg offspring. Alcohol exposure decreases proliferation in a variety of tissues (Di Rocco et al., 2019). For example, proliferation of neural progenitor cells is transiently decreased in mice chronically exposed to alcohol (Rice et al., 2004). Similarly, neurospheres from fetal rat ganglionic eminence regions exhibit decreased proliferation and differentiation when cultured at a high concentration of ethanol (Vemuri and Chetty, 2005). Mammary glands from FAE offspring showed a decrease in Ki67 but an expansion of a MEC subpopulation that may be susceptible to tumorigenesis. This suggests that decreased tumor latency may have resulted from the shift in cell composition versus an increase in overall proliferation.

In conclusion, the results of this study are the first to show that FAE affects MEC composition, suggesting a possible mechanism by which it may promote tumorigenesis. The increase in the luminal progenitor population and subsequent increase in the basal cell population of the tumors is important given that luminal progenitors are the cell of origin for some basal-like breast cancers (Lim et al., 2009; Visvader and Stingl, 2014). Future work will be aimed at determining the mechanism by which FAE elicits this effect.