Hypomorphic PKD1 Alleles Impact Disease Variability in Autosomal Dominant Polycystic Kidney Disease

Key Points

• Autosomal dominant polycystic kidney disease (ADPKD) manifesting earlier than expected on the basis of family history can identify clinically tolerant PKD1 alleles with reduced expression.

• Hypomorphic PKD1 alleles can cause mild kidney disease or liver cysts in the absence of clinically manifest kidney involvement.

• The presented data highlight pleiotropic ADPKD clinical presentations and varying severity of kidney disease from PKD1 allele combinations.

Introduction

Autosomal dominant polycystic kidney disease (ADPKD) is widely recognized as the most common single-gene cause of progressive CKD. Genetic investigation of cohorts with ADPKD detects heterozygous pathogenic variants in PKD1 (chromosome 16p13.3) encoding polycystin-1 (PC1) (approximately 77% cases) or PKD2 (chromosome 4q22.1) encoding polycystin-2 (PC2) protein (approximately 15% cases); approximately 8% cases remain as genetically unresolved mostly because of the technical complexity of PKD1 variant detection or presence of conditions that can phenocopy ADPKD.^1,2 A well-established genic and allelic correlation with the ADPKD disease course is observed, with average age at ESKD being two decades earlier with PKD1—approximately 54 years as compared with approximately 74 years with PKD2—with protein-truncating variants in PKD1 exhibiting the most aggressive CKD progression.³

This genotype-phenotype model is, however, unable to explain the disease course variability in the context of the same familial ADPKD pathogenic variant. The kidney disease variability in ADPKD most commonly presents as variable age at ESKD onset ranging from young to mid or late adulthood in members within families affected with ADPKD.⁴ Very early-onset ADPKD that comes to clinical attention in the perinatal or early childhood period represents an extreme form of such kidney disease course variability.^5,6 Another commonly observed phenotypic variability in ADPKD is the manifestation of non–kidney-related complications such as cysts mostly in the liver.⁷ Partially functioning PKD1 alleles by their varied contribution to the total effective polycystin dose reduction have the potential to explain disease course variability in ADPKD.^8,9 The biological effect of these hypomorphic PKD1 alleles that are often classified as variants of uncertain significance¹⁰ is difficult to ascertain, thus making allele-specific clinical case reporting valuable.

Methods

Patients

Clinical, familial, and genetic information was collected on five probands with ADPKD and early significant kidney involvement manifesting in the perinatal to early adulthood (<21 years) period with or without an apparent family history of ADPKD. Significant kidney involvement included ultrasound or computed tomography or magnetic resonance imaging showing multiple (5–20) or innumerable (>20) kidney cysts on each side and at least one of the following: fetal oligohydramnios, perinatally detected large echogenic kidneys, abnormal kidney function (≥CKD stage 2), and hypertension.

Genetic Testing and Variant Pathogenicity Prediction

Whole-exome sequencing and PKD1 gene-specific long-range PCR¹¹ were performed in Family Y1 as part of an institutional review board–approved research protocol and informed consent process. Families C1, C2, Y2, and C3 had clinical genetic testing. Details of the genomic DNA testing method used are included in Figure 1. Other prominent gene causes of bilateral large and hyperechoic kidneys including PKHD1 and HNF1B-related disease⁵ were excluded in probands presenting perinatally or during childhood (Families Y1, C1, C2, C3).

Figure 1.

Clinical, familial, and genetic characteristics of probands (black arrowhead) with biallelic ADPKD. Family pedigree shown with individual identifiers (age at evaluation) and clinical information (black dot represents ADPKD-affected status on the basis of clinical/radiological criteria). PKD1 variant status^# for tested individuals and their associated kidney imaging where available (ultrasound or CT; kidney cysts—multiple: >4 to <20 on each side; innumerable: >20 count on each side; kidney size^##) and/or histopathology is shown. ^#Gene testing method used: Family Y1: whole-exome sequencing for Y1-I and Y1-II, including analysis of variants in a manually curated list of 113 genes known to associate with isolated or syndromic early renal cystic phenotype. PKD1 gene-specific long-range sequencing for identified PKD1 variants (Y1-I, Y1-II, Y1-III). Family C1: next-generation sequencing panel with CNV detection (genes included: DNAJB11, DZIP1L, GANAB, HNF1B, PKD1, PKD2, PKHD1) and PKD1 gene-specific Sanger sequencing for specific variant testing in both parents. Family C2: next-generation sequencing panel with CNV detection (genes included: DNAJB11, DZIP1L, GANAB, HNF1B, PKD1, PKD2, PKHD1) and PKD1 gene-specific Sanger sequencing for specific variant testing in both parents. Family Y2: gene-specific Sanger sequencing (genes included PKD1, PKD2). Family C3: next-generation sequencing panel with CNV detection (genes included: PKD1, PKD2, PKHD1, HNF1B), followed by PKD1 gene-specific Sanger validation of identified variants. *(red asterisk) means no genetic testing available. ^##Kidney size: defined as large for demographic parameters and normative data on the basis of “Sonographic Assessment of Renal Length in Normal Children,” Am J Roentgenol. 1984;142:467-469. CT, computed tomography.

Characterization and pathogenicity prediction for PKD1 missense variants was performed using a combination of bioinformatics tools (Table 1 footnotes) including the PC1 domain involved, general population-based minor allele frequency, SIFT, PolyPhen-2, residue conservation on the basis of the Grantham Matrix Score, and previously reported variant associations, if any. Splice site predictions were assessed using the neural network modeling NNSplice and SpliceAI scores.

Table 1.

PKD1 missense variant details from presented families with biallelic autosomal dominant polycystic kidney disease

Family	Overall Predicted Variant Rolea	PKD1 Amino Acid Change	PC1 Domain Involved12	Population Frequencyb (EU Non-Finnish) Overall	SIFT13	PolyPhen-214	Residue Conservationc (Grantham Matrix Score)	ADPKD Database15	ClinVar Interpretation16	Prior Reports
Y1	Likely hypomorphic	Thr2250Met	REJ	(3.5×10−3) 2.3×10−3	Deleterious	Probably damaging	Moderately conservative (81)	Likely neutral	Hypomorphic (rs139971481)	Irazabal M et al. (2011); PMID: 21551026 Reiterova et al.19 Paul BM et al. (2014); PMID: 23760289
Y1	Likely hypomorphic	Val1971Met	PKD 15 domain	(1×10−4) 6×10−5	Deleterious	Probably damaging	Conservative (21)	Absent	Uncertain significance	—
C1, C2	Likely hypomorphic	Asp1332Asn	PKD 8 domain	(8×10−5) 1.9×10−4	Deleterious	Probably damaging	Conservative (23)	Likely neutral	Uncertain significance	Audrèzet et al.6
Y2	Likely hypomorphic	Val1611Ile	PKD 11 domain	(2×10−5) 2×10−5	Deleterious	Probably damaging	Conservative (29)	Likely pathogenic	Absent	—
Y2	Likely pathogenic	Arg2767His	REJ	(9×10−6) 4×10−6	Deleterious	Possibly damaging	Conservative (29)	Indeterminate	Uncertain significance	—
C3	Likely pathogenic	Arg3750Gln	MOTIF	(0) 4×10−6	Deleterious	Possibly damaging	Conservative (43)	Highly likely pathogenic	Conflicting interpretations of pathogenicity	Audrézet MP et al. (2012); PMID: 22508176 Bataille S et al. (2011); PMID: 22008521 Hoefele J et al. (2011); PMID: 21115670
C3	Likely hypomorphic	Phe2132Cys	PKD 17 domain	(0) 5×10−6	Deleterious	Probably damaging	Radical (205)	Likely pathogenic	Conflicting interpretations of pathogenicity	O'Brien et al.21

None of the variants were predicted to affect splicing using https://spliceailookup.broadinstitute.org and NNSplice (http://www.fruitfly.org/seq_tools/splice.html). ADPKD, autosomal dominant polycystic kidney disease; PKD, polycystic kidney disease.

^aOverall predicted variant role based on parameters mentioned in the table above and correlation with clinical findings presented in this report. Variant role is classified as likely hypomorphic (variant assessed to be incapable of causing typical adult-onset ADPKD alone but could cause milder disease or act as a modifier) or likely pathogenic (variant assessed to be capable of causing typical adult-onset ADPKD by itself).

^bNone of these variants were present in homozygous form, except T2250M (three homozygotes in gnomAD).²²

^cResidue conservation on the basis of the Grantham Matrix Score designated conservative (0–50), moderately conservative (51–100), moderately radical (101–150), or radical (≥151).²³

Results

Clinical, familial, and genetic characteristics of five probands (Y1-I, C1-I, C2-I, Y2-I, C3-I) with ADPKD and early significant kidney involvement than expected on the basis of family history are included in Figure 1. Family history consisted of either a member with typical ADPKD presenting with large cystic kidneys diagnosed during mid to late adulthood (>35 years) with varying CKD stages 1–4 (Figure 1, Families C1, C2, and C3) or members with clinically missed ADPKD manifestations of fewer kidney cysts (Figure 1, Family Y1) or liver cysts only (Figure 1, Family Y2) that were apparent on imaging performed after proband diagnosis was made.

Each of the five probands carry a unique combination of PKD1 variants with in trans inheritance either supported by both parental testing (Families C1, C2) or favored by a combination of single-parent testing with family history (Families Y1, Y2, C3). PKD1 variants coexisted as combinations of a likely partially functioning allele, also referred to as hypomorphic allele with another hypomorphic allele (Figure 1A; proband Y1-I), a protein truncating allele (Figure 1, B and C; probands C1-I, C2-I), or a likely pathogenic missense allele (Figure 1, D and E; probands Y2-I, C3-I). Each of the hypomorphic and missense PKD1 alleles is predicted to result in reduced PC1 function to a varying degree on the basis of the criteria listed in Table 1. The probands and available family members depict a spectrum of the ADPKD disease course and onset dictated by the functional effect of the underlying individual PKD1 variants or their combinations as depicted in Figure 2.

Figure 2.

Pictorial representation of PKD1 alleles that correlate with the reported ADPKD disease spectrum. (A) Gradation of the functional PKD1 gene dosage or polycystin-1 (PC1) level ranging from two healthy PKD1 alleles (PKD1^wt/wt; normal PC1 level) to complete absence of PC1 protein with PKD1 knock-out alleles (PKD1^null/null) is pictorially shown. This is on the basis of the permissive hypothesis that PKD1 hypomorphic variants and the resulting allele combinations lie across a gene-dosage spectrum that translates into a gradation of functional PC1 dosage with clinical correlations. Hypomorphic PKD1 alleles in the heterozygous form depending on their individual functional effect may associate with no cysts (Val1971Met; Asp1332Asn), liver cysts only (Val1611Ile), or fewer kidney cysts (Thr2250Met) than would be expected in a typical age-matched ADPKD course because of a definite pathogenic variant (Thr2192Alafs*18; His526Thrfs*32). In trans combinations of a hypomorphic allele with/a definite pathogenic variant (Asp1332Asn/Thr2192Alafs*18; Asp1332Asn/His526Thrfs*32), another likely pathogenic missense variant (Val1611Ile/Arg2767His; Phe2132Cys/Arg3750Gln) or another hypomorphic allele (Thr2250Met/Val1971Met) can result in early-onset ADPKD. Clinical/Imaging information not available for wt/Arg2767His or wt/Phe2132Cys combinations in this series. (Figure adapted from: Gallagher AR, Germino GG, Somlo S. Molecular advances in autosomal dominant polycystic kidney disease. Adv Chronic Kidney Dis. 2010;17(2):118-130).

Family Y1 (Figure 1A): Perinatal-onset cystic kidney disease was seen in the 1-year-old proband (Y1-I) and stillborn fetus (Y1-II) with the same combination of two hypomorphic PKD1 alleles (Thr2250Met/Val1971Met). The maternal Thr2250Met variant in itself resulted in mild cystic kidney disease; the clinically unaffected father has not been evaluated by imaging or genetics. The second PKD1 variant (Val1971Met) is likely to be paternally inherited rather than de novo given its presence in both affected siblings, favoring its trans inheritance in Y1-I and Y1-II.

Family C1 (Figure 1B): Perinatal demise was seen in one of the affected fraternal twins, both of which carried a combination of a PKD1 hypomorphic allele (Asp1332Asn) in trans with an inactivating PKD1 allele that in itself resulted in typical adult-onset disease in the affected parent (C1-III).

Family C2 (Figure 1C): Significant kidney disease manifested in 3- and 6-year-old siblings as focal cystic disease and hypertension with the underlying in trans combination of a hypomorphic PKD1 allele (Asp1332Asn) with an inactivating PKD1 allele that in itself resulted in typical adult-onset disease in the affected parent (C2-IV).

Family Y2 (Figure 1D): The proband (Y2-I) evaluated at the age of 20 years for CKD stage 2 and hypertension with cystic kidney disease was found to have a combination of PKD1 variants (Val1611Ile/Arg2767His). The maternal (Y2-II) Val1611Ile hypomorphic allele associates with the phenotype of liver cysts without any kidney cysts. The Arg2767His variant could either be inherited from the untested father or can be de novo. Marked difference with maternal disease manifestation and severity favors a trans phase for the two variants in the proband.

Family C3 (Figure 1E): The proband (C3-I) progressed to ESKD at the age of 2 years and had associated congenital hepatic fibrosis (CHF) and bile duct proliferation. Genetic testing showed PKD1 variants (Phe2132Cys/Arg3750Gln). The clinically unaffected mother carries the PKD1 hypomorphic allele (Phe2132Cys). The genetically untested but clinically affected father with ADPKD is likely to have transmitted the Arg3750Gln variant favoring an in trans inheritance in the proband.

Discussion

Genomic PKD1 variant detection and pathogenicity classification is complicated by its higher mutability related to a large open reading frame spanning approximately 13 kb (NM_001009944.3) and GC richness and because of the presence of six pseudogenes that are highly homologous to their exons 1–33.¹¹ Despite rare definite pathogenic variation in PKD1, a constrained gene under negative selection, hypomorphic PKD1 alleles with milder or clinically missed phenotypes may not qualify as rare in population cohorts, a usual requirement for Mendelian disease-causing variants. Thus, in the absence of available in vitro models for functional PC1 testing or creating cumbersome variant-specific knock-in models, the biological significance of a hypomorphic allele can only be ascertained by examining its role in contributing to clinically significant biallelic ADPKD as demonstrated by the presented cases.

A threshold level of functional polycystin protein is known to be essential for survival.¹⁷ As such, at the genetic level, both PKD1 and PKD2 pathogenic variants are inherited in the heterozygous form. However, a prevalent hypothesis is that somatic inactivation of the healthy polycystic kidney disease (PKD) allele in ADPKD occurs during the course of a lifetime and results in typical adult-onset CKD.^8,18 This PC1 dosage-sensitive phenomenon aligns with the observation that in trans arrangements combining PKD alleles of varied pathogenicity can result in germ line biallelic yet viable ADPKD.⁹ To the same effect, the hypomorphic alleles can dictate PKD1-related pleiotropic phenotypes ranging from clinically unapparent milder to clinically notable severe variations of the typical adult-onset ADPKD course depending on the underlying molecular architecture as demonstrated by the presented families. The unique in trans combination of two hypomorphic PKD1 alleles (Thr2250Met and Val1971Met) resulting in perinatal-onset cystic kidney disease in Family Y1 is interesting because most previously described in utero and early childhood-onset patients involve at least one inactivating PKD1 allele⁹ as was the case in Families C1 and C2. Consistent with prior reports,^11,19 Thr2250Met is a hypomorphic variant associated with mild PKD in Family Y1. The biological effect of the Val1971Met variant that lacks any robust pathogenicity prediction can be on the basis of the associated clinical observation of perinatal-onset ADPKD with the biallelic state (Thr2250Met and Val1971Met) in Family Y1.

In situations where the in trans combination of partially functioning PKD1 alleles results in a diffuse homogeneous PKD akin to that seen in autosomal recessive PKD (ARPKD) (e.g., Families Y1 and C3), it is plausible that a critical PC1 dosage reduction results in severe cystic disease without relying on the age-related second somatic hit to the healthy allele, which is the proposed mechanism in this otherwise dominant disease model. However, in the focal-appearing cystic disease (e.g., Families C1, C2 and Y3), it can be proposed that the biallelic PKD1 disease state may only represent an increased susceptibility of either of the PKD1 alleles to additional second somatic hits for exaggerated cyst formation.

The maternal phenotype of liver cysts without any kidney cysts with the Val1611Ile variant in individual Y2-II is a unique finding that is consistent with the knowledge that some of the autosomal dominant polycystic liver disease genes that intersect with polycystin pathways result in a milder PC1 functional dose reduction and tend to cause a cystic liver phenotype with few kidney cysts, if any.⁷ Another rarely described observation is the manifestation of CHF and bile duct proliferation with early-onset severe ADPKD in individual C3-I. Liver involvement in the form of CHF that is typically a part of the PKHD1-related ARPKD spectrum, but an extremely rare association with ADPKD further brings to fore the incompletely understood complex pathophysiology of renal cyst formation in ADPKD and contributions from PKD1-PKHD1 interactions.²⁰ Of note, similar association of CHF with ADPKD has been described once in the context of the same Phe2132Cys variant as in Family C3.²¹ However, ascertainment of any such variant-specific association will need mechanistic investigation. Another distinctive observation of intrafamilial variability wherein the same PKD1 allelic combination results in variable clinical severity between siblings (Family Y1) or fraternal twins (Family C1) remains unexplained but may be influenced by the genetic background and other intrauterine environmental factors influencing pregnancy outcomes.

A positive family history in either of the parents is the most reliable clinical marker favoring a diagnosis of early-onset ADPKD, but is often absent in the context of inherited hypomorphic variants, and thus, parental screening for asymptomatic kidney cysts and liver cysts may be useful. In addition to aiding molecular characterization, understanding the clinical context of hypomorphic PKD1 alleles serves the clinical purpose of evaluating apparently healthy family members as carriers for purposes of genetic counseling or for living-related kidney donor assessment. Similarly, molecular characterization of the subset of ADPKD that comes to clinical attention because of early significant kidney disease as demonstrated by the patients presented here can aid management of the early-onset disease course and provide useful information on the underlying genetic mechanisms in ADPKD.

Disclosures

N.K. Dahl reports the following: Consultancy: Otsuka Pharmaceuticals and Vertex; Research Funding: I am a PI for clinical trials sponsored by Reata and Vertex; Honoraria: AstraZeneca and Otsuka Pharmaceutical; Advisory or Leadership Role: Natera Scientific Advisory Board and PKD Foundation; Speakers Bureau: I am on the unbranded speakers bureau for Otsuka; and Other Interests or Relationships: Medical Advisory Board and NKF NE Chapter. J. Goodwin reports the following: Patents or Royalties: Patent pending US 16/864,521. E.A. Hartung reports the following: Advisory or Leadership Role: PKD in Children Council/ARPKD Task Force and Polycystic Kidney Disease Foundation—Scientific Advisory Committee; and Other Interests or Relationships: American Board of Pediatrics—Nephrology Subboard Member and American Society of Pediatric Nephrology—Member. S. Somlo reports the following: Consultancy: BridgeBio Pharma and Maze Therapeutics; Ownership Interest: Goldfinch Bio; and Patents or Royalties: Yale University. The remaining authors have nothing to disclose.

Funding

This work was supported by the American Heart Association Award to A. Gulati and the Yale Center for Mendelian Genomics (NIH-U54-HG006504).

Author Contributions

Data curation: Neera K. Dahl.

Investigation: Ashima Gulati.

Methodology: Ashima Gulati, Stefan Somlo.

Writing – original draft: Ashima Gulati.

Writing – review & editing: Stephanie L. Clark, Neera K. Dahl, Julie Goodwin, Erum A. Hartung, Asha Moudgil, Stefan Somlo.

Data Sharing Statement

All data are included in the manuscript and/or supporting information.