The cohort of 99 pregnant Chinese women with their fetuses diagnosed with a solitary functioning kidney was enrolled from 2008 until 2017. Informed consent was obtained from all participants, and institutional approval (No. 2019190) was given by the medical ethical committee of the affiliated Huai’an No. 1 People’s Hospital of Nanjing Medical University, China. All pregnancies with a diagnosis of solitary functioning kidney by using ultrasound examinations were included. Those patients who did not have solitary functioning kidney were excluded from the study. The ultrasound phenotypes of a solitary functioning kidney include (1) RHD: the shadow of one side of the kidney significantly shrunken, with a smaller renal pelvis and calyxes instead of normal shapes; compensatory hypertrophy of the contralateral kidney; the normal position of both sides of the ureteric orifices. (2) Unilateral RA: non-visualization of one kidney with normal bladder and amniotic fluid; single renal artery; compensatory hypertrophy of the contralateral kidney. Bilateral RA: non-visualization of both kidneys and bladder in combination with anhydramnios; failure to visualize renal arteries. (3) MCDK: the presence of interfaces between cysts; non-medial location of the largest cyst; absence of an identifiable renal sinus or parenchymal tissue; multiplicity of oval or round cysts that do not communicate. If a prenatal ultrasound scan revealed that the fetus had solitary functioning kidney, fetal samples were obtained using amniocentesis or tissue or cord blood sampling.

Genomic DNA was extracted from samples following standard protocols of QIAamp DNA Mini Kit (Qiagen, Germany). Human cyto12 SNP array (Illumina, San Diego, CA, United States) covering approximately 300,000 SNP probes was applied for the whole-genome scan. Molecular karyotype analysis was performed using KaryoStudio V 1.4.3.0 (Illumina). Results were interpreted using the UCSC¹ GRCh37/hg19 assembly and were compared with data recorded in the Database of Genomic Variants (DGV²), DatabasE of genomiC varIation and Phenotype in Humans using Ensembl Resources (DECIPHER³), and Online Mendelian Inheritance in Man (OMIM⁴). According to the American College of Medical Genetics (ACMG) standards and guidelines for interpretation of copy number variants, chromosomal microdeletions/microduplications were classified into three categories: pathogenic CNVs, variants of uncertain significance (VOUS), and benign CNVs. CNVs were defined as pathogenic if (1) the CNVs were documented as clinically significant in multiple peer-reviewed publications, regardless of its penetrance and expressivity, or (2) the CNVs overlapped with well-established clinical significance. CNVs that have been reported as benign variants in multiple peer-reviewed publications or databases were considered benign. CNVs that did not fit any of the above criteria were considered as VOUS. Fetuses with pathogenic CNVs were suggested for termination of pregnancy, while the remaining fetuses were required for follow-up until birth.

Immunohistochemistry was performed on slides containing paraffin sections. Slides were deparaffinized by xylene followed by serial rehydration in ethanol. Antigen retrieval was performed by heating slides to 95°C for 10 min in 0.1 mol/L citric acid repair solution (pH 6). Endogenous peroxidase activity was blocked by incubating slides in 3% hydrogen peroxide for 10 min. After rinsing, the slides were blocked in 5% BSA solution and incubated in a wet box at room temperature for 30 min. The slides were incubated in primary antibody (anti-QPRT: 25174-1-AP; Proteintech Group) overnight at 4°C. Next, the biotinylated secondary antibody [Goat Anti-Rabbit IgG H&L (Biotin): ab6720; Abcam] was added to the slides for 20 min at room temperature. DAB peroxidase Substrate Kit was used for visualization, and hematoxylin was used as a counterstain. Slides were mounted following serial dehydration.

Quantitative RT-PCR experiments were performed to estimate QPRT copy number on control adult and fetal human tissues. Each sample was assayed in triplicate for each gene by using SYBR Green PCR Master Mix on a BioRad CFX96 qPCR System. Two primer couples were employed to amplify QPRT and reference gene 36B4 (Table 1). The copy number of QPRT was calculated by dividing the yield of the QPRT gene by that of the reference gene. The reaction mix for each sample was prepared in a 96-well plate, and the plate was placed in the Biorad CFX Connect^TM Real-Time PCR machine. Samples were run as recommendations. Samples were pre-denatured at 95°C for 10 min, followed by 40 cycles of 95°C for 10 s, 58°C for 30 s, and 72°C for 30 s. At the end of the thermal profile, the temperature was increased slowly from 60° to 95°C to generate a dissociation curve.

We used chopchop to design the target siRNA, and then delivered for synthesis. HEK293T were plated in 6-well plates at 50,000 cells per well, and transfected using lipofectamine RNAi-MAX reagent with 10 nM final concentration of siRNA following the manufacturer’s instructions. After transfection, knockdown was validated by qRT-PCR. RNA was extracted and cDNA was synthesized according to the reverse transcription kit. The cDNA was used as a template and GAPDH was used as an internal reference. After the reaction was completed, the relative quantitative analysis of Cq value was performed by the 2^{–△ △} Ct method to calculate the relative expression of QPRT in each group of cells. Primer sequences (QPRT validation/GAPDH) were presented in Table 1.

The QPRT knockdown (QPRT siRNA), negative control (QPRT siRNA NC), and wildtype (control) HEK293T cells were seeded in 6-well plates. Cells were lysed in lysis buffer for 30 min on ice, and then insoluble materials were removed after 15 min centrifugation at 10,000 rpm/min. Protein concentrations were determined by BCA assay (Pierce, Rockford, United States). Lysate supernatant was separated using 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto a PVDF membrane (IPVH00010, Millipore). The membrane was incubated with Rabbit monoclonal antibody against QPRT (ab171944; Abcam) or Mouse Anti-GAPDH antibody (1/2000; TA-08; ZSGB-BIO) followed by incubation with peroxidase-conjugated goat anti-rabbit IgG (H + L) (1/2000; ZB-2301; ZSGB-BIO) and peroxidase-conjugated goat anti-mouse IgG (H + L) (1/2000; ZB-2305; ZSGB-BIO) secondary antibody, respectively. Western blots were developed using ECL solution. Grayscale results were analyzed by “Quantity one” software.

After trypsin digestion, HEK293T cells were incubated in 1 ml DNA Staining solution and 10 ul Permeabilization solution mixture at room temperature for 30 min in the dark. Then flow cytometry was used for cell-cycle analysis. For apoptosis study, cells were resuspended in 300 ul of pre-chilled 1× binding buffer, and treated with 3 ul of Annexin V-FITC and 5 ul of propidium iodide (PI) at room temperature for 10 min in the dark. After mixing with 200 ul of pre-chilled 1× binding buffer, cells were analyzed by flow cytometry (Ex = 488 nm; Em = 530 nm) using FITC signal detector (usually FL1) and the phycoerythrin emission signal detector (usually FL2). Annexin V-FITC indicates loss of membrane asymmetry/phosphatidylserine exposure and PI denotes nuclear condensation. When cells are double stained with Annexin V-FITC and PI, three different cell populations may be observed: live cells that do not stain with either Annexin V-FITC or PI, necrotic cells that stain with both reagents, and apoptotic cells that stain with Annexin V-FITC only. After treatment with both dyes, apoptotic cells show green fluorescence, dead cells show red and green fluorescence, and live cells show little or no fluorescence.

The GraphPad Prism (version 8.0.1) was used for statistical analysis. One-way ANOVA was chosen from the list of column analyses. The difference in the mean between the groups was analyzed by the t-test. p < 0.05 was considered for significant differences.

During 2008 to 2017, a total of 99 pregnant Chinese women with their fetuses diagnosed of a solitary functioning kidney by ultrasound were recruited in this study. Among them, seven fetuses were diagnosed as acquired solitary functioning kidney, and the remaining fetuses were matched with congenital solitary functioning kidney. As shown in Table 2, the mean gestational age of these women recruited at the time of CMA test was approximately 25.1 weeks. Accordingly, two fetuses (2.02%) were diagnosed with aneuploidy (one case with trisomy 21 and one with Triple X syndrome), and 34 fetuses (36.36%) were diagnosed with CNV (Supplementary Table 1). The mean maternal age for pregnant women with aneuploidy fetuses was 26.5 years old while that for pregnant women with CNV fetus was 28.5 years old. Moreover, the gestational age of pregnant women with CNV fetus is comparably longer than that of pregnant women with aneuploidy fetuses. The mean gestational age of these women recruited for CNV analysis was approximately 25.1 weeks. And the sex ratio among the CNV fetuses was 19:15 (female:male).

All samples collected from the fetuses were analyzed by CMA. As shown in Figure 1, CMA testing upon 99 subjects identified a total of 45 CNVs out of 34 subjects (defined as CNV size > 66 kb; details are listed in Supplementary Table 1). The majority of these large CNVs were duplications (n = 38, 80.85%) while the rest were deletions (n = 7, 15.6%). According to UCSC GRCh37/hg19, these fetuses carried eight genomic imbalances with significant overlap with known genomic disorders, accounting for 8.1% of the cohort. Median large CNV size was 318,347 bp (range 66,842–5,299,999 bp). The most frequently identified CNVs were Xp22.33 duplication (N = 4), 2q13 duplication (N = 2) and 5q35.3 duplication (N = 2). By comparing with the data recorded in the Database of Genomic Variants (DGV), only one CNV was interpreted as benign CNV. Moreover, CNVs from 12 fetuses, which accounted for 11.1% of the cohort, matched with the records in the Database of Decipher or relevant published works for known genomic disorders or overlapped with pathogenic CNVs with a similar phenotype, which were therefore interpreted as pathogenic CNVs.

Among the pathogenic CNVs (n = 14) from 12 fetuses, eight types of CNVs from seven fetuses were found to overlap with the records in Decipher database with the same mutation type (duplication/deletion), as listed in Table 3. The size of the identified pathogenic CNVs in this study ranged from 84 kb to 521 kb. It is noteworthy that two subjects, fetus 10 and fetus 82, were diagnosed with carrying two different pathogenic CNVs. Fetus 82 had a syndrome of ectopic kidney, with 7q36.2 duplication and 9p24.1p23 deletion. Fetus 10 had a syndrome of left multicystic renal dysplasia, with 10p12.31 duplication and 12p11.23 deletion. Meanwhile, fetus 77 carried a 4p16.1 CNV which matched with three records in the Decipher database. Similarly, fetus 16 carried an 8p23.2 duplication and fetus 47 carried a 16p11.2 deletion which respectively matched with two and four records in the Decipher database. Furthermore, fetus 45 and fetus 46 both carried 22q11.2 duplications, with the phenotype being malrotated kidneys. Among the pathogenic CNVs (n = 14), fetus 47 carried a 16p11.2 deletion which matched with the highest records in the Decipher database. Therefore, we chose 16p11.2 for our prioritization research.

To define the candidate genetic drivers for the phenotype of the eight types of CNVs (Figure 2 and Table 3), a total of 37 genes were identified out of seven types of CNVs, and no genes was found in 8p23.2 locus (Table 4). To be specific, only one gene, namely PTPRD (Protein Tyrosine Phosphatase Receptor Type D), was identified in 9p24.1p23. In addition, DPP6 (Dipeptidyl Peptidase Like 6) and TOP3B (DNA Topoisomerase III Beta) were respectively identified within 7q36.2 and 22q11.22 locus. For 4p16.1 region, three genes, ACOX3, HTRA3, and SH3TC1, were identified. Moreover, we identified miRNAs (microRNAs) in the CNVs as well. For example, miR-4675 was the only gene identified in 10p12.31. Besides, two miRNAs (DGCR5 and DGCR9), as well as PRODH (Proline Dehydrogenase 1), were found to be located within the 22q11.21 region. Notably, mutation in TOP3B is associated with 22q11.2 Duplication Syndrome. DGCR5, DGCR9, and PRODH in 22q11.21 have been reported to be associated with Digeorge Syndrome. Unlike other CNVs, 16p11.2, which was selected for prioritization research, was identified to harbor 26 genes, among which seven genes were previously reported to be associated with kidney development.

Combining with a retrospective into the literature about 16p11.2 duplication/deletion syndromes, we listed the phenotype and pathogenicity of another eight kidney abnormal patients harboring 16p11.2 dosage variations besides fetus 47 in our study (Table 5). By compiling the CNVs of these nine patients into a CNV overlap map, it is easily visible that eight out of nine patients harbored a genomic space in 16p11.2 between coordinator 29,673,954 and 30,085,308 (Figure 3). The molecular mechanisms of five genes specifically involved in the diseases, MAZ, PAGR1, MVP, TBX6, and MAPK3, have been studied (Supplementary Table 2). Meanwhile, among the other three genes, only QPRT was reported to be associated with the pathogenesis of neurodegenerative disorders, which was also a syndrome of 16p11.2. And the expression profiles provided by GTEx Portal⁵ indicated that QPRT was highly expressed in kidney with an FPKM value of 37.95. Therefore, we chose QPRT as a candidate gene associated with susceptibility for solitary functioning kidney.

Immunohistochemical staining of the healthy fetal kidney demonstrated that QPRT was distinctly detected in renal tubules while being barely observed in renal interstitial and glomeruli (Figure 4). Furthermore, we randomly selected 20 fetuses (from the cohort of solitary functioning kidney, as case group) and 20 healthy adults (as control group) for quantitative real time polymerase chain reaction (qRT-PCR) to analyze the CNV of QPRT. As shown in Figure 5, the average QPRT copy number of control group was significantly higher than that of case group. Moreover, according to the Exome Aggregation Consortium database (ExAC⁶), QPRT has a loss-of-function variant frequency of <1:1,000, and the pLI score which means the gene’s tolerance to a loss of function mutation is 0.59. Taken together, the above results implied that QPRT mutation might have an indirect role in nephrogenesis during the fetal period.

To verify the potential functional link between the inactivation of QPRT and renal cell development, QPRT was knocked down using siRNA in human embryonic kidney cells, and the successful knockdown of QPRT was further confirmed by western blot (Figures 6A,B) and qRT-PCR (Figure 6C). Moreover, CCK8 assay demonstrated that QPRT knockdown cells significantly exhibited reductions in proliferation (Figure 6D). Flow cytometry data revealed that the apoptosis rate of QPRT knockdown cells was significantly higher than the control groups. The QPRT knockdown cells cannot enter the S phase with the same efficiency as control cells, showing a 21% reduction in the number of cells able to enter S phase over the same time frame. Compared with the control group, the percentage of cells in the G2 phase was more significantly increased than that of the G1 phase. The reduction in the number of cells in S phase and the increased efficiency of entering G2 phase indicated that cells transfected with QPRT siRNA cannot divide properly from tetraploid to diploid (Figure 6E). Hence, cell proliferation was inhibited. Together, these findings suggest that QPRT is involved in the regulation of cell-cycle progression and proliferation of HEK cells.