GMECs were isolated from three healthy Wanlin white goats at peak lactation period (2 years old, average of about 52 kg, second parturition, 45 d postpartum). The protocol for purification and cultured procedure as previously described [11]. Briefly, mammary tissue was collected using a surgical approach and washed with D-Hank’s solution. Tissue was then minced into about 1 mm³ cubes, plated in 60 mm culture dishes and cultured in 5% CO₂ at 37°C. The medium was replaced every 2 days until the epithelial cells digested from the tissue block with 0.25% trypsin- ethylene diamine tetraacetic acid solution. Subsequently, fibroblasts were eliminated by differential adherence method, which adhered to culture dishes faster than GMECs. Cells were transferred to new culture dishes and purified 5 passages for subsequent experiments. The complete medium for GMECs was incorporated with 10% fetal bovine serum (11011-8615; Every Green) and 90% Dulbecco’s modified eagle medium/Ham’s F-12 (SH30023; Hyclone) medium with 10 μg/L epidermal growth factor (PHG0311; Invitrogen), 100 kU/L penicillin/streptomycin (080092569; Harbin Pharmaceutical), 5 mg/L bovine insulin (16634; Sigma-Aldrich), and 1 mg/L hydrocortisone (H0888; Sigma-Aldrich).

To determine the purity GMECs, immunofluorescence and oil red O staining were performed. Cells were washed with phosphate buffer solution, fixed with 4% paraformaldehyde, permeabilized with Triton, and blocked with 5% bovine serum albumin at room temperature for 1 h. Subsequently, the cells were incubated with Cytokeratin 18 (66187; Proteintech; 1:800) and β-casein antibody (bs-0466R; Bioss; 1:500) for 1 h in the dark. Thereafter, the cells were incubated with Alexa Fluor 488-conjugated goat anti-mouse immunoglobulin G (IgG) (CW0102; CWBIO; 1:5,000). To ascertain whether GMECs possess normal secretory function, cells were stained with the oil red O staining kit (G1262; Solarbio) and observed through a microscope.

Based on goat ELF5 mRNA sequence (GenBank accession no. XM_005690080.3), the forward primer and reverse primer are CGCGGATCCCTCACGGTGATGTTGGAC and GATGGAGCAGATCATAGCT, respectively. Employing cDNA samples derived from the mammary gland of goats at peak lactation. The coding sequence of the goat ELF5 gene was cloned into the BamHI/EcoRI sites of pcDNA3.1 using T4 DNA ligase (Takara), generating the pcDNA3.1-ELF5 expression vector. By utilizing Invitrogen online software, specific small interfering ribonucleic acid (siRNA) targeting goat ELF5 gene was designed, concurrently with an irrelevant negative control (NC) sequence siRNA-NC (Sangon, China). The sequence of siRNA-ELF5 is as follows: sense: GCAGAGCUCUGAGGUACUATT, antisense: UAGUACC UCAGAGCUCUGCTT. The sequence of siRNA-NC is: sense: UUCUCCGAACGUGUCACGUT, antisense: ACG UGACACGUUCGGAGAATT.

For ELF5 gene overexpression assay, pcDNA3.1-NC or pcDNA3.1-ELF5 was transfected into GMECs by Lipofectamine 3000 (Invitrogen). For the knockdown of ELF5 gene, siRNA-NC or siRNA-ELF5 (100 nM) was transfected into cells by Lipofectamine RNAiMAX (Invitrogen). For the inhibition of JAK2/STAT5 pathway, JAK2 inhibitor AG490 (30 μM, MedChemExpress) or STAT5 inhibitor STAT5-IN-1 (50 μM, MedChemExpress) were treated with cells. For rescue experiment with STAT5 inhibitor and ELF5 overexpression, cells seeded in 6-well plates were grown to 80% confluence. After pre-treatment with STAT5-IN-1 (50 μM) for 1 h, transfection was performed with pcDNA3.1-NC or pcDNA3.1-ELF5 using Lipofectamine 3,000 according to the manufacturer’s protocol. Total protein was collected 48 h after transfection for Western blot analysis. Similarly, after pre-treatment with STAT5-IN-1 (50 μM) for 1 h, transfection was performed with siRNA-NC or siRNA-ELF5 (100 nM) using Lipofectamine RNAiMAX for 48 h. For rescue experiment with JAK2 inhibitor and ELF5 overexpression, cells seeded in 6-well plates were grown to 80% confluence. After pre-treatment with AG490 (30 μM) for 1 h, transfection was performed with pcDNA3.1-NC or pcDNA3.1-ELF5 using Lipofectamine 3,000 for 48 h. Similarly, after pre-treatment with AG490 (30 μM) for 1 h, transfection was performed with siRNA-NC or siRNA-ELF5 (100 nM) using Lipofectamine RNAiMAX for 48 h. Besides, for the interaction assay of ELF5 and STAT5, cells were transfected with pCMV-STAT5a or pCMV-NC. After 48 h treatment, cells were harvested to extract total RNA and protein.

Cell viability was carried out by cell counting kit-8 (CCK-8) assay in GMECs. After ELF5 gene overexpression or interference for 24, 48, and 72 h, the cells were incubated with CCK-8 reagent (5 g/L, CA1210; Solarbio) at 37°C for 2 h. A microplate reader (TECAN Spark) was employed to measure the absorbance of cell culture medium at 450 nm. Then, cell apoptosis was carried out by flow cytometry. The cells were transfected with pcDNA3.1-ELF5 or si-ELF5 for 48 h, subsequently cell apoptosis was determined by using the Annexin V-FITC/PI dual staining cell apoptosis detection kit (KGA1102-100; KeyGen) by a flow cytometer (BD Accuri C6puls).

Total RNA of GMECs was extracted by SPARKeasy (AC0-202-B; Sparkjade). The quality of RNA was evaluated by employing a NanoDrop One spectrophotometer (Thermo Fisher Scientific), where the ratio of absorbance at 260 nm to 280 nm ranged from 1.8 to 2.1. The integrity of RNA was determined through the analysis of 28S and 18S rRNA via agarose gel electrophoresis, with a ratio of 28S/18S approximately 2:1. For the quantitative analysis of cell proliferation and apoptosis marker genes as well as casein genes, the first strand cDNA was generated by utilizing purified total RNA through the PrimeScriptRT reagent kit (PC7001; Aidlab), which incorporates a gDNA eraser. The synergetic binding reagent green kit (PC3301; Aidlab) was used for quantitative real-time polymerase chain reaction (qPCR). Employing ubiquitous expression transcript (UXT) as an internal control gene, the qPCR primer sequences are presented in Table 1. All qPCR reactions were conducted on the CFX96 sequence detection system (Bio-Rad), and the data were analyzed by relative quantification (2^−ΔΔCt) method.

Total protein of GMECs was harvested after treated with radio immunoprecipitation assay lysis buffer (R0010; Solarbio) containing protease inhibitors (G2006-250UL; Servicebio). The protein assay kit (P0010S; Beyotime) was employed to detect protein concentration. Proteins were separated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride membrane (IPVHO0005; Merck) by BG-transBLOT. The membrane was blocked with 5% skim milk for 2 h (D8340; Solarbio) and subjected to incubation with the primary antibody at 4°C overnight. Then the membrane incubated with the secondary antibody. Primary antibodies contains ELF5 (sc-166653, SantaCruz; 1:1,000), αS2-casein (bs-10034R, Bioss; 1:2,000), κ-casein (bs-10031R; Bioss; 1:1,000), αS1-casein (bs-10033R; Bioss; 1:1,000), β-casein (sc-166530; SantaCruz; 1:1,000), p-JAK2 (ab32101; Abcam; 1:1,000), p-STAT5 (9351S; 1:1,000), JAK2 (3230; CST; 1:1,000), STAT5 (610191; BD; 1:1,000), proliferating cell nuclear antigen (PCNA, 10205-2-APl Proteintech, 1:5,000), cyclin dependent kinase 2 (CDK2, 10122-1-AP; Proteintech, 1:10,000), B-cell lymphoma-2-associated X (BAX, 50599-2-lg; Proteintech, 1:10,000), Caspase3 (19677-1-AP; Proteintech, 1:2,000), β-Tubulin (CW0098M; CWBIO, 1:5,000). Secondary antibody contains horseradish peroxidase-conjugated goat anti-rabbit-IgG (CW0102; CWBIO; 1:5,000) and goat anti-mouse-IgG (CW0103; CWBIO; 1:5,000). Signals were measured using an enhanced chemiluminescent Western blot system (AllianceQ9Advanced). The intensity of indicated bands was quantified by densitometry using ImageJ software (http://imagej.nih.gov/ij/). The relative abundance of proteins is normalized to β-Tubulin. The relative abundance of p-JAK2 and p-STAT5 is normalized to total JAK2 or STAT5.

Following 48 h co-transfection with pcDNA3.1-ELF5-MYC and PCMV-STAT5-3xFLAG vectors, GMECs were lysed for 20 min by immunoprecipitation lysis buffer. Subsequently, the mixture was centrifuged at 13,000×g for 10 min, and the supernatant was taken as the total protein. The Protein G beads were washed with pre-cooled phosphate buffer solution and diluted to a 50% suspension. Subsequently, Protein G beads (L-1202A; Biolinkedin) pre-washing was employed to eliminate non-specific adsorption from the cell lysate. Then centrifuged total cell protein and Protein G beads suspension were mixed and incubated at 4°C for 1 h. Centrifugation was performed at 600×g for 3 min, and the supernatant was collected. The antibody and Protein G beads were pre-incubated. The diluted solution was added Mouse-MYC-tag antibody (60003-2-Ig; Proteintech), Mouse-FLAG-tag antibody (66008-4-Ig; Proteintech) and Protein G beads suspension. The mixtures were incubated at 4°C for 1 h and centrifuged at 600×g for 3 min to retain the precipitates for three times. Subsequently, the samples were collected for Western blot detection.

The data were conducted by SPSS 22.0 (IBM) and presented as mean±standard error of the mean (SEM) of three biological replicates. When only two groups were compared, the data were analyzed through the t-test (two-tailed). For multiple comparisons, a general linear model program was utilized for ANOVA with Tukey’s test. A difference was regarded as statistically significant (* p<0.05, ** p<0.01).

Primary GMECs were migrating out from the goat mammary tissue, and purified cells were attained after 5 passages of purified culture (Figure 1A). Oil red O staining showed that numerous lipid droplet milk globules were dispersed around the GMECs, signifying that cells possess normal secretory function (Figure 1B). Immunofluorescence assays showed mammary epithelial cell markers cytokeratin 18 and β-casein were positive staining (Figure 1C).

To explore the regulatory mechanism of ELF5 on casein synthesis, overexpression vector and siRNA of ELF5 gene were performed. qPCR and Western blot results indicated that pcDNA3.1-ELF5 vector could significantly enhance the mRNA and protein expression of ELF5 compared with the pcDNA3.1-NC group (p<0.001, Figures 2A, 2B, Supplement 1). Similarly, compared with the siRNA-NC group, siRNA-ELF5 markedly decreased the mRNA (p<0.001) and protein expression (p = 0.005) of ELF5 (Figures 2C, 2D, Supplement 2).

The CCK-8 results showed that cell viability was significantly enhanced after ELF5 gene overexpression for 24, 48, and 72 h (p<0.01, Figure 3A). Flow cytometry assay revealed cell apoptosis rate was decreased by overexpressing ELF5 gene (p = 0.043, Figure 3B). The results of qPCR and Western blot showed that ELF5 gene overexpression markedly increased the mRNA and protein expression of proliferation marker genes PCNA and CDK2, and conspicuously decreased the expression of apoptosis marker genes Bax and Caspase3 (p< 0.05, Figures 3C, 3D, Supplement 3).

The CCK-8 results showed that cell viability was significantly decreased after ELF5 gene interference for 24, 48, and 72 h (p<0.01, Figure 4A). Flow cytometry assay revealed cell apoptosis rate was increased by knocking down ELF5 gene (p = 0.035, Figure 4B). qPCR and Western blot results showed that ELF5 gene interference markedly decreased the mRNA and protein expression of PCNA and CDK2, and increased the expression of Bax and Caspase3 (p<0.05, Figures 4C, 4D, Supplement 4).

In comparison with control group, ELF5 gene overexpression conspicuously elevated the mRNA and protein expression of αS1-casein, αS2-casein, β-casein, and κ-casein as well as the activity of phosphorylated JAK2 and STAT5 (p<0.05, Figures 5A, 5B, Supplement 5). In addition, ELF5 gene interference significantly decreased the expression of αS1-, αS2-, β-, and κ-casein, and reduced the activity of JAK2/STAT5 signaling pathway (p<0.05, Figures 5C, 5D; Supplement 6).

Regulation of casein synthesis by JAK2/STAT5 signaling pathway was confirmed by Western blot assay. The protein abundance of p-JAK2, p-STAT5, αS1-casein, αS2-casein, β-casein, and κ-casein were reduced after JAK2 inhibitor AG490 treatment (p<0.01, Figure 6A; Supplement 7). Then, the protein abundance of p-STAT5 and four caseins were significantly reduced in STAT5 inhibitor STAT5-IN-1 group (p<0.01, Figure 6B, Supplement 8).

Moreover, to validate whether ELF5 promotes casein synthesis via JAK2/STAT5 signaling pathway, the rescue experiments were carried out. Overexpression of ELF5 gene upregulated p-STAT5 activity and casein expression, while co-treatment with STAT5 inhibitor abrogated the increase of p-STAT5 activity and casein expression (Figure 7A, Supplement 9). Simultaneously, knockdown of ELF5 gene downregulated p-STAT5 activity and casein expression, whereas co-treatment with STAT5 inhibitor further reduced activity of p-STAT5 and abundance of caseins (Figure 7B, Supplement 10). Furthermore, co-treatment with ELF5 gene overexpression and JAK2 inhibitor abrogated the increase of JAK2/STAT5 activity and casein expression (Figure 8A, Supplement 11). Besides, co-treatment with ELF5 gene interference and JAK2 inhibitor further decreased JAK2/STAT5 activity and casein expression (Figure 8B, Supplement 12).

To verify whether ELF5 interacts with STAT5, an immunoprecipitation assay was carried out. We demonstrated that ELF5 interacted with STAT5 in GMECs by Western blot assay (Figure 9A). Moreover, qPCR result showed that the mRNA level of ELF5 was highly elevated after STAT5a gene overexpression (p<0.001, Figure 9B, Supplement 13). These results further validate that ELF5 could regulate casein synthesis via JAK2/STAT5 signaling pathway by interacting with STAT5 (Figure 10).