Infant Formula Feeding Changes the Proliferative Status in Piglet Neonatal Mammary Glands Independently of Estrogen Signaling

ABSTRACT

Background

Soy infant formula contains isoflavones, which are able to bind to and activate estrogen receptor (ER) pathways. The mammary gland is sensitive to estrogens, raising concern that the use of soy formulas may promote premature development.

Objective

We aimed to determine if soy formula feeding increases mammary gland proliferation and differentiation in comparison to other infant postnatal diets.

Methods

White-Dutch Landrace piglets aged 2 d received either sow milk (Sow), or were provided milk formula (Milk), soy formula (Soy), milk formula supplemented with 17-beta-estradiol (2 mg/(kg·d); M + E2), or milk formula supplemented with genistein (84 mg/L of diet; M + G) until day 21. Mammary gland proliferation and differentiation was assessed by histology, and real-time RT-PCR confirmation of differentially expressed genes identified by microarray analysis.

Results

Mammary terminal end bud numbers were 19–31% greater in the Milk, Soy, and M + G groups relative to the Sow and M + E2, P <0.05. Microarray analysis identified differentially expressed genes between each formula-fed group relative to the Sow (±1.7-fold, P <0.05). Real-time RT-PCR confirmed 2- to 4-fold increases in mRNA transcripts of genes involved in cell proliferation, insulin-like growth factor 1 (IGF1), fibroblast growth factor 10 (FGF10), and fibroblast growth factor 18 (FGF18), in all groups relative to the Sow, P <0.05. In contrast, genes involved in cell differentiation and ductal morphogenesis, angiotensin II receptor type 2 (AGTR2), microtubule associated protein 1b (MAP1B), and kinesin family member 26b (KIF26B), were significantly upregulated by 2-, 4-, and 13-fold, respectively, in the M + E2 group. Additionally, mRNA expression of ER-specific gene targets, progesterone receptor (PGR), was increased by 12-fold, and amphiregulin (AREG) and Ras-like estrogen regulated growth inhibitor (RERG) expression by 1.5-fold in the M + E2 group, P <0.05. In the soy and M + G groups, mRNA expressions of fatty acid synthesis genes were increased 2- to 4-fold.

Conclusions

Our data indicate soy formula feeding does not promote ER-signaling in the piglet mammary gland. Infant formula feeding (milk- or soy-based) may initiate proliferative pathways independently of estrogenic signaling.

Introduction

Breastfeeding is considered the optimal choice for early infant nutrition. The American Academy of Pediatrics (AAP) advocates exclusive breastfeeding for at least the first 6 mo of life (1, 2). When women are unable to exclusively breastfeed, the AAP recommends the use of infant formula from birth ≤12 mo. Even with these recommendations, many women will choose infant formula as a reasonable nutritional alternative to breast milk. In a recent NHANES, 2003–2010, >81% of infants, aged 0–12 mo consumed some type of infant formula; cow milk formula was the most widely used (69%) followed by soy formula (12%) (3). Soy formula was introduced into the market as an alternative for infants with milk allergies and has been used for over 100 y (4). Although clinical studies in human infants consuming soy formula ≤12 mo of age report normal growth and development when compared with breast-fed and milk formula-fed counterparts (4–8), other studies suggest soy formula use can result in adverse reproductive health outcomes in women (9–12). As recently as 2011, the National Toxicology Program raised its level of concern about the potential reproductive toxicity of soy formula from negligible to minimal (13).

Isoflavones are naturally occurring compounds in plants, structurally similar to 17-beta-estradiol (E2), and are commonly classified as phytoestrogens (14). Genistein and other soy-associated isoflavones, namely daidzein and equol are able to bind weakly to estrogen receptors (ERs) α (ESRA) and β (ESRB), with a preference for ESRB (15–17). In this regard, these compounds fit the definition of endocrine-disrupting compounds, which have the potential to interfere with reproductive development and fertility (18). Several prospective and retrospective studies have linked any soy formula use with long-term adverse outcomes, including increased risk of endometriosis and uterine fibroid burden (11, 19). With respect to breast development, some report a positive association between soy feeding and breast tissue size in female infants and toddlers (10, 12), whereas others observed no changes in breast tissue volume with formula (5, 7).

In rodent models, pure genistein treatment is known for estrogenic-like effects, including increased uterine weight and reproductive toxicity (20, 21). Additionally, neonatal exposure in female rats to either pure isoflavones or soy protein isolate via the maternal diet reduced the number of mammary terminal end buds (TEBs), a marker for increased mammary gland differentiation (22–25). It has been proposed that these effects are the result of genistein binding to ERs and activating ER-associated signaling pathways during early development (24, 25). Yet, the application of these findings to infants receiving soy formula during the postnatal feeding period is not clear. An alternative animal model is a neonatal piglet, which is a good surrogate for many aspects of infant physiology and nutritional requirements (26, 27). Piglets can be fed commercially available infant formulas immediately after birth, which more readily models the feeding paradigm of human infants. In this study, we utilized a neonatal piglet model to determine if postnatal formula feeding, specifically soy formula, activates ER-signaling pathways thus promoting morphological changes resulting in premature development. Therefore, groups of female piglets aged 2 d, were fed either cow milk formula, soy formula, or sow milk until postnatal day (PND) 21. To serve as positive controls, additional groups were fed cow milk formula supplemented with known ER agonists, E2, or pure isoflavone genistein, to activate ER signaling pathways in the mammary gland for comparison purposes.

Methods

Animal experiments

Animals were housed in the animal facilities of the Arkansas Children's Hospital Research Institute. Animal maintenance and experimental treatments were conducted in accordance with the ethical guidelines for animal research established and approved by the Institutional Animal Care and Use Committee at the University of Arkansas for Medical Sciences. A detailed description of protocol and animal care has been reported (28), and data from this animal study has been previously published (28–30). Litters containing 8–11 piglets from multiparous Dutch Landrace Duroc sows fed a soy-free diet were used. At birth, female piglets were allowed to suckle at the farm for 2 d, then randomly assigned to continue sow-feeding [Sow, n = 6] at the farm, or transferred to Arkansas Children's Nutrition Center and provided either a cow milk-based formula [Milk, n = 6] (Similac Advance powder, Ross Products, Abbott Laboratories), or soy-based formula [Soy, n = 6] (Enfamil Prosobee Lipil powder, Mead Johnson Nutritionals) or the same cow milk-based formula (Similac Advance, Ross Products, Abbott Laboratories) supplemented with E2 [M + E2, n = 6], or with pure genistein [M + G, n = 6] for 21 d. Formula-fed piglets were housed individually and trained to drink from small bowls on a fixed schedule: the first week was every 2 h, the second week was every 4 h, and the third week was every 6 h, to provide 1.04 MJ/(kg·d) until day 21. In the M + E2 group, the estradiol stock was made daily in ethanol and added to the food bowls at each feeding using the required dosing concentration to achieve the serum E2 concentrations achieved in peripubertal piglets (15.6–30 pmol/L) (31). In the M + G group, the genistein stock was made daily and added to the food bowls to achieve a concentration (84 mg/L of diet) that was comparable to the genistein concentrations in Prosobee, 33 mg/L of diet, and soy protein isolate, 54 mg/L, which is the sole protein source of all soy infant formulas (Dupont). Diet composition for milk and formulas, including estimation of sow milk, have been published (30). Piglets were killed by exsanguination after anesthetization, 6–8 h after final feeding period. Tissue samples were snap-frozen in liquid nitrogen and stored at −70°C until analysis.

Serum soy isoflavones

Individual- and total-serum isoflavone concentrations were extracted and analyzed by LC-MS as previously described (32).

Serum hormones

E2 concentrations were measured using a customized MULTI-SPOT® 96-well Custom Steroid Hormone Panel Plate (Lot # Z0055407, Meso Scale Diagnostics) using the MESO SQ120 QuickPlate instrument. Growth hormone (GH) was measured by ELISA (MBS026403 My BioSource).

Mammary gland histology

Mammary gland whole mounts were prepared (17, 25, 33, 34) and TEBs were counted as previously described (25). Two mammary glands per piglet, a total of 12 glands per group, were used for the TEB assessment; distal portions of the gland were counted using an Olympus BX40 microscope; numbers were determined in 5 contiguous 100× fields (diameter 2 mm), and averaged per diet group.

Mammary microarray analysis

Microarray analysis of gene expression was performed as previously described (34). For each group, individual piglet cDNAs were grouped into pools (n = 2 piglets/pool), hybridized to an Affymetrix GeneChip porcine genome array #900,624. The probe array was washed and stained using the Affymetrix GeneChip fluidics station 450 and scanned using GeneChip Scanner 3000 (Affymetrix). Data on probe-level intensities were extracted using GeneChip operating software; CEL(raw) files were generated.

Gene enrichment analysis

Gene ontology (GO) enrichment analysis was performed on differentially expressed genes between each formula group versus the Sow (David Bioinformatics Resources, 6.8). Significant pathways were determined using the Fisher's Exact test followed by Benjamini–Hochberg correction for multiple corrections.

Microarray validation by real-time RT-PCR

For each sample, total RNA was reverse-transcribed into cDNA and subsequent real-time RT-PCR analysis was carried out using SYBR green and ABI 7500 sequence detection system (Applied Biosystems) as described previously (33). Gene expression was analyzed for each piglet using the 2^−ΔΔct method relative to a housekeeping gene 18S, or beta-2-microglobulin (B2M), gene amplification and expressed as fold change compared with the Sow group. Pig primer assays, amphiregulin (AREG), angiotensin II receptor type 2 (AGTR2), fibroblast growth factor 10 (FGF10), and B2M were purchased (Biorad Laboratories, Inc.); primer sequences for all other genes are detailed in Supplemental Table 1.

Western blot analysis

Nuclear protein fractions were extracted from 100 mg of frozen mammary tissue as previously described (35). Proteins (30 μg) were separated by SDS-PAGE by standard methods using anti-ESRA rabbit polyclonal antibody (Santa Cruz Technology) or anti-ESRB rabbit polyclonal antibody (Abcam) against blotted proteins. Protein bands were quantified using a densitometer and band densities were corrected for total protein loaded by staining the membrane with 0.1% amido black.

Statistical analysis

For microarray analysis, the .CEL format data obtained from the GeneChip Scanner in .CEL format was analyzed using the Affymetrix Expression Console wherein the raw data were subjected to the normalization of signal intensity and quality control. The RMA method (Robust Multichip Analysis) was used for background adjustment, quantile normalization, and summarization. The data were then analyzed for differentially expressed genes between groups using Affymetrix Transcriptome Analysis Console (TAC), following the software guidelines. The criteria for selecting differentially expressed genes between each formula group compared with Sow were a ± 1.7-fold change and a t-test P value <0.05. A Venn diagram was generated using jvenn (36). All other experimental outcomes are summarized as mean ± SE. Differences in isoflavone concentrations between Soy and M + G groups were determined by 2-tailed Student's t-test, P <0.05. With respect to body weight, tissue weight, mammary gland morphology, hormone concentrations, real-time RT-PCR and protein expression, comparisons between diets were accomplished by 1-factor ANOVA followed by a Student–Newman–Keuls post hoc analysis or by Kruskal–Wallis 1-factor ANOVA on ranks followed by Dunn post hoc analysis, P <0.05.

Results

Sow-, Milk-, Soy-, M + E2-, and M + G-fed piglets had similar body weights at PND 21

Mean body weights after the final feeding (PND 21) for the Sow, Milk, and Soy groups were 7.8 ± 0.26 kg, 8.3 ± 0.21 kg, and 7.8 ± 0.35 kg, respectively (28, 29). Mean body weights for M + E2 and M + G were 7.7 ± 0.29 and 7.7 ± 0.21 kg, respectively. On PND 21, mean body weights did not significantly differ by group, P = 0.41.

Serum soy isoflavone concentrations in piglets are detected in the Soy- and M + G-fed piglets

In the Soy-fed group we detected, genistein, daidzein, glycitein, and o-desmethylangolansin in the serum (Table 1). We detected genistein in the M + G group. Genistein concentrations between Soy and M + G groups did not differ (Student's t-test, P = 0.29). These concentrations are comparable to circulating soy isoflavones, genistein, and daidzein concentrations reported in infants aged 4 mo consuming soy formula (37, 38).

TABLE 1.

Individual and total-isoflavone concentrations measured in serum from female piglets receiving Sow or Milk, Soy, M + E2, and M + G formulas during PND 2–21¹

Isoflavones (ng/mL)	Sow	Milk	Soy	M + E2	M + G
Genistein	n.d.	n.d.	487 ± 107	n.d.	736 ± 121
Daidzein	n.d.	n.d.	306 ± 72.6	n.d.	n.d.
Glycitein	n.d.	n.d.	20.3 ± 5.47	n.d.	n.d.
O-desmethylangolensin	n.d.	n.d.	299 ± 107	n.d.	n.d.
Equol	n.d	n.d	n.d	n.d	n.d
Total			1105 ± 244		736 ± 121

¹Values are mean ± SE, n = 6. E2, 17-beta-estradiol; Milk, milk formula; M + E2, milk formula supplemented with E2; M + G, milk formula supplemented with genistein, n.d., not detected; PND, postnatal day; Sow, sow milk; Soy, soy formula. For genistein, daidzein, glycitein, o-desmthylangolensin, and equol the limit of detection (LOD) is 1.28 ng/mL, 0.206 ng/mL, 2.96 ng/mL, 0.585 ng/mL, and 1.59 ng/mL respectively.

Serum E2 concentrations and reproductive tissue weights increased in the M + E2-fed piglets

In the M + E2 group, the E2 concentration measured was comparable to E2 concentrations previously reported in the peripubertal period in female pigs (31). E2 concentration in the M + E2 group was 10- to 20-fold higher in comparison to all other groups (Table 2). E2 concentrations were also 1.6-fold higher in the Milk and M + G groups and 2-fold higher in the Soy group relative to the Sow group. GH concentrations were 2-fold higher in the Milk, M + E2, and Soy groups, relative to M + G and the Sow group. In the mammary gland, TEBs were increased (30%) in Milk, Soy, and M + G groups relative to the Sow and M + E2 groups. In the M + E2 group, ovary and uterus weights increased 3- to 4-fold, respectively, relative to all other groups.

TABLE 2.

Serum hormone concentrations and reproductive organ weights from Sow, Milk, Soy, M + E2, or M + G fed female piglets¹

	Sow	Milk	Soy	M + E2	M + G	P value
Serum hormones
E2, pg/mL	0.13 ± 0.01a	0.21 ± 0.03b	0.31 ± 0.04b	2.8 ± 0.96c	0.21 ± 0.02b	<0.001
GH, ng/mL	6.9 ± 0.9a	19 ± 0.40c	12 ± 0.34b	17 ± 0.40c	7.8 ± 0.40a	<0.001
Reproductive organs
TEBs, structures/mm2	5.9 ± 0.38a	7.8 ± 0.34b	7.7 ± 0.37b	6.3 ± 0.53a	7.7 ± 0.33b	0.002
Ovary weight, g	0.44 ± 0.02a	0.37 ± 0.03a	0.35 ± 0.01a	1.5 ± 0.09b	0.42 ± 0.02a	<0.001
Uterus weight, g	4.4 ± 0.32a	4.1 ± 0.11a	3.6 ± 0.11a	20 ± 1.26b	4.1 ± 0.10a	<0.001

¹Values are mean ± SE, n = 6. Labeled means in a row without a common letter differ, P <0.05. E2, 17-beta-estradiol; GH, growth hormone; Milk, milk formula; M + E2, milk formula supplemented with E2, M + G, milk formula supplemented with genistein; Sow, sow milk; Soy, soy formula; TEB, terminal end bud.

Formula feeding, E2 supplementation, and soy isoflavones alter mammary gland gene expression

Microarray analyses identified a total of 249, 226, 435, and 301 differentially expressed transcripts between Milk, Soy, M + E2, and M + G groups, respectively, versus the Sow group. Of these genes, 125 were held in common between the Milk, Soy, and M + E2 groups, with 224 genes expressed exclusively in the M + E2 group, 44 genes expressed exclusively in the Milk group, and 43 genes expressed solely in the Soy group (Figure 1A). Similarly, in Figure 1B, common genes held between Milk, Soy, and M + G were 141 with 65, 38, and 40 expressed exclusively to the M + G, Milk, and Soy groups, respectively. A list of the differentially expressed genes common and specific to these diet groups can be found in Supplemental Tables 2 and 3. GO analysis clustered the differentially expressed genes from each diet group, relative to Sow, into the biological pathways listed in Table 3. For all groups, the biological processes (BP) pathway, cell proliferation, was enriched. In the M + E2 group BP pathways: response to hormone, cell morphogenesis during differentiation, and regulation of extrinsic apoptotic signaling, were enriched. Considering the BP pathway, the lipid biosynthetic process was enriched in the Soy and M + G groups. Confirmation of these pathways was performed by real-time RT-PCR analysis on selected genes associated with each GO term. In Table 4, fibroblast growth factors 10 and 18 (FGF 10, 18) were increased in groups relative to the Sow. Insulin-like growth factor (IGF1) mRNA expression was increased in all groups relative to Sow with the greatest expression (2-fold) in the M + E2 group, P <0.05. Androgen receptor (AR) mRNA was increased in all groups, with the highest expression (6-fold) in the M + E2 relative to Sow. In contrast, kinesin family member 26b (KIF26B) and AGTR2 mRNA were increased in the M + E2 group, relative to all other groups, P <0.05 (Table 4). Microtubule associated protein 1b (MAP1B) mRNA expression was increased in all groups relative to Sow, with expression in the M + E2 group higher in comparison to the Milk, Soy, and M + G groups. Gene expression of transforming growth factor β2 (TGFB2) was increased in M + E2 and M + G groups relative to Milk, Soy, and Sow groups (Table 4).

FIGURE 1.

Venn diagrams of (A) the overlap among significantly differentially expressed genes in the mammary gland between the Milk, Soy, and M + E2 groups, and (B) the overlap among significantly differentially expressed genes between the Milk, Soy, and M + G groups at PND 21. Milk, milk formula; Soy, soy formula; M + E2, milk formula supplemented with E2; E2, 17-beta-estradiol; M + G, milk formula supplemented with genistein.

TABLE 3.

GO terms of biological pathways enriched in mammary glands from piglets fed Milk, Soy, M + E2, or M + G formula relative to the Sow group¹

	P value
GO term	Milk	Soy	M + E2	M + G
GO:0,008283-Cell proliferation	0.001	0.012	0.001	<0.001
GO:0,009725-Response to hormone	0.267	0.108	0.003	0.055
GO:0000904-Cell morphogenesis during differentiation	0.575	n.d.	0.001	0.145
GO:2,001,236-Regulation of extrinsic apoptotic signaling	n.d.	n.d	<0.001	0.075
GO:0,008610-Lipid biosynthetic process	0.306	0.001	0.156	0.011

¹GO terms among significant transcripts differentially expressed between Milk, Soy, M + E2, and M + G groups versus the Sow group determined by DAVID functional annotation analysis. Significance determined by the Benjamini–Hochberg correction, P value <0.05. E2, 17-beta-estradiol; GO, gene ontology; Milk, milk formula; M + E, milk formula supplemented with E2, M + G, milk formula supplemented with genistein; n.d., pathway not detected; Sow, sow milk; Soy, soy formula.

TABLE 4.

Real-time RT-PCR confirmation of genes associated with GO terms: cell proliferation and cell morphogenesis during differentiation in mammary tissue from piglets in the Sow, Milk, Soy, M + E2, and M + G groups¹

Genes	Sow	Milk	Soy	M + E2	M + G	P value
GO:0,008283-Cell proliferation
FGF10 mRNA, fold of Sow	1.0 ± 0.34a	2.9 ± 0.49b	3.3 ± 0.82b	4.3 ± 0.92b	3.2 ± 0.83b	0.038
FGF18 mRNA, fold of Sow	1.0 ± 0.08a	5.8 ± 0.41b	6.8 ± 0.98b	4.8 ± 0.43b	6.1 ± 0.36b	<0.001
IGF1 mRNA, fold of Sow	1.0 ± 0.14a	1.5 ± 0.08b	1.6 ± 0.15b	2.3 ± 0.25d	1.7 ± 0.10c	<0.001
GO:0000904-Cell morphogenesis during differentiation
AR mRNA, fold of Sow	1.0 ± 0.12a	2.1 ± 0.32b	1.8 ± 0.22b	6.3 ± 0.82c	2.1 ± 0.32b	<0.001
KIF26B mRNA, fold of Sow	1.0 ± 0.26a	1.3 ± 0.33a	1.5 ± 0.21a	13 ± 3.2b	1.4 ± 0.29a	<0.001
AGTR2 mRNA, fold of Sow	1.0 ± 0.26a	0.41 ± 0.09a	0.48 ± 0.11a	2.2 ± 0.33b	0.68 ± 0.20a	<0.001
MAP1B mRNA, fold of Sow	1.0 ± 0.17a	2.8 ± 0.23b	2.6 ± 0.39b	4.5 ± 0.44c	2.8 ± 0.27b	<0.001
TGFB2 mRNA, fold of Sow	1.0 ± 0.08a	1.3 ± 0.15a	1.1 ± 0.28a	1.8 ± 0.13b	1.6 ± 0.07b	0.007

¹Values are means ± SE, n = 6. Labeled means in a row without a common letter differ, P <0.05. AGTR2, Angiotensin II receptor type 2; AR, Androgen receptor; E2, 17-beta-estradiol; FGF10, fibroblast growth factor 10; FGF18, fibroblast growth factor 18; GO, gene ontology; IGF1, insulin-like growth factor 1; KIF23B, Kinesin family member 26b; Milk, milk formula; M + E2, milk formula supplemented with E2; M + G, milk formula supplemented with genistein; PND, postnatal day; MAP1B, microtubule associated protein 1b; Sow, sow milk; Soy, soy formula; TGFB2, transforming growth factor β2.

Nuclear expression of ESRB was increased in the mammary tissue of Soy-, M + G-, and M + E2-fed groups

As illustrated in Supplemental Figure 1, mammary ESRA and ESRB protein expression was observed in the nucleus by Western blot. In Figure 2A, ESRA expression increased (∼20%)  in the M + E2 group, relative to all other g roups. In   Figure 3A, B, mRNA expression of ESRA target genes progesterone receptor (PGR), and AREG also increased in the M + E2 group in comparison to the Milk, Soy, M + G, and sow groups, P <0.05 (39, 40). ESRB expression increased ∼2-fold in the M + E2, Soy, and M + G groups relative to the Milk and Sow groups (Figure 2B). The ESRA: ESRB ratio was also lower in the M + E2, Soy, and M + G groups in comparison to the Milk and Sow groups (Figure 2C). In Figure 3C, mRNA Ras-like estrogen-regulated growth inhibitor (RERG) was increased in the M + E2 group in comparison to Milk, Soy, M + G, and Sow groups, P <0.05; RERG is a potential ERSB gene target (41, 42).

FIGURE 2.

Fold expression of nuclear ESRA (A) and ESRB (B) proteins and the ESRA: ESRB ratio (C) in mammary tissue of piglets fed Milk, Soy, M + E2, and M + G formula relative to Sow. Values are mean ± SE, n = 5–6. Bar graphs without a common letter differ, P <0.05. ESRA, estrogen receptor α; ESRB, estrogen receptor β; Milk, milk formula; Soy, soy formula; M + E2, milk formula supplemented with E2; E2, 17-beta-estradiol; M + G, milk formula supplemented with genistein; Sow, sow milk.

FIGURE 3.

Fold mRNA expression of genes involved in E2-induced ESRB and ESRB-signaling pathways, PGR (A), AREG (B), and RERG (C), in the mammary tissue of piglets fed Milk, Soy, M + E2, or M + G formula relative to Sow. Values are mean ± SE, n = 6. Bar graphs without a common letter differ, P <0.05. ESRA, estrogen receptor α; ESRB, estrogen receptor β; Milk, milk formula; Soy, soy formula; M + E2, milk formula supplemented with E2; E2, 17-beta-estradiol; M + G, milk formula supplemented with genistein; Sow, sow milk; PGR, progesterone receptor; AREG, amphiregulin; RERG, Ras-like estrogen regulated growth inhibitor.

Soy formula feeding upregulated genes involved in lipid biosynthetic pathways in mammary tissue

GO analysis suggested lipid biosynthetic processes are enriched in the Soy and M + G groups (Table 2). Genes associated with this BP pathway include those related to de novo cholesterol and fatty acid synthesis. Real-time RT-PCR confirmed mRNA expression of acetyl-coenzymeA synthetase 2 (ACSS2) and long-chain fatty acid-coenzyme A ligase 1 (ACSL1) was increased in Soy and M + G groups, relative to the Milk, M + E2, and Sow groups (Figure 4A, B). Additional genes associated with de novo fatty acid synthesis, fatty acid synthase (FASN) and stearoyl-CoA desaturase (SCD1) were also increased 2- to 3-fold, respectively, in the Soy group, P <0.05 (Figure 4C, D). SREBP1 mRNA was higher in the Milk, Soy, and M + G groups in comparison to the Sow and M + E2 groups, with the highest expression in the Soy and M + G groups, P <0.05 (Figure 4E). Similarly, peroxisome proliferator activated receptor γ (PPARG) mRNA was increased in the Soy and M + G groups relative to M + E2 and Sow groups (Figure 4F).

FIGURE 4.

Fold mRNA expression of genes involved in fatty acid synthesis, ACSS2 (A), ACSL1 (B), FASN (C), SCD1 (D), SREBP1 (E), and PPARG (F), in the mammary tissue of piglets fed Milk, Soy, M + E2, or M + G formula relative to Sow. Values are mean ± SE, n = 6. Bar graphs without a common letter differ, P <0.05. Milk, milk formula; Soy, soy formula; M + E2, milk formula supplemented with E2; E2, 17-beta-estradiol; M + G, milk formula supplemented with genistein; Sow, sow milk; ACSS2, acetyl-coenzyme A synthetase 2; ACSL1, long-chain fatty acid coenzyme A ligase 1; FASN, fatty acid synthase; PPARG, peroxisome proliferator activated receptor γ; SCD1, stearoyl-CoA desaturase; SREBP1, sterol regulatory element-binding transcription factor 1.

Discussion

Soy isoflavones such as genistein are considered to have estrogen-like properties; therefore, concerns have been voiced regarding the safety of soy protein-based infant formula as a nutritional alternative to breast milk (13). Several clinical studies have suggested any soy formula use during the postnatal feeding period (0–12 mo) may result in adverse outcomes in E2-sensitive organs, i.e., uterus, breast (9–12). Previous work in weanling male and female rats consuming soy protein has shown no marked increases in E2-responsive gene expression or altered mammary gland morphology as was observed in rats that were exogenously exposed to E2 (33). Yet, there are other reports in the literature suggesting soy protein consumption does have effects on mammary gland development in rodents, however, these effects are dependent on several different factors, including developmental stage of exposure, presence or absence of endogenous estrogens, and the type of soy product used, i.e., pure isoflavones versus soy protein (24, 25, 43–45). It is unknown if there are any potential interactions of soy isoflavones with estrogen receptor signaling pathways in the mammary gland during the postnatal feeding period. Therefore, to address this question, we fed 2-d old neonatal female piglets either Soy, Milk formulas, or sow milk until weaning. We compared mammary gene expression and mammary gland morphology in relation to 2 additional groups receiving milk formula supplemented with either E2 (M + E2) or the pure isoflavone genistein (M + G). In the M + E2 group, we reported increased uterine and ovarian size relative to the other groups. These findings are consistent with the literature reporting increased size in the uterus in response to postnatal E2 exposure in pig neonates (46) and indicate that the E2 treatment was successful as a positive control in triggering ER-associated changes.

Mammary gland development occurs in 3 stages: embryonic, pubertal, and adult (47, 48). After birth, the rudimentary epithelial tree developed during embryogenesis remains quiescent until puberty when the onset of sex steroids, i.e., E2 and progesterone, in conjunction with GH becomes the master regulators of mammary gland outgrowth. Of these, E2 promotes cell proliferation and ductal elongation into the mammary fat pads (47). As expected, feeding piglets in the M + E2 group had elevated serum E2 concentrations (20-fold). Surprisingly, E2 concentrations were also elevated (2-fold) in the Milk, Soy, and M + G groups relative to the Sow group. In the M + E2 group, mRNA expression of genes involved in cell proliferation, ductal elongation, cell growth arrest, and apoptosis was upregulated in the mammary tissue; all of which are consistent with an E2 primary role in initiating branching morphogenesis during puberty (47, 48). Mammary glands express both ESRA and ESRB (49, 50). In the M + E2 group both ESRA-responsive genes (AREG, PGR), which associate with TEB formation and branching morphogenesis, and ESRB-responsive genes (TGFB, RERG), which associate with growth inhibition are upregulated, thereby potentially limiting the effect of E2 on cell proliferation (39–41, 51). Alternatively, ESRB may have decreased ESRA transcriptional activity by acting as a trans-dominant repressor through competition for DNA binding sites, as indicated by the decrease in the ESRA to ESRB ratio in the M + E2 group (52). Isoflavone consumption increased ESRB nuclear expression in the Soy and M + G groups. However, we observed no increased mRNA expression of selected ESRA- and ESRB-targets in the Milk, Soy, and M + G groups. We have previously reported such inhibition of ER-mediated gene expression in the mammary gland of rats fed soy protein isolate and treated with E2 (34). Taken together, these findings demonstrate soy isoflavones consumed during the postnatal feeding period do not induce E2-associated changes in mammary gene expression in neonatal piglets.

Although the mammary gland is considered a lipid-synthesizing organ during lactation, neonatal glands would not be expected to display significant lipogenesis. Interestingly, we observed soy-specific, ER-independent mRNA responses related to lipid biosynthesis following postnatal feeding. Microarray analysis revealed an upregulation of 2 de novo synthesis pathways, cholesterol and fatty acid synthesis. Previously, we reported similar increases in cholesterol synthesis markers in the livers of the Milk- and Soy-formula-fed groups and suggested that this hepatic upregulation of cholesterol synthesis reflects differences in cholesterol content between breast milk and infant formula (29). In rodents, we previously reported soy protein isolate feeding effects on fatty acid homeostasis in the liver and mammary gland (17, 34, 53); in particular, activation of PPAR mRNAs and regulated pathways (53). In the current study, mRNA expression of transcription factors (SREPB1, PPARG) associated with increased lipid synthesis were upregulated in the Milk, Soy, and M + G groups. However, mRNA expression of fatty acid synthesis genes were only upregulated in the Soy and M + G groups. These findings complement previous work in the field suggesting soy isoflavones are capable of binding to and activating nuclear receptors regulating lipogenesis, in particular, liver X receptor (LXR), and PPARG (54). Soy formula feeding had a greater impact on de novo fatty acid synthesis gene expression (i.e., FASN, SCD1) than genistein alone. This difference in response to soy formula versus pure isoflavone feeding is frequent in the literature, and expected since soy protein isolate (used in soy formula) is a complex mixture of proteins, peptides, isoflavones, and additional phytochemicals suggested to have bioactive abilities (55).

Noteworthy is the heightened proliferation responsein mammary gland as indicated by increased TEBs in the Milk, Soy, and M + G groups. GO analyses showed significant differences in cell proliferation pathways and real-time RT-PCR confirmed a modest upregulation of FGF10, FGF18, and IGF1, irrespective of E2 supplementation. FGF10 and IGF1 are key proliferative markers associated with murine mammary ductal outgrowth observed at puberty (56, 57). Consistent with these findings TEB numbers were increased in the Milk, Soy, and M + G groups. Our data are consistent with previous reports in older infants (≥ 4 mo) citing increased plasma IGF1 concentrations in response to formula feeding (58). Our findings suggest formula feeding may influence mammary gland morphology via the GH/IGF1 axis during postnatal feeding. Long-term effects associated with the upregulation of IGF1 signaling in the mammary gland during postnatal feeding are unknown and require further investigation.

In conclusion, soy formula feeding in neonatal piglets did not elicit an estrogenic response in the mammary gland during the postnatal feeding period. This null finding is in stark contrast to the mammary histology and gene expression changes observed in the M + E2 group, which served as a positive control. These findings are particularly relevant to the risk assessment of soy infant formula given the relevance of the neonatal piglet model to infant feeding practices, especially considering the circulating isoflavones measured in our piglet model are similar to those observed in soy formula-fed infants (37, 38). Our data are consistent with several clinical studies reporting no increases in breast bud growth in formula-fed female infants during the first year of life (7, 10, 12). Soy isoflavones may have organ-specific effects so additional studies in other reproductive organs are warranted to rule out the possibility of estrogenic actions or developmental toxicity from soy infant formula.

Notes

This work is supported by the USDA-ARS Project 6026-51000-012-06-S; LY is partially funded by NIH 1P20GM121293 and NIH 1R21AI146521.

Author disclosures: The authors report no conflicts of interest.

Supplemental Tables 1–3 and Supplemental Figure 1 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used: AAP, American Academy of Pediatrics; AGTR2, angiotensin II receptor type 2; AR, androgen receptor; AREG, amphiregulin; BP, biological processes; B2M, beta-2-microglobulin; ER, estrogen receptor; ESRA, estrogen receptor α; ESRB, estrogen receptor β; E2, 17-beta-estradiol; FASN, fatty acid synthase; FGF10, fibroblast growth factor 10; FGF18, fibroblast growth factor 18; GH, growth hormone; GO, gene ontology; IGF1, insulin-like growth factor 1; M + E2, milk formula supplemented with 17-beta-estradiol; M + G, milk formula supplemented with genistein; Milk, milk formula; PGR, progesterone receptor; PND, postnatal day; PPARG, peroxisome proliferator activated receptor γ; RERG, Ras-like estrogen regulated growth inhibitor; SCD1, stearoyl-CoA desaturase; Soy, soy formula; TEB, terminal end bud; TGFB, transforming growth factor β.