• ADPKD
• PKD2
• uniparental disomy
• homozygosity
• aneuploidy
• chromosomal
• clinical genetics
• copy-number
• genetics
• movement disorders (other than parkinsons)
• neurology
• renal medicine

Autosomal dominant polycystic kidney disease (ADPKD) is a systemic disease with progressive development of fluid-filled cysts in both kidneys as a major characteristic. The formation of numerous cysts together with interstitial fibrosis usually causes chronic renal failure beyond mid life. Heterozygous mutations in the polycystic kidney disease 1 (PKD1) gene and polycystic kidney disease 2 (PKD2) gene account for most cases. Although the clinical manifestations of mutations in both gene types overlap completely, a mutation in PKD2 is associated with a milder disease than a mutation in PKD1, with smaller kidneys and later onset of end stage renal disease.1 However, marked intra-familial renal disease variability is well documented in ADPKD. In general, ADPKD patients carry one germline mutation in PKD1 or PKD2 and one normal PKD1 or PKD2 allele, respectively. In renal epithelial cells, dosage reduction of the gene products below a critical threshold, either by somatic inactivation or by stochastic influences, is presumed to trigger cyst formation.2 3

Recently, several families have been reported with non- or incompletely penetrant alleles of PKD1.4 5 At a heterozygous state, they produce no or a very mild phenotype. However, when these hypomorphic alleles are inherited at a homozygous state, or in trans with another mutation, they can produce mild atypical or severe polycystic kidney disease (PKD). Even more, two hypomorphic PKD1-alleles were identified in trans in patients with an ARPKD (autosomal recessive polycystic kidney disease)-like renal cystic phenotype identified in utero, in families with no history of PKD.5

In this report we describe a male patient with neonatal presentation of PKD, who is homozygous for a PKD2 missense variant due to uniparental disomy (UPD). No abnormalities were detected on prenatal ultrasound at the age of 8 weeks, but no additional prenatal ultrasounds were made.

At birth, the boy presented with a distended abdomen; an ultrasound examination revealed enlarged kidneys 10 cm and 9 cm in length with bilateral microcysts (<5 mm) and disrupted differentiation between the cortex and medulla, as seen in severe ARPKD. A biopsy showed polycystic renal tissue with many glomerular cysts matching early onset ADPKD (figure 1A–D). Renal function was abnormal in the first week of life (oliguria, maximum creatinine 222 μmol/l), but improved over the next few weeks. Until the age of 13 years the estimated glomerular filtration rate (eGFR) was about 80 ml/min/1.73 m², without proteinuria or haematuria. Antihypertensive medication, however, had to be started at the age of 2 months. He developed normally except for growth delay (P10).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A–D). Light and electron microscopically views of a renal biopsy of the index patient, taken 4 days after birth. Panel A shows a light microscopically overview. Distinctive features in the biopsy are the prominent glomerular cyst with dilation of Bowman space and collapse of the glomerular capillary tuft. Cysts of the ducts are also seen*. Panel B shows vacuolisation in the tubular cells, probably as result of disturbed glomerular function. Panel C shows an electron microscopic view of vacuoles of a proximal tubule. In panel D collapse of the glomerular tuft is shown. (A and B, original magnification ×400; C and D, original magnification ×4000). (E) Renal ultrasound of the proband at age 9 years showing his right kidney in a longitudinal section. The kidney is enlarged (12.7 cm, >P95), diffusely increased in echogenicity, and shows loss of corticomedullary differentiation. Microcysts and macrocysts are present. Reniform shape is preserved. (F) Pedigree of the family showing the index patient with neonatal onset of polycystic kidney disease, denoted by the filled symbol, and his healthy parents, denoted with open symbols. PKD2 analysis was performed in all individuals indicated by an asterisk.

Follow-up ultrasounds showed cysts growing in number and size (figure 1E, age 9 years), as well as cysts in the prostate and epididymis. At the patient's last ultrasound examination, at the age of 16 years, both kidneys were about 12.5 cm in length; the ultrasound revealed microcysts as well as macrocysts (>10 mm) with increasing diffuse hyperechogenicity in due course, which could also fit ADPKD. No liver fibrosis or liver cysts were detected by ultrasonography or laboratory investigations. Magnetic resonance angiography (MRA) of the cerebral vessels was completely normal. At last follow-up at 18 years of age, he had no symptoms despite occasional bilateral flank pain. Renal function was disturbed with an eGFR of 52 ml/min/1.73 m². Mild proteinuria (0.15 g/l) was present with an albumin/creatinine ratio 21.6 mg/mmol; α1-microglobulin was normal. No haematuria, leucocyturia or glucosuria were detected. His blood pressure was normal (he is being treated with the ACE inhibitor enalapril, 30 mg once daily). His liver function was also normal.

The family (pedigree in figure 1F) has no history of renal cystic disease. Renal ultrasound of the parents at the age of 42 years (mother) and 36 years (father) revealed no abnormalities consistent with PKD. Laboratory investigations showed normal renal function and absence of proteinuria or haematuria in both mother (eGFR (MDRD) >90) and father (eGFR (MDRD)=85) at age 42 years and 44 years, respectively.

Initial mutation analysis of the complete coding region of the ARPKD gene, PKHD1, did not reveal a mutation. Given the previous description of patients with hypomorphic PKD1 alleles and early onset PKD, we screened the complete coding region of both ADPKD genes for mutations.5 A homozygous variant, c.1967T>G p.Leu656Trp, was detected in the PKD2 gene (figure 2A). No pathogenic mutation was found in PKD1 (screening did not include exon 1 and 12). Multiplex ligation dependent probe amplification (MLPA) analysis of the PKD1 and PKD2 genes (MRC-Holland kits P351-B1 and P352-B1) revealed no copy number variations. DNA analysis of the parents showed that the patient's mother was a carrier of the PKD2 variant and that the variant was absent in the paternal DNA (figure 2A). A test using multiple single tandem repeat markers on different chromosomes (Powerplex 16 kit, Promega, Madison, WI, USA) confirmed paternity and maternity. However, the chromosome 4 marker in this kit (FGA; 4q28) suggested that the patient inherited two maternal alleles. This was confirmed by subsequent single nucleotide polymorphism (SNP) array analysis (GeneChip Human Mapping 250K NspI Array; Affymetrix, Santa Clara, CA, USA), which revealed a homozygous region from band 4q12 to 4q26, encompassing 58.6 Mb (chr4: 56 471 882–115 075 420, Ensembl v54) (figure 2B). The parental SNP genotyping data for this region illustrated the occurrence of maternal isodisomy, as only a single maternal allele was found to be inherited by the patient. Comparison of patient and parental SNP genotypes across the remainder of chromosome 4 was consistent with maternal heterodisomy. These results indicate that the patient inherited the unclassified variant in PKD2 in a homozygous state because of the occurrence of maternal UPD of chromosome 4 (UPD4), specifically due to segmental isodisomy at the PKD2 locus. The most probable mechanism of UPD4 formation is a trisomy 4 rescue event initiated by the occurrence of maternal non-disjunction during meiosis II.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

(A) Sequence data of exon 9 of the PKD2 gene in the index patient and his parents. The mutation (c.1967T>G p.Leu656Trp) is indicated with an arrow. (B) Single nucleotide polymorphism array data of chromosome 4 in the index patient. The log R ratio shows two copies of chromosome 4. The homozygous region is indicated by the box at the bottom of the plot. (C) Sequence alignment of PC2 orthologues. The mutation is indicated with an arrow. Pathogenicity prediction results for three different software programs: Align GVGD: Class C15 (GV 21.82 – GD 39.66); SIFT: deleterious (score 0.01); Polyphen: probably damaging with a score of 0.997 (sensitivity 0.40; specificity 0.98).

The p.Leu656Trp substitution is located in a highly conserved region in the PKD2 protein polycystin-2 (PC-2), and scored as a likely pathogenic mutation using prediction programs (SIFT, PolyPhen-2 and GVGD; see caption for figure 2C). This variant has not been reported previously (PKD mutation database:http://pkd.mayo.edu) and has no predicted effect on splicing (programs Splice Site Finder-like, MaxEntScan, NNSPLICE and Gene Splicer).

This is the first description of a homozygous PKD2 variant resulting in a severe neonatal renal cystic phenotype. The phenotype of the heterozygous mother shows that the mutation is not fully penetrant since no cysts were observed by ultrasound examination at middle age. A biopsy taken from the patient at birth indicated many glomerular and proximal tubular cysts indicative of ADPKD. It should be noted, however, that taking a biopsy only for diagnosis is no longer a recommended procedure.

The leucine at position 656 is in the extracellular region at two amino acids distance from the predicted last trans membrane domain. We propose that the variant tryptophan, a large bulky amino acid, causes reduced function by changing the topology of the protein, since the score for the trans membrane domain is slightly increased (data not shown). Previously, Rossetti et al4 reported that incomplete penetrant PKD1 alleles when inherited at a homozygous state, or in trans with another mutation, can produce mild atypical or severe PKD. They proposed that incomplete penetrant alleles can be important in modulating cyst development and propose that the p.Arg2765Cys allele, with a frequency of approximately 1%, could be a major modulator of disease. The only PKD1 variants detected in this patient were a homozygous neutral polymorphism in exon 5, a polymorphism in intron 12, and a likely neutral variant in exon 22 (c.8087T>G p.Leu2696Arg).

Moreover, a homozygous PKD2 variant, p.Phe482Cys, was suggested as a modifier of a splice site mutation in PKD1, and it is likely that more PKD1 and PKD2 disease modifying alleles will exist.6 In addition, a PKD2 family with perinatal death in two severely affected infants carrying a de novo PKD2 frameshift mutation, c.1934_1935delinsT (p.Asn645fs), was described previously. It will be interesting to analyse the PKD1 gene for potential hypomorphic alleles in this particular family.7

Mouse models with homozygous missense mutations or low levels of PKD1 or PKD2 gene expression also survived the embryonic stage and presented with a severe renal cystic phenotype.8–10 Similar to our patient and patients homozygous for incomplete penetrant PKD1 alleles, these models did not show congenital hepatic fibrosis, thereby distinguishing them from the ARPKD phenotype.5 11

Maternal UPD4 is a rare finding. In the literature, two cases with complete maternal isodisomy 4 and one with segmental isodisomy 4 have been described so far.12–14 The first case was a 5-year-old girl with congenital afibrinogenaemia but who was otherwise phenotypically normal, with a homozygous 15 kb deletion at the fibrinogen Aα locus due to maternal UPD4. The second case was a healthy adult woman with a depressive disorder. The authors suggested that a recessive allele for mood disorders was expressed due to UPD. The third case was a patient with typical limb girdle muscular dystrophy 2E and a homozygous mutation in SGCB due to maternal UPD4. Since there were no other clinical features present in these reported patients, it is unlikely that critical maternally imprinted genes are located at chromosome 4. These papers suggest that UPD for chromosome 4 is clinically silent. In the patient presented in our report, however, a mutation in PKD2 resulted in a severe renal cystic phenotype.

In conclusion, a unique combination of UPD of chromosome 4 with meiotic recombination involving the PKD2 locus resulted in severe neonatal PKD due to homozygosity for a not fully penetrant, likely hypomorphic, missense PKD2 mutation.