Article Text

Download PDFPDF
Original research
Haploinsufficiency in non-homologous end joining factor 1 induces ovarian dysfunction in humans and mice
  1. Guoqing Li1,
  2. Xi Yang1,
  3. Lingbo Wang1,2,
  4. Yuncheng Pan1,
  5. Siyuan Chen1,
  6. Lingyue Shang1,
  7. Yicheng Zhang1,
  8. Yucheng Wu1,
  9. Zixue Zhou1,
  10. Qing Chen1,2,
  11. Xue Zhang1,
  12. Ling Zhang1,
  13. Yingchen Wang1,
  14. Jinsong Li3,
  15. Li Jin1,
  16. Yanhua Wu1,2,4,
  17. Xiaojin Zhang1,2,
  18. Feng Zhang1,2
  1. 1 Obstetrics and Gynecology Hospital, NHC Key Laboratory of Reproduction Regulation (Shanghai Institute of Planned Parenthood Research), State Key Laboratory of Genetic Engineering at School of Life Sciences, Institute of Metabolism and Integrative Biology, Fudan University, Shanghai, China
  2. 2 Shanghai Key Laboratory of Female Reproductive Endocrine Related Diseases, Shanghai, China
  3. 3 State Key Laboratory of Cell Biology, Shanghai Key Laboratory of Molecular Andrology, CAS Center for Excellence in Molecular Cell Science, Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai, China
  4. 4 National Demonstration Center for Experimental Biology Education, School of Life Sciences, Fudan University, Shanghai, China
  1. Correspondence to Professor Feng Zhang, Obstetrics and Gynecology Hospital, NHC Key Laboratory of Reproduction Regulation (Shanghai Institute of Planned Parenthood Research), State Key Laboratory of Genetic Engineering at School of Life Sciences, Institute of Metabolism and Integrative Biology, Fudan University, Shanghai, China; zhangfeng{at}; Professor Xiaojin Zhang; zxjgcp{at}; Professor Yanhua Wu; yanhuawu{at}


Background Premature ovarian insufficiency (POI) is a common disease in women that leads to a reduced reproductive lifespan. The aetiology of POI is genetically heterogeneous, with certain double-strand break (DSB) repair genes being implicated in POI. Although non-homologous end joining (NHEJ) is an efficient DSB repair pathway, the functional relationship between this pathway and POI remains unknown.

Methods and results We conducted whole-exome sequencing in a Chinese family and identified a rare heterozygous loss-of-function variant in non-homologous end joining factor 1 (NHEJ1): c.532C>T (p.R178*), which co-segregated with POI and irregular menstruation. The amount of NHEJ1 protein in the proband was half of the normal level, indicating a link between NHEJ1 haploinsufficiency and POI. Furthermore, another rare heterozygous NHEJ1 variant c.500A>G (p.Y167C) was identified in one of 100 sporadic POI cases. Both variants were predicted to be deleterious by multiple in silico tools. In vitro assays showed that knock-down of NHEJ1 in human KGN ovarian cells impaired DNA repair capacity. We also generated a knock-in mouse model with a heterozygous Nhej1 variant equivalent to NHEJ1 p.R178* in familial patients. Compared with wild-type mice, heterozygous Nhej1-mutated female mice required a longer time to first birth, and displayed reduced numbers of primordial and growing follicles. Moreover, these mice exhibited higher sensitivity to DSB-inducing drugs. All these phenotypes are analogous to the progressive loss of ovarian function observed in POI.

Conclusions Our observations in both humans and mice suggest that NHEJ1 haploinsufficiency is associated with non-syndromic POI, providing novel insights into genetic counselling and clinical prevention of POI.

  • DNA repair
  • genetic variation
  • gynecology
  • high-throughput nucleotide sequencing
  • reproductive medicine

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Premature ovarian insufficiency (POI), referring to a continuum of declining ovarian function in young women, is an important cause of menstrual disturbance and reduced female fertility. The diagnostic criteria of POI include oligomenorrhea/amenorrhea for at least 4 months, and follicle-stimulating hormone (FSH) is ≥25 IU/L on two occasions >4 weeks apart before 40 years of age.1 POI affects approximately 1% of women.2 Known aetiologies of POI include chromosomal, genetic, autoimmune, metabolic, infectious and iatrogenic causes,3 but the majority of patients with POI do not receive any definitive aetiological diagnosis.4 Approximately 10%‒15% of patients with POI have first-degree relatives also affected with the same disease,5 suggesting an important role of genetic aetiologies in POI.

Different approaches have been used to identify susceptible loci or genes involved in POI, including linkage analysis in familial POI with multiple affected members, genome-wide association study in a large population and whole-exome sequencing (WES) in familial and sporadic POI. To date, over 70 causative genes of POI have been identified,6 7 with some of them known to be enriched in DNA damage response (DDR), such as MCM8/9 and BRCA2 in homologous recombination (HR) during meiosis,8–15 MSH4/5 in double-strand break (DSB) repair,16 17 FANCA/L/M/U in interstrand crosslink repair18–25 and ERCC6 in nucleotide excision repair.26

DSBs are the most dangerous type of DNA damage, and could be induced by ionising radiation, radiomimetic chemicals, reactive oxygen species, DNA replication errors or inadvertent cleavage by nuclear enzymes.27 28 Cells have different and independent pathways to repair DSBs. HR faithfully repairs DSBs using an intact DNA molecule as the template, but this is restricted to the late S and G2 phases of the cell cycle.29 Unlike HR, the non-homologous end joining (NHEJ) mechanism does not require a homologous template and can directly ligate the broken DNA ends, allowing it to function in all phases of the cell cycle. Therefore, NHEJ is considered to be the major DSB repair pathway in mammalian cells.30 The basic steps of the NHEJ pathway include: (i) recognition and binding of Ku proteins (heterodimers of Ku70 and Ku80) to the DSB ends for protection28; (ii) recruitment and stabilisation of the NHEJ complex, consisting of DNA-PKcs, Artemis, XRCC4, DNA ligase IV (LIG4) and non-homologous end joining factor 1 (NHEJ1) to the DSB ends31; (iii) processing of the DSB ends by endonucleases32; (iv) ligation of the DSB ends by the XRCC4/LIG4 complex. NHEJ1, also called the XRCC4-like factor or Cernunnos, was originally identified in 2006 and demonstrated to participate in the DSB end joining step by enhancing the efficiency of the XRCC4/LIG4 complex.33–35

In this study, a Chinese family with POI and irregular menstruation across three generations was subjected to WES analysis, and an extremely rare heterozygous nonsense variant of NHEJ1 was identified to be co-segregating with POI and related phenotypes in this family. Protein expression analysis of peripheral blood cells of the proband revealed an approximately 50% reduction in NHEJ1 expression. We also generated a knock-in mouse model with a heterozygous nonsense mutation in Nhej1 (Nhej1+/‒ ) for in vivo functional assays. The ovaries of the Nhej1+/‒ female mice were found to be sensitive to DNA damage stimulus, exhibiting impaired DNA repair functions and reduced ovarian functions. Our experimental observations in both humans and mice suggest that heterozygous NHEJ1 loss-of-function (LoF) variants and NHEJ1 haploinsufficiency are associated with POI through increased sensitivity of ovaries to DNA damage.


Study participants

Familial and sporadic patients with POI were diagnosed and recruited at the Obstetrics and Gynecology Hospital of Fudan University (Shanghai, China). Inclusion criteria included oligomenorrhea/amenorrhea for at least 4 months before the age of 40 and abnormal serum FSH levels (≥25 IU/L) in two measurements >4 weeks apart. Women who underwent ovarian surgery, radiotherapy or chemotherapy intervention were excluded. The karyotypes of all patients were normal (46,XX). Genomic DNA was extracted from peripheral blood samples using a QIAamp DNA Blood Mini Kit (Qiagen). The CGG repeats in the 5'-untranslated region of FMR1 were investigated in Suzhou PerkinElmer Medical Laboratory (Suzhou, China).36

WES and data processing

Genomic DNA (1.5 μg) was used to prepare a captured library using a SureSelectXT Human All Exon V6 kit (Agilent Technologies). Next-generation sequencing of familial cases and sporadic cases was performed at Shanghai Genesky Bio-Tech (Shanghai, China) and iGeneTech Bioscience (Beijing, China), respectively. The processing of WES and data analysis were carried out, as previously described.37 Briefly, raw data of approximately 10 Gb per exome were mapped to a human reference genome sequence (GRCh37/hg19) using the Burrows-Wheeler Alignment tool. Variant calling was performed using the Genome Analysis Toolkit. All variants were further annotated using the ANNOVAR software. Minor allele frequencies were estimated according to the 1000 Genomes (1KG) project, ExAC or gnomAD databases. Variant pathogenicity was assessed by the SIFT, PolyPhen-2 and MutationTaster tools. OMIM, mouse model studies, gene ontology terms and gene expression patterns were used to analyse the potential function of the identified genes in ovaries.

Sanger sequencing

Specific primers were designed for both flanks of NHEJ1 or Nhej1 variation sites and synthesised by Generay Biotech Company (Shanghai, China). PCR products were sequenced by Majorbio Bio-pharm Technology Company (Shanghai, China). Primers for amplification and Sanger sequencing are shown in online supplemental table 1.

Supplemental material

Q-PCR and relative quantification of NHEJ1 expression

Total RNA was extracted from cultured cells or ovaries using Invitrogen TRIzol reagent (Thermo Fisher Scientific), and 1000 ng of RNA per sample was used for reverse transcription using the PrimeScript RT reagent Kit (Takara). Q-PCR analysis was performed using the TB Green Master Mix Kit (Takara) with the CFX Connect Real time PCR Detection system (Bio-Rad Laboratories). Reaction mixtures with no cDNA template were tested as negative controls to rule out the possibility of contamination. The specificity of the PCR products was confirmed via melting curve analysis. The expression levels of NHEJ1/Nhej1 were normalised to those of GAPDH/Gapdh, and the relative gene expression levels were calculated by the 2−ΔΔCt method as previously described.38 Primers for Q-PCR are shown in the online supplemental table 1.

Western blot analysis

Protein samples were collected in sodium dodecyl sulfate lysis buffer and boiled in water for 10 min. Protein samples (15‒30 μg) were separated via SDS-PAGE on a 10% or 12% gel and then transferred onto polyvinylidene fluoride membranes (Pall). After blocking in 5% fat-free milk for 1 hour, these membranes were incubated with specific primary antibodies at 4°C overnight. Afterwards, the membranes were washed and incubated with secondary antibodies at room temperature for 1 hour. Membranes were visualised using an enhanced chemiluminescence kit (GE Healthcare Life Science). The images acquired were representatives of several independent experiments with consistent results, and the densitometric values were quantified with GeneTools from Syngene software (Frederick). β-Actin was used as a loading control. The antibodies used included anti-NHEJ1 (cat. no. ab189917, Abcam; cat. no. 11 888-1-AP, Proteintech), anti-Nhej1 (cat. no. A300-730A, Bethyl Laboratories), anti-γH2AX (cat. no. ab26350, Abcam), HRP-labelled anti-β-actin (cat. no. HRP-60008, Proteintech), HRP-labelled goat antimouse IgG (cat. no. IH-0031, DingGuo Changsheng Biotech) and HRP-labelled goat antirabbit IgG (cat. no. IH-0011, DingGuo Changsheng Biotech).

Cell culture and RNA interference

The KGN cell line was cultured in Dulbecco’s Modified Eagle Medium (Gibco) with 10% fetal bovine serum (Gibco) in a 37°C incubator with 5% CO2. Small interfering RNAs (siRNAs) were purchased from GenePharma (Shanghai, China), and the corresponding sequences are listed in the online supplemental table 2. KGN cells were transfected with siNHEJ1 or a negative siRNA control using Lipofectamine 3000 (Invitrogen) 48 hours before experiments. Cells were then treated with phleomycin (Sangon Biotech) for the indicated time.

Animal use and care

All animal experiments were performed adhering to the ethical guidelines of the Animal Care and Use Committees at Fudan University and Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences.

Mouse model construction and breeding assays

The Nhej1-mutated mouse model was generated by direct injection of CRISPR-Cas9 reagents into zygotes. Briefly, B6D2F1 (C57BL/6×DBA/2) female mice aged 8 weeks were superovulated and mated with C57BL/6 male mice to collect zygotes. Cas9 and sgRNA were prepared as previously described.39 After Cas9/sgRNA injection, the zygotes were further cultured in EmbryoMax KSOM Mouse Embryo Media (Merck Millipore) at 37°C under 5% CO2 until the 2-cell stage, followed by embryo transfer into oviducts of pseudopregnant ICR females at 0.5-day postcoitum as previously described.39 The founders were genotyped and mated with C57BL/6 for several generations before analysis. Primers used for mouse mutant construction and genotyping are detailed in the online supplemental table 1. In the breeding assays, each adult C57BL/6 wild-type male was mated with one Nhej1-mutated female or wild-type female aged 8 weeks. The ages of the first litter and litter sizes were recorded.

H&E staining and follicle counting

Ovaries were fixed in 4% paraformaldehyde overnight at room temperature, dehydrated and embedded in paraffin before being sectioned at a thickness of 5 μm. Every fifth section was stained with H&E for histological analysis. Follicles containing an oocyte with a clearly visible nucleus and normal morphology were counted and classified based on the system by Pedersen.40 The final count included all H&E-stained sections through the entire ovary. Each group had three mice, and only one ovary from each mouse was counted.


Female mice aged 3 months were injected with 5 IU pregnant mare serum gonadotropin (Sansheng Pharmaceutical), followed by 5 IU human chorionic gonadotropin (Sansheng Pharmaceutical) after 48 hours as previously described.39 Superovulated oocytes were then collected from the ampulla of the oviduct. Bleomycin (BLM, Universal Biotech) was injected 24 hours after the pregnant mare serum gonadotropin treatment.

Statistical analysis

Values are expressed as the mean±SD. Comparisons of quantitative data were performed using Student’s t-test. P value <0.05 was considered to be significantly different.


Co-segregation of a heterozygous LoF variant of NHEJ1 with POI and irregular menstruation in a Chinese family

The proband (family 1, II-3; figure 1A) from a Chinese family had menarche at the age of 15, and presented irregular menstruation at the age of 25. She was administered with femoston for maintenance of menstruation, and was diagnosed with POI at the age of 31 (table 1). The affected mother (I-2; figure 1A) had a history of irregular menstruation, and was menopausal at the age of 37.

Figure 1

Identification of a rare non-homologous end joining factor 1 (NHEJ1) nonsense variant in a Chinese family with premature ovarian insufficiency (POI). (A) A heterozygous nonsense variant (M1) c.532C>T (p.R178*) of NHEJ1 was identified in a Chinese family with POI. The proband is indicated by a black arrow. Black solid circles indicate cases affected with POI, and the grey solid circle indicates a case with incipient symptoms of POI. (B) Sanger sequencing confirmed the heterozygous NHEJ1 variant identified using WES in patients. The variant position is indicated by a red arrow. (C) Schematic representations of human NHEJ1 (NM_024782.3) and NHEJ1 protein (NP_079058.1). The variant positions are indicated by red arrows. The gene coding region is in grey. The head and stalk domain of the NHEJ1 protein comprises amino acids 1–135 and 136–233, respectively. (D) The amino acid residue of NHEJ1 affected by the identified mutation is conserved across species.

Table 1

Clinical characteristics of the subjects with POI

To identify the potential pathogenic variant in this POI family, WES analysis was performed in both affected subjects (I-2 and II-3). The workflow for genetic analysis is shown in online supplemental figure 1, and the variant filtering steps employed are summarised in online supplemental table 3. Briefly, only the variants shared between subjects I-2 and II-3 were considered during the analysis. Variants with minor allele frequencies ≥0.001 according to the 1KG Project or ExAC databases were excluded, and only those in coding regions or splicing sites were preserved. Next, missense variants predicted to be deleterious by in silico tools and LoF variants were chosen as candidate pathogenic variants (detailed in online supplemental table 4). As there were no rare variants located in any known POI causative/candidate genes,6 7 variants in genes that are functionally related to ovary development and function were preferred for the analysis. This led to the identification of a rare nonsense variant of NHEJ1 (c.532C>T, p.R178*; NM_024782.3) shared between subjects I-2 and II-3.

This variant was further confirmed via Sanger sequencing (figure 1B). The proband’s daughter (III-1; figure 1A) is only 13 years old, but she has already exhibited irregular menstruation. Sanger sequencing confirmed that she carried the same heterozygous NHEJ1 LoF variant as her mother.

The NHEJ1 c.532C>T variant was absent in the 1KG Project or ExAC databases, and the gnomAD database showed this variant to be extremely rare (8.1×10−6) in human populations (table 2). The corresponding amino acid residue R178 was located in the stalk domain of NHEJ1 and was found to be conserved across species (figure 1C,D). The NHEJ1 c.532C>T variant was predicted to be deleterious by MutationTaster, CADD and DANN bioinformatics tools (table 2).41 Notably, this NHEJ1 variant could be classified as pathogenic based on the American College of Medical Genetics and Genomics guidelines (table 2).42

Table 2

Heterozygous NHEJ1 variants identified in subjects with POI

Identification of an additional rare deleterious NHEJ1 variant in sporadic POI

To investigate the genetic contribution of NHEJ1 to Chinese patients with POI, genetic analysis based on WES was further performed in 100 sporadic POI cases. We identified another rare deleterious NHEJ1 variant (c.500A>G, p.Y167C) in one patient F077 (online supplemental table 5). The variant was also confirmed via Sanger sequencing (online supplemental figure 2A). This patient had menarche at the age of 15, and was diagnosed with POI at the age of 33 (table 1).

The c.500A>G variant is located in the exon 4 of NHEJ1, and the corresponding amino acid residue p.Y167 was also located in the stalk domain of NHEJ1 (online supplemental figure 2B,C). The ExAC and gnomAD databases showed the NHEJ1 c.500A>G variant to be extremely rare (8.24×10−6 and 1.22×10−5, respectively) in human populations (table 2). This variant also showed a high degree of evolutionary conservation according to PhastCons and phyloP scores, and was predicted to be deleterious by all five bioinformatics tools used, including SIFT, PolyPhen-2, MutationTaster, CADD and DANN (table 2). These findings collectively provide further evidence for the involvement of rare NHEJ1 variants in human POI pathogenesis.

Haploinsufficiency of NHEJ1 decreased cell viability and DDR function in vitro

To investigate the biological effect of the heterozygous LoF NHEJ1 variant in familial POI, we first conducted protein expression analysis using peripheral blood samples from the proband of family 1. Western blot analysis revealed that the amount of NHEJ1 protein in the proband was nearly half of that in normal individuals (figure 2A). Additionally, no truncated NHEJ1 protein was detected.

Figure 2

Reduced non-homologous end joining factor 1 (NHEJ1) expression levels decreased cell viability and DNA repair function. (A) Western blot analysis of NHEJ1 protein levels in the peripheral blood of a wild-type (WT/WT) control and the heterozygous NHEJ1-mutated (WT/M1) proband in family 1. β-Actin was used as a loading control. Representative figures are shown in the left panel and densitometric values expressed as the mean±SD (n=3; ***p<0.001) are on the right panel. (B, C) Q-PCR analysis of NHEJ1 mRNA levels (B) and western blot analysis of NHEJ1 protein levels (C) in KGN cells transfected with the indicated small interfering RNAs (siRNAs). Values are expressed as the mean±SD (n=3; ***p<0.001). (D) Survival rate of KGN cells after 20 μg/mL phleomycin (Phleo) treatment for the indicated time period. Values are expressed as the mean±SD (n=4; ***p<0.001). (E) Western blot analysis of γH2AX protein levels in KGN cells. Cells were treated with 50 μg/mL Phleo for 1 hour before cell lysates were collected at the indicated time points.

It has been previously reported that downregulation of NHEJ1 expression could sensitise human somatic cells (293T, HeLa and so on) to DNA damage, leading to reduced cell viability.43 44 To further investigate whether reduction in NHEJ1 levels could affect DNA repair efficiency and cell viability in ovarian cells, NHEJ1 expression was knocked down using a siRNA against NHEJ1 (siNHEJ1) in the human ovarian granulosa-like tumour cell line KGN. After endogenous NHEJ1 expression in the siNHEJ1 group was successfully reduced relative to the control group (figure 2B,C), KGN cells in the siNHEJ1 group were found to be more sensitive (than control cells) to DSB damage induced by the DNA-cleaving agent phleomycin. As shown in figure 2D, the siNHEJ1 group exhibited decreased survival ratios in a 3-day culture period under phleomycin treatment. Consistent with these results, the amount of γH2AX in the siNHEJ1 group was substantially higher than that of the control group under drug treatment (figure 2E), suggesting the dysfunction of DSB repairs on NHEJ1 knock-down in KGN cells. These findings demonstrated that haploinsufficiency of NHEJ1 increased cell sensitivity to DSB damage and reduced cell viability in KGN cells.

Next, expression analysis and functional study of the NHEJ1 variant c.500A>G were carried out. There was no obvious difference at the protein level in the peripheral blood sample between the sporadic case F077 and a normal (control) individual (online supplemental figure 3A). To confirm this result, we constructed recombinant wild-type (Myc-NHEJ1-WT) and mutant (Myc-NHEJ1-Y167C) NHEJ1 plasmids. Western blot analysis revealed that the amount of Y167C-altered protein was consistently similar to that in the wild-type (online supplemental figure 3B). To further investigate whether this missense variant affects the biological function of NHEJ1 in DNA repair, the amount of γH2AX in KGN cells transfected with wild-type or mutant NHEJ1 was estimated. In contrast to the wild-type group, a higher γH2AX level was observed in cells overexpressing NHEJ1-Y167C after DNA damage stimulus (online supplemental figure 3C), indicating that the Y167C variant identified in the sporadic case F077 also displayed compromised DNA repair capacity.

Haploinsufficiency of Nhej1 leads to impaired fertility in female mice

To investigate the in vivo effects of NHEJ1 haploinsufficiency on ovarian function, we constructed a heterozygous Nhej1-mutated (Nhej1+/− ) mouse model using CRISPR/Cas9. This mutant mouse harboured a heterozygous nonsense Nhej1 mutation (c.532C>T, p.R178*; NM_029342.4) (figure 3A) equivalent to the human NHEJ1 variant in POI family 1. Relative quantification of Nhej1 mRNA expression revealed that the mRNA levels of Nhej1 in the ovaries of Nhej1+/− mice were approximately half of those in wild-type mice (figure 3B). Subsequent cDNA sequencing further demonstrated that the expression level of the mutant allele was much lower than that of the wild-type allele in Nhej1+/− (online supplemental figure 4), indicating that the mutant mRNA might undergo nonsense-mediated decay. Western blot analysis further revealed reduced protein level of Nhej1 in mutant mice (figure 3C), along with an absence of truncated isoform. These experimental observations in Nhej1+/− mice are consistent with the alteration in protein expression observed in the patient with POI with heterozygous LoF NHEJ1 variant studied here.

Figure 3

Ovaries from non-homologous end joining factor 1 (Nhej1) +/− exhibited reduced number of primordial and growing follicles. (A) Sanger sequencing confirmed the heterozygous nonsense mutation (Nhej1 +/ ) in the knock-in mouse model. The red arrow indicates the variant position. (B, C) Q-PCR analysis of Nhej1 mRNA levels (B) and western blot analysis of Nhej1 protein levels (C) in mouse ovaries. Values were expressed as the mean±SD (n=3; **p<0.01; ***p<0.001). (D) Representative H&E-stained images of ovary from wild-type and mutant mice aged 1 month and 6 months, respectively. Scale bars represent 100 μm. Primordial follicles are indicated by yellow arrows. (E) Numbers of primordial, primary, secondary and antral follicles were counted and expressed as the mean±SD (n=3; *p<0.05).

Breeding assays were performed by mating a wild-type C57BL/6 male mouse with a Nhej1+/− female mouse and a wild-type female mouse since 8-week-old from the same cage. As shown in table 3, the time interval between mating and first birth for the Nhej1+/− mice was significantly longer than that of wild-type female controls, highlighting the decreased fertility of Nhej1+/− mice. However, no obvious difference was found in the number of first litter sizes.

Table 3

Fertility in Nhej1+/− mice

To investigate whether Nhej1+/− mice exhibited ovarian dysfunction, an anatomical analysis of the female reproductive system was performed. There was no apparent difference in the size of ovaries between mutant and wild-type mice. H&E staining revealed all types of growing follicles in Nhej1+/− mice (figure 3D). However, follicle counting revealed reduced average numbers of both primordial and growing follicles in mutant mice when compared with those in wild-type ovaries (figure 3E). Particularly, the number of primordial follicles in Nhej1+/− exhibited a rapidly decreasing pattern with ageing, with approximately 50% primordial follicles remaining at 6 months of age compared with that in wild-type mice (p<0.05). These histological results suggest that haploinsufficiency of Nhej1 might affect the ovarian reserve and subsequent follicular development in mice, which resembles the POI phenotype in humans. To further investigate whether Nhej1 haploinsufficiency could modulate the sensitivity of ovary to DNA damage, we carried out a superovulation experiment in the presence or absence of a DNA DSB damage stimulus. The number of superovulated oocytes was similar in Nhej1+/− and wild-type female mice under normal conditions. However, a significantly decreased number of superovulated oocytes in Nhej1+/− female mice, compared with that in the wild-type controls, was observed after treatment with the chemotherapeutic drug BLM (figure 4A,B). Lysates of ovaries were further prepared at the indicated time after BLM injection in ovaries and subjected to DNA damage analysis. As shown in figure 4C,D, the amount of γH2AX decreased rapidly at 12 hours post-BLM injection in wild-type females, whereas it remained at a constant high level for a much longer time in Nhej1+/− mice. Taken together, these results strongly suggest that DSB repair capacity was impaired in ovaries from the Nhej1+/− mice, which could potentially contribute to decreased female fertility on external DNA damage stimulus.

Figure 4

Ovaries from non-homologous end joining factor 1 (Nhej1) +/− mice exhibited increased sensitivity to DNA damage stimulus. (A, B) Representative images (A) and numbers (B) of superovulated oocytes collected from Nhej1+/+ and Nhej1+/− mice with or without bleomycin (BLM) treatment. Representative images of oocytes are shown at a higher magnification. Scale bars represent 200 μm. Values are expressed as the mean±SD (n=3; ns, not significant; *p<0.05). (C) Western blot analysis of γH2AX level in lysates of mouse ovaries. Ovaries were collected at the indicated time points after BLM injection. β-Actin was used as a loading control. (D) Statistically significant changes in γH2AX levels in ovaries, expressed as the mean±SD (n=3; *p<0.05; **p<0.01).


In this study, we reported a heterozygous NHEJ1 LoF variant c.532C>T (p.R178*) in a Chinese family with POI. The variant was characterised by a significantly decreased level (by ~50%) of NHEJ1 protein in human peripheral blood samples. Several lines of evidence from the present study link NHEJ1 haploinsufficiency to POI pathogenesis. First, the identified NHEJ1 LoF variant in familial POI is extremely rare in human populations. Second, knock-down of NHEJ1 expression in KGN cells decreased DNA repair efficiency and cell viability on DNA damage stimulus. Third, the heterozygous knock-in mutant mouse exhibited reduced fertility, decreased number of primordial follicles and enhanced sensitivity to DNA damage stimulus consistently.

The NHEJ pathway is the main mechanism employed by mammalian cells for repairing DNA DSB damages. This pathway is specifically responsible for V(D)J recombination during T-cell and B-cell maturation. NHEJ deficiency in humans caused radiosensitivity and severe combined immunodeficiency.45 As most patients with NHEJ deficiency do not survive till puberty, the reproductive phenotype was only occasionally described in grown-up patients. For example, a female patient with homozygous missense variants in XRCC4 had progressive growth failure, failed to enter puberty spontaneously and was diagnosed with severe gonadal failure at the age of 16.46 Additionally, bi-allelic truncated LIG4 variants were identified in 11 patients with microcephalic primordial dwarfism, including the two eldest girls who were diagnosed with syndromic POI at the age of 17 and 11, respectively.47 These clinical reports preliminarily indicate that the NHEJ pathway might be implicated in human female fertility. Bi-allelic variants in NHEJ1 have also been reported to cause combined immunodeficiency, microcephaly and growth retardation in clinical settings.48 No description of fertility has been reported in NHEJ1-deficient patients mainly due to their early death. The present patients with a heterozygous NHEJ1 LoF variant did not have any of the above symptoms throughout their life, probably because they had a low level (approximately half) of NHEJ1 that was sufficient for DNA DSB repairs in somatic cells.

Some mouse models with defects in the NHEJ pathway are known to exhibit an early onset of ageing, including decreased ovarian function, providing additional evidence for the involvement of the NHEJ pathway in female fertility. For example, the mice with homozygous Lig4 R278H mutation have phenotypes of a decreased life span and a severely compromised fertility.49 Additionally, a conditional knockout of PP6c, which is required for the efficient activation of DNA-PKcs, caused premature ovarian failure due to increased levels of γH2AX, abolished DDR and induced defects in follicular activation and growth.50

Previous work on Nhej1-mutated mice has reported a suboptimal V(D)J recombination and a moderate lymphopenia, with no information on female fertility.51 In our present work, female subfertility and reduced ovarian reserve were observed in the heterozygous Nhej1-mutated female mice carrying the equivalent NHEJ1 LoF variant of the proband. Our Nhej1-mutated ovarian cells exhibited higher sensitivity to DSB damage than the wild-type controls. By contrast, previous studies have reported that the DSB repair capacities of heterozygous NHEJ1-mutant HCT116 human colon cancer cells and murine embryonic stem (ES) cells were similar to those of corresponding wild-type cells.52 53 These discrepancies might be partially explained by different cell lines in different organisms. For example, while Nhej1-dificient murine ES cells could not support V(D)J recombination, Nhej1-dificient pro-B cell lines performed V(D)J recombination at nearly wild-type levels.51 Recently, NHEJ1-mediated filament formation has been demonstrated to be involved in bridging and stabilising the broken ends prior to repair.54 This function of NHEJ1 is also variable in different cell types.55 Multiple reports have demonstrated the functional redundancy between NHEJ1 and other DNA response factors (such as ATM, 53BP1 and DNA-PKcs) in V(D)J recombination and DSB repair.56–58 Double knockout mice (such as Nhej1 and Atm) exhibited more profound deficits than single Nhej1 knockout mice. Therefore, variable levels of these compensatory mechanism might contribute to the variable function of NHEJ1 in different cell types or organisms.

The discrepancies between our findings and previous reports maybe also partially attribute to the different sensitivities to DSB-inducing drugs of different cells in different stages. It is interesting to find that while Nhej1-mutant young mice are immunocompetent, premature ageing of haematopoietic stem cells were found in the ageing mice due to functional decline during ageing process.59 As to the mammalian ovary, DNA DSBs are accumulated in arrested follicles for decades during maternal ageing as a consequence of cellular metabolism, or exposure to radiation, chemotherapy or environmental toxicants. Therefore, primordial follicles exhibited lower tolerance for DNA damage and lower threshold to activate apoptosis on DNA damage stimulus than somatic cells.60–62 Since NHEJ pathway is important for maintaining high DSB repair capacity and cell viability in ovaries to protect female fertility and offspring health,63 64 the remaining level of NHEJ1 in our familial patients is probably not sufficient for the ovarian cells, rendering an increased risk of early ovarian dysfunction during maternal ageing. Whether NHEJ1 haploinsufficiency affects the genome stability of ovarian cells and how the NHEJ pathway regulates DSB repair and cell apoptosis in arrested follicles remain to be explored in the future. Furthermore, our findings suggest that environmental factors may cooperate with genetic variations in Nhej1 in female fertility regulation. In our study, only mild differences between Nhej1+/− and wild-type female mice were observed under normal conditions; however, Nhej1+/− female mice exhibited significantly reduced DNA repair capacity in ovaries on BLM treatment. As for our patients, although they have not been exposed to typical exogenous DSB stress such as chemotherapy, there exist many environmental toxicants, including ionising radiation and endocrine disruptors, that are capable of inducing DSBs.65 We, therefore, propose that the interaction between genetic variations in DNA repair genes and environmental DNA damage stimulus during the lengthy progression of POI could contribute to the complex pathogenesis of this disease.

In conclusion, the present study reported a rare heterozygous NHEJ1 LoF variant in a Chinese non-syndromic family with POI. Haploinsufficiency in NHEJ1 may render oocytes vulnerable to exogenous DNA damage stimulus (eg, ionising radiation and endocrine disruptors), leading to a high risk of ageing-induced premature female infertility. Defining a clear genotype-phenotype correlation in patients with POI harbouring different variants of NHEJ pathway genes is, therefore, of great importance in gaining more insights into POI aetiology. Notably, the proband in our family has a young daughter, who carries the same heterozygous deleterious LoF variant of NHEJ1 and has already exhibited irregular menstruation. The molecular mechanism of the NHEJ pathway in POI pathogenesis also requires further investigation to allow the development of novel treatment modalities for patients with POI.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Ethics approval

This study was approved by the institutional review boards of the centres participating in this study and followed the Declaration of Helsinki ethical principles.


The authors would like to thank the patients for participating and supporting this study.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • GL, XY and LW contributed equally.

  • Contributors FZ, XiZ, YaW, GL, XY and LW designed the study. XiZ, QC, XuZ and YiW provided patients’ data and performed clinical assessments. GL, XY, LW, YP, SC, LS, YZ, YuW, YiW, LZ and ZZ conducted experiments. GL, XY, LW, JL, LZ, LJ, YaW, XiZ and FZ analysed data. GL, YaW, XY and FZ wrote the manuscript. FZ, XiZ and YaW supervised the study.

  • Funding This work was supported by National Key Research and Development Program of China (2017YFC1001100), National Natural Science Foundation of China (31625015 and 31521003), Shanghai Municipal Science and Technology Major Project (2017SHZDZX01), Natural Science Foundation of Shanghai (20ZR1407000), State Key Laboratory of Reproductive Medicine (SKLRM-K202002), Science and Technology Major Project of Inner Mongolia Autonomous Region of China (zdzx2018065) and Innovative Research Team of High-level Local Universities in Shanghai (SSMU-ZLCX20180500).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.