Novel function of LHFPL2 in female and male distal reproductive tract development

Congenital reproductive tract anomalies could impair fertility. Female and male reproductive tracts are developed from Müllerian ducts and Wolffian ducts, respectively, involving initiation, elongation and differentiation. Genetic basis solely for distal reproductive tract development is largely unknown. Lhfpl2 (lipoma HMGIC fusion partner-like 2) encodes a tetra-transmembrane protein with unknown functions. It is expressed in follicle cells of ovary and epithelial cells of reproductive tracts. A spontaneous point mutation of Lhfpl2 (LHFPL2G102E) leads to infertility in 100% female mice, which have normal ovarian development, ovulation, uterine development, and uterine response to exogenous estrogen stimulation, but abnormal upper longitudinal vaginal septum and lower vaginal agenesis. Infertility is also observed in ~70% mutant males, which have normal mating behavior and sperm counts, but abnormal distal vas deferens convolution resulting in complete and incomplete blockage of reproductive tract in infertile and fertile males, respectively. On embryonic day 15.5, mutant Müllerian ducts and Wolffian ducts have elongated but their duct tips are enlarged and fail to merge with the urogenital sinus. These findings provide a novel function of LHFPL2 and a novel genetic basis for distal reproductive tract development; they also emphasize the importance of an additional merging phase for proper reproductive tract development.


Results
Lhfpl2 and Lhfpl2 point mutation. Since very limited information is available about the expression of Lhfp gene family in mice, RT-PCR (see Supplementary Fig. S1) was performed on several adult tissues, including heart, spleen, liver, kidney, thymus, testis, brain, lung, intestine, vagina, ovary and uterus. Among the six members of the Lhfp gene family, Lhfp and Lhfpl2 were ubiquitously expressed in these tissues, Lhfpl1 was detected in most tissues, Lhfpl3, Lhfpl4, and Lhfpl5 had more tissue specific expression patterns with the highest expression in the brain. Lhfpl2 was highly detected in the female reproductive system, such as ovary, uterus and vagina. The expression levels in ovary and uterus seemed comparable between 3 weeks old and 8 weeks old.
There are 7 mouse Lhfpl2 transcripts indicated in Ensembl (http://www.ensembl.org/index.html). They encode proteins of 222aa (three transcripts), 139aa (one transcript), 34aa (one transcript), or no protein product (two transcripts). The full-length LHFPL2 protein (222aa) is predicted to be a tetra-transmembrane protein with a molecular weight of ~23.9 kD (see Supplementary Fig. S2a). It has both N and C termini in the cytoplasm, one intracellular loop between transmembrane domains 2 and 3, and two extracellular loops between transmembrane domains 1 and 2 as well as 3 and 4 (see Supplementary Fig. S2b). Based on the full-length Lhfpl2 transcript (Lhfpl2-001), the Lhfpl2 mutant mice (MUT/ LHFPL2 G102E ) used in this study have a spontaneous mutation in the 676 th nucleotide from G to A (see Supplementary Fig. S3a), resulting in a missense mutation of the 102 nd amino acid from nonpolar glycine (G) to negatively charged glutamic acid (E) (see Supplementary Fig. S3b) located on the predicated second transmembrane domain (see Supplementary Fig. S2c). This transition mutation creates a BglII restriction site (see Supplementary Fig. S3b) and can be confirmed by both restriction digestion coupled with PCR (see Supplementary Fig. S3c) and SNP genotyping (see Supplementary Fig. S3d).
The average litter size from heterozygous (HET) Lhfpl2 females mated with HET males was 7.8 ± 2.3 (N = 16). The offspring from HET × HET crosses followed the Mendelian inheritance, with 21.8% WT, 52.8% HET, and 25.4% MUT at weaning. The LHFPL2 G102E mice generally appeared healthy except imperforate vagina in all females and corneal opacity and exophthalmos in both genders. The Jackson Laboratory also reported abnormal pupillary reflex, thin retinal inner nuclear layer, hearing problem, and abnormal perineum morphology in the MUT mice. Defects in male reproductive tract was identified later via functional analyses in our study. When LHFPL2 G102E mutant mice expressed BAC mouse Lhfpl2 transgene, the female MUT/BAC+ mice had normal vagina and vaginal opening and the male MUT/BAC+ mice were fertile, confirming that this Lhfpl2 point mutation was responsible for the phenotypes in both female and male reproductive tracts.
Vaginal agenesis caused by defective distal MD development in Lhfpl2 mutant females. Vaginal closure was observed in 100% of the MUT but none of WT or HET females at 8-week old ( Fig. 1a-c). Vaginal closure prevented mating. Compared to the control (WT/HET) female reproductive tracts at the estrus stage when the uterus had the biggest size during estrous cycle (Fig. 1d), all the MUT female reproductive tracts were bigger with 30% of them having a dark appearance in the uterus and the upper vagina (Fig. 1e,f). Besides distended uterine horns and upper vagina, each MUT vagina lacked the lower vagina and had double lumens in upper vagina visualized by blue dye (Fig. 1e-g). Histology of 8-week old WT uterine horn and vagina at estrous stage showed a typical edematous stromal compartment and an open lumen in the uterus (Fig. 1h) as well as a single lumen vagina (Fig. 1j). Each MUT uterine horn had an enlarged lumen filled with a viscous fluid and had a compressed uterine wall (Fig. 1i). Each MUT upper vagina had enlarged double lumens filled with a viscous fluid and had a compressed vaginal wall (Fig. 1k), consistent with the double lumens indicated by blue dye infusion (Fig. 1g).
Analysis of postnatal/prepubertal vaginas showed distal vagina agenesis at all these ages examined from PND3 to PND28, so was double-lumen upper vagina in the MUT females ( Fig. 2a-e). PAX2 staining of E15.5 female embryos revealed that the two WT MDs had partially fused to form a Y-shape and entered the urogenital sinus (Fig. 2f), while the two MUT MDs had enlarged tips that failed to merge with each other and failed to enter the urogenital sinus (Fig. 2g). The two MUT WDs also had enlarged tips that failed to enter the urogenital sinus that were not seen in the WT WDs (Fig. 2f,g). These data demonstrated that the MUT MDs had gone through initiation and elongation phases to reach the urogenital sinus, but they had defective tip development and failure of merging with urogenital sinus during embryonic stage, eventually resulting in unmerged upper vagina and lower vaginal agenesis (Fig. 2e).
Expression of Lhfpl2 in the female reproductive system. In situ hybridization of female reproductive system at 4 weeks old indicated that Lhfpl2 mRNA was mainly detected in follicle cells of the ovary (Fig. 3a), epithelial cells of the oviduct (Fig. 3b), both luminal and glandular epithelial cells of the uterus (Fig. 3c,e), and epithelial cells of the vagina (Fig. 3d, Normal ovarian function and uterine response to E2 in Lhfpl2 mutant females. The high expression levels (see Supplementary Fig. S1) and cell type specific expression patterns of Lhfpl2 in the ovary and the uterus (Fig. 3) promoted us to investigate potential roles of LHFPL2 in the ovary and the uterus. Although distal vaginal agenesis in 100% of Lhfpl2 MUT females ( Fig. 1) prevented investigation of ovarian and uterine functions through pregnancy, histology and E2 treatment provided important information about the functions of ovary and uterus in the MUT females. Histology showed that WT and MUT females had comparable ovary histology at 4-week old (data not shown) and 8-week old, when follicles at different stages and corpora lutea were present in both WT and MUT ovaries (Fig. 4a,b), indicating that the MUT ovary had normal development and ovulation. E2 treatment on newly weaned females (prior to vaginal opening and estrous cyclicity) revealed that E2 treatment accelerated vaginal opening in the WT females but the MUT vaginas remained closed (Fig. 4d). Both WT and  MUT female reproductive tracts had comparable responses to E2 treatment, such as enlarged uteri and vaginas ( Fig. 4c,e,f), fluid accumulation in the reproductive tracts (Fig. 4c,e), and taller uterine luminal epithelium ( Fig. 4f), demonstrating that the MUT uterus had normal development and was functional in responding to E2 treatment. These data also confirmed that the MUT MDs had normal triphasic process of initiation, elongation, and differentiation.

Impaired fertility in MUT males.
To speed up the breeding of MUT females, MUT males were mated with HET females but many did not sire offspring despite normal mating activity (Fig. 5a). Male fertility test by mating control and MUT males with control females revealed that ~70% of MUT males were infertile (Fig. 5b). The fertile MUT males produced normal litter sizes (Fig. 5c). Although the Lhfpl2 MUT infertile males at 6-month old had slightly lower body weight and testis weight than their age-matched control males (see Supplementary  Fig. S4a,b), their relative testis weight was comparable with that of the control (see Supplementary Fig. S4c) and there was no significant difference in the sperm counts from cauda epididymis (see Supplementary Fig. S4d) or the testis histology (Fig. 5d,e) between WT and sterile MUT males. However, none of the infertile males delivered sperms to the reproductive tract of their mated control females (Fig. 5f). These data demonstrated that the cause for MUT male infertility was not in the testis or epididymis, but beyond the cauda epididymis. Immunohistochemistry using an anti-LHFPL2 antibody (Fig. 5g) localized LHFPL2 in the epithelium of PND7 vas deferens (Fig. 5h,i).

Distal vas deferens blockage caused by defective distal WD development in Lhfpl2 mutant males.
There was a smooth transition at the junctions of vas deferens and urethra with no visible "knots" in WT adult males (0/8) (Fig. 6a). In the 19 MUT adult male examined, 15 (15/19 = 79%) had clearly visible "knots" at the junctions of vas deferens and urethra (Fig. 6c,d), and the other four (4/19 = 21%) had small "knots" that can be identified with bright light under a dissecting scope (Fig. 6b). These "knots" indicated distal vas deferens convolution. In addition, 6 (6/19 = 32%) MUT males with visible "knots" also had abnormal seminal vesicles (e.g., 1-2 extra or tiny lobes of seminal vesicles) on one side or both sides ( Fig. 6d and data not shown). To reveal the passage from vas deferens to urethra, blue dye was injected through one vas deferens. The blue dye traveled through vas deferens to bladder and urethra but not the seminal vesicle in the WT (Fig. 6e). The blue dye traveling paths in the MUT males varied: no pass beyond vas deferens (Fig. 6f), traveling to seminal vesicle but not urethra or bladder (Fig. 6g), or traveling mainly to seminal vesicle and very limited to the bladder and urethra (Fig. 6h). The MUT males in the first (Fig. 6f) and second (Fig. 6g) scenarios were infertile while those in the third scenario (Fig. 6h) were fertile. Interestingly, it seemed that the blue dye could travel through the "knot" (Fig. 6h,h1) only in the fertile males. These data indicated that distal vas deferens convolution led to complete blockage in the infertile MUT males and incomplete blockage in the fertile MUT males. Histology of adult vas deferens showed that the MUT lumen ( Fig. 6j) but not WT lumen (Fig. 6i) was filled with dense sperm. Sperm were detected in the MUT seminal vesicles (Fig. 6l) that the blue dye traveled to (Fig. 6g,h,h1) but not in WT seminal vesicles (Fig. 6k). PAX2 staining of E15.5 male embryos revealed that the two WT WD tips had individually entered the urogenital sinus (Fig. 6m), while the two MUT WD tips were enlarged and flattened and failed to enter the urogenital sinus (Fig. 6n). In addition, the tips of the MUT MDs were enlarged and apart while the WT MDs had merged and without enlargement (Fig. 6m,n). These data (Figs 5 and 6) demonstrated that the MUT WDs had no obvious defect in the triphasic process involving initiation, elongation, and differentiation; and that the defect in MUT WDs was at the WD tips that failed to merge with urogenital sinus, leading to convolution at the junctions of vas deferens and urethra (Fig. 6o) and abnormal seminal vesicles in some MUT males (Fig. 6d).

Discussion
Vaginal opening has been used as an early indicator of pubertal onset in rodents. It can be regulated by estrogen signaling because estrogenic endocrine disruptors could promote vaginal opening [19][20][21][22][23] , while anti-estrogens could  34 . Estrogen-induced apoptosis has been proposed as a mechanism for vaginal opening during pubertal development 34 . Since exogenous E2 did not induce vaginal opening in Lrp2/Megalin-deficient mice 31 similar as in this LHFPL2 G102E mouse strain with lower vaginal atresia, it is possible that vaginal atresia might be responsible for vaginal imperforation in Lrp2/Megalin-deficient mice.
Interestingly, β -catenin C429S mice 32 have several similar phenotypes with LHFPL2 G102E mice. For example, they both have normal oogenesis and spermatogenesis. They both have double-lumen upper vagina and no lower vagina, which occur in 91% β -catenin C429S females and 100% LHFPL2 G102E females. β -catenin C429S males are 100% infertile mainly due to abnormal seminal vesicles 32 . LHFPL2 G102E males are 70% infertile, but 100% of them have distal vas deferens convolution with complete or incomplete blockage, in addition to 32% of them having abnormal structure of seminal vesicles. These phenotypes in males point to defective development in the caudal WD, which differentiates into distal vas deferens and seminal vesicles 5,45 . Since LHFPL2 G102E WDs elongated but couldn't merge with the urogenital sinus thus the WD tips became enlarged, and the remaining male reproductive tract had normal differentiation, it would suggest that the abnormal differentiation of distal vas deferens and seminal vesicle in the LHFPL2 G102E males resulted from failed merging with urogenital sinus. It has been estimated that congenital vaginal atresia occurs in 1 of ~4000-5000 live female births in humans 46 . The mechanisms in human congenital vaginal atresia are unknown. Animal studies have provided clues for distal vaginal development. Lineage tracing study has demonstrated that the vagina is solely derived from MD in mice 47 . At birth, the vagina consists of "Müllerian vagina" (upper vagina) and "sinus vagina" (lower vagina). During postnatal development, the "Müllerian vagina" extends caudally following the migration of the "sinus vagina" towards the posterior end of the body. Although genetic fate mapping reveals that WD does not contribute cells to the MD 7 , there is caudal residual WD near MD 47 that might guide the extension of MD to form the entire vagina 48 . The extension of MD depends on the cell proliferation of the caudal tip of the developing MD to deposit a cord of cells with mesoepithelial nature 7 .
Based on current knowledge, at least four mechanisms may explain vaginal atresia in the LHFPL2 G102E female and potentially in human. First, "sinus vagina" origin: "sinus vagina" fails to migrate towards the posterior end of the body so the "Müllerian vagina" could not extend to form the entire vagina. Second, MD origin: the caudal tip of the developing MD does not proliferate when it gets close to the urogenital sinus. However, the enlarged MD tips in both male and female E15.5 LHFPL2 G102E embryos would suggest that the MD tips could proliferate but could not extend to the urogenital sinus thus became enlarged. Third, WD origin: because the WD does not extend to the urogenital sinus in the LHFPL2 G102E embryos, the MD does not have a guide to extend. Fourth, defective communication: besides the three physical origins described above, it is possible that signaling molecules communicating between the urogenital sinus and MD might be altered. These mechanisms might also prevent the timely fusion of the upper vagina thus resulted in longitudinal upper vaginal septum in the LHFPL2 G102E females. Male distal reproductive tract obstruction, such as distal vas deferens convolution and ejaculatory duct obstruction, is responsible for 1-5% of male infertility 49 . Although the molecular and cellular mechanisms for male distal reproductive tract development are also not well understood, mutations or deletion of CFTR (cystic fibrosis transmembrane conductance regulator), a membrane protein, are associated with a range of defects in vas deferens in mice 50 , pigs 51 , and humans 52,53 , indicating the critical role of CFTR in vas deferens development. In the LHFPL2 G102E males, the main anomaly in the reproductive tract is confined to the distal connection of the vas deferens, seminal vesicles, and urethra. The obvious defect is failed fusion of caudal WD with urogenital sinus, leading to the enlarged WD tips on E15.5 and distal vas deferens convolution shown as visible "knot" at the junction of vas deferens and urethra in adult.
According to a microarray database from GenitoUrinary Development Molecular Anatomy Project (GUDMAP), Lhfpl2 mRNA is expressed in the fetal male reproductive tract (WD) at all the time points checked between E12 and E18 (http://www.gudmap.org/gudmap/pages/mastertable_browse.html? gene = Lhfpl2&geneId = MGI%3A2145236&masterTableId = 4_7&cleartabs = true). Since Lhfpl2/LHFPL2 is mainly detected in the epithelium of the postnatal and adult reproductive tracts, it is possible that Lhfpl2/LHFPL2 is also expressed in the epithelium of fetal reproductive tracts.
LHFP family members are tetraspan membrane proteins. It has been demonstrated that LHFPL5/TMHS is an integral component of the mechanotransduction machinery of cochlear hair cells regulating mechanotransduction in these cells and mutations in LHFPL5 cause hear loss 15,17 . Since a mutation in Lhfpl2 (LHFPL2 G102E ) also leads to hear loss, it is possible that LHFPL2 might have a similar function as LHFPL5 in regulating mechanotransduction. Because mechanotransduction is a driving mechanism for organ morphogenesis 54 , LHFPL2 may play a role in mechanotransduction between the WDs and the urogenital sinus when the WD tips reach the urogenital sinus. The failed extension of WDs to the urogenital sinus in the LHFPL2 G102E embryos may simultaneously block the WD-guided extension of MDs to the urogenital sinus, eventually leads to defective distal reproductive tract development in both males and females.
Taken together, this study reveals that LHFPL2 is essential for distal reproductive tract development in both female and male mice and provides a novel genetic basis for distal reproductive tract development. This study also demonstrates that besides initiation, elongation and differentiation of MDs and WDs, an additional merging phase with urogenital sinus is essential for proper distal reproductive tract development. This LHFPL2 G102E mouse strain is a good model for studying mechanisms in distal reproductive tract development. It may also provide a genetic etiology for clinical syndromes, such as Park-Jones Syndrome associated with absent vagina and auditory anomaly 55 . It is interesting to investigate if LHFPL2 mutations are associated with longitudinal vaginal septum, vaginal agenesis, or obstructive azoospermia in humans.

Methods
Animals. Lhfpl2 mutant mice (Stock No: 013716, LHFPL2 G102E ) in C57BL/6J background were purchased from the Jackson Laboratories (Bar Harbor, ME, USA) to establish a colony in Coverdell Vivarium at University of Georgia. All mice were housed in polypropylene cages with free access to food and water. The animal facility was maintained on a 12-hour light/dark cycle (0600 h to 1800 h) at 23 ± 1 °C with 30-50% relative humidity. All methods used in this study were approved by the University of Georgia Institutional Animal Care and Use Committee (IACUC) and conform to National Institutes of Health guidelines and public law.
Genotyping. Genomic DNA was extracted using DirectPCR (Viagen, USA) according to manufacturer's protocol. Two methods were used to determine the genotypes. One was PCR coupled with restriction digestion. A primer pair (Supplementary Table S1) flanking the point mutation on exon 3 amplified a 461 bp PCR product, which was subsequently digested by BglII (New England Biolabs, MA, USA) at 37 °C overnight. Undigested and digested products were run in 10% agrose gel. The digested wild type (WT) allele was expected to be 461 bp and the digested Lhfpl2 mutant (MUT) allele was expected to be 189 bp and 272 bp. The other genotyping method was Custom Taqman SNP (single nucleotide polymorphism) genotyping assay (Life Technology, CA, USA). Taqman SNP genotyping primers were designed according to manufacturer's instruction. VIC fluorescence represented WT allele; FAM fluorescence indicated MUT allele.
Generation of Lhfpl2 mutant transgenic mice. Transgenic mice in C57BL/6J background harboring wild type mouse Lhfpl2 gene were generated at Cyagen Biosciences (CA, USA) using BAC ID#23-403I16. Three female and one male founders were produced. These female founders (WT/BAC+ ) were mated with pre-identified fertile Lhfpl2 mutant males (MUT/BAC− ) to generate Lhfpl2 transgenic Lhfpl2 heterozygous offspring (HET/BAC+ ). Female and male HET/BAC+ mice were mated to generate Lhfpl2 transgenic Lhfpl2 mutant offspring (MUT/BAC+ ). To determine the rescue effect of the transgene, the female MUT/BAC+ mice were monitored for vaginal opening, the male MUT/BAC+ mice were either tested for fertility or injected with blue dye via vas deferens to visualize the passage of the distal reproductive tract as described below.
Visualization of reproductive tract. To visualize MUT vagina, the MUT female reproductive tract was dissected and Evans blue dye (0.05% in 1 × PBS) was injected via a blunt needle into both uterine horns and was accumulated in the two compartments of the upper vagina. To visualize the passage in male distal reproductive tract, the male reproductive system together with urethra were dissected and submerged into sesame oil in a petri dish. Under a dissecting microscope, Evans blue dye was injected into one vas deferens using a blunt needle. The Scientific RepoRts | 6:23037 | DOI: 10.1038/srep23037 distribution of the blue dye upon injection revealed whether or not there was a blockage in the male distal reproductive tract. Images were taken with a Nikon digital camera.
17β-estradiol (E2) treatment. Newly weaned WT and MUT females (3 weeks old) were i.p. injected with 100 μl sesame oil (vehicle) or 12.5 μg E2 (Sigma-Aldrich, USA) in vehicle for 3 consecutive days. Vaginal opening was monitored daily. All females were sacrificed the day after the third injection. Body weight, the weight of the female reproductive system (FRS, including ovary, oviduct, uterus, fluid, and vagina), the weight of the fluid in the female reproductive tract, which was determined by the difference between the FRS weight above and the FRS weight after the fluid was drained, and the weight of the uterine horns. Uterine tissues were fixed in formalin for histology. Relative weight was determined by dividing each weigh with body weight. N = 5-9.
Collection of vaginas. WT  Perfusion. WT females at estrus stage, which was determined by vaginal smear 56 , WT and MUT females at 8 weeks old were perfused transcardially as previously described 57 . Cross sections of the uterus and upper vagina were obtained.
RT-PCR. The gene specific primers encompassing different exons were listed in Supplementary Table S1. Total RNAs from the following tissues (8 weeks old males and females, and 3 weeks old females), heart, spleen, liver, kidney, thymus, testis, brain, lung, intestine, vagina, ovary and uterus, were extracted using TRIzol (Invitrogen, Carlsbad, CA, USA). RT-PCR was done as previously described 58,59 . PCR products were visualized by electrophoresis in 10% agrose gel. Hprt1 (hypoxanthine phosphoribosyltransferase 1) was used as a loading control.
In situ hybridization. In situ hybridization was performed as previously described in frozen sections from female reproductive system 60 . Sense and antisense probes for Lhfpl2 were synthesized from cDNA fragment amplified with a specific Lhfpl2 primer pair (Table S1) and confirmed by sequencing (Genewiz, USA).
Male mating activity, fertility, sperm delivery test, and sperm count. To determine the mating activity of the males (2 months old), control (WT/HET, N = 8) and MUT (N = 10) males were each housed with 3 virgin control (WT/HET, 2-4 months old) females, which were checked for a vaginal plug next morning. Plugging latency (the number of days between the start of cohabitation and the appearance of the first vaginal plug from each male) was recorded. To determine male fertility, each male (4 months old; control, N = 15; MUT, N = 17) was housed with 3 virgin control (WT/HET, 2-4 months old) females for 2 months to determine if it could sire offspring. The percentage of fertile males in each group was determined. Litter sizes were recorded from 16 litters sired by 6 WT males and 10 litters sired by 5 fertile MUT males. To examine sperm delivery, young control males (N = 4) and infertile MUT males (N = 6) were mated with control females. Upon identification of the first vaginal plug from a mated male, the female reproductive tract was flushed with 1 × PBS and the flushing was examined for the presence of sperm. Sperm counts were from cauda epididymis of 6 months old WT and infertile MUT males (N = 7), whose body weight and testis weight were also recorded.
Western blot. Uterine samples (N = 3) from gestation day 3.5 WT mice were homogenized in RIPA buffer with protease inhibitor (1:100). Protein concentrations were measured by Bradford assay with Nanodrop. Samples (30 μg) were loaded into polyacrylamide gel. Separated proteins were transferred onto PVDF membrane and blocked with 5% non-fat milk for 1 hour at room temperature on an orbital shaker. After washed with the Tris-buffered saline containing 0.1% tween (TBST), the membrane was incubated in the primary rabbit polyclonal anti-mouse LHFPL2 antibody (1:200, PA5-25166, Thermo Fisher Scientific) over 2 nights at 4 °C in TBST containing 1% BSA (bovine serum albumin) on an orbital shaker. The membrane was then washed and incubated with the appropriate peroxidase-labeled secondary antibody (1:3000) diluted in 5% non-fat milk. The image was developed on the film after incubation with Pierce ECL western blotting substrate.
Immunohistochemsitry. Immunohistochemistry was performed as previously described 61 . Briefly, PND7 vas deferens frozen sections (10 μm) were mounted, fixed, and subjected to antigen retrieval. Non-specific staining was blocked. Sections were then incubated with rabbit polyclonal anti-mouse LHFPL2 antibody (1:250, PA5-25166, Thermo Fisher Scientific) in blocking reagent at 4 °C for overnight, washed in 1 × PBS and incubated with biotinylated goat anti-rabbit secondary antibody (1:200 Santa Cruz Biotechnology, USA) for 30 min at room temperature. Sections were then incubated with ABComplex/HRP (Santa Cruz Biotechnology, USA) for 30 min at room temperature, washed, incubated with 3, 3′ -diaminobenzidine tetrahydrochloride, counterstained with hematoxylin, and mounted for imaging. Negative control using serial sections was processed exactly the same except that the primary antibody was replaced with normal rabbit IgG.
Statistical analyses. Two-tail, unequal variance Student's t test was used to compare means and chi-square test was used to compare rates. The significant level was set at p < 0.05.