Epithelial Xbp1 Is Required for Cellular Proliferation and Differentiation during Mammary Gland Development

Generation and breeding of Xbp1^flox/flox; GFAP-Cre mice.

Xbp1^flox/flox mice were a gift from Laurie H. Glimcher (Weill Cornell Medical College). Xbp1^flox/flox mice, which had the C57BL/6 background and which were described previously (19), were crossed with a transgenic FVB line expressing Cre recombinase under the control of the hGFAP promoter (GFAP-Cre) (15) to generate Xbp1^flox/flox; GFAP-Cre mice. For cell fate studies, hGFAP-Cre mice were crossed with B6.129(Cg)-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J reporter mice (20). A breeding strategy that ensured that control Xbp1^flox/flox and experimental mice Xbp1^flox/flox; GFAP-Cre mice were always littermates was followed. Xbp1^flox/flox and Xbp1^flox/flox; GFAP-Cre dams that delivered a litter within 24 h of each other were used in cross-fostering experiments. The litters born to Xbp1^flox/flox; GFAP-Cre and Xbp1^flox/flox dams were swapped as described previously (21). All experimental mice were anesthetized (ketamine at 100 mg/kg of body weight, xylazine at 20 mg/kg) and humanely euthanized prior to collection of tissue; the conduct of our experiments has complied with the highest international criteria for studies with animals. All studies were approved by the Institutional Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai and followed the National Institutes of Health guidelines for animal care.

Antibodies.

The antibodies used in this study were as follows: anti-GFP antibody (1:1,000; catalog number ab6556; Abcam), anti-keratin 8/18 (anti-K8/18) antibody (1:200; Progen Biotechnik), anti-smooth muscle actin α (anti-α-SMA) conjugated with Cy3 antibody (1:500; catalog number C6198; Sigma, St. Louis, MO), anti-Xbp1 antibody (1:200; catalog number M-186; Santa Cruz), anti-protein disulfide isomerase (anti-PDI) antibody (catalog number 3501; Cell Signaling), anti-Bip antibody (1:1,000; catalog number 3177; Cell Signaling), anti-CHOP antibody (1:1,000; catalog number 2895; Cell Signaling), anti-IRE1α antibody (1:1,000; catalog number 3294; Cell Signaling), anti-GAPDH (anti-glyceraldehyde-3-phosphate dehydrogenase) antibody (1:5,000; catalog number ab9482; Abcam), anti-Ki67 antibody (1:2,500; catalog number ab15580; Abcam), anti-phospho-histone H3 (Ser10) antibody (1:300; catalog number 9701; Cell Signaling), anti-cleaved caspase 3 (Asp175) antibody (1:300; catalog number 9661; Cell Signaling), anti-phospho-Stat5 (Tyr694; C71E5) antibody (1:400 for immunohistochemistry and 1:1,000 for Western blotting; catalog number 9314; Cell Signaling), anti-Stat5 antibody (1:1,000; catalog number 9363; Cell Signaling), and anti-mouse milk-specific protein antibody (1:5,000; Accurate Chemical).

Preparation of tissue protein and immunoblot analysis.

For total protein extraction, mouse mammary gland tissues and cells were lysed in radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.2% SDS) containing protease and phosphatase inhibitors (Roche). Protein content was measured by use of the Bio-Rad protein assay dye reagent (catalog number 500-0006). Proteins were resolved by the use of NuPAGE 4 to 12% bis-Tris gels (Invitrogen) and transferred to polyvinylidene difluoride membranes. Blots were probed with the primary antibody for overnight at 4°C, washed, and incubated with horseradish peroxidase-linked secondary anti-rabbit or anti-mouse IgG antibody (catalog numbers 7074 and 7076, respectively; Cell Signaling) for 1 h at room temperature. Blots were developed using the HyGlo quick spray reagent (Denville).

Genotyping PCR.

Xbp1^flox/flox; hGFAP-Cre and ROSA26^mT/mG; hGFAP-Cre mice were genotyped via PCR analysis, as previously described (19). The presence of the floxed Xbp1 allele was detected via PCR using mouse tail DNA as a template. A PCR product specific for the floxed Xbp1 allele was amplified by using the following primers: 5′-ACTTGCACCAACACTTGCCATTTC-3′ (forward) and 5′-CAAGGTGGTTCACTGCCTGTAATG-3′ (reverse). The PCR conditions were as follows: 94°C for 3 min, followed by 35 cycles at 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s and then 72°C for 5 min (the PCR product was 141 bp for wild-type [WT× mice and 183 bp for floxed mice). The presence of the hGFAP-Cre transgene was detected by using the following PCR primers: 5′-CCTGGAAAATGCTTCTGTCCG-3′ (forward) and 5′-CAGGGTGTTATAAGCAATCCC-3′ (reverse). The PCR conditions were as follows: 94°C for 4 min, followed by 40 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 1.5 min and then 72°C for 10 min (size of the PCR product, 400 bp).

Real-time PCR.

PCR primers for real-time PCR are summarized in Table S1 in the supplemental material. Fresh isolated mammary tissues were frozen and stored at −80°C until specimens were homogenized by bead milling using a TissueLyser LT apparatus (Qiagen, Hilden, Germany), which rapidly disrupts up to 12 samples simultaneously via high-speed vertical shaking (oscillation frequency, 50 Hz) with a bead (7 mm) in a sealed tube for 3 min. Then, mammary tissue RNA was extracted using Qiagen minicolumns and RNeasy minikits (Qiagen, Germantown, MD) with an on-column DNase treatment. One microgram of RNA was reverse transcribed using an RT Complete Double PrePrimed kit (Clontech, Mountain View, CA). FastStart SYBR green master mix (Roche, Indianapolis, IN) was used for PCR. The samples were analyzed in triplicate, and the data were analyzed by the use of Microsoft Excel software (Microsoft Corp., Redmond, WA), with the levels of gene expression being normalized to the level of β-actin gene expression.

Whole-mount analysis and histological analysis.

Whole-mount analysis of mouse mammary glands was performed as described previously (22). Mammary glands were excised and spread on microscope slides. The tissues were fixed in 10% formalin at 4°C, defatted in Carnoy's fixative, washed in 70% ethanol, hydrated by passage through decreasing ethanol concentrations, and stained with carmine alum stain (0.2% carmine, 0.5% aluminum potassium sulfate) overnight at room temperature. For histological analysis, mammary gland specimens were fixed overnight in 10% formalin at 4°C, dehydrated, and embedded in paraffin. Tissue blocks were sectioned into 5-μm-thick slices and stained with hematoxylin and eosin (H&E). Sirius red (Sigma) was used to determine collagen deposition. Sections were stained with Sirius red solution (saturated picric acid containing 0.1% Direct Red 80 and 0.1% fast green) to visualize collagen deposition. The relative fibrosis area was assessed on the basis of the findings for 20 fields of Sirius red-stained mammary gland sections from each animal. Whole-field areas were from the mammary luminal area, and each field was acquired at a ×400 field magnification and then analyzed using a computerized Bioquant morphometry system (R & M Biometrics, Nashville, TN). To evaluate the relative fibrosis area, the measured collagen area was divided by the net field area.

Immunohistochemistry and immunofluorescence.

Tissues were immediately fixed in 10% neutral buffered formalin for 24 h. After fixation, the tissue was immersed in 70% ethanol until processing. All tissues were processed simultaneously. The fixed tissues were dehydrated in ethanol, cleared in xylene, and embedded in paraffin blocks, which were cooled before sectioning. The mammary gland tissues were sectioned into 4-μm-thick slices, and the sections were mounted on silane-coated slides. For antigen activation, the slides were incubated in target retrieval solution (Dako) and heated for 30 min in a microwave oven. The slides were allowed to cool. Then, sections were used for immunohistochemistry with a Dako EnVision detection system using the immunoperoxidase method according to the Dako EnVision kit manual.

All samples subjected to immunohistochemistry staining were probed with the primary antibody overnight at 4°C. The slides were then washed three times in phosphate-buffered saline and incubated with secondary anti-rabbit or anti-mouse antibody for 1 h at room temperature. For immunofluorescence analysis, Alexa Fluor 488–donkey anti-rabbit IgG (H+L) antibody (1:500; Invitrogen) and Alexa Fluor 647–goat anti-guinea pig IgG (H+L) antibody (1:500; Life Technologies) were used as secondary antibodies. An isotype control was used to assess nonspecific binding. The slides were mounted with DAPI (4′,6-diamidino-2-phenylindole) ProLong Gold antifade reagent (catalog number P36931; Life Technologies, Carlsbad, CA) and examined in a Zeiss Axiophot photomicroscope (Carl Zeiss, Oberkochen, Germany). The cells in the stained sections that by diaminobenzidine staining were positive for Ki67, phosphorylated histone H3, and cleaved caspase 3 were counted under a microscope at a ×200 or ×400 field magnification, using 20 randomly selected microscopic fields per section.

Isolation of mouse mammary epithelial cells.

Mouse mammary epithelial cells were isolated as previously described, with minor modifications (23). Inguinal glands from ROSA26^mT/mG; hGFAP-Cre and Xbp1^flox/flox; hGFAP-Cre mice were removed aseptically, minced, and digested with collagenase at 37°C for 45 min. The digested glands were subsequently centrifuged at 1,000 rpm for 5 min, and the fat layer and supernatant were removed. The pellet was washed and incubated in red blood cell lysis buffer (catalog number R7757; Sigma, St. Louis, MO) for 2 min before centrifugation. The cells were then plated in Dulbecco modified Eagle medium (DMEM) with 10% fetal bovine serum and 1% penicillin-streptomycin and incubated for 30 min at 37°C to allow the selective attachment of fibroblasts. The supernatant was transferred into bovine serum albumin-coated tubes and washed once before resuspension in MCF10A medium (DMEM–F-12 medium supplemented with antibiotics, epidermal growth factor [20 ng/ml], hydrocortisone [0.5 μg/ml], insulin [10 μg/ml], cholera toxin [0.1 μg/ml], and 5% heat-inactivated horse serum) on a 10-cm dish coated with Matrigel (BD, San Jose, CA).

Mammary gland transplantation.

Transplantation experiments were performed as described previously (24). Pubescent 3- to 4-week old donor and host female mice were anesthetized, and midsagittal and oblique cuts were made in order to expose the 4th inguinal mammary gland. The fat pad in the mammary gland was cleared from the host epithelium by removing the tissue anterior to the lymph node. Epithelial fragments (2 mm³) of the donor mice were transplanted into the remaining cleared fat pad of the host (the fat pad from Xbp1^flox/flox; GFAP-Cre mice was transplanted into Xbp1^flox/flox mice and the fat pad of Xbp1^flox/flox mice was transplanted into Xbp1^flox/flox; GFAP-Cre mice) and the skin was closed. After 6 weeks, pregnant females that had received the transplants were sacrificed at day 7 of lactation. The transplanted glands were excised and processed for whole-mount analysis as described above.

Evaluation of serum prolactin level.

Xbp1^flox/flox; GFAP-Cre (n = 3) and Xbp1^flox/flox (n = 5) littermate mice were mated after mammary gland transplantation and sacrificed at day 7 of lactation. Blood was collected via inferior vena cava puncture and allowed to clot overnight at 4°C. Samples were centrifuged at 4,000 × g for 15 min, and serum was collected and stored at −80°C. Prolactin concentrations were assessed by prolactin mouse enzyme-linked immunosorbent assay (ELISA) by use of an ELISA kit (Abcam).

RT-PCR for Xbp1 splicing assay.

Reverse transcription-PCR (RT-PCR) analysis was performed using a 2× PCR master mix (Thermo). A PCR product specific for the spliced Xbp1 was amplified by using the following primers: 5′-ACACGCTTGGGAATGGACAC-3′ (forward) and 5′-CCATGGGAAGATGTTCTGGG-3′ (reverse). For spliced Xbp1 and cyclophilin detection by standard RT-PCR, the following program was used: (i) 94°C for 3 min, (ii) 30 cycles of 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s, and (iii) 72°C for 10 min. The PCR products were separated by agarose gel electrophoresis to resolve the 171-bp (unspliced) and 145-bp (spliced) amplicons. RT-PCR was performed with cyclophilin mRNA as a loading control to validate cDNA synthesis.

Evaluation of serum cholesterol and triglyceride concentration.

Quantitative determination of serum cholesterol and triglyceride levels was performed by use of a kit from Pointe Scientific, Inc. (Canton, MI) and spectrophotometer analysis.

Statistical analysis.

Statistical analysis was performed using a commercially available software package (Prism, version 5.0; GraphPad Software). Data were tested by Student's t test. Differences were considered statistically significant at a P value of <0.05. All results were derived from at least 3 independent experiments.

Pups suckling on Xbp1^flox/flox; GFAP-Cre dams fail to thrive.

In the process of breeding and expanding the Xbp1^flox/flox; GFAP-Cre colonies to generate mesenchymal cell-specific Xbp1-knockout mice, we observed an unexpected postnatal loss of several pups. Xbp1^flox/flox; GFAP-Cre mice were distinguished from control littermates by PCR with Cre and Xbp1 genotyping primers (Fig. 1A) and were born at Mendelian ratios (data not shown). However, in pups of Xbp1^flox/flox; GFAP-Cre dams, the body size and weight were clearly decreased at day 10 compared to those of the pups of Xbp1^flox/flox dams (Fig. 1B). Xbp1^flox/flox; GFAP-Cre female mice had normal pregnancies and gave birth to pups of the same number and sex as those of Xbp1^flox/flox control female mice (Table 1).

To investigate pup growth in detail during nursing, we recorded pup weights daily from the time of birth to day 20 (Fig. 1C). The weights of the pups from Xbp1^flox/flox; GFAP-Cre dams were significantly decreased from postnatal 3 day compared to those of the pups from Xbp1^flox/flox dams (Fig. 1C and D), which began to die at day 12 in the initial pregnancies, and all pups were dead by day 20 (Fig. 1E). Overall, these data indicate that pups born to Xbp1^flox/flox; GFAP-Cre dams failed to thrive.

Next, we confirmed that the failure to grow was due to the dam phenotype, because pups of Xbp1^flox/flox; GFAP-Cre dams were rescued by cross fostering by Xbp1^flox/flox dams. In contrast, the pups of Xbp1^flox/flox dams had impaired growth when fostered by Xbp1^flox/flox; GFAP-Cre dams (Fig. 1F and G). These data suggest that Xbp1^flox/flox; GFAP-Cre dams were unable to sufficiently nourish the newborn pups, arguing for a potential mammary gland phenotype.

GFAP-Cre labels mammary epithelial cells in the mammary gland.

Based on the observations described above, we sought to localize the specific Cre-expressing cells in female mammary glands. We generated ROSA26^mT/mG; GFAP-Cre mice for lineage tracing using Cre reporter mice expressing double-fluorescent membrane-targeting tdTomato fluorescent protein (mT) and membrane-targeting green fluorescent protein (GFP) (mG) crossed with GFAP-Cre mice. This reporter mouse ubiquitously expresses mT prior to Cre activation and mG upon Cre activation (Fig. 2A). We first visualized Cre-mediated recombination in situ in whole-mount mammary gland tissue without fixation at the virgin stage and on day 7 of lactation. Although all cells, including mammary epithelium and adipocytes, expressed dtTomato red in ROSA26^mT/mG control mice, Cre expression was observed in mammary branches only in ROSA26^mT/mG; GFAP-Cre mice at the virgin stage and on day 7 of lactation, as indicated by GFP expression (Fig. 2B).

Furthermore, to determine the Cre-expressing cell type(s) in mammary branches in ROSA26^mT/mG; GFAP-Cre mice, we analyzed the colocalization of membrane-targeting GFP, the epithelial cell marker keratin 8/18 (K8/18), and the myoepithelial cell marker α-SMA in mammary gland sections and isolated primary epithelial cells by multiple-labeling fluorescent immunohistochemistry. Surprisingly, GFP expression was almost completely colocalized with K8/18 but not with α-SMA in these virgin and lactating female ROSA26^mT/mG; GFAP-Cre mice (Fig. 2C and D). These data indicate that in these animals GFAP-Cre is expressed specifically in epithelial cells, most likely of luminal origin, in the mammary gland.

Xbp1 loss impairs mammary gland development and induces abnormal stromal fibrosis.

Next, to determine the effects of Xbp1 depletion on mammary morphogenesis, we performed whole-mount analysis and H&E staining of paraffin-embedded sections of mammary glands from Xbp1^flox/flox; GFAP-Cre mice at the virgin stage, during early pregnancy (pregnancy day 9 [P9]) and late pregnancy (pregnancy day 16 [P16]), and during lactation (lactation day 7 [L7]). The mammary glands of Xbp1^flox/flox; GFAP-Cre mice showed poor branching morphogenesis, impaired terminal end bud (TEB) formation at the virgin stage, and strongly impaired lobuloalveolar development during the early lactation period (Fig. 3A and B). Interestingly, during pregnancy there were no morphological differences between control and Xbp1^flox/flox; GFAP-Cre mice (Fig. 3A and B). These data suggest that the loss of Xbp1 in mammary epithelial cells during lactation renders the gland unable to respond to the pups' suckling. Double immunofluorescence staining for K8/18 (red) and α-SMA (green) also demonstrated a lack of epithelial cell cohesiveness, with mammary epithelial cells being delocalized from the basement membrane and decreased cell numbers being noted in the luminal layer in virgin Xbp1^flox/flox; GFAP-Cre mice (Fig. 3C, arrowheads). Furthermore, the stromal border of the ducts exhibited increased matrix deposition in Xbp1^flox/flox; GFAP-Cre mice (Fig. 3D, top, arrowheads). To confirm collagen accumulation in the mammary glands of Xbp1^flox/flox; GFAP-Cre mice, Sirius red staining, which specifically identifies collagen, was performed on sections of mammary gland from these animals at 12 weeks. Stromal fibrosis thickness was significantly higher in Xbp1^flox/flox; GFAP-Cre mice than in control mice at the virgin stage (Fig. 3D, bottom, and E). These data suggest that epithelial Xbp1 deletion in the epithelium induces abnormal stromal fibrosis in the mammary gland during adulthood. To investigate whether macrophage infiltration is altered as part of an aberrant immune response in the abnormal stromal fibrosis of Xbp1^flox/flox; GFAP-Cre mice, we performed F4/80 staining of mammary gland sections from Xbp1^fl/fl; GFAP-Cre mice. Overall, the percentage of cells staining positive for F4/80 was low (<0.5% in each section), and their frequency and expression levels of F4/80 were not different between in Xbp1^flox/flox control and Xbp1^flox/flox; GFAP-Cre mice (data not shown).

During lactation, Xbp1^flox/flox; GFAP-Cre dams also dramatically lacked expansion of the mammary epithelial compartment (Fig. 3B). Although between pregnancy and parturition the large cytoplasmic lipid droplets (CLDs) are usually replaced by small lipid droplets outside the cells on the luminal surface (Fig. 3B) (25), CLDs remained in the luminal alveolar cells in the Xbp1^flox/flox; GFAP-Cre mouse mammary glands (Fig. 3B). These data suggest that the physiological response to mammary epithelium differentiation may be impaired by the loss of Xbp1.

The sXbp1 protein is highly induced during lactation in mammary epithelium.

Next, we investigated whether sXbp1 is expressed in mammary glands in Xbp1^flox/flox and Xbp1^flox/flox; GFAP-Cre mice during adult mammary gland development by assessing whole mammary gland proteins at progressive developmental stages (at the virgin stage, during early pregnancy [P9] and late pregnancy [P16], and on L7). Interestingly, sXbp1 was detectable only during lactation and was almost completely absent in Xbp1^flox/flox; GFAP-Cre mice (Fig. 3F). To confirm the activity of Cre recombinase on Xbp1 by GFAP-Cre in mammary glands, we also performed quantitative PCR (qPCR) using primers that detect either wild-type (WT) Xbp1 or the Xbp1 deletion (the ΔXbp1 mutant) only. To quantify WT Xbp1 mRNA, we used a reverse primer that binds to a region in the floxed exon 2, and to detect the ΔXbp1 deletion, we used a primer that binds to a sequence present only in mice with theΔXbp1 deletion. Relative to the level of expression of Xbp1 mRNA in whole mammary gland tissues from control, WT mice on lactation day 7, in Xbp1^flox/flox; GFAP-Cre mice, the WT Xbp1 mRNA level was decreased 99% and the level of expression of the allele with the Xbp1 mRNA deletion was increased 1,000-fold compared to that in WT mice (Fig. 3G and H).

We further performed Western blotting of isolated mammary epithelial cell proteins from Xbp1^flox/flox and Xbp1^flox/flox; GFAP-Cre virgin and lactating mice (Fig. 3I). sXbp1 was completely absent in mammary epithelial cells isolated from lactating Xbp1^flox/flox; GFAP-Cre mice (Fig. 3I). These data confirm adequate Cre recombination in the mammary epithelium, consistent with the findings for GFAP-Cre reporter mice.

Impaired lobuloalveolar development during lactation occurs only when Xbp1 is absent in mammary epithelial cells.

We used a mammary gland transplantation technique to determine whether the observed lactation defect results directly from epithelial Xbp1 loss or is an indirect, systemic effect due to Xbp1 loss in the brain or other cells of Xbp1^flox/flox; GFAP-Cre mice. Transplantation was performed, in which gland tissue fragments from which Xbp1 was deleted were placed into cleared fat pads of Xbp1^flox/flox recipients and Xbp1^flox/flox gland tissue fragments were placed into Xbp1^flox/flox; GFAP-Cre recipients to establish that the lactation defect is present only when the mammary gland lacks Xbp1 (Fig. 4A). At postnatal day 7, the weights of pups of Xbp1^flox/flox; GFAP-Cre dams into which the Xbp1^flox/flox pad was transplanted were significantly decreased compared to those of pups of Xbp1^flox/flox dams into which the Xbp1^flox/flox; GFAP-Cre pad was transplanted (Fig. 4B). Transplanted Xbp1^flox/flox mammary glands developed normal alveologenesis during lactation in Xbp1^flox/flox; GFAP-Cre mice (Fig. 4C). However, the mammary glands transplanted from Xbp1^flox/flox; GFAP-Cre mice into Xbp1^flox/flox mice demonstrated impaired alveologenesis during lactation (Fig. 4C). These data suggest that the lactation defect is present only when the mammary gland lacks Xbp1.

We also measured by ELISA serum prolactin levels in transplanted mice on lactation day 7. There were no significant differences in prolactin levels between Xbp1^flox/flox and Xbp1^flox/flox; GFAP-Cre mice (Fig. 4D). These data suggest that the phenotype did not result from the altered prolactin levels resulting from the loss of Xbp1 in the central nervous system.

Epithelial Xbp1 deficiency leads to impaired mammary alveologenesis via reduced proliferation.

Next, we investigated whether Xbp1 loss in luminal epithelial cells affected their proliferation by immunostaining for Ki67 on lactation day 7. There was a significant reduction in the number of Ki67-positive cells in Xbp1^flox/flox; GFAP-Cre mice (Fig. 5A and D). The Ki67 mRNA level was also significantly reduced in Xbp1^flox/flox; GFAP-Cre mice compared to that in the controls, as assessed by quantitative PCR (Fig. 5E). Furthermore, the number of phosphorylated histone H3 protein-positive cells in mammary glands was decreased in Xbp1^flox/flox; GFAP-Cre mice (Fig. 5B and F). These findings indicate that the delay in mammary gland development in Xbp1-deficient mice during early lactation is caused by the reduced proliferation of epithelial cells.

To examine whether the hypomorphic state of mammary gland tissue was linked to an increase in apoptosis, we performed cleaved caspase 3 staining on day 7 of lactation. The percentage of cleaved caspase 3-positive cells was low (<0.5% in each section). They were commonly luminal, and their frequency and expression levels were not different between Xbp1^flox/flox control mice and Xbp1^flox/flox; GFAP-Cre mice (Fig. 5C and G). During normal lactation, only a small proportion of apoptotic cells are present in the mammary gland (26). Taken together, our data provide evidence that the loss of mammary gland tissue upon Xbp1 deletion is due to reduced proliferation and not increased apoptosis. This is surprising, as the loss of Xbp1 in the developing liver leads to apoptosis (27), which is likely due to the persistent ER stress that occurs in cells with a high secretory output when this important UPR regulator is missing.

ER stress in the mammary epithelium is provoked by Xbp1 deficiency.

In mammals, activation of IRE1α results in splicing of the Xbp1 mRNA to generate an active form of Xbp1 (spliced Xbp1 [sXbp1]) that initiates an arm of the UPR program (19). Therefore, to investigate the effect of Xbp1 loss on the UPR pathway, we analyzed the expression of UPR mRNAs and the IRE1α protein in the mammary glands of Xbp1^flox/flox; GFAP-Cre mice by qPCR and immunoblotting, respectively. The expression level of the Xbp1 target chaperone protein protein disulfide isomerase (PDI) was significantly decreased in Xbp1^flox/flox; GFAP-Cre mice, as were the mRNA levels of Edem1 and Hrd1, which are Xbp1 target genes in the endoplasmic reticulum-associated degradation (ERAD) pathway (Fig. 6A and B).

In contrast, the ER stress marker proteins BiP, CHOP, and IRE1α were significantly induced in Xbp1^flox/flox; GFAP-Cre mice (Fig. 6B). Although the IRE1α protein was induced in Xbp1^flox/flox; GFAP-Cre mice, as previously reported in other systems (19), the Ire1α mRNA expression level was not changed between control and Xbp1^flox/flox; GFAP-Cre mice (Fig. 6A), which is not surprising, given that this gene is not regulated at the transcriptional level. To identify the localization of IRE1α induction, we performed immunostaining for IRE1α on lactating mammary glands from Xbp1^flox/flox; GFAP-Cre mice. IRE1α was highly expressed in the cells of the mammary gland ducts and alveoli during lactation (Fig. 6C). In addition, we confirmed by RT-PCR ΔXbp1 mRNA mutant splicing in Xbp1^flox/flox; GFAP-Cre mice using extracts from whole mammary gland tissue (Fig. 6D). The splicing of the mutant ΔXbp1 mRNA was detected in Xbp1^flox/flox; GFAP-Cre mice (Fig. 6D). These data suggest that epithelial Xbp1 loss leads to chronic ER stress and constitutive activation of its upstream activator IRE1α in the mammary epithelium. The fact that BiP and CHOP were also significantly upregulated suggests that ATF6 and PERK signaling might be induced. The induction of CHOP and possibly of PERK could explain the strong growth arrest phenotype, as both pathways can induce arrest at G₀/G₁ phase (28, 29).

Reduced quality and quantity of milk in the Xbp1^flox/flox;GFAP-Cre mouse mammary gland during lactation.

The maturation and secretion of essential protein constituents of milk are dependent on the capacity of the ER to properly function, and Xbp1 is known to be required for the secretory ability of exocrine glands (11, 12). Thus, we measured the level of milk in isolated primary epithelial cells during lactation using an anti-mouse milk-specific antibody. The mouse milk protein production was significantly decreased in Xbp1^flox/flox; GFAP-Cre mice (Fig. 7A). Furthermore, casein, which is a major component of mouse milk, has a significant effect on maturation of milk secretion and growth of the offspring (30). Therefore, we assessed the expression of genes for mouse milk constitutive proteins in whole mammary lactating gland tissue by qPCR. Interestingly, mRNA levels for major components of mouse milk, α/β-casein and whey acidic protein (WAP), were decreased by 75% in Xbp1^flox/flox; GFAP-Cre mice (Fig. 7B). These data suggest that Xbp1 deficiency impairs the functional differentiation involved in the control of the quality and quantity of mouse milk.

sXbp1 regulates the expression of genes encoding lipogenic enzymes, including Acc2 and Scd1, which are direct sXbp1 targets. The levels of Acc2 and Scd1 expression were significantly decreased in Xbp1^flox/flox; GFAP-Cre mice (Fig. 7C). However, the expression of upstream master factors that control lipid metabolism, Srebp-1, Chrebp, Ppar-γ, and Adrp/adipophilin, were not altered (Fig. 7C). In addition, there was no compensatory induction of Atf6 mRNA, which is involved in de novo lipogenesis (31–33) and which is involved in another branch of UPR (Fig. 7C); however, it is possible that cleavage and activation of the Atf6 protein were altered by Xbp1 deletion. There were no significant differences between Xbp1^flox/flox; GFAP-Cre and control mice in total body weight and isolated inguinal fat pad weight (Fig. 7D and E) or in serum cholesterol and triglyceride levels at lactation day 7 (Fig. 7F and G). Taken together, these data suggest that epithelial Xbp1 is also essential for the expression of genes involved in the maturation of milk production but not basal metabolic parameters.

Xbp1 loss disrupts Stat5 phosphorylation due to reduced prolactin and ErbB4 receptor expression.

Next, we studied the mechanism by which Xbp1 could be affecting mouse milk synthesis. We focused on the prolactin-Stat5 pathway, which regulates mouse milk synthesis, because Prlr (prolactin receptor) (34), ErbB4 (a tyrosine-protein kinase receptor) (35), and Stat5 (36) are essential for the development of secretory mammary epithelial cells during lactation, proliferation, and milk production (37). When we analyzed whole mammary gland tissue for the expression of the genes for Prlr, ErbB4, and Cav1 (caveolin-1, which is involved in JAK2 activation), the levels of the transcripts of these genes were significantly decreased in Xbp1^flox/flox; GFAP-Cre mice; there was no difference in the level of Cav1 gene expression (Fig. 8A). These data show an association between Xbp1 loss and the disruption of Prlr/ErbB4 gene expression.

Because Prlr and ErbB4 regulate Stat5 tyrosine phosphorylation (pStat5) and nuclear translocation in secretory mammary epithelial cells during lactation, we compared Stat5 tyrosine phosphorylation levels in mammary gland sections from Xbp1^flox/flox and Xbp1^flox/flox; GFAP-Cre animals by immunostaining and immunoblot analysis (Fig. 8B and C). Interestingly, epithelial pStat5 was dramatically decreased on lactation day 7 according to both the number of nuclei stained and the protein levels obtained by immunoblotting (Fig. 8B and C). There were no differences in the expression of total Stat5 protein and Stat5a/b mRNA (Fig. 8C and D). These data suggest that Xbp1 loss disrupts Stat5 phosphorylation due to reduced prolactin and ErbB4 receptor expression in the mammary epithelium.