Novel inactivating mutations of the DCAF17 gene in American and Turkish families cause male infertility and female subfertility in the mouse model

Abstract

Loss-of-function DCAF17 variants cause hypogonadism, partial alopecia, diabetes mellitus, mental retardation, and deafness with variable clinical presentation. DCAF17 pathogenic variants have been largely reported in the Middle Eastern populations, but the incidence in American families is rare and animal models are lacking. Exome sequencing in 5 women with syndromic hypergonadotropic hypogonadism from 2 unrelated families revealed novel pathogenic variants in the DCAF17 gene. DCAF17 exon 2 (c.127–1G > C) novel homozygous variants were discovered in 4 Turkish siblings, while 1 American was compound heterozygous for 1-stop gain variant in exon 5 (c.C535T; p.Gln179*) and previously described stop gain variant in exon 9 (c.G906A; p.Trp302*). A mouse model mimicking loss of function in exon 2 of Dcaf17 was generated using CRISPR/Cas9 and showed female subfertility and male infertility. Our results identify 2 novel variants, and show that Dcaf17 plays a significant role in mammalian gonadal development and infertility.

1 |. INTRODUCTION

Woodhouse-Sakati syndrome (WSS, MIM 241080) is characterized by a combination of female/male hypogonadism, partial alopecia, diabetes mellitus, mental retardation, and deafness since 1983.¹ WSS is an autosomal recessive disorder and to date around 69 individuals, 34 males and 35 females, have been identified with WSS. Out of these patients, 12 males and 1 female were reported to be hypogonadotropic, whereas 2 males and 16 females as hypergonadotropic. Familial genetic analysis revealed that the loss-of-function variants in the DCAF17 (Ddb1- and Cul4-associated factor 17, MIM 612515) associate with WSS. So far, 10 homozygous variants and 2 compound heterozygous variants have been identified in the DCAF17 gene.^2–5 Like WSS, female/male hypogonadism due to elevated/reduced serum follicle-stimulating hormone (FSH) and luteinizing hormone (LH) is also seen in other rare inherited syndromes and include Kallmann syndrome (MIM 308750), Richards-Rundle syndrome (MIM 245100), Perrault syndrome 5 (MIM 616138), and mitochondrial DNA depletion syndrome (MIM 271245). Here, we evaluated 2 families from different ethnic populations whose members presented with primary amenorrhea and normal karyotype, and samples were referred for genomic analysis. We used whole-exome sequencing (WES) to identify underlying genetic etiology of hypogonadism in these 2 families. We identified homozygous and compound heterozygous pathogenic variants in the DCAF17 gene, predicted to cause truncated form of the DCAF17 protein in all 5 affected probands. We also utilized CRISPR-Cas9 to generate mice with loss of Dcaf17 function and assess its effects in gonadal development.

2 |. METHODS

2.1 |. Study approval

The IRB (#PRO09080427) approval to study the genetic basis of syndromic and non-syndromic hypogonadism was obtained from the University of Pittsburgh. All experimental and surgical procedures complied with the Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh.

2.2 |. Whole exome sequencing

Whole exome sequencing (WES) approach was used to identify pathogenic variants in these 2 families as previously described.⁶ The sequencing was performed at the Pittsburgh Clinical Genomics Laboratory. Exons and splice sites were captured using the Haloplex Exome Target Enrichment System, and 2× 100 bp paired-end WES was performed on an Illumina HiSeq 2500 San Diego, CA, USA. We prepared reads for analysis with freeware: Cutadapt version 1.2.1 to remove the adapters and with the FASTx-Toolkit version 0.0.13.2 to trim the first 5 bp at the left end of reads. We aligned data to the UCSC Genome Browser hg19 human reference sequence using Burrows-Wheeler Aligner version 0.7.5a-r405 MEM (maximal exact match). Local realignment around insertions and deletions, recalibration of read base quality, and variant calling were conducted with freeware Genome Analysis Toolkit (GATK) version 2.8.1. GATK HaplotypeCaller was used for calling variants. Sanger sequencing was used to confirm WES variants.

2.3 |. CRISPR-Cas9 modification of mouse genome

We used CRISPR-Cas9, to generate loss-of-function mutations in the mouse Dcaf17 gene. We targeted second exon of the Dcaf17 gene, due to the presence of acceptor splicing mutation (c.127–1 G > C) in family A, predicted to cause mis-splicing and Dcaf17 loss of function. sgRNAS were designed by online CRISPR Design Tool (http://crispr.mit.edu). The selected sgRNA oligonucleotides were cloned into kanamycin resistant pENTR-pSpCas9(BB)-2A-GFP vector (Addgene plasmid ID: 48138). Sequence verified sgRNA plasmids were further used for cell transfection and microinjection. Neuro2A cell line (ATCC CCL-131) was used to validate the sgRNA function in vitro. Sequence-verified sgRNA ligated into pENTR-pSpCas9(BB)-2A-GFP was transfected into Neuro2A cell line using Lipofectamine 2000. DNA was extracted from transfected cells, and the targeted region was amplified by PCR using HotStart PCR Kit (KK2502, Kapa Biosystem) and region specific primers (Forward, 5′-3′ [TTTACGCGCTCTGATACCCC]; Reverse, 5′-3′ [CACAGCACCGGTAATTGTCAAA]). We used SURVEYOR Variation Detection Kit (Catalog No. 706020, Transgenomic) to detect region-specific Cas9 activity. SURVEYOR nuclease digestion products were visualized on a 2% (wt/vol) agarose gel (Figure S1, Supporting information). Mouse genome modification was achieved by injecting zygotes of C57BL/J6 mice with sgRNA ligated pENTR-pSpCas9(BB)-2A-GFP plasmid DNA. The cleaved high-quality blastocyst-stage embryos were transferred into the oviduct of pseudopregnant C57BL/J6 mice. Twenty seven F0 mice were initially genotyped by Sanger sequencing.

2.4 |. Histology and morphology

Testes/ovaries were dissected from wild-type (WT), heterozygous and homozygous Dcaf17 mutant mice at different postnatal ages up to 11-month-old adults. For histological examination, dissected tissues were fixed in Bouin (testes) or 10% neutral buffered formalin (ovary) and embedded in paraffin. Five micrometer sections were stained with hematoxylin and eosin (H&E) or Periodic acid-Schiff (PAS) for gross morphology as previously described.⁷ Epididymides were dissected from adult males, 12 to 14 weeks of age, minced with scissors, and incubated in media for 10 minutes at 37°C. The swim-up sperm were counted using a hemocytometer. Morphology of sperm was evaluated under light microscopy. Ovarian follicle counts and immunohistochemistry (IHC) with anti-NOBOX antibodies were performed as previously described.⁷

2.5 |. Statistics

All the values provided are the means standard deviation (SD) for at least 3 independent experiments. Statistical significance was determined by unpaired t-test or by one-way ANOVA where applicable. Differences were considered significant at P ≤ .05.

3 |. RESULTS

3.1 |. Family-based WES

Two families were referred for WES to determine genetic etiology of their hypergonadotropic hypogonadism. In family A, recruited from Turkey, 4 of the 5 sisters (Figure 1A, II-1, II-2, II-3, II-4) presented with primary amenorrhea (Figure 1A). Their parents were healthy and distant relations. The 4 affected daughters had normal karyotype (46, XX). Their basal estradiol levels were low, with elevated FSH and LH levels, consistent with hypergonadotropic hypogonadism (Table 1). Their pubic and axillary hair was Tanner stage 1, and alopecia was observed at the time of recruitment. On physical examination, they all presented with excessive hair loss, webbed neck, and widely spaced nipples. Ultrasonography and magnetic resonance imaging revealed ovarian agenesis and hypoplastic uterus in the 4 affected sisters. The two affected sisters, II-1, II-2 also presented with speech impairment and intellectual disability, however, this was not observed in II-3 and II-4. A second family, Family B (Figure 1B), was recruited at Magee-Women’s hospital in Pittsburgh, PA. The proband (III-4) presented with primary amenorrhea at the age of 16 years, with normal karyotype (46, XX), type II diabetes mellitus, anxiety, pervasive developmental delay (speech, fine motor skills), and alopecia. Her basal FSH and LH levels were within the normal range (Table 1). Physical examination revealed course facial features, wide spaced eyes, and down-slanting palpebral fissures. Ultrasound and MRI findings revealed pre-pubertal uterus with the absence of ovaries.

FIGURE 1.

Exome sequencing and DCAF17 pathogenic variants in Families A and B. (A, B) Pedigrees of Family A (A) and family B (B). Family members and generations are designated by Arabic numerals. Double horizontal line in (A) indicates consanguinity. Individuals who underwent whole exome sequencing are labeled with an asterisk. Squares denote male family members and circles female family members. Black circles represent affected individuals. Arrows denote affected individuals that underwent whole exome sequencing. Deceased individuals are slashed. (C, D) DNA sequence chromatograms obtained by Sanger sequencing to confirm exome sequencing variants, WT, wild-type variant; MT, pathogenic variant. (E, F) The location of the pathogenic variants (arrows) on the Dcaf17 gene schema is shown for Families A and B. The homozygous c.127–1G > C variant, predicted to disrupt conserved splice acceptor site AG to AC in Family A, is shown in (E). Compound pathogenic variants in Family B, c.C535T and c.G906A variants, in Exon 5 and Exon 9, respectively, are shown in (F).

TABLE 1.

Clinical and laboratory attributes of affected individuals with Woodhouse-Sakati syndrome

	Family A (II-3)	Family A (II-4)	Family B (III-4)
Age of diagnosis (y)	12	13	16
Ethnicity	Turkey	Turkey	Caucasian
Karyotype	46, XX	46, XX	46, XX
Ovary size (6.6 ± 0.19 cm3)	Not detected	Not detected	Not detected
Uterine size (75–200 cm3)	Hypoplastic	Hypoplastic	Pre-pubertal (2.99 × 0.85 cm)
Amenorrhea	Primary	Primary	Primary
Alopecia	Yes	Yes	Yes
Diabetes mellitus	No	No	Yes
BMI (18.5–24.9)	12.5	9.5	14.6 (>97%)
FSH (3–20 IU/L)	33.93	43.02	16.9
LH (<7 IU/L)	12.37	8.93	7
E2 (25–75 pg/mL)	11.8	11.8	<2
Prolactin (27 ng/mL)	3.75	5.39	<2
TSH (0.3–5 mIU/L)	2.96	2.75	-
Free T4 (0.7–1.9 ng/dL)	0.76	1.08	-
ACTH (10–50 pg/mL)	34.9	24.4	-
Insulin (<25 mU/L)	64.4	46.7	43
HbA1C (4%–5.6%)	5.8%	6.4%	10.2%
C-peptide (0.8–3.1 ng/mL)	6.18	4.34	4.5
Glucose (70–100 mg/dL)	130	129	-
Cortisole (5–23 μg/dL)	14.57	11.08	-
Creatinine (0.6–1.2 mg/dL)	0.7	0.7	-
IGF-1 (100–700 ng/mL)	294	158	-
Growth hormone (<10 ng/mL)	0.283	0.383	-

Abbreviations: ACTH, adrenocorticotropic hormone; BMI, body mass index; E2 estradiol; FSH, follicle-stimulating hormone; IGF-1, insulin-like growth factor 1.

Family A WES was carried out on 2 affected daughters (II-3, II-4) and parents. WES results indicated that parents were first cousins (daughter genomes with 6.5% homozygosity) and autosomal recessive inheritance was assumed. Out of nine homozygous recessive variants that segregated with 2 affected daughters, only the homozygous variant in DCAF17 (c.127–1G > C) was interpreted as pathogenic per ACMG guidelines⁸ and corroborated by Sanger sequencing (Figure 1C). The DCAF17 (c.127–1G > C) variant changes the highly conserved splice acceptor AG dinucleotide of exon 2 into AC (Figure 1E). The DCAF17 (c.127–1G > C) variant was not previously described in the literature.

In family B, WES was performed on the affected proband (III-4) alone, and 2 heterozygous pathogenic variants were identified in the DCAF17 gene (c.C535T; p.Gln179* and c.G906A; p.Trp302*). Both variants are predicted to introduce a stop codon and cause loss of function of DCAF17, and both were confirmed by Sanger sequencing (Figure 1D,F). Family B parental sequencing determined that parents were carriers for the two variants. The c.C535T (p.Gln179*) was novel, while c.G906A; p.Trp302* was previously reported as pathogenic.^3,5 Pathogenic variants in the DCAF17 gene associate with WSS. WSS is an autosomal recessive disorder that primarily presents with alopecia and hypogonadism in more than 90% of affected individuals, while less frequent associations include diabetes mellitus, intellectual disability, deafness and extrapyramidal disorders. The association between WSS and DCAF17 gene was previously established thru the study of consanguineous families,¹ and subsequent studies identified additional individuals throughout Middle East, Europe, as well as other ethnicities.^2,3,5,9 A total of 12 different pathogenic variants leading to protein truncations including 6 deletions, 3 nonsense and 3 splicing errors have been described in the past. No missense mutations have been associated with the WSS phenotype up to date. DCAF proteins have been implicated to function as a substrate receptor for CUL4-DDB1 E3 ubiquitin-protein ligase complex.¹⁰ DCAF proteins vary in structure and have diverse functions including transcription, DNA methylation, detection of UV-induced genomic lesions, cell cycle, cell death, and embryonic development. DCAF17 mechanism of action is hampered by a lack of animal model. We therefore generated a loss-of-function model of Dcaf17 in the mouse to assess its effects on hypogonadism, one of the most commonly encountered phenotypes in the WSS.

3.2 |. Dcaf17 loss-of-function mice

The 3 DCAF17 variants identified in our current study (c.127–1G > C, p.Gln179* and p.Trp302*), as well as in previous studies, are loss-of-function DCAF17 variants. We targeted mouse exon 2 to model the c.127–1G > C acceptor splice site loss-of-function variant in family A. We utilized a CRISPR/Cas9 system to generate insertions and/or deletions within exon 2 of Dcaf17 (Figure S1, Supporting information). We designed sgRNA targeting exon 2 and injected it into 1-cell stage zygotes of C57BJ6 mice. Ten out of 27 (37%) F0 mice showed deleterious variants in Dcaf17. The 3 deleterious variants in the Dcaf17 exon 2 were 2 deletions (−AG, chr2:71056553–71056554, and –CCTATGAG, chr2:71056547–71 056 554) and 1 insertion (+A, chr2:71056553), all predicted to cause loss of Dcaf17 function via frameshift introducing premature stop codon. F0 animals were mosaic and were crossed with WT animals to eliminate mosaicism. We initially analyzed all 3 different loss-of-function variants in F1 mutant mice, and all displayed the same phenotype. The bulk of the data were obtained from mice homozygous for the 8 base pair deletion (−8 bp), and these mice are referred to as Dcaf17-KO mice, while mice heterozygous for the deletions are referred to as Dcaf17-Het.

3.3 |. Dcaf17-KO male mice are infertile with defective spermatogenesis and spermiogenesis

Dcaf17-KO males were initially crossed with WT females to determine infertility defects, if any. No liveborn pups were observed after 6 months of mating Dcaf17-KO males with WT females (Figure 2E). Dcaf17-Het animals were fertile and did not show statistically significant difference from WT in measures such as pups/litter, litter/month, and pups/month. The testes weights of the sterile Dcaf17-KO homozygous (2.8 ± 0.24 mg) were significantly less than WT (3.9 ± 0.65 g) mice (Figure 2C), consistent with a diagnosis of hypogonadism in knockouts. The total number of sperm isolated from epididymis of Dcaf17-KO (0.005 ± 0.0070×10⁶) mice were significantly less from both Dcaf17-Het (11.925 ± 7.48×10⁶) and WT mice (25.25 ± 4.59×10⁶) (Figure 2D). We performed PAS stains on testis sections to assess spermatogenesis (Figure 2A). The histology showed numerous vacuoles and degeneration of a subset of seminiferous tubules in Dcaf17-KO males. Defects in spermatogenesis were observed at the round elongate spermatids (stages VI-VII) and zygotene spermatocyte (stages XI-XII)¹¹). The few sperm that were retrieved from the epididymis showed abnormal morphology characterized by defective head, middle piece, and abnormal acrosome hook structure(Figure 2B).¹²

FIGURE 2.

Dcaf17 loss-of-function transgenic males are infertile. (A) Testes from wild type (WT) and Dcaf17 loss-of-function 14-week-old mice (Dcaf17-KO), were fixed in Bouin, sectioned and stained with Periodic acid-Schiff. Dcaf17-KO testes show diffuse pathology, with block at round spermatids, stages VI-VII: elongate spermatids (stage IX); zygotene spermatocytes (stages XI-XII); as well as degeneratin seminiferous tubules. (B) Morphology comparison of epididymal sperm from WT and Dcaf17-KO mice. Epididymal sperm numbers were low in Dcaf17-KO testes. Defective head and middle pieces were universal in sperm recovered from Dcaf17-KO epididymides (black arrowheads). (C) Testes/body weight ratios were calculated for WT, mice heterozygous for the Dcaf17 knockout (Dcaf17-het) and Dcaf17-KO mice. Dcaf17-KO as well as Dcaf17-het testes were significantly smaller than WT mice. Significance of P < .05 is indicated by asterisk. (D) Epididymal sperm recovered from WT, Dcaf17-het and Dcaf17-KO was counted and compared among different genotypes. Epididymal sperm counts in Dcaf17-KO mice were significantly less when compared to Dcaf17-het and wild-type mice. Significance of P < .01 is indicated by asterisk. (E) Fertility analysis of Dcaf17-KO, Dcaf17-Het and WT male mice for a 6-month duration. Dcaf17-KO male mice were infertile. No significant difference in pups per litter, pups per month, and litters per month were calculated between WT and Dcaf17-Het mice.

3.4 |. Dcaf17-KO female mice are subfertile and have abnormal folliculogenesis

Dcaf17-KO female mice crossed with WT males produced significantly less pups/l (1.8 ± 0.4) and pups/month (1.6 ± 0.6) when compared to WT females (pups/l: 7 ± 1.14, and pups/month: 5.3 ± 1.3). Females heterozygous for the Dcaf17 deletion had normal fertility (data not shown). The total weight of the Dcaf17-KO female mice was significantly lower as compared to WT mice (data not shown). Further, the follicle count at 6 months revealed that the overall number of follicles at each stage: primordial, primary, secondary and antral were reduced at 6 months in the Dcaf17-KO mice (Figure 3B). Histologically, as the size of ovaries shrank in older Dcaf17-KO mice, the corpora lutea were significantly more prominent (Figure 3D).

FIGURE 3.

Dcaf17 deficiency causes female subfertility. (A) Dcaf17-KO and wild-type (WT) mice fertility with stud males was evaluated over a 1-year period. Dcaf17-KO matings produced significantly less pups per litter and pups per month when compared with WT females. (B) The number of ovarian follicles at different stages from 6-month-old ovaries was compared between Dcaf17-KO and WT mice. The follicle pool was significantly lower in Dcaf17-KO mice as compared to the WT. (C) WT and Dcaf17-KO 3-week-old mice serum estradiol levels were measured before and after PMSG treatment. The serum level of estradiol was significantly lower in Dcaf17-KO female mice with and without PMSG treatment when compared to the WT. Significance at P < .05 was represented by an asterisk.

PMSG induces estradiol synthesis in granulosa cells and promotes maturation of small follicles into larger, pre-ovulatory follicles.¹³ We tested the effect of PMSG on WT and Dcaf17-KO ovaries. Prior to PMSG administration, the serum estradiol levels were significantly lower in Dcaf17-KO mice as compared to the WT (Figure 3C). Upon the injection of PMSG, there was a sharp increase of estradiol in WT mice as expected, however, the estradiol response to PMSG was blunted Dcaf17-KO mice (Figure 3C).

4 |. DISCUSSION

Previous studies on consanguineous families associated DCAF17 pathogenic variants with the WSS.¹ WSS is a rare syndrome which primarily presents with alopecia and hypogonadism. In the current study, we identified additional two families, one Turkish and 1 American with WSS. The consanguineous Turkish family presented with a novel loss-of-function variant inherited in homozygous state (c.127–1G > C), while the American family presented as compound heterozygous, with one novel (c.C535T; p.Gln179*), and one known (c.G906A; p.Trp302*) variant in Dcaf17 gene in women with primary amenorrhea. It is interesting to note that all the WSS associated variants in the Dcaf17 gene, including the two novel variants identified in the current study (c.127–1G > C, c.C535T; p.Gln179*), are predicted to cause loss of Dcaf17 function.⁴ Moreover, all the 3 affected females in the current study presented with the clinical phenotypes that are common in patients with WSS, such as alopecia, hypogonadism and diabetes mellitus. The intra-familial phenotypic variation in expressivity was observed among the 4 affected siblings in family A, with 2 siblings affected with speech impairment and intellectual disability, while the other two were not. Variable expressivity has been previously reported in WSS patients, and other genetic modifiers involved in the pathogenesis of WSS likely exist.¹⁴

DCAF17 is highly conserved and shares 91% identity with the mouse Dcaf17. We are not aware of a previous animal model for Dcaf17. We used CRISPR/Cas9 and Dcaf17-specific guide RNA to create loss-of-function Dcaf17 mouse model. As expected, we generated multiple different variants predicted to cause loss of function, and we further characterized a mouse model with 8 nucleotide deletion in exon 2. The mouse model mimicks the human gonadal disease. Dcaf17-KO males are infertile and hypogonadal with diffuse abnormalities in the seminiferous tubules and abnormal spermiogenesis. The mouse results support the crucial role of Dcaf17 in mammalian spermatogenesis and male reproduction, as also observed in males affected with WSS. The mechanism of Dcaf17 action in male testes is unclear and remains to be further elucidated.

Female phenotype in mice was similar to the human phenotype. Mating studies showed that Dcaf17-KO females are sub-fertile, with significantly reduced number of pups per litter and pups per month. We further investigated ovarian histomorphology in Dcaf17-KO mice. At 6 months, there is a significant depletion of various follicle types, including primordial follicles. These results are consistent with the interpretation that Dcaf17 deficiency causes follicle depletion in adulthood. The number of pre-ovulatory follicles and corpora lutea are not significantly decreased, possibly due to the sufficient number of small follicles available for recruitment. It is therefore possible that subfertility is due to abnormal ovulation, oocyte quality and/or implantation. Dcaf17 is expressed in both somatic and germline component of the mouse ovary and whether both contribute to the observed phenotype is unknown. The role of pituitary in the genesis of hypogonadism is unclear, as some WSS patients present with hypergonadotropic hypogonadism (suggestive of primarily ovarian defect), while others have relatively normal FSH and LH levels.¹⁴ Experiments with PMSG are consistent with the interpretation that Dcaf17 deficiency causes intrinsic ovarian defects. PMSG induces estradiol synthesis in granulosa cells and promotes maturation of small follicles into larger, pre-ovulatory follicles.¹³ Upon the injection of PMSG, there was a sharp increase of estradiol in WT mice as expected, however, the estradiol response to PMSG was blunted Dcaf17-KO mice. The blunted response to PMSG is consistent with intrinsic defect in ovarian function. Dcaf17 mouse uteri were not hypoplastic like human uteri, likely due to persistent estradiol production in the mouse knockouts and stimulatory effect on the mouse uterus.

Unlike women, where hypogonadism and infertility are universal among individuals with WSS, mice were subfertile. The difference between subfertility and total infertility could be due to the basic difference between human and mouse reproductive systems. Mice have large litter sizes as compared to humans, and subfertility in mice may translate into total infertility in humans. Alternatively, Dcaf17, despite its conservation, may have different functions in the mouse ovaries as compared to the human.

Mechanisms behind Dcaf17 actions in the humans are unknown. Dcaf17 is expressed in many tissues, including gonads, brain, liver, skin, embryo and localizes to nucleoli.² Dcaf family of genes regulates cell cycle, apoptosis, DNA methylation and cellular aging.¹⁰ At least 18 different Dcaf genes exist.² Fourteen out of 18 Dcaf genes encode a conserved WD40 domain required for the interaction of DCAF proteins with Ddb1-Cul4-associated ubiquitin ligase complex.^10,15 Dcaf genes have been implicated in reproduction, and Dcaf1 is an example (16,17). Oocyte-specific deletion of Dcaf1 results in rapid oocyte loss and infertility, presumably by regulating the hydroxymethylation of genomic DNA as well as oocyte maturation.^16,17 Dcaf17 gene does not encode the WD40 motif, just like Dcaf15 and Dcaf16). Dcaf15 was recently implicated in regulation of pre-mRNA splicing by degradation of the RBM39 protein.¹⁸ Further studies are needed to elucidate mechanisms of Dcaf17 action in male and female gonadal development.

Supplementary Material

Additional Supporting Information may be found online in the supporting information tab for this article.