Loss of glutaredoxin 3 impedes mammary lobuloalveolar development during pregnancy and lactation

Abstract

Mammalian glutaredoxin 3 (Grx3) has been shown to be important for regulating cellular redox homeostasis in the cell. Our previous studies indicate that Grx3 is significantly overexpressed in various human cancers including breast cancer and demonstrate that Grx3 controls cancer cell growth and invasion by regulating reactive oxygen species (ROS) and NF-κB signaling pathways. However, it remains to be determined whether Grx3 is required for normal mammary gland development and how it contributes to epithelial cell proliferation and differentiation in vivo. In the present study, we examined Grx3 expression in different cell types within the developing mouse mammary gland (MG) and found enhanced expression of Grx3 at pregnancy and lactation stages. To assess the physiological role of Grx3 in MG, we generated the mutant mice in which Grx3 was deleted specifically in mammary epithelial cells (MECs). Although the reduction of Grx3 expression had only minimal effects on mammary ductal development in virgin mice, it did reduce alveolar density during pregnancy and lactation. The impairment of lobuloalveolar development was associated with high levels of ROS accumulation and reduced expression of milk protein genes. In addition, proliferative gene expression was significantly suppressed with proliferation defects occurring in knockout MECs during alveolar development compared with wild-type controls. Therefore, our findings suggest that Grx3 is a key regulator of ROS in vivo and is involved in pregnancy-dependent mammary gland development and secretory activation through modulating cellular ROS.

Postnatal mammary epithelial cell growth and gland development are controlled by a complex of hormones and signaling pathways (1, 6, 28, 43, 48, 63, 71). During pubertal growth, growth hormone, ovarian estrogen, progesterone, and IGF-I are required to initiate branching morphogenesis to generate a ductal tree that fills fat pads (63, 71). Upon pregnancy, the combined actions of progesterone and prolactin promote the proliferation and the differentiation of specialized structures, i.e., alveoli, where milk is synthesized and secreted during lactation (7, 43, 48). After weaning, the lack of demand for milk triggers the process of involution in which the gland undergoes remodeling to remove the milk-producing epithelial cells and subsequent adipocyte differentiation (43, 70). The different stages of mammary gland development are regulated by numerous signaling pathways (57).

Reactive oxygen species (ROS) are considered by-products of aerobic metabolism in all oxygenic organisms (18). Cells also actively generate ROS as signals through the activation of various oxidases and peroxidases in response to internal developmental cues and external stresses (36). However, when in excess, ROS induce oxidative stress, causing a wide range of damage to macromolecules and eventually leading to apoptotic or necrotic cell death (59, 66, 67), which are often associated with human diseases (22–24, 60, 62, 78).

Intracellular ROS and redox balance have been implicated in mammary epithelial cell growth and differentiation (34, 65). Recent studies indicate that ROS-mediated signaling pathways play an important role in the process of epithelial-to-mesenchymal transition, which is crucial for tumor metastasis (5, 16, 55). Thioredoxin (Trx) and glutaredoxin (Grx) are the two major antioxidant systems that play a vital role in redox homeostasis (2, 14). Grxs are ubiquitous, small heat-stable disulfide oxidoreductases which are conserved in both prokaryotes and eukaryotes (31, 32). Grxs appear to be involved in many cellular processes and play an important role in protecting cells against oxidative stress (39). Grxs can be categorized into dithiol Grxs, which contain two cysteine residues in their active motifs, and monothiol Grxs, which contain a single cysteine residue in their putative motifs (39). There is a growing body of evidence suggesting that monothiol Grxs may have multiple functions in biogenesis of iron-sulfur clusters, iron trafficking and homeostasis, protection of protein oxidation, cell growth, and proliferation (13, 14, 29, 30, 38, 42, 44, 50, 53, 58). In mammalian cells there are two monothiol Grxs, Grx3 and Grx5 (33, 73). Mammalian Grx5 is a mitochondrial Grx that plays a critical role in iron-sulfur cluster biogenesis and heme synthesis in red blood cells (8, 75) and appears to be crucial for protecting osteoblasts from oxidative stress-induced apoptosis (40). Grx3, also termed thioredoxin-like 2 (Txnl2) or PICOT (protein kinase C interacting cousin of thioredoxin), was originally identified through a yeast two-hybrid screen, in which Grx3 physically interacted with the protein kinase C theta isoform (73). Genetic studies also demonstrate that Grx3 is essential for early embryonic growth and development, as the deletion of Grx3 causes embryonic lethality (10, 15).

Our previous work indicated that the knockdown of Grx3 in human breast cancer cells dramatically increased ROS levels and apoptosis, inhibited the proliferation and invasion of breast cancer cells, and diminished mammary tumor growth and metastasis in xenograft models (54). These studies further demonstrated that Grx3 played a critical role in human breast cancer cell growth and metastasis via regulating redox homeostasis and NF-κB signaling (54).

However, the precise function of Grx3 in mammary epithelial cell growth and differentiation and gland development in vivo remains to be fully elucidated. In this study, we investigated the expression of Grx3 during mammary gland developmental stages. We created Grx3 floxed mice by flanking the exon 2 of Grx3 with two loxP sites and further generated Grx3 MEC-specific knockout (KO) mice (Grx3MKO) by intercrossing Grx3 floxed mice with MMTV-Cre transgenic ones. We functionally characterized the Grx3 MKO mice and documented the molecular mechanism underlying Grx3 function in mammary epithelial cell growth and differentiation. Taken together, these findings suggest that the presence of Grx3 in mammary epithelial cells may be critical for lobuloalveolar development during pregnancy and lactation.

MATERIALS AND METHODS

Reagents.

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless stated otherwise. Dulbecco’s modified Eagle medium (DMEM), DMEM/F12, and HyClone Newborn Bovine Calf Serum (CS) were obtained from Thermo Scientific (Waltham, MA). Fetal bovine serum (FBS) was from ATLANTA Biologicals (Lawrence, GA). The 0.25% Trypsin EDTA was from Mediatech (Manassas, VA). Penicillin-Streptomycin Solution was from Global Cell Solutions (Charlottesville, VA). Collagenase type I was purchased from Worthington (Lakewood, NJ). Anti-cyclin D1, anti-PCNA, and anti-cyclin B1 antibodies were purchased from Cell Signaling Technology. Anti-β-actin antibody was bought from Sigma-Aldrich. Anti-GAPDH was bought from Chemicon International. Anti-Keratin 14 polyclonal antibody was from BioLegend (San Diego, CA). Anti-milk specific protein antibody was purchased from Nordic Immunological Laboratory (Tilburg, The Netherlands), and secondary antibodies were from Cell Signaling Technology (Beverly, MA). Grx3 monoclonal antibody was made in-house using full-length human Grx3 recombinant protein. This antibody was validated and used in our previous studies, in which Grx3 only detected a predicted 38-kDa protein in mouse embryos (and embryonic fibroblasts), but not in KO embryos (12). It was also validated when used in normal human and breast cancer patient breast tissues (54).

Generation of a conditional knockout allele for the Grx3 gene in mice.

To generate the targeting vector for Grx3 floxed mice (Fig. 1A), a bacterial artificial chromosome (BAC) clone, RP23–349D22, was obtained, which contained the mouse Grx3 gene from Children’s Hospital Research Center at Oakland (CHRCO, Oakland, CA). Two DNA fragments, 5.0 kb (left arm) and 5.4 kb (right arm) around exon 2 (Fig. 1A), were subcloned from this BAC by recombineering (37) and used for homologous recombination. A 356-bp DNA fragment containing the targeted exon 2 with its immediate 5′- and 3′-introns (partial) was amplified by PCR and inserted in between two loxP sites of the NeoFrtLoxP vector (Fig. 1A) (74). Two TK cassettes were inserted into the 5′-end of the targeting vector. R1 mouse embryonic stem cells (46) were transfected by electroporation with the linearized targeting construct. Positive embryonic stem (ES) cells were selected with G418 (Invitrogen) and ganciclovir. About 200 ES clones were screened by a long-PCR approach and several ES clones were obtained with the targeting construct (Fig. 1B). Two ES clones (no. 2 and no. 6) were expanded and used for the production of chimeras. Blastocyst injection and germline transmission were done by standard techniques. All cloning and genotyping primers are listed in Table 1.

Fig. 1.

Generation of Grx3 floxed mice. A: diagram of genomic structure (partial) of Grx3 gene and target vector. Purple arrowheads indicate two lox P sites. Red arrows indicate two primers for genotyping PCR. B: genotypic PCR identifies positive ES clones. 1, wild-type ES cells; 2 to 6, Grx3 ES clones; nos. 2, 3, 4, and 6 are positive; no. 5 is negative. C and D: genotyping PCR to identify wt, heterozygous, and homozygous progenes and Cre positive mice. E: deletion of Grx3 genomic DNA was detected in KO MG by genotyping PCR.

Table 1.

Primer sequences for cloning and genotyping

	Sequence
Exon2-F	gcc taa gga tcc gca cac tgc agt gga gtt ctt tgt ttt cta gg
Exon2-R	gcc taa gga tcc aga gtt caa cac aaa gtt caa g
Left arm
Left fra-F	gcc taa gga tcc gcc tgt cta tgt aag ttt tat gtt t
Left fra-R	gcc taa gaa ttc gcc cac aca ccc cac cac tga gct
Right fra-F	gcc taa gaa ttc gcc agt gac tcc acc gtc agc cca c
Right fra-R	gcc taa gcg gcc gcc ttc ctg tct gga taa aca gtc aag
Right arm
Left fra-F	gcc taa Gtt taa acc gtt ata ttt gca gca agc tat t
Left fra-R	gcc taa gaa ttc gcc gtt tcg caa cat act acc ttc
Right fra-F	gcc taa gaa ttc gcc ccc aaa tga gtg aga ccc cct
Right fra-R	gcc taa ggc gcg ccc tcc ttc cta aaa caa atc tta c
Primer 1 Forward	gca agt tca agg tca gtt tga
Primer 2 Reverse	gcc taa gaa ttc gcc gtt tcg caa cat act acc ttc
Primer 3 Forward	ctg ggc acc atg ggc tcc ac
Primer 4 Reverse	tgt cct ttg ccc att ctg tag c
Cre Forward	ctg cat tac cgg tcg atg ca
Cre Reverse	cag cat tgc tgt cac ttg gt

F, forward; R, reverse; Fra, fragment.

Mouse strains.

FVB/NJ mice were purchased from Jackson Laboratory. Grx3 MG knockout mice with mixed background were generated by intercrossing Grx3 floxed mice (B6/129Sc mixed background) with MMTV-Cre mice (Line D-B6129F1). Previous reports indicated that mammary gland lactation in Line D was not affected by expression of MMTV-Cre (56, 76). There was no difference in MG development between Grx3^flox/flox and MMTV-cre transgene alone mice. We used mixed background littermates as controls in our studies. All mouse strains were housed in temperature-controlled environment and fed a standard chow diet (LabDiet 5V5R chow; LabDiet, St. Louis, MO) ad libitum unless stated otherwise. The virgin mice used for these studies were housed together before analysis. All animal protocols were approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine.

Mammary gland whole mounts.

Mammary gland whole mounts and images were obtained following the procedure as previously described (17). In brief, MMTV-Cre/Grx3^flox/flox (Grx3 MKO) and aged-match Grx3^flox/flox and MMTV-cre mice as controls at age of 4 wk, 6 wk, 10 wk, pregnancy (P) at day 14–15, and lactation (L) at day 1 were euthanized. Both inguinal mammary gland fat pads (no. 4) were excised, mounted on the Superfrost slides (Thermo Fisher Scientific. Waltham, MA), and fixed for a minimum 2 h in Carnoy’s solution (60% ethanol, 30% chloroform, and 10% glacial acetic acid). The fixed glands were washed in 70% ethanol for 15 min and then rinsed in water for 5 min. The mammary glands were stained overnight at 4°C in carmine alum stain (1 g carmine and 2.5 g aluminum potassium sulfate in 500 ml water). The glands were then dehydrated progressively in 70–95%-100% ethanol, and cleared in xylene for 1 h. Mammary whole mounts were photographed using a MZ6 dissecting microscope (Leica, Deerfield, IL) and an Olympus (Lake Success, NY) Oly 760 video camera. Morphometric analysis of virgin and pregnant mammary ductal development was conducted as previously described (27). After segmentation of the ductal tree, mammary ductal specific end points included ductal area, ductal perimeter, total branch count, and branch density. Morphometric analysis of lobuloalveolar development consisted of determining epithelial area in H&E-stained tissue sections prepared from glands collected at day 1 of lactation.

H&E staining and immunohistochemistry.

Sections of no. 4 mammary glands from Grx3 MKO and wild-type controls were stained with H&E staining (26). Immunohistochemistry was performed using a highly sensitive streptavidin-biotin-peroxidase detection system with an in-house made Grx3 monoclonal antibody (54). Bromodeoxy-uridine labeling was conducted by intraperitoneal administration of a dose of 3 mg/kg body wt. Animals were then euthanized at 2 h after dosing and BrdU staining was conducted as previously reported (11).

Isolation of mammary epithelial cells.

Mammary epithelial cells (MECs) were isolated from control and knockout MGs at P15–18 or L1–2 following the protocol as previously described (72). In brief, MG no. 3 and no. 4 glands (without muscle and lymph nodes) were collected and minced with scissors in 10 cm culture plate. Ten milliliters of digestion buffer (DMEM/F12 medium containing 100 μg/ml gentamycin, 5% FBS, 1X anti-mycotic solution, 2 mg/ml collagenase type A) were added and incubated in 37°C water bath for 2–3 h with gentle shaking until no clumps remained. Digested tissues were transferred to 50-ml tubes and centrifuged at 1,400 rpm for 3 min at 4°C. Supernatant was discarded, and pellets were washed 3 times with 1X PBS at differential speeds (i.e., 1,400, 1,300, 1,200 rpm) until supernatants were clear. One milliliter of 0.25% trypsin was added into each digested tube and incubated for 10 min at 37°C; then 5 ml wash buffer (DMEM/F12 medium containing 5% FBS and 50 μg/ml gentamycin) was added to stop digestion, and centrifuged at 1,400 rpm for 3 min at 4°C. Pellets were then treated with 1 ml of Dispase II (5 mg/ml) for 3 min at 37°C, then washed with 5 ml of wash buffer to stop digestion. Purified MECs were resuspended in 1–2 ml of wash buffer and stored at −80°C until use.

Western blotting analysis.

Mammary tissue homogenates or MEC lysates were run on SDS-PAGE gel and Western blot analysis was conducted following an established procedure (12). Apoptosis was assessed by examining for cleaved caspase 3 in WT and KO MECs extracted from MGs collected at both day 14 of pregnancy and day 1 of lactation.

Quantitative real-time PCR analysis.

Total RNA was extracted from either mammary gland tissues or MECs of wild-type controls and Grx3 MKO mice at 4, 6, 10, and 12 wk of age, pregnancy (14.5 day), and lactation (1 day). Purified RNA samples underwent reverse transcription to yield cDNA. qPCR was performed using the SYBR Green-based system on the Bio-Rad CFX96. Differences in gene expression of control and knockout mice were quantified by the comparative CT method. Keratin 5 was selected to serve as the internal control because its expression level was consistent between WT and KO MECs and relatively stable across different stages compared with keratin 8, whose expression was somewhat affected by Grx3 deletion at different developmental stages.

Detection of ROS by DHE staining.

Dihydroethidium (DHE) staining to measure ROS production in MECs followed published procedures with modification (34, 77). In brief, the wax-embedded MG slides were immersed in xylene twice for 10 min each, then immersed in 100% ethanol twice for 10 min each, in 95% ethanol once for 5 min, in 75% ethanol once for 5 min, in 50% ethanol once for 5 min, and finally in distilled water once for 10 min. Slides were washed 2 times for 10 min each in 1x PBS; placed in ice-cold 4% PFA with PBS for 10 min; washed with PBS 3 × 10 min; incubated for 5 min with 5% Triton X-100 in 1x PBS, and washed twice with PBS for 10 min each. They were then incubated with 5 μM DHE for 1 h at RT, washed 3 x 10 min with PBS, and covered. They were allowed to dry overnight at RT in a cabinet away from light. Vectashield mounting media was placed on the slides and the slides were covered, allowed to dry overnight again at RT away from light, and then imaged under confocal microscope with the same setting.

Statistical analysis.

All results are shown as means ± SE. Student’s t test was used to compare 2 groups. P < 0.05, P < 0.01, and P < 0.001 were used as indicators of the level of significance.

RESULTS

Grx3 expression during mammary gland development.

Immunohistochemical analysis of Grx3 expression in the developing mouse mammary glands (MGs) revealed that Grx3 was expressed in all stages of mammary development (Fig. 2). In the terminal end buds (TEBs) of the peripubertal MG, Grx3 was detected in different cellular subtypes including cap cells, myoepithelial cells, luminal epithelial cells, most of body cells, and stroma cells as well (Fig. 2, A and B). It appears that Grx3 was predominantly localized in the cytoplasm of those cells in TEBs (Fig. 2B). However, more Grx3 proteins were detected in ductal and lobuloalveolar epithelial cells in the mature virgin (Fig. 2C), pregnant (Fig. 2D), and lactating (Fig. 2E) MGs, whereas less Grx3 proteins were detected in involuting MGs (Fig. 2F). It was also evident that Grx3 was present in both cytoplasm and nuclei of luminal cells but was rarely detected in other subtypes of cells (Fig. 2, C–F). In agreement with these observations, double immunolabeling with antibodies against Grx3 and CK14, a marker to label the myoepithelial cells, demonstrated that Grx3 is localized in luminal epithelial cells (Fig. 2, G–I). The quantification of Grx3 expression in the developing MG using q-PCR analysis of Grx3 mRNA levels in isolated mammary epithelial cells (MECs) demonstrated that Grx3 expression was upregulated in MECs of pregnant and lactating MGs compared with those of virgin MGs (Fig. 2J). Taken together, these findings suggest that Grx3 may play an important role in MG development during pregnancy and lactation.

Fig. 2.

Grx3 expression at different mammary gland developing stages. A and B: terminal end bud of 7-wk-old virgin mice; Grx3 proteins were detected in cap, myoepithelial, luminal epithelial, most of body cells, and stroma cells. Grx3 proteins were examined in myoepithelial cells and luminal epithelial cells in gland ducts at 10th week of age in virgin mice (C), at pregnancy day 14 (D), at lactation day 1 (E), at involution day 7 (F). 200× magnification for A, 400× magnification for B–F. G–I: double immunolabeling of mammary gland tissue sectioned from MG at P14.5 with antibodies against Grx3 (G) and CK14 (H). I: shown is the merged image with Grx3 in luminal epithelial cells (arrow) and CK14 is present in myoepithelial cells (arrowhead). Scale bars, 50 µm. J: q-PCR analysis of Grx3 mRNA levels in mammary epithelial cells (MECs) isolated from glands at 10th week, pregnancy day 14.5, and lactation day 1. Cytokeratin 5 (CK5) was used as an internal control. Student’s t-test; n = 3. **P < 0.01, L1 vs. 10th week; ***P < 0.001, P14.5 vs. 10th week.

Generation of mammary epithelial cell-specific Grx3 knockout mice.

The disruption of Grx3 in germline transgenic mice causes embryonic lethality at 12.5 days postgestation (12). To understand the role of Grx3 in the mature MGs, a conditional knockout mouse line was generated, in which the second exon of Grx3 was deleted through Cre-mediated recombination. The details of the targeting vector and the generation of Grx3 floxed mice are described in materials and methods (Fig. 1). Homozygous mice carrying two floxed Grx3 alleles grew and developed normally; both female and male mice were fertile. When Grx3 floxed mice were intercrossed with MEC specific MMTV-Cre mice (Line D) to generate the MMTV-Cre/ Grx3^flox/flox mice, termed as Grx3 MKO mice, the DNA sequence of the second exon of Grx3 was specifically deleted in MECs, which was revealed by PCR analysis of genomic DNA extracted from different tissues (Fig. 1, C–E). Immunohistochemical analysis indicated that Grx3 protein levels were significantly reduced in Grx3 MKO, but remained detected in certain MECs in the ducts of virgin MGs compared with those of control mice (Fig. 3, A and B). In contrast, the reduction of Grx3 expression was profound in the MECs of pregnant and lactating Grx3 MKO MGs (Fig. 3, C, D, and I). In agreement with these findings, q-PCR analysis demonstrated that Grx3 mRNA levels were reduced ~60% in virgin Grx3 MKO MECs (Fig. 3, E and F), whereas more than 95% reduction of Grx3 mRNA levels occurred in pregnant and lactating Grx3 MKO MECs compared with that of control MECs (Fig. 3, G and H), suggesting that MMTV-Cre mediated recombination sufficiently generates Grx3 KO in MECs of pregnant and lactating MGs.

Fig. 3.

Grx3 deletion in MMTV-Cre/ Grx3^flox/flox mammary glands. A–D: immunohistochemical staining at 6th week of age in virgin mice (A and B) and at pregnancy day 14 mice (C and D) shows Grx3 expression in ducts of floxed (A and C) and knockout (B and D) MGs. 400× magnification. E–H: q-PCR analysis of Grx3 expression in MECs from 6th week (E), 10th week (F), pregnancy day 14.5 (G), and lactation day 1 (H). Student’s t-test; n = 3. ***P < 0.001, significant difference between WT and KO. Cytokeratin 5 (CK5) was used as an internal control. I: Western blotting analysis of Grx3 protein levels in WT and KO MECs isolated from 10th week, P15, and L1 of MGs.

Loss of Grx3 impairs pregnancy-dependent lobuloalveolar development.

The Grx3 MKO females were viable, fertile, and able to give birth to normal-sized litters and to nurse their youngsters (Fig. 4). Examination of mature Grx3 MKO female mice revealed no dramatic developmental defects. To investigate the specific effects of Grx3 on ductal growth, MG whole mounts of early pubertal (6 wk age) and mature (12 wk age) virgin Grx3 MKO mice were compared with control littermates. Although quantitative morphometric analysis revealed modest developmental differences in ductal outgrowth between Grx3 MKO and control mice at 6 wk of age (Fig. 5, A–C), these differences were not evident at 12 wk of age (Fig. 5, D–F). The results suggest that the loss of Grx3 neither impacts mammary ductal development in the virgin gland nor occurs in a sufficient number of cells to cause consequence.

Fig. 4.

Grx3 MKO female mice have normal litter size and body weight of pups. A: average litter size of MKO female mice (n = 5) and controls (n = 4) was recorded and there is no significant difference between MKO and control mice. B: average body weight of each individual pup of total 42 and 37 pups from MKO mice and their littermate controls, respectively, was monitored for 14 days after birth. There is no significant difference between MKO and control mice.

Fig. 5.

Effects of Grx3 on ductal-branching in virgin mice. A–C: compared with controls (A), Grx3 MKO mice (B) displayed more tertiary branching and branch density at 6th week of age. C: quantification of parameters measured. Student’s t test; *P < 0.05, significance between controls (n = 9) and Grx3 MKO (n = 9) MGs. D–F: compared with controls (D), Grx3 MKO mice (E) displayed no difference in ductal tertiary branching and branch density at 12th week. F: quantification of parameters measured. Controls (n = 4) and Grx3 MKO (n = 6) MGs. 1.1× magnification; 40× magnification for insets.

Because the morphogenesis of the ductal tree in mature Grx3 MKO mice was similar to that in the control littermates and because the number of cells undergoing Cre-dependent deletion of Grx3 during pregnancy was much more extensive than that in virgin mice, lobuloalveolar development was evaluated at mid pregnancy and at day 1 postpartum. The visual comparison of whole-mount MGs from Grx3 MKO mice at midpregnancy indicated that although pregnancy-dependent alveolar budding was present, the expansion of this budding was less profound than that of control littermates (Fig. 6, A–D). Quantitative measurement of epithelial area further revealed a lower density of lobuloalveolar units in those MGs derived from Grx3 MKO mice (Fig. 6E; P < 0.05, one-tailed). On day 1 of lactation, H&E staining showed that in control littermates, the extent of lobuloalveolar development completely filled the mammary fat pad with dilated ducts, indicating active milk secretion (Fig. 6F). In contrast, the mammary glands of lactating Grx3 MKO mice appeared to have normally differentiated alveoli, but at significantly reduced density (Fig. 6, G and H; P < 0.01) supporting the conclusion that alveolar development during pregnancy is compromised in the Grx3 MKO females.

Fig. 6.

Lobuloalveolar growth impairment as a result of Grx3 loss. Whole mount staining of MGs at pregnancy day 14.5 reveals reduced lobuloalveolar density in Grx3 MKO mice compared with control MGs. MGs were examined by whole-mount staining (A and B). 1.1× magnification; 40× magnification for insets. C and D: background subtracted images of the ductal tree from A and B). 1.1× magnification. E: quantification of parameters measured. Student’s t-test, *P < 0.05, one tailed, indicates the significance between controls (n = 6) and Grx3 MKO (n = 7) MGs. F–H: reduced lobuloalveolar growth in Grx3 KO MGs at lactation. H&E staining of lactating MGs from wild-type control (F) and Grx3 KO mice (G). 100× magnification. H: quantification of parameters measured. Student’s t-test; ***P < 0.001, significance between controls (n = 3) and Grx3 MKO (n = 3) MGs.

In addition to the morphological defects, the mRNA levels of lactogenic genes in Grx3 MKO mammary epithelial cells were significantly reduced in comparison to that of control MECs (Fig. 7A). Furthermore, the production of major milk proteins, such as caseins, was also decreased in Grx3 MKO MECs (Fig. 7B). Taken together, these results suggest that not only does Grx3 play a supporting role in normal lobuloalveolar development, but it may also be required for complete differentiation of the mammary epithelium during secretory activation.

Fig. 7.

Effects of Grx3 on MG lactogenesis. A: significant decreases in lactogenic gene expression in MECs of Grx3 MKO mice compared with control mice. Student’s t-test, n = 3; ***P < 0.001, significant difference between WT and KO. Cytokeratin 5 (CK5) was used as an internal control. B: major milk protein productions, such as caseins, were decreased in Grx3 MKO MECs. Student’s t-test, n = 3; *P < 0.05, significant difference between WT and KO.

Increased ROS levels in Grx3 MKO MECs.

Our previous studies indicated that the knockdown of Grx3 in human cancer cells resulted in ROS accumulation, inhibition of cell proliferation, and eventually cell death (52, 54). Our studies further demonstrated that deletion of Grx3 in mouse embryonic fibroblasts (MEFs) inhibited cell proliferation through controlling redox states and cell cycle progression at G2/M phase (12, 52). Given that a group of Grx and thioredoxin (Trx) genes exist in the genome, it will be interesting to determine if there is a compensatory increase in the expression of those genes in the Grx3 MKO mouse MECs. Our q-PCR analysis indicated that both Grx and Trx gene expression was enhanced during pregnancy (Fig. 8A). In contrast, only Grx2 was induced in Grx3 MKO mouse MECs, whereas other Grx and Trx gene expression was, in fact, decreased at lactation (Fig. 8B). To further determine whether the deletion of Grx3 could affect cellular redox homeostasis that contributes to defective cell proliferation in MECs, ROS (superoxide) levels in WT and KO MECs were measured using dihydroethidium (DHE) staining. As shown in Fig. 9A, the signals from lactating KO MECs were stronger and brighter compared with those of wild-type controls, indicating a significant increase of ROS production in the absence of Grx3 (Fig. 9B). This finding suggests that Grx3 is critical for maintaining redox states in MECs during mammary development.

Fig. 8.

Effects of Grx3 deletion on redox gene expression. q-PCR analysis of redox gene mRNA levels in mammary epithelial cells (MECs) isolated from glands at pregnancy day 14.5 (A) and lactation day 1 (B). Cytokeratin 5 (CK5) was used as an internal control. Student’s t-test, n = 3; *P < 0.05, ***P < 0.001.

Fig. 9.

ROS levels were increased in Grx3 MKO alveoli. A: representative images showed strong DHE-staining signals as indicators of ROS levels in the lactating MGs of Grx3 MKO mice compared with wild-type controls. B: quantification of signal intensity per area indicated that ROS levels was increased ~3-fold in Grx3 MKO compared with wild-type controls. Student’s t-test, n = 3; ***P < 0.001, significance between controls and Grx3 MKO MGs.

Altered expression of proliferating genes in Grx3 MKO MECs.

The decreased alveolar development in Grx3 MKO mice could be due to defects in the pregnancy-dependent proliferation of MECs. In support of this notion, levels of Cyclin D1, a key regulator in cell proliferation of mammary gland development (19, 61), were significantly reduced in Grx3 MKO MECs compared with those in control MECs in both pregnant and lactating MGs as shown by q-PCR analysis (Fig. 10, A and B). Furthermore, Cyclin D1 protein levels were also decreased in Grx3 MKO MECs as revealed by Western blot analysis (Fig. 10C). Interestingly, cyclin B1 protein levels were significantly increased in Grx3 MKO MECs in comparison to those of control MECs (Fig. 10D), suggesting that Grx3 MKO MECs may have defects in cell cycle progression at G2/M phase.

Fig. 10.

Expression of proliferation genes in Grx3 MKO MECs. q-PCR analysis of Cyclin D1 mRNA levels in Grx3 MKO MECs compared with controls. A: pregnancy day 14.5. B: lactation day 1. Student’s t-test, n = 3; ***P < 0.001, significant difference between WT and KO. Cytokeratin 5 (CK5) was used as an internal control. C–D: Western blot analysis of Cyclin D1 (C) and cyclin B1 (D) protein levels in Grx3 MKO MECs at pregnancy and lactation compared with controls.

Pgr-RANKL-RANK pathway is attenuated in Grx3 MKO MECs.

Receptor activator of nuclear factor κB (RANK) and its ligand (RANKL) are key osteoclast differentiation/activation factors for bone metabolism (47). Signaling by this pathway is also essential for mammary gland development (9, 20, 21, 25), through mediating progesterone-driven epithelial cell proliferation (4, 45, 49, 64). Measurement of mRNA levels of progesterone receptor (Pgr) indicated a significant decrease in Grx3 MKO MECs from pregnant mice compared with wild-type controls (Fig. 11), whereas Pgr levels were increased in lactating Grx3 MKO MECs (Fig. 11). RANKL levels were decreased during pregnancy but remained unchanged during lactation (Fig. 11). In contrast, RANK levels did not change during pregnancy but were significantly decreased during lactation in Grx3 MKO MECs compared with those of wild-type controls (Fig. 11). Together, with the significant decrease of Cyclin D1 levels in Grx3 MKO MECs during both pregnancy and lactation (Fig. 10), these results indicate that Grx3 may play an important role in the Pgr-RANKL-RANK pathway regulating epithelial cell proliferation and lobuloalveolar development.

Fig. 11.

Expression of progesterone receptor (Pgr) and NF-κB pathways in Grx3 MKO MECs. q-PCR analysis of Pgr, RANK, RANKL mRNA levels in Grx3 MKO MECs compared with controls. A: pregnancy day 14.5. B: lactation day 1. Student’s t-test, n = 3; *P < 0.05, **P < 0.01, ***P < 0.001, significant difference between WT and KO. Cytokeratin 5 (CK5) was used as an internal control.

DISCUSSION

In this study, we generated a mammary gland-specific Grx3 KO mouse model and demonstrated that Grx3 may play an important role in mammary gland development through modulating the RANKL-RANK-NF-κB signaling pathway to control epithelial cell proliferation.

Our previous studies indicated that Grx3 is ubiquitously present in multiple tissues and organs in mice (12). The current work showed that Grx3 is expressed in many cell types of MGs during mammary gland development (Fig. 2). The temporal and spatial expression of Grx3 in MGs is dependent on the developmental stages, in which Grx3 expression is significantly increased in the pregnant and lactating MGs (Fig. 2J). This suggests that the expression of Grx3 in MGs is regulated by the same developmental cues that control lobuloalveolar development and lactogenesis. This result is consistent with the finding that Grx3 expression in human breast cancer cells is induced by growth hormone, IGF-I (54). Also, the Grx3 proteins appear to localize in both the cytoplasm and the nuclei during developmental stages (Fig. 2, A–F). Interestingly, our recent report demonstrated that the nuclear accumulation of Grx3 induced by oxidative stress was partially inhibited by a synthetic anti-progestogen, mifepristone (RU-486) (52). Whether and how the subcellular location of Grx3 in MECs during MG development is modulated by hormones, such as progesterone, during mammary gland development remains to be further investigated.

By crossing with the MMTV-Cre transgenic line, Grx3 gene (exon 2) was efficiently deleted (Fig. 1) and the expression of Grx3 was specifically reduced in MECs (Fig. 3). There are multiple MMTV-cre lines available, such as Lines A, D, and F (69), and several reports indicated that MG growth and development were compromised in lines A and F (56, 76). There are no reports indicating the defective effect of MMTV-Cre transgenic line D, which was used to generate Grx3 MKO. The MMTV-cre transgene was also found to express in the spleen, which is consistent with the previous report (35). However, how Grx3 expression in spleen is affected and whether this impacts mammary gland development have not been evaluated in the present study. Importantly, the efficacy of Grx3 deletion in the mammary ductal tree was found to vary depending on developmental stage. In the virgin gland, the deletion of Grx3 was incomplete resulting in a significant level of residual expression. This residual expression could very well be the reason that developmental effects in the virgin mammary gland were minimal in comparison to those observed during pregnancy, where the loss of expression is almost complete in the epithelial compartment.

In contrast to virgin MECs, pregnant and lactating MECs showed significantly enhanced Grx3 expression (Fig. 2J), suggesting an important function that Grx3 may play during those developmental stages. In support of this notion, the deletion of Grx3 in pregnant and lactating MECs in Grx3 MKO mice displayed an impairment of lobuloalveolar development (Fig. 6), indicating the requirement of Grx3 in this process. Furthermore, Grx3 appeared to be required for complete differentiation of secretory epithelial cells since the expression of milk protein genes was significantly decreased in the Grx3 MKO (Fig. 7A). These results suggest that Grx3 may affect prolactin (PRL)/prolactin receptor (PRLR)-mediated JAK2-STAT5 signaling in regulating epithelial proliferation and differentiation. However, in contrast to Jak2 and Stat5 mammary gland-specific deletion mice that are unable to nurse their youngsters because of no/reduced milk production (41, 68), the effect of Grx3 on milk production seems more modest (Fig. 7B). In this regard, although Grx3 MKO mice were able to nurse their litters, the possibility that milk production is compromised even more modestly in these animals has not been definitively evaluated under more stringent conditions when comparing the growth of size- and weight-matched cross-foster litters. It is unlikely that Grx3 effect on lobuloalveolar development is mediated by the JAK2-STAT5 pathway because there was no difference in phosphorylation of STAT5 observed in isolated MECs of Grx3 MKO MGs and cultured MECs stimulated with PRL compared with controls (data not shown).

In Grx3 KO MECs, both cyclin D1 mRNA and protein levels were significantly decreased (Fig. 10, A–C), suggesting that Grx3 may regulate cell cycle progression in MECs and is an important regulator in epithelial cell proliferation during pregnancy and lactation. Surprisingly, the results from BrdU labeling assays were not consistent with cyclin D1 gene expression analyses (Fig. 10; Fig. 12, A and B). Our previous studies with MEFs also indicated that there were no differences in the BrdU labeling of Grx3 KO MEFs and wild-type controls; instead, Grx3 KO MEFs had higher levels of cyclin B1 and increased numbers of cells with double nuclei that accounted for G2/M arrest of MEFs during proliferation (12). Interestingly, in Grx3 KO MECs, more cyclin B1 accumulated compared with controls (Fig. 10D), suggesting that Grx3 may play a similar role in MECs.

Fig. 12.

BrdU labeling to measure proliferating cells in pregnant and lactating MGs. BrdU positive cells for controls and Grx3 MKOMGs were counted and expressed as a ratio of positive nuclei to the total number of nuclei counted in the percentage. A: ducts in pregnancy day 14.5; 18–28 ducts were counted. B: ducts in lactation day 1; 18–24 ducts were counted.

The RANKL-RANK signaling pathway is important for mammary gland development. Although the overexpression of either RANKL or RANK gene in mammary glands promotes proliferation of MECs, ductal outgrowth, and alveologenesis (21, 25), the deletion of either one significantly impairs epithelial cell proliferation and gland development (20). It is also evident that RANKL-RANK signaling activates the NF-κB pathway to regulate cyclin D1 expression responsible for cell proliferation in the gland (9). Here, we found that the RANKL/RANK expression was downregulated in Grx3 MECs compared with controls (Fig. 11). Given the RANKL-RANK-axis functions in the same regulatory pathway during mammary gland development, one would expect that the regulation of RANKL and RANK expression mediated by Grx3 could be simultaneous. Surprisingly, in our study, we observed that RANK and RANKL expression are affected at different stages in the MKO mice. One plausible explanation is that RANKL and RANK expressions are regulated by multiple hormones and factors. Also, previous studies have found that RANKL and RANK show differential expression, with inhibitory effects on each other, during mammary gland development (21, 25). Furthermore, Grx3 effects on RANKL and RANK expression could be distinct in different types of luminal epithelial cells, like ER⁺/PR⁺ cells vs. ER^-PR⁻ cells. Therefore, the molecular mechanism by which Grx3 regulates expression of RANKL and RANK and the downstream signaling pathway remain to be explored in future studies.

When Grx3 was knocked down in human breast cancer cells, neither Grx1 nor Trx1 protein levels were changed (54). In Grx3 MKO mouse MECs, however, both Grx and Trx gene expression was enhanced during pregnancy (Fig. 8A), suggesting that a compensatory effect occurs. This could explain why the impact of Grx3 deletion was less profound at pregnancy (Fig. 6, A–E). In contrast, other Grx and Trx gene expression was decreased in Grx3 MKO mouse MECs during lactation (Fig. 8B). This is consistent with the result that more ROS were significantly accumulated compared with wild-type controls during lactation (Fig. 9). In our current study, we hypothesize that the RANKL-RANK-NF-κB signaling pathway is involved in controlling ROS in MECs and that the dysregulation of this pathway could cause defective cell cycle progression in MECs. The direct link between RANKL/RANK signaling and ROS scavenging in mammary epithelial cell proliferation remains to be explored. However, a recent study reported that RANKL/RANK signaling promoted FoxO-regulated antioxidant pathways to attenuate ROS production during osteoclastogenesis (3). Furthermore, dysregulated redox states in mammary epithelial cells can cause developmental arrest in mammary glands as seen in mice with loss of mitochondrial superoxide dismutase (51). Given that Grx3 has been shown to be essential for cancer cell proliferation and survival (12, 54), and reduced Grx3 in human breast cancer cells attenuates NF-κB activation under oxidative stress (54), our findings suggest that a potential redox-dependent RANKL-RANK-NF-κB signaling pathway, mediated by Grx3, may play a unique role in regulating mammary epithelial growth and gland development.

GRANTS

This work was supported by the United States Department of Agriculture/Agricultural Research Service under Cooperation Agreement 6250–51000–054 (to N. Cheng); National Institutes of Health (NIH) Grant R01-GM115622 and Cancer Prevention and Research Institute of Texas Grant CPRIT R1104 (to J. Wang); NIH Grant CA151610, the Avon Foundation Grant 02–2014–063, David Salomon Translational Breast Cancer Research Fund Grant, the Fashion Footwear Charitable Foundation of New York, Inc. Grant, and the Margie and Robert E. Petersen Foundation Grant (to X. Cui and Y. Qu); and National Heart, Lung, and Blood Institute Grants R01-HL-089598, R01-HL-091947, R01-HL-117641, R41-HL-129570 and American Heart Association (AHA) Established Investigator Grant 13EIA14560061(to X. H. Wehrens).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

K.P., J.D., X.J., Y.Q., and H.Y. performed experiments; K.P., J.D., X.J., Y.Q., H.Y., W.O., J.C.M., J.W., X.H.W., X.C., Y.L., D.L.H., and N.C. analyzed data; K.P., J.D., Y.Y., Y.L., D.L.H., and N.C. interpreted results of experiments; K.P., X.J., Y.Q., H.Y., and N.C. prepared figures; K.P. and N.C. drafted manuscript; K.P., J.D., X.J., Y.Q., H.Y., Y.Y., W.O., J.C.M., L.C., J.W., X.H.W., X.C., Y.L., D.L.H., and N.C. approved final version of manuscript; J.C.M., L.C., X.H.W., X.C., Y.L., D.L.H., and N.C. edited and revised manuscript.