Downstream Genes of Sox8 That Would Affect Adult Male Fertility

Abstract

Sertoli cells provide nutritional and physical support to germ cells during spermatogenesis. Sox8 encodes a high mobility group transcription factor closely related to Sox9 and Sox10. Sertoli cells produceSOX8 protein, and its elimination results in an age-dependent deregulation of spermatogenesis resulting in male infertility. This suggests that Sox8 is a critical regulator of Sertoli cell function for the maintenance of male fertility beyond the first wave of spermatogenesis. To better understand the roles of Sox8 in testicular development and maintenance of male fertility, we have performed a detailed analysis to explore its downstream genes. We have used mRNA expression profiling to identify affected genes in Sertoli cells in the mutant testes of 2-month-old mice. Expression profiling of the heterozygous and homozygous Sox8 mutant testes indicates alterations in several important spermatogenesis and blood-testis barrier genes, providing insight into the molecular basis of the defects in Sox8^–/– testes beyond the first wave of spermatogenesis.

Introduction

Sox8 isa member of the SRY-box containing gene family and encodes a high mobility group (HMG) transcription factor that possesses the ability to bind and bend DNA, thus allowing the trans-activation of target genes (Giese et al., 1994; Pontiggia et al., 1994). The Sox family has been divided into 10 subgroups, A–J. Sox8 along with Sox9 and Sox10 compose the SoxE subgroup (Bowles et al., 2000). Sox8 is a product of Sertoli cells and is critical for the maintenance of adult male fertility (O’Bryan et al., 2008). Apart from their obvious supportive role for the seminiferous epithelium, the Sertoli cells harbor clones of differentiating germ cells in cytoplasmic crypts and provide the germ cells with nutrients and growth factors throughout their development (Petersen and Soder, 2006). Sertoli cells regulate the flow of nutrients, growth factors, and mitogens by tight junctions between adjacent cells, creating a structure called blood-testis barrier (Lui and Cheng, 2007). Disruption of Sertoli cell function is suggested to be the cause of pathogenesis in the testicular dysgenesis syndrome (TDS) (Skakkebaek et al., 2001).

Based on its expression pattern in developing testes and ability to induce Amh expression in vitro, Sox8 was initially believed to be a regulator of male sex determination, differentiation, or germ cell development (Schepers et al., 2002, 2003). However, genetic ablation of Sox8 in mice does not result in abnormal sexual development. The lack of defects in sex determination in the mutant mice could be the result of a functional compensation by other SoxE family members (Chaboissier et al., 2004; Koopman, 2005). Detailed analysis revealed that the Sox8^–/– mice showed decreased adiposity (Sock et al., 2001) and premature osteoblast differentiation, which resulted in poor tarsal development and low bone density (Sock et al., 2001; Schmidt et al., 2005). We independently generated a Sox8 mutant mouse strain and found that Sox8^–/– males rarely produced litters, while Sox8^+/– males and Sox8^–/– females appeared to be reproductively normal (O’Bryan et al., 2008). In embryonic stages, no overt morphological disruption of cord formation was noted in Sox8^–/–embryos (O’Bryan et al., 2008). Very small numbers of young Sox8^–/– mice can sire pups with reduced litter size suggesting that at least some sperm resulting from the first few waves of spermatogenesis can be fertile and then they progressively become sterile. Testes weight was normal up to 35 days after birth, but by 2 months of age, the testes of Sox8^–/– mice were significantly smaller than those of control animals. Sox8^–/– adult mice (35 days) displayed an obvious infertility phenotype after the first wave of spermatogenesis (O’Bryan et al., 2008). A failure of spermiation was evident in stage IX tubules in 1-month-old mice and was seen more frequently in 2- and 5-month-old mutant animals, and a complete loss of the cycle of the seminiferous epithelium was evident by 9 months. In the gonad, Sox8 is expressed in Sertoli cells and not in germ cells (Kennedy et al., 2007). SOX8 localization varied within Sertoli cells based on the stage of the spermatogenic cycle. Specifically, within stage I–IX tubules SOX8 protein is localized in nucleo- and cytoplasm, whereas in stage X–XII tubules SOX8 protein was seen only in Sertoli cell cytoplasm (O’Bryan et al., 2008). Sertoli cells usually create 4 to 5 different and ever changing microenvironments concurrently around germ cells. Taken together this suggests that Sox8 expression in Sertoli cells regulates a set of genes that are essential for male germ cell differentiation and spermatogenesis. Sertoli cells in Sox8^–/– mice lose this ability in an age-dependent manner, which results in infertility due to the inability to establish full spermatogenesis and shows substantial deregulation of spermatogenesis at 2 months of age.

To gain insight into the molecular basis of this phenotype found in Sox8^–/– males, we have performed a detailed expression analysis using testes of 2-month-oldhomozygousand heterozygous mutant mice to identify the downstream genes of Sox8 that would affect male fertility. By microarray analysis we have determined that loss of Sox8 affects the expression levels of mRNAs that are important for the functions of Sertoli cells, germ cells, and the blood-testis barrier. The abnormal expression of these genes suggests an involvement of Sox8,in part, in the regulation of testicular genes required for male fertility.

Materials and Methods

Mice

The generation of Sox8 mutantmice has been previously described (O’Bryan et al., 2008). In brief, a 2.6-kb locus of Sox8 including exons 1 and 2 was replaced with the lacZ expression cassette and the floxed Pgk-neo cassette. The Pgk-neo cassette was removed after germline transmission of the mutation by breeding with Mox2-Cre mice. All animal studies were approved by the National Institute of Environmental Health Sciences Animal Care and Use Committee. All experimental data were collected from a minimum of 3 animals of each genotype and stage.

Genotyping

Genotypes of pups were determined by PCR analysis of genomic DNA with primers S8a: 5′-GAG GAC AAA GAT TGG GTC CTG C-3′ and S8b: 5′-GAA GCG TTC GTC TGC TGC C-3′ to detect the wild-type allele (299 bp); and primers S8a: 5′-GAG GAC AAA GAT TGG GTC CTG C-3′ and S8c: 5′-GAT GAA ACG CCG AGT TAA ACG C-3′ to detect the mutant allele (553 bp).

Microarray Analysis

Total RNA was isolated from 2-month-old heterozygous and homozygous mutant mice using TRIzol (Invitrogen). Gene expression analysis was conducted using Agilent Whole Mouse Genome 4×44 multiplex format oligo arrays (014868) (Agilent Technologies) following the Agilent 1-color microarray-based gene expression analysis protocol. Starting with 500 ng of total RNA, Cy3 labeled cRNA was produced according to the manufacturer's protocol. For each sample, 1.65 mg of Cy3 labeled cRNA was fragmented and hybridized for 17 h in a rotating hybridization oven. Slides were washed and then scanned with an Agilent Scanner. Data were obtained using the Agilent Feature Extraction software (v9.5), using the 1-color defaults for all parameters.

The Agilent Feature Extraction Software performed error modeling, adjusting for additive and multiplicative noise. The resulting data were processed using the Rosetta Resolver® system (version 7.0) (Rosetta Biosoftware, Kirkland, WA). In order to identify differentially expressed genes, analysis of variance (ANOVA) was used to determine if there was a statistical difference between the mean heterozygous and homozygous Sox8 mutants. Specifically, an error-weighted ANOVA with multiple test corrections was performed using Rosetta Resolver. The Benjamini Hochberg False Discovery Rate multiple test correction was used to reduce the number of false positives. Probes with p <0.01 were considered to be differentially expressed.

Quantitative PCR

Heterozygous and homozygous mutant mice were euthanized on postnatal day 15 (P15), P20, P25, P35, and at 2, 3, and 5 months of age. RNA was isolated from testes using TRIzol (Invitrogen), and cDNA was synthesized using the SuperScript First-Strand Synthesis System (Invitrogen) with random hexamer primers. Quantitative real-time reverse transcriptase PCR (Q-RT-PCR) measurements of individual cDNAs were performed with the ABI Prism 7700 sequence detection system. Gene-specific primers for Pde4D (Mm01304777_m1), Cd164 (Mm01189607_m1), Cep2 (Mm01329789_m1), Cldn3 (Mm00515499_s1), and Cldn23 (Mm00510971_s1) were purchased as pre-designed TaqMan gene-specific probe and primer mixtures (PE Applied Biosystems). The TaqMan rodent glyceraldehyde-3-phosphate dehydrogenase (Gapdh) control reagent (PE Applied Biosystems) was used as an internal control. All measurements were performedin triplicate. Values were normalized to Gapdh using the 2^–ΔΔCt method and expressed as means ± SEM.

Results

Expression Profiling and Validation

Sox8 is critical for the maintenance of male fertility (O’Bryan et al., 2008). Homozygous mutant mice were infertile except for a few that produced small litters of pups while young and then subsequently became infertile (O’Bryan et al., 2008). Weights of testes from Sox8^–/– mice were normal at 20 and 35 days after birth, suggesting that alteration of gene expression may not be apparent at this age. The obvious infertility phenotypes were apparent in homozygous mutant mice at 2 months of age; testes were significantly smaller, the seminiferous epithelium showed histological abnormalities, and the number and mobility of sperm were significantly reduced (O’Bryan et al., 2008). Based on these previous observations, we chose 2-month-old testes for gene expression profiling to identify the gene alteration in Sertoli cells.

We performed three independent microarray comparisons of mRNA from testes of heterozygous and homozygous mice at 2 months of age. Statistical analysis identified 1,601 probes as differentially expressed between the homozygous Sox8 mutant testes and the heterozygous testes (fig. 1A). Of these, 1,314 probes were upregulated, and 108 thereof showed a 2-fold induction or more in the mutant testes (fig. 1B, table 1). 286 probes were downregulated, and 46 of these showed a 2-fold reduction or more in mutant testes at 2 months of age (fig. 1C, table 2). The targeted disruption of the Sox8 locus resulted in a complete absence of Sox8 mRNA. Neither the targeted exons (exon 1 and 2) nor the residual coding exon (exon 3) were expressed in the mutant animals (O’Bryan et al., 2008). As expected, Sox8 was listed as the most reduced gene in the mutant testes by the microarray comparison (26-fold reduction at 2 month of age).

Fig. 1.

Microarray comparison of the levels of gene expression in heterozygous and homozygous mutant testes. A Induced genes are indicated in shades of red, and repressed genes are indicated in shades of green for the 3 comparisons shown. B Representative graph of the upregulated genes in testes of Sox8^–/–mice. C Representative graph of the downregulated genes in testes of Sox8^–/–mice. Genes listed in B and C that are known to be involved in testicular development were selected from the results shown in A.

Table 1.

Genes with at least 2-fold increased expression in Sox8^−/− testes at 2 months of age

Sequence code	Gene symbol	Sequence description	Fold change	ANOVA p value (p <0.01)
A_51_P370363	1700030L22Rik	RIKEN cDNA 1700030L22 gene	77.4	4.2E-14
A_51_P433789	Agr3	Anterior gradient homolog 3 (Xenopus laevis)	26.3	4.2E-14
A_51_P433796	Agr3	Anterior gradient homolog 3 (Xenopus laevis)	14.0	4.2E-14
A_52_P731333	5730526G10Rik	3 days neonate thymus cDNA, RIKEN full-length enriched library, clone:	10.3	4.2E-14
		A630054F14 product: unclassifiable, full insert sequence
A_52_P499523	Scfd1	Sec1 family domain containing 1	9.3	4.2E-14
A_52_P739568	Npas3	0 day neonate cerebellum cDNA, RIKEN full-length enriched library, clone:	7.0	7.0E-06
		C230053P15, product: unclassifiable, full insert sequence
A_52_P333567	Adam33	A disintegrin and metallopeptidase domain 33	6.3	2.0E-10
A_51_P335555	Snap25	Synaptosomal-associated protein 25	5.3	7.7E-04
A_51_P464420	4921508M14Rik	RIKEN cDNA 4921508M14 gene	5.0	5.1E-04
A_51_P458638	Spink8	Serine peptidase inhibitor, Kazal type 8	4.8	4.2E-14
A_51_P180423	Camp	Cathelicidin antimicrobial peptide	4.2	4.3E-09
A_51_P514177	Lrrc2	Leucine rich repeat containing 2	4.0	8.1E-07
A_52_P420608	Gm1631	Gene model 1631NCBI	3.9	4.0E-05
A_52_P188295	A930021C24Rik	RIKEN cDNA A930021C24 gene	3.9	5.9E-10
A_51_P131408	Tnfrsf12a	Tumor necrosis factor receptor superfamily, member 12a	3.8	1.0E-05
A_51_P452533	D1Ertd471e	DNA segment, Chr 1, ERATO Doi 471, expressed	3.8	9.0E-05
A_51_P340456	Ela3	Elastase 3, pancreatic	3.7	4.2E-14
A_51_P324633	Elovl3	Elongation of very long chain fatty acids (FEN1/Elo2, SUR4/Elo3, yeast)-like 3	3.6	2.0E-05
A_52_P339959	Rab3b	RAB3B, member RAS oncogene family	3.4	4.2E-14
A_52_P298002	Gch1	GTP cyclohydrolase 1	3.4	6.0E-05
A_51_P465211	Wfdc2	WAP four-disulfide core domain 2	3.3	4.2E-14
A_52_P420357	Slc15a1	Solute carrier family 15 (oligopeptide transporter), member 1	3.3	3.3E-03
A_51_P180091	Cyp2s1	Cytochrome P450, family 2, subfamily s, polypeptide 1	3.3	4.2E-14
A_51_P302942	Rasl10a	RAS-like, family 10, member A	3.3	6.2E-11
A_52_P261184	Il1rl2	Interleukin 1 receptor-like 2	3.2	1.0E-05
A_52_P520788	Scfd1	Sec1 family domain containing 1	3.2	3.9E-09
A_51_P340699	Rasl11a	RAS-like, family 11, member A	3.2	4.8E-09
A_51_P304478	Tmem28	Transmembrane protein 28	3.2	5.0E-05
A_51_P153042	Psg16	Pregnancy specific glycoprotein 16	3.2	1.5E-11
A_52_P159365	Sall3	Sal-like 3 (Drosophila)	3.2	5.8E-06
A_52_P21550	Gcnt1	Glucosaminyl (N-acetyl) transferase 1, core 2	3.1	9.9E-06
A_51_P385030	Svs6	Seminal vesicle secretory protein 6	3.1	4.2E-14
A_52_P452256	NAP055974-1	Unknown	3.1	1.1E-03
A_51_P464387	Hspb8	Heat shock protein 8	3.0	4.2E-12
A_52_P418952	Krt79	Keratin 79	3.0	4.2E-14
A_52_P214347	Cst11	Cystatin 11	3.0	1.6E-10
A_51_P138939	Robo4	Roundabout homolog 4 (Drosophila)	2.9	1.6E-06
A_51_P239203	Mapk13	Mitogen activated protein kinase 13	2.9	4.6E-09
A_51_P308048	Cmtm8	CKLF-like MARVEL transmembrane domain containing 8	2.8	4.2E-14
A_51_P496905	Cfi	Complement component factor i	2.8	1.4E-08
A_52_P180933	Defb8	Defensin beta 8	2.7	5.0E-05
A_52_P10622	Emb	Embigin	2.7	4.2E-14
A_52_P211418	6030408C04Rik	RIKEN cDNA 6030408C04 gene	2.7	1.0E-03
A_51_P465449	Mybpc3	Myosin binding protein C, cardiac	2.7	5.0E-07
A_51_P382849	Emb	Embigin	2.6	1.6E-12
A_51_P510466	Pldn	Pallidin	2.6	4.2E-14
A_52_P964651	2310033K02Rik	RIKEN cDNA 2310033K02 gene	2.6	1.0E-02
A_51_P157255	Sdc2	Syndecan 2	2.4	4.2E-14
A_51_P362638	Trf	Transferrin	2.4	4.2E-14
A_52_P244682	5430435G22Rik	RIKEN cDNA 5430435G22 gene	2.4	2.8E-06
A_52_P644297	Pafah1b2	Adult male spinal cord cDNA, RIKEN full-length enriched library, clone:	2.4	4.4E-03
		A330108B13, product: unclassifiable, full insert sequence
A_52_P65237	Zbtb7c	Zinc finger and BTB domain containing 7C	2.4	4.2E-14
A_51_P440365	Frrs1	Ferric-chelate reductase 1	2.4	2.0E-05
A_51_P100327	Tap1	Transporter 1, ATP-binding cassette, sub-family B (MDR/TAP)	2.4	3.0E-05
A_52_P407049	Hoxd10	Homeo box D10	2.4	1.8E-07
A_51_P267754	Icam2	Intercellular adhesion molecule 2	2.4	1.3E-07
A_52_P217796	1500035H01Rik	RIKEN cDNA 1500035H01 gene	2.3	3.0E-05

Table 2.

Genes with at least 2-fold decreased expression in Sox8^−/− testes at 2 months of age

Sequence code	Gene symbol	Sequence description	Fold change	ANOVA p value (p <0.01)
A_52_P269667	Sox8	SRY-box containing gene 8	−25.65	6.1E-09
A_52_P462657	TC1060275	Unknown	−14.25	2.3E-03
A_51_P259848	Ccdc32	Coiled-coil domain containing 32	−11.68	1.0E-03
A_52_P421413	Cd164	Mus musculus adult male thymus cDNA, RIKEN full-length enriched library, clone: 5830453P09, product: CD164 antigen, full insert sequence	−10.20	4.2E-14
A_51_P245324	4931402G19Rik	RIKEN cDNA 4931402G19 gene	−7.57	4.2E-14
A_52_P269672	Sox8	SRY-box containing gene 8	−5.31	4.2E-14
A_51_P438711	Tbx18	T-box18	−4.99	4.3E-08
A_51_P120990	Pde4d	Phosphodiesterase 4D, cAMP specific	−4.51	5.3E-03
A_51_P458707	ENSMUST00000055793	OLFACTORY RECEPTOR GA_X6K02T2NHDJ-9643949-9642957	−4.07	7.7E-03
		[Source: SPTREMBL; Acc: Q7TS05] [ENSMUST00000055793]
A_51_P157902	Cldn3	Claudin 3	−3.88	4.2E-14
A_52_P392216	Dab1	Disabled homolog 1 (Drosophila)	−3.87	4.2E-14
A_52_P529790	TC1034687	AK122264 mKIAA0377 protein (Mus musculus), partial (7%) [TC1034687]	−3.74	4.2E-14
A_51_P115178	Scara3	Scavenger receptor class A, member 3	−3.70	4.2E-14
A_51_P131494	Foxk2	RIKEN cDNA 6230415M23 gene	−3.64	4.0E-05
A_51_P326994	2810048G17Rik	RIKEN cDNA 2810048G17 gene	−3.57	6.2E-11
A_51_P497295	Qpct	Glutaminyl-peptide cyclotransferase (glutaminyl cyclase)	−3.45	3.6E-13
A_51_P176811	Tspan10	Tetraspanin 10	−3.33	3.5E-09
A_52_P1202828	6620401J10Rik	RIKEN cDNA 6620401J10 gene	−3.32	9.9E-04
A_52_P1027837	9330175M20Rik	RIKEN cDNA 9330175M20 gene	−3.15	5.0E-05
A_52_P246831	Cep27	Centrosomal protein 27	−3.10	4.2E-14
A_52_P192085	Kcng2	Potassium voltage-gated channel, subfamily G, member 2	−2.94	4.2E-14
A_51_P271200	Slco1a5	Solute carrier organic anion transporter family, member 1a5	−2.81	6.8E-10
A_51_P302831	Dbc1	Deleted in bladder cancer 1 (human)	−2.77	8.0E-03
A_52_P377838	AY172399	Cell-line M2-86 immunoglobulin heavy chain variable region	−2.68	8.5E-03
A_52_P110604	Hmg20a	High mobility group 20A	−2.67	3.3E-03
A_52_P421543	Cldn23	Claudin 23	−2.56	2.5E-12
A_52_P253044	Syt13	Synaptotagmin XIII	−2.47	4.2E-14
A_51_P114094	Clstn3	Calsyntenin 3	−2.44	4.2E-14
A_51_P128807	4930485G23Rik	MAX gene associated	−2.42	2.6E-12
A_52_P541095	Unc5a	Unc-5 homolog A (C. elegans)	−2.39	4.2E-14
A_52_P116102	Tpi1	Triosephosphate isomerase 1	−2.39	5.0E-09
A_51_P496253	Slc6a4	Solute carrier family 6 (neurotransmitter transporter, serotonin), member 4	−2.34	1.2E-04
A_51_P155152	Ank	Progressive ankylosis	−2.34	1.1E-03
A_52_P322096	Patl2	RIKEN cDNA 4930424G05 gene	−2.31	3.8E-04
A_51_P258894	Chst2	Carbohydrate sulfotransferase 2	−2.27	4.7E-04
A_52_P1149183	AI841723	Transcribed locus	−2.22	6.0E-03
A_51_P249335	Sds	Serine dehydratase	−2.20	1.9E-07
A_52_P185705	Acn9	Mus musculus adult male corpora quadrigemina cDNA, RIKEN full-length enriched library, clone: B230384K14, product: hypothetical ACN9 containing protein, full insert sequence	−2.15	8.6E-03
A_51_P424739	Nmnat3	Nicotinamide nucleotide adenylyltransferase 3	−2.15	3.6E-03
A_52_P304105	Ppp1r16b	Protein phosphatase 1, regulatory (inhibitor) subunit 16B	−2.13	4.8E-10
A_52_P723436	D130062J10Rik	RIKEN cDNA D130062J10 gene	−2.08	1.3E-03
A_52_P674189	4930512B01Rik	RIKEN cDNA 4930512B01 gene	−2.08	9.9E-03
A_52_P67270	4930515G01Rik	Mus musculus RIKEN cDNA 4930515G01 gene (4930515G01Rik), mRNA	−2.06	6.3E-03
A_52_P168575	EG434197	Predicted gene, EG434197	−2.06	2.0E-05
A_52_P599317	Hs6st2	Heparan sulfate 6-O-sulfotransferase 2	−2.01	4.2E-14
A_51_P386069	Rab9b	RAB9B, member RAS oncogene family	−2.00	4.2E-14

Among the genes identified, some are known regulators of testicular development. These include Sex comb on midleg-like 4 (Scml4, 2.1-fold induction); SRY-box containing gene 9 (Sox9, 1.4-fold induction); cytochrome P450, family 2, subfamily s, polypeptide 1 (Cyp2s1, 3.3-fold induction); DMRT-like family C1a (Dmrtc1a, 1.3-fold induction); regulator of sex limited protein 1 (Rsl1, 1.3-fold induction); and Anti-Mullerian hormone type 2 receptor (Amhr2, 1.5-fold induction). Some of the chromatin remodeling factors known for gene repression were also upregulated in the testes of homozygous mice. These include histone deacetylases 6 (Hdac6, 1.3-fold induction) and chromodomain helicase DNA binding protein 6 (Chd6, 1.7-fold induction). In addition, the expression of spermatogenesis-associated genes was altered in the mutant testes, including phosphodiesterase 4D (Pde4D, 4.5-fold reduction), centrosomal protein 27 (Cep27, 3.1-fold reduction), sperm motility kinase 3A (Smok3a, 1.2-fold induction), and spermatogenesis associated, serine-rich 1 (Spats1, 1.2-fold reduction). Furthermore, there were changes in the expression of genes for proteins perhaps associated with the blood-testis barrier, including occludin (Ocln, 1.4-fold induction), claudin 11 (Cldn11, 1.4-fold induction), claudin 12 (Cldn12, 1.4-fold induction), claudin 10 (Cldn10, 1.5-fold reduction), claudin 3 (Cldn3, 3.9-fold reduction), and claudin 23 (Cldn23, 2.6-fold reduction) (table 1 and 2).

Expression Levels of the Downstream Genes for Sox8 in the Mutant Testes at 2 Months of Age

To validate the microarray results we further analyzed mRNA expression levels of genes by quantitative real-time reverse transcriptase PCR (Q-RT-PCR) in 2-month-old testes. Cldn23, Cep27, Cldn3, and Pde4D were chosen because their expression levels were downregulated more than 2-fold in the mutant, and they were known for their involvement in spermatogenesis and formation of the blood-testis barrier. We also included Cd164 because it encodes a cell surface antigen expressed higher in developing testes than in developing ovaries (McClive et al., 2003) and found it to be highly downregulated (10.2-fold reduction) in the 2-month-old mutant testes (table 2).

Studies of human sperm centrioles showed a greater incidence of centriolar abnormalities in nonprogressively motile spermatozoa when compared with normally motile sperm (Sathananthan, 1994). Males homozygous for a Sox8 mutation displayed an age-dependent decrease in the percentage of progressively motile sperm (O’Bryan et al., 2008). Together, these data suggest that downregulation of Cep27 in the mutant testes could be the cause of immotile sperm. Claudins are known as the major component of tight junctions in several tissues (Tsukita and Furuse, 2000; Van Itallie and Anderson, 2004). A number of claudins have been shown to be expressed in testes (Lui et al., 2003). Cldn3 was concentrated in the tight junctions near the basal lamina by P20 (Meng et al., 2005). Pde4D is expressed in pachytene spermatocytes and during the spermatid elongation phase (Salanova et al., 1999). Based on the localization and function of these repressed genes, we assayed mRNA expression by Q-RT-PCR. As expected, expression levels of all 5 genes were reduced in the mutant testes at 2 months of age (fig. 2A). The expression level of Cldn23 was reduced most in mutant testes compared to the other 4 genes. Cep27,Cldn3, and Cd164 also showed reduced expression levels in the homozygous mutant testes compared to the heterozygous testes. The change in expression level of Pde4D was relatively minor in the homozygous mutant testes at 2 months of age compared to the heterozygous testes.

Fig. 2.

Temporal pattern of expression of putative downstream genes of Sox8 in adult testes. Quantitative RT-PCR determined the relative abundance of Pde4D, Cep27, Cldn3, Cldn23, and Cd164 transcripts compared to Gapdh mRNA in heterozygous (black bars) and homozygous (gray bars) mutant testes for Sox8 at 2 months (A), 3 months (B) and 5 months (C) after birth. The data are expressed as mean ± SEM.

Age-Dependent Expression of the Downstream Target Genes for Sox8 in Mutant Testes

Previously, we have shown that spermiation failure was evident in 1-month-old homozygous mutant mice but becomes more apparent in 2- and 5-month-old homozygous mutant mice, resulting in complete sterility (O’Bryan et al., 2008). This led us to investigate the expression of these selected genes at 3 and 5 months of age. At 3 months, the expression of all 5 genes was reduced in homozygous mutant testes compared to heterozygous testes. The expression level of Cldn3 in mutant testes was greatly reduced at 3 months of age relative to the 2-month-old animals (fig. 2B). The level of Cldn23 expression was higher at 3 months relative to the level in testes of 2-month-old mutants but reduced in comparison to the level in testes of heterozygous animals. The Cep27 mRNA expression level was reduced at 3 months and 5 months of age in homozygous testes compared to heterozygous testes. In contrast, the expression levels of Pde4D, Cldn3, and Cd164 at 5 months of age were unchanged or showed only small changes in homozygous compared to heterozygous testes (fig. 2C), suggesting that these genes are required for male fertility in a particular time frame during spermatogenesis. Expression levels of Cldn23 in 5-month-old testes were too low to detect.

Sertoli cells proliferate during fetal life until puberty and then cease to divide. In wild type mice, Sertoli cells cease proliferation by about P15 and undergo maturation (Sharpe et al., 2003; Walker, 2003). At the start of puberty (P25) the tight junctions of the barrier are formed (Skinner and Griswold, 1980). Following the establishment of full spermatogenesis at P35 in the mouse (Kramer and Erickson, 1981), each Sertoli cell concurrently supports 4 or 5 different types of germ cells. Our previous study demonstrated that 80% of Sox8 mutant males were sterile, and ∼20% were able to produce small litters at young age (O’Bryan et al., 2008). Where specific expression changes were seen in microarray expression analysis at 2 months of age, we investigated expression at earlier time points, prior to the time that changes in spermatogenesis are observed in Sox8 mutant testes. To investigate the expression of downstream genes of Sox8, RNA wasextracted from P15, P20, P25, and P35 heterozygous and homozygous mutant testes.

In contrast to the findings from adult mice, the expression levels of Pde4D, Cep27, and Cldn3 were elevated at P15 in the Sox8 mutant testes before puberty and blood-testis-barrier formation (fig. 3A). Similar results were obtained from P20 samples for Pde4D, Cep27, Cldn3, and Cd164 (fig. 3B). The expression levels of Pde4D, Cep27, and Cd164 were increased in Sox8^–/– mice at P25, at the time of blood-testis-barrier formation (fig. 3C). At this stage, Cldn3 expression was decreased in Sox8 mutant testes (fig. 3C). At the establishment of full spermatogenesis (P35) expression levels of Pde4D,Cldn3, and Cd164, but not Cep27 were decreased in the homozygous mutant testes (fig. 3D). These results suggested that the loss of Sox8 in Sertoli cells might be the cause of differential changes in expression of these genes in an age-dependent manner and in different stages of testicular maturation.

Fig. 3.

Effect of Sox8mutation on expression of selected downregulated genes before puberty (A and B, P10 and P20, respectively) at puberty (C, P25) and during the establishment of full spermatogenesis (D, P35). The relative abundance of Pde4D, Cd164, Cep27, Cldn3, and Cldn23 transcripts in heterozygous (black bars) and homozygous (gray bars) mutant testes compared to Gapdh mRNA as an external control is shown (mean ± SEM).

Sequence analysis 2.0 kb upstream of the Pde4d and Cldn23 promoter revealed the presence of a putative SOX binding site (cacacac) between −748 and −742 bp (Pde4d) as well as a SOX4 binding site (tctttctctc) between −2095 and −2086 bp and a SOX2 binding site (cacacacaca) between −940 and −931 bp (Cldn23) from the initiation codon. Quantitative mRNA expression analysis of the downregulated genes in Sox8 heterozygous and homozygous mutant testes revealed age-dependent expression changes in the homozygous mutant testes. These data strongly suggest that ablation of Sox8 results in an age-dependent and within a definite time frame progressive decrease in the expression level of the selected downregulated genes.

Discussion

In the present study we have investigated transcriptional changes in the Sox8 mutant testis during the stages immediately preceding extensive germ cell loss in the Sox8^–/–mice. Expression profiling identified 108 mRNAs whose expression is significantly (>2-fold) elevated, and 46 mRNAs whose expression is significantly reduced (>2-fold) at 2 months of age, providing a suite of candidate effectors of Sox8 function for further investigation. We reported earlier that Sox8 mutant mice displayed an obvious mutant phenotype after the first wave of spermatogenesis in adult mice (35 days), and this became progressively more pronounced in 2-month- and 5-month-old Sox8 mutant animals (O’Bryan et al., 2008). Some of the genes shown in table 1 and 2 might become apparent due to the block in spermiogenesis or germ cell loss and their effect on whole tissue microarray expression data. Therefore, we chose genes for further analyses based on their expression patterns, i.e., genes specifically expressed in Sertoli cells and repressed more than 2-fold.

Among the genes altered in the Sox8^–/–testes there are several proteins important for testicular development, spermatogenesis, and formation of tight junctions.

Pde4D is expressed during spermatogenesis in round spermatids and pachytene spermatocytes (Salanova et al., 1999) in addition to Sertoli cells (Levallet et al., 2007). As we reported, in Sox8 mutant testes spermatogenesis is arrested at stage IX and the formation of round spermatids to elongated spermatids failed (O’Bryan et al., 2008). In Sox8 mutant testes, expression of Pde4D measured by quantitative RT-PCR decreased from 2 to 3 months of age, but then remained unchanged at 5 months of age. This finding suggested that Pde4D might be required for spermatogenesis at a specific age. At the same time, it appears that Sox8 is required for specific timing of Pde4D expression. The decrease of the expression of Pde4D in the mutant testes may be, at least in part, explained by a reduction of elongated spermatids where Pde4D is expressed. However, presence of a SOX binding site strongly suggests that SOX8 might directly regulate the expression of Pde4D in Sertoli cells. Further investigation is required to define molecular mechanisms on how SOX8 regulates expression of Pde4D.

We have found a number of genes up- and downregulated in Sox8 mutant testes that are important for tight junctions and the blood-testis barrier, for instance Cldn3 and Cldn23 which have a reduced expression between puberty and adulthood. Formation of the blood-testis barrier starts around P11. CLDN3 is first detected at P15 in a diffusive manner and then in the blood-testis barrier at P20 in newly formed tight junctions that regulate the permeability of the barrier as germ cells move from the basal to the adluminal compartment (Meng et al., 2005). Tight junctions in other tissues are shown to regulate the selective transport of ions and nutrients from the blood stream and interstitial fluid (Goodenough, 1999). We also found that Cldn11 and Cldn12, which are crucial for testis tight junction formation (Gow et al., 1999; Bronstein et al., 2000), were upregulated 1.4-fold in the microarray analyses. Lower expression of Cldn3 and Cldn23 in the Sox8 mutant testes suggested that this might be the cause of an altered function of the blood-testis barrier and contributes to male infertility. However, we previously found no overt changes in the expression of Espin and Vinculin (markers for the blood-testis barrier) in Sox8 mutant testes relative to control testes (O’Bryan et al., 2008). It is possible that the blood-testis barrier structurally forms in the Sox8 mutant testes, but with an altered function. Thus, further investigation of the blood-testis barrier function in Sox8 mutant testes is needed.

Spermatogenesis in seminiferous tubules proceeds if an appropriate environment is provided by Sertoli cells (Clermont, 1972). This notion is supported by the study reported here, which demonstrates that the loss of Sox8 in Sertoli cells downregulates the expression of a group of genes resulting in male infertility. By using microarray analysis and quantitative RT-PCR in a complementary strategy, we have identified expression changes associated with the loss of Sox8 and consequent loss of spermatogenesis in the mutant mice. Some of these genes are expressed in Sertoli cells and may be direct or indirect targets of SOX8. Further investigation into the function of these genes will help to explain their specific roles in testicular development, male fertility, and its maintenance in an adult age.