Identification of a New Theca/Interstitial Cell-Specific Gene and Its Biological Role in Growth of Mouse Ovarian Follicles at the Gonadotropin-Independent Stage

Theca/interstitial cells are responsible for the growth and maturation of ovarian follicles. However, little is known about the theca/interstitial cell-specific genes and their functions. In this study, we explored transcriptomes of theca/interstitial cells by RNA-seq, and the novel biological roles of a theca cell marker, asporin (Aspn)/periodontal ligament-associated protein 1 (PLAP-1). RNA-seq detected 432 and 62 genes expressed specifically in theca/interstitial cells and granulosa cells isolated from 3-weeks old mouse ovaries. Gene ontology analysis demonstrated that these genes were largely categorized into four major groups: extracellular matrix organization-related terms, chemotaxis-related terms, the angiogenesis-related terms, and morphogenesis-related terms. In situ hybridization demonstrated that the newly detected representative gene, Aspn/PLAP-1, was detected specifically in the outer layer of theca cells in contrast with the expression of the basal lamina-specific gene, Nidgen-1. Intriguingly, an Aspn/PLAP-1 antibody completely arrested the growth of secondary follicles that is the gonadotropin-independent follicle developmental stage. Furthermore, transforming growth factor-β (TGF-β)-triggered signaling was induced by the Aspn/PLAP-1 antibody treatment, which is consistent with the inhibitory effect of Aspn/PLAP-1 on TGF-β. Altogether, these results suggest that theca cells are classified into subpopulations on the basis of new marker genes and their biological functions, and provide evidence that Aspn/PLAP-1 is expressed exclusively in the outer layer of theca cells and plays a pivotal role in the growth of secondary follicles via downregulation of the canonical TGF-β signaling cascade.


INTRODUCTION
In dioecious species, generation of fertile oocytes is an essential biological process. Oocytes are grown and matured in follicles, and ovulated from the ovary. Follicle development is a multi-step and complexed event controlled by a wide variety of endogenous factors. In mammals, a follicle is composed of a single oocyte, granulosa cells, and theca cells, and follicle development is largely classified into two stages: gonadotropin-dependent and independent stages. The former, namely, the antral follicle, is regulated by the hypothalamus-pituitary-gonad (HPG) axis via secretion of gonadotropins, LH and FSH, to the ovary in an endocrine fashion only at sexually mature individuals. In response to LH, theca cells produce androgen that is converted into estrogen by specific aromatases in response to FSH in granulosa cells, indicating that a major biological role of theca cells is production of sexual steroid hormones at the antral follicles (1)(2)(3). The latter is initiated at the fetus, where primordial follicles are generated. After birth, the primary, secondary and preantral follicles grow sequentially via the regulation exerted by diverse endogenous factors in the intraovarian paracrine/endocrine networks, in a manner independent of gonadotropins (1)(2)(3)(4)(5). In secondary follicles, theca cells are differentiated from interstitial (stromal) cells, and start the formation of the thecal layers and the basal lamina, indicating that theca cells are responsible for maintenance of physical structure of follicles (1)(2)(3)(4)(5)(6)(7). Moreover, interstitial cells, differentiating into theca cells, were found to be necessary for the normal growth of mouse secondary follicles in vitro (5,8). However, the net molecular mechanisms of follicle growth at gonadotropin-independent stages remain largely unknown.

Animals
Female ICR mice (Japan SLC, Inc., Shizuoka, Japan) were maintained under controlled light conditions (14L:10D) Abbreviations: Aspn, Asporin; BMP, bone morphogenetic protein; Nid1, Nidogen 1; PLAP-1, periodontal ligament-associated protein 1; TGF-β, transforming growth factor-β. and were given food and water ad libitum. The day of birth was designated as Day 0. On Day 1, each mother was left with eight pups to equalize the growth of pups between litters. Our investigations were conducted in accordance with the Animal Care Committee of Nara Women's University.
Preparation of Follicles, Theca/interstitial Cells, and Granulosa Cells Secondary follicles, theca/interstitial cells, and granulosa cells were collected from the ovaries of 20 3-weeks old different mice as previously described (6).

RNA-Seq
Total RNA was extracted from ovarian interstitial cells and granulosa cells using RNeasy Mini Kit (Qiagen Gmbh., Hilden, Germany). The quality of the RNA samples was evaluated by using BioAnalyzer (Agilent Technologies, CA, USA) with RNA6000 Nano Chip. A 1-µg aliquot of total RNA from each sample was used to construct cDNA libraries using TruSeq Stranded mRNA Sample preparation kit (Illumina, CA, USA), according to the manufacturer's instructions. The resulting cDNA library was validated using BioAnalyzer with DNA1000 Chip and quantified using Cycleave PCR Quantification Kit (TAKARA BIO INC., Shiga, Japan). Single end sequencing using 101 cycles was performed using HiSeq1500 (Illumina) in the rapid mode. Total reads were extracted with CASAVA v1.8.2 (Illumina). The resulting fastq files were deposited in Sequence Read Archive (SRA) under accession numbers SRR5097519 (Theca/Interstitial Cell, rep. 1), SRR5097520 (Theca/Interstitial Cell, rep. 2), SRR5097517 (Granulosa cells, rep. 1), and SRR5097518 (Granulosa cells, rep. 2) ( Table 1). Then, PCR duplicates, adaptor sequences and low-quality reads were removed from the extracted reads as follows. Briefly, if the first 10 bases of the two reads were identical and the entire reads exhibited >90% similarity, the reads were considered PCR duplicates. Remaining reads were then aligned with Bowtie version 2.2.3 to the mouse transcriptome mm10 with refseq ID, which was downloaded from ENSEMBL database. The difference of expression levels was calculated with edgeR (25), which is the Bioconductor package based on negative binomial generalized linear models.

GO Enrichment Analysis
To predict the function of ovarian interstitial cells and granulosa cells -specific transcripts, the sequences with lower p-values than 0.025 were analyzed using GO enrichment analysis, as described previously (26). The GO terms for the transcripts were downloaded from Mouse Genome Informatics database (MGI). Briefly, to quantify the enrichment of GO terms, we calculated enrichment scores as follows: where Nx(GO) indicates the frequency of each GO term for theca/interstitial cell (Gr<<In)-specific or whole transcriptome genes (i.e., x), and Ntotal(X) indicates the frequency of theca/interstitial cell (Gr<<In)-specific or whole mouse transcriptome genes (i.e., X) mapped to each GO term in MGI database. For correction, a pseudo-count was set as fixed value (0.05).

Real-Time PCR
Theca/interstitial cells and granulosa cells were isolated from the ovaries of 30 individuals of 3-weeks-old mice as previously described (5

In situ Hybridization in the Mouse Ovary
The open reading frames of Nid1 (nt 1,246-1,643) and Aspn/Plap-1 (nt 267-668) were amplified using ovary cDNA and gene-specific primers (Supplemental Data 2), respectively. The PCR program was 94 • C for 3 min, 35 cycles of 94 • C for 30 s, 50 • C for 30 s, and 72 • C for 45 s, and final extension at 72 • C for 7 min. Each of the PCR products was inserted into the pCRII-TOPO dual promoter vector (Thermo Fisher Scientific) according to the manufacturer's instruction and supplied to preparation of RNA probes as a template. Digoxigenin-labeled RNA antisense and sense probe for each gene was prepared using a digoxigenin-labeled RNA labeling kit (Roche Diagnostics, Switzerland). Ovaries of 3-weeks female mice were dissected and fixed in 4% paraformaldehyde in PBS at 4 • C overnight. The fixed tissues were soaked in a refrigerated sucrose solution (10% in PBS) for 30 min at room temperature. The tissues were then soaked in a 15% sucrose solution at 4 • C overnight until they sank. The ovaries were embedded in Super Cryoembedding Medium-L1 (Leica Microsystems Japan, Tokyo, Japan) and sectioned at a 10 µm thickness with a CryoStar NX70 cryostat (Thermo Fisher Scientific Inc.) at −18 • C. The sections were placed onto FRONTIER-coated slides (FRC-04; Matsunami Glass Ind., Ltd., Osaka, Japan). Hybridization, washing and detection were carried out as previously reported (27)(28)(29). No positive signals were observed when sense probes were used, confirming the specificity of hybridization.

Western Blotting of Smad Proteins in Aspn/PLAP-1 Antibody-Treated Ovaries
Ovaries were isolated from 3-weeks old mice, and cut longitudinally into symmetrical half-pieces. Each half-portion of the mouse ovaries was incubated with the culture medium containing anti-Aspn/PLAP-1 antibody, normal rabbit IgG (Wako, Japan), TGF-β1 (R&D systems, MN, USA), or BMP2 (R&D systems) at 37 • C for 0-3 h. The cultured ovaries were homogenized in 100 µl of RIPA buffer (Thermo Fisher Scientific Inc.) containing cOmplete, EDTA-free (Roche) and PhosSTOP (Roche) at 4 • C for 1-2 min at 28 Hz using Tissue lyser II (Qiagen Gmbh). The homogenate was incubated on ice for 10 min, followed by centrifugation at 15,300 g for 10 min at 4 • C.
The protein amount in the supernatant was measured using BCA protein assay kit (Thermo Fisher Scientific Inc.). 40 µg of the resultant ovarian protein lysate was separated on 5-20% gradient SDS-PAGE gel, and was electroblotted onto PVDF membranes. The membranes were blocked by Block-Ace (DS pharma biomedical, Osaka, Japan) for 1 h at room temperature, and were treated with a 1:1,000-2,000 dilution of an antibody against Smad2/3 (SantaCruz Biotechnology, CA, USA), phospho-Smad2/3 (p-Smad2/3) [Cell Signaling Technology (CST), MA, USA] Smad1/5/9 (Abcam, Tokyo, Japan) or p-Smad1/5/9 (CST) in the Can get signal 1 solution (TOYOBO, Osaka, Japan) overnight at 4 • C. After washing with 0.05% Tween TBS (TBST), the membranes were incubated with a 1:2,000 HRP-conjugated donkey anti-rabbit secondary antibody (GE healthcare, IL, USA) in Can get signal 2 solution (TOYOBO) for 1 h at room temperature. The anti-GAPDH antibody (Abcam) and the corresponding secondary antibody (GE healthcare) in TBST were used as an internal control. Signals were detected using a chemiluminescent detection system, ECL Select Detection Reagent according to the manufacturer's instruction (GE healthcare).

Statistical Analysis
Results are shown as mean ± SE. Data were analyzed by one-way ANOVA with Turkey's multiple comparison tests. Differences were accepted as significant for P < 0.05.

Detection of Mouse Theca/Interstitial
Cell-Specific Genes mRNA library of mouse ovarian theca/interstitial cells and that of granulosa cells were sequenced using Hiseq 1500, and 26-32 million reads for 101 single end reads were obtained from two replicates for each samples ( Table 1). After mapping to the mouse transcriptome, 77-82% of the reads were mapped, displaying 21,028 and 20,425 expressed genes with more than 10 reads on granulosa cell and theca/interstitial cell, respectively. Subsequently, we compared the gene expression levels between theca/interstitial cell and granulosa cell using EdgeR pipeline, detecting 432 and 62 upregulated genes with lower p-values than 10 −6 in theca/interstitial cell and granulosa cell, respectively (Figure 1). CYP17A1, BMP-4, and -7 genes, known theca/interstitial cell markers were shown to be intensely expressed in theca/interstitial cells ( Table 2). These data are consistent with the previous findings (1 -3, 11, 30), confirming the detection of differentially expressed genes by the RNA-seq method.
The GO enrichment analysis against upregulated genes in each cell detected 35 and 9 GO terms with GO-levels at 4-5 as enriched GO terms with more than 2 in GO enrichment score in granulosa cell and theca/interstitial cell, respectively ( Table 3). As depicted in Figure 2, 39 theca/interstitial cell-enriched GO terms were grouped to four groups; extracellular matrix organization-related  (Figures 1, 2, and Table 3). These results are in good agreement with the previous studies demonstrating that theca/interstitial cells participate in the follicle development via production of various extracellular matrix and cell adhesion proteins (1-3, 6, 31-33). Particularly, the high expression of Fn1 and Nid1, which are basal lamina-specific adhesive molecules in extracellular matrix (1,2,30,31), suggests the biological roles for the theca/interstitial cells in production of basal membranes of follicles. Chemoattractants Ccl2, Ccl21a, and Cxcl12 were 8.75-, 20.37-, and 3.20-fold upregulated in theca/interstitial cell, respectively (Figures 1, 2, and Table 3) which are contained in various GO terms related to chemotaxis; for instance, movement of cell or subcellular component (GO:0006928); cell chemotaxis (GO:0060326).
The GO terms related to morphogenesis, including regulation of anatomical structure morphogenesis (GO:0022603); regulation of developmental process (GO:0050793); and regulation of multicellular organismal process (GO:0051239), were enriched in theca/interstitial cell. Aquaporin 1 (Aqp1), Aspn/PLAP-1, and Cadherin 5 (Cdh5), which were categorized into all of these GO terms, were 20.48-, 11.46-, and 6.70-fold upregulated in theca/interstitial cell, respectively (Figures 1, 2, and Table 3), suggesting the role of theca/interstitial cells in some developmental processes. Furthermore, the expression of aforementioned 12 genes in theca/interstitial cells were examined by real-time PCR. As shown in Figure 3, real-time PCR demonstrated that the 11 genes were more intensely expressed in theca cells/interstitial cells than granulosa cells, validating the RNA-Seq data.

Localization of Nid1 and Aspn in the Mouse Ovary
TGF-β is found to exhibit some biological effects on isolated theca cells and granulosa cells (35,36), and Aspn/PLAP-1, a  small leucine-rich repeat protein, plays endogenous inhibitory roles for the TGF-β signaling (37)(38)(39)(40)(41)(42). These findings suggest some biological roles of Aspn/PLAP via regulation of TGF-β signaling in the ovary. However, localization of their mRNAs has yet to be examined. Subsequently, we thus performed in situ hybridization of Nid1 (as a reference to typical theca-specific localization) and Aspn/Plap-1 in the ovary. Nid1 was shown to be expressed in theca cells and interstitial cells of secondary, preantral and antral follicles but not in granulosa cells or oocytes ( Figure 4A). Interestingly, expression of Aspn/Plap-1 was detected specifically in the outer layer of theca cells and the adjacent interstitial cells but not in the inner layer of theca cells of secondary, preantral, and antral follicles ( Figure 4B). Altogether, these results provide evidence for the specific and differential expression of Nid1 and Aspn/Plap-1 in theca/interstitial cells in the ovary.

Biological Roles for Aspn/PLAP-1 in the Follicle Growth
The unique localization of the Aspn/Plap1 gene expression in the outer layer of theca cells (Figure 4B) suggested the involvement of Aspn/PLAP1 in the formation of the theca cell layers and the follicle growth. We thus evaluated the effect of Aspn/PLAP-1 on the growth of secondary follicles in a collagen gel medium (4). Aspn/PLAP-1 was formerly shown to suppress a TGF-β-Smad signaling cascade via interaction with TGF-β (11,(37)(38)(39)(40)(41)(42). As previously reported (5,6), all secondary follicles grew to preantral follicles, and interstitial cells developed to theca cell layers in 5 days (Figures 5A,D,E). Hence, we examined whether Aspn/PLAP1 affected the growth of secondary follicles in the presence of an anti-Aspn/PLAP-1 antibody. A striking feature is that treatment of secondary follicles with the anti-Aspn/PLAP-1 antibody resulted in complete loss of the growth of all secondary oocytes and follicles and of the theca cell layer formation (Figures 5C-E), whereas normal IgG is devoid of any effects (Figures 5B,D,E). Moreover, ∼70% of the non-grown secondary oocytes and follicles displayed shrunk shape in the presence of the anti-Aspn/PLAP-1 antibody (Figures 5C-E), but the basal lamina remains (Figure 5C). Altogether, these results revealed that Aspn/PLAP-1 plays a crucial role in the growth of secondary follicles.

DISCUSSION
Theca/interstitial cells are believed to participate in the follicle development and maturation via a wide range of endocrine/paracrine pathways. However, the biological significance of theca/interstitial cells in the growth of follicles largely remains unclear. In particular, these cells are expected to play crucial roles in the growth of secondary and preantral follicles as well as antral and preovulatory ones, given that such early-stage follicles are not regulated by gonadotropins from the HPG axis. Thus, a series of identification of theca/interstitial cell-specific marker genes and elucidation of their localization in follicles and their intra-follicular functions are expected to contribute to the verification of the regulatory mechanisms underlying the development of follicles by theca/interstitial cells at gonadotropin-independent stages.
RNA-seq and real-time PCR revealed the specific expression of genes encoding component proteins of extracellular matrix and basal lamina adhesive molecules, including Col3a1 and Nid1, in theca/interstitial cells (Figures 1-3). These expression profiles are compatible with the previous studies (2, 3, 6, 17-21, 32, 33) showing that some theca cells are responsible for construction of the basal lamina and extracellular matrix in various mammalian follicles. Gene expression of a typical basal lamina component, Nid1, was detected in multiple, but not all, the inner and outer layers of theca cells in secondary, preantral, and antral follicles ( Figure 4A). These results indicated that some populations of theca cells participate in formation of the basal lamina.
Various signaling molecule-related genes were also shown to be expressed specifically in theca/interstitial cells: including Aspn/PLAP-1 (Figures 1-3). To our knowledge, this is the first report showing the expression of Aspn/PLAP-1 in theca/interstitial cells. Furthermore, the gene expression of Aspn/PLAP-1 was shown to be limited to the outer layer of the theca of the secondary, preantral, and antral follicles ( Figure 4B). Such unique and unprecedented localization led to the identification of the Aspn/PLAP-1 gene as a novel specific marker for the thecal outer layer. Of particular significance is that neutralization of Aspn/PLAP-1 with an anti-Aspn/PLAP-1 antibody arrested the growth of the secondary follicles; 70% of them showed an abnormal shrunk morphology (Figure 5). Although Aspn/PLAP-1 has been shown to be involved in various biological and pathogenic processes including periodontal ligament mineralization and osteoarthritis (37)(38)(39)(40)(41)(42), we originally provide evidence for a novel biological role of Aspn/PLAP1 as a key factor for the normal growth of the secondary follicles. It is also noteworthy that the growth of the secondary follicles is not regulated by gonadotropins and that its rigorous mechanisms has been poorly investigated, compared with gonadotropin-dependent follicle maturation and ovulation of antral stage onward (1)(2)(3)(4)(5)(6)8). In other words, combined with the specific expression of the Aspn/PLAP-1 gene in the outer layer of theca cells, we have also elucidated a novel endocrine/paracrine role of the outer layer of theca cells in gonadotropin-independent growth of secondary follicles (Figure 5), and also presume that theca cells are categorized as various subpopulations on the basis of the specific gene expression and their biological functions.
Aspn/PLAP-1 has been shown to bind to only TGF-β and BMP2 among the TGF-β superfamily including activins and Growth/differentiation factors, and to down-regulate TGF-β-Smad2/3 signaling cascade via binding to TGF-β (37)(38)(39)(40)(41)(42). Moreover, blockade of Aspn/PLAP-1 triggered activation of this signaling cascade (Figure 6), leading to disruption of the growth of secondary follicles ( Figure 5). Altogether, these results verified that Aspn/PLAP-1 participates in the regulation of normal growth of secondary follicles via inactivation of TGFβ-Smad2/3 signaling cascades (Figure 6). In other words, TGFβ signaling cascades are negative factors of the growth of secondary follicles. The involvement of TGF-β in proliferation of granulosa cells and theca cells and LH-induced androgen production has been extensively documented (1)(2)(3)(9)(10)(11)(12). However, such studies have been limited to theca cells from antral follicles. Consequently, biological roles of TGF-β in secondary follicles, which are LH-insensitive, have yet to be investigated. Moreover, the functional correlation of these factors with antral follicle maturation has been controversial. This is due to variable distribution of gene and protein expression of TGF-β and their receptors among animal species and/or developmental stages of follicles (1)(2)(3)(9)(10)(11)(12). For instance, the expression of the TGF-β gene was detected in granulosa cells and theca cells of preantral follicles in human and mice, but only in theca cells in pig and bovine, and these expression levels are temporarily altered (10,11). Likewise, TGF-β was shown to suppress LH-induced androgen production in human theca cells, but not in their mouse counterparts (1,2,11). Such confounding findings regarding the effects of TGF-β may result from administration of TGF-β to cultured ovaries, given that the ovary includes various developmental stagefollicles, and thus, TGF-β may exhibit differential activities on follicles at the respective developmental stages. For instance, the effect of administrated TGF-β may over-exhibit, conflict, or suppress those of endogenous TGF-β. In keeping with these findings, studies using isolated theca cells may show biological functions from those residing in follicles, leading to the misunderstanding of authentic endogenous roles of TGF-β in theca cells. In this context, the specific localization of Aspn/PLAP-1 to the outer layer of theca cells ( Figure 4B) and our culture system using isolated secondary follicles (Figure 5) eliminate these possible artifacts, leading to the elucidation of a novel biological role of Aspn/PLAP-1 in regulation of the normal growth of secondary follicles via suppression of TGFβ-Smad2/3 signaling cascades. In addition, Aspn/PLAP-1 was found to be expressed in the outer layer of theca cells and the adjacent interstitial cells of preantral and antral follicles as well as those of secondary follicles (Figure 4B). These results suggest that biological functions of Aspn/PLAP-1, namely those of TGF-β, in preantral and antral follicles are different from those in secondary follicles. Investigation of the Aspn/PLAP-1 regulatory mechanism in the preantral and antral development and maturation is underway.
In conclusion, the present study has provided fundamental gene expression profiles in theca/interstitial cells in mice, identified a novel theca/interstitial cell marker, and explored the novel biological role of Aspn/PLAP-1 in the growth of mouse secondary follicles at the gonadotropin-independent developmental stages.

DATA AVAILABILITY
The datasets generated for this study can be found in the Sequence Read Archive.