Identification of Pathogenic PKHD1 Variants in Infants with Autosomal Recessive Polycystic Kidney Disease from the Dhofar Region, Oman

Ethical approvals and patients’ inclusion and clinical evaluation

This case series includes six families from the Dhofar region of Oman who were referred to the National Genetic Centre at the Royal Hospital, Oman, following antenatal diagnoses of ARPKD. All families had neonates with multiple congenital anomalies, including bilateral renal abnormalities, pulmonary hypoplasia, and oligohydramnios. This study was ethically approved by the Royal Hospital Research Etical Committee, Ministry of Health, Oman (MOH/CSR/21/24412). For genetic studies of patients, written informed consent was provided by the parents of all affected neonates.

Patient selection

The six families included in this study were referred to the National Genetic Centre at Royal Hospital, Oman, based on antenatal findings of congenital anomalies consistent with ARPKD. The clinical features included: oligohydramnios, bilateral kidney enlargement, pulmonary hypoplasia, and multiple congenital anomalies (e.g., dysmorphic features such as low-set ears, box-shaped head, and sandal gap deformity). The clinical presentation, including prenatal ultrasound findings, birth details, and postnatal progress, was recorded. Neonatal deaths occurred within 48 hours postpartum in all cases, which provided an opportunity for genetic analysis from post-mortem DNA samples. Additionally, clinical investigations included antenatal ultrasound scanning where the prenatal diagnosis of ARPKD was based on ultrasound findings showing enlarged and echogenic kidneys, absent bladder, and oligohydramnios. Postnatal clinical observations were performed on arrival at the Special Care Baby Unit (SCBU), where the neonates were monitored for respiratory distress (grunting, cyanosis, and decreased oxygen saturation), and vital signs were recorded. The clinical progress for each family such as gestational age at delivery, birth weight and vital signs (e.g., heart rate, blood pressure, oxygen saturation, and temperature) was documented.

DNA isolation, library preparation and next generation sequencing

DNA samples for genetic analysis were extracted from cord blood/post-mortem blood specimens and archived at the National Genetic Centre for subsequent analysis. Clinical information was gathered from medical records, including antenatal ultrasound findings, birth outcomes, and postnatal management. Genetic testing was performed using both Sanger sequencing and next-generation sequencing (NGS) to identify mutations in the PKHD1 gene ( Al Alawi, Al Salmi et al. 2019). For Sanger sequencing, target regions were amplified using AmpliTaq Gold 360 Master Mix kit (Applied Biosystems) using gene specific oligonucleotide primers. For NGS, a multiplex PCR approach amplified 4898 amplicons from 60 target genes (including GANAB, HNF1B, NOTCH2, PKD1, and PKD2, among others) using a custom QIAseq panel (Qiagen). Subsequent NGS was performed on an Illumina MiSeq system. The resulting data (vcf files) was processed through Variant Studio software to allow nucleotide variant and copy number variant calling. All variant nomenclature followed the standards set by the Human Genome Variation Society guidelines and confirmed via Sanger sequencing. PKHD1 frameshift variants leading to predicted downstream nonsense alleles were modelled using MutationTaster2025 (https://www.genecascade.org/MutationTaster2025/) and sequences are shown in Extended Data.

Computational protein structure prediction

The wild-type structure of fibrocystin was predicted with AlphaFold3 (Abramson, Adler et al. 2024) where the overall structure was of high confidence (pLDDT = 73.59). To model the frameshift and duplication variants, we predicted their complete structure and superimposed it on the wild-type. PyMOL was used for visualization (Schrödinger and DeLano 2020).

Clinical and molecular genetic findings

We analyzed six families with neonates affected by ARPKD. The clinical presentation of the affected neonates included signs of oligohydramnios, bilateral enlarged kidneys, pulmonary hypoplasia, and other dysmorphic features. All neonates died within 48 hours postpartum, with no survival beyond seven days. Table 1 illustrates the clinical presentation and outcomes of the affected neonates. The clinical progress for each family such as gestational age at delivery, birth weight and vital signs (e.g., heart rate, blood pressure, oxygen saturation, and temperature) was documented. The gestational age at delivery in our families ranged between 32 and 36 weeks, while the birth weight ranged from 2.3 kg to 3.4 kg. Table 2 summarizes these critical postnatal data.

Each of the six families described had a molecular genetic diagnosis of ARPKD secondary to biallelic PKHD1 disease causing variants ( Figure 2) and are summarized below.

Figure 2. Genetic analysis of novel PKHD1 variants.

(a) The functional regions of fibrocystin, including the IPT, PA14, G8, PbH1, CST, and NLS domains. (b) Sanger sequencing of the c.340C>T, c.6111_6112delTT, c.7011dupT, c.9550dupT variants. The sequence for c.7011dupT of the patient is unavailable and instead displayed for the heterozygote parent. (c) The pedigree diagrams for all six Omani families. Squares, males; Circles, females; Shaded, affected. (d) Position of novel mutations on the wild-type structure of fibrocystin.

Family 1 included a female infant born at 34 weeks of gestation to consanguineous parents presented antenatally with bilateral enlarged kidneys and pulmonary hypoplasia. The infant died within 48 hours. Genetic testing identified a homozygous frameshift mutation, PKHD1 c.7011dupT; p.(G2338WfsX41), classified as likely pathogenic. There was a previous history of 3 neonatal deaths with similar clinical presentations.

Family 2 included a female infant, born at 33 weeks of gestation via breech delivery, had bilateral enlarged kidneys and pulmonary hypoplasia and died within 48 hours. Genetic analysis confirmed a homozygous frameshift mutation, PKHD1 c.7011dupT; p.(G2338WfsX41). There was a previous history of an early miscarriage.

Family 3 included a male infant, born at 35 weeks of gestation, was diagnosed antenatally with polycystic kidneys and suspected Potter syndrome. The neonate died within 7 hours due to spontaneous pneumothorax. Genetic analysis confirmed a frameshift mutation (not present in gnomAD, ClinVar, Leiden Open Variation Database) PKHD1: c.6111_6112delTT; p.(L2039fsX13).

Family 4 included a male infant born at 32 weeks of gestation to consanguineous parents presented with bilateral enlarged kidneys and other dysmorphic features. Genetic analysis revealed a novel homozygous deletion spanning exons 58-60 of the PKHD1 gene, classified as likely pathogenic.

Family 5 included a second twin from a consanguineous couple, born at 36 weeks, presented with oligohydramnios and bilateral kidney enlargement. The neonate died within 22 hours postpartum. A novel frameshift variant, PKHD1 c.9550dupT; p.(Y3184LfsX18), was identified.

Family 6: A male infant was born at 32 weeks of gestation to consanguineous parents presented with oligohydramnios and bilateral kidney enlargement. A homozygous nonsense variant, PKHD1 c.340C>T; p.(Q114*) was identified.

In silico modelling of novel PKHD1 variants

In Figure 3, we have displayed the predicted protein structures for the five novel PKHD1 variants superimposed onto the wild-type fibrocystin structure. The c.340C>T results in a nonsense Q114X variant that encodes for the complete IPT 1 domain ( Figure 3a) but causes a loss of most of its wild-type structure ( Figure 3b). The c.6111_6112delTT is predicted to cause a L2039fsX13 frameshift, impacting the G8 1 domain ( Figure 3c). The wild-type G8 1 domain is composed of 10 β-strands and one α-helix (He, Liu et al. 2006). In our structural model ( Figure 3d), we observed these structural characteristics, but an additional β-strand is formed resulting from the frameshift. The c.7011dupT mutation is predicted to result in a G2338WfsX41 frameshift, where the PbH1 domain is affected ( Figure 3e). The wild-type PbH1 domain is composed of 3 β-strands stacked into six repeats (Mayans, Scott et al. 1997). In the mutant structure, PbH1 1 to PbH1 4 is normal, but the frameshift results in a loss of the remaining two repeats and instead a structure composed of a loop and α-helix is formed ( Figure 3f). Lastly, the c.9550dupT (Y3184LfsX18) is located on an unknown region of fibrocystin ( Figure 3g), where the frameshift does not form secondary structure, and the C-terminal region is lost ( Figure 3h).

Figure 3. In silico modelling of PKHD1 variants.

(a) The c.340C>T mutation is predicted to cause a Q114X nonsense variant, (b) where the mutant structure only partially encodes for the IPT one domain compared to the wild-type structure. (c) The c.6111_6112delTT leads to L2039fsX13, impacting the G8 1 domain, and (d) encoding nearly half of the wild-type structure. (e) The c.7011dupT cause a G2338WfsX41 frameshift, (f ) where PbH1 is partially lost. (g) Lastly, the c.9550dupT leads to Y3184LfsX18, encoding an unknown domain of PKHD1 (h) close to the C-terminal region.