The Position of the Polycystic Kidney Disease 1 (PKD1) Gene Mutation Correlates with the Severity of Renal Disease

PKD1 Pedigrees

The proband from each family was recruited through the Oxford Renal Unit (68 pedigrees), other adult nephrology departments (5 pedigrees), or various pediatric nephrology units (7 pedigrees). All individuals involved in the study gave informed consent, and the project had ethics committee approval. All probands had ADPKD by defined ultrasound criteria (29). Family histories were taken, and all available family members contacted. Medical histories and blood samples for DNA isolation were collected from family members wishing to take part in the study. Where possible, clinical details were verified from medical records or death certificates. Abdominal ultrasound was organized for at-risk patients who wished to be diagnosed. The onset of ESRD was defined as the time when renal replacement therapy began or death from uremia occurred. Serum creatinine (SC; μmol/L) measurements were used to assess renal function in patients not having ESRD. Hypertension was defined as when hypertensive therapy was required.

Clinical Details of PKD1 Pedigrees with Atypical Presentations

Atypical renal presentations of PKD1 were defined as early onset in cases with enlarged hyperechoic kidneys before 2 yr of age, or late onset in patients with adequate renal function at >75 yr.

• P117: The clinical details of members of this pedigree have previously been described in reference 19 (mutation Y3818X).

• P118: The proband died immediately postnatally with pulmonary hypoplasia and grossly enlarged kidneys. The mother was diagnosed by ultrasound examination and had normal renal function at 25 yr (8126ins20).

• P148: The proband was diagnosed in utero with greatly enlarged kidneys. At 19 yr, her kidneys remain enlarged but renal function is normal. Her sister has impaired function (SC, 218 μmol/L) at 14 yr and is hypertensive. Their mother (46 yr) has normal renal function but enlarged kidneys and polycystic liver disease. The aunt (46 yr) has normal kidney function and an affected daughter with normal function (18 yr). The affected grandmother started dialysis at 54 yr (E2771K). Reference 30.

• P159: The proband was diagnosed at 2 yr with gross hematuria and ultrasound examination showing multiple cysts and renal enlargement. His father has multiple renal cysts (30 yr) (S75F).

• P169: The proband was diagnosed in the first week of life with enlarged kidneys but has normal renal function at 7 yr. Her sister (2 yr) and mother (30 yr) have asymptomatic ADPKD. The aunt has hypertension, and affected grandmother died from metastases of renal carcinoma at 49 yr. The affected great-grandmother died of chronic renal failure at 40 yr (7211ins7).

• P190: The proband presented at birth with enlarged kidneys, had early hypertension, and has slowly progressing renal failure at 15 yr. The affected father has multiple renal cysts but normal renal function (39 yr) and a negative family history (8507ins12).

• P499: The proband was diagnosed by intravenous pyelography after minor hematuria at 70 yr and ultimately developed ESRD at 80 (S225X). He has an affected son (49 yr) with normal renal function (SC, 78 μmol/L).

• P228: The proband had a SC of 130 μmol/L at 81 yr, and her affected children and niece had only slightly elevated SC of 135, 121, and 111 μmol/L at 63, 57, and 46 yr, respectively (L2816P).

Molecular Data

DNA isolation procedures, mutation screening methods, and the mutations identified in these pedigrees have been described elsewhere (28). Mutations were defined as follows: truncating (nonsense, frameshifting deletions, insertions, or splice events); in-frame (insertions and splicing events that maintain the reading frame, ranging from loss of a single codon to deletion of 97 residues; reference 28); or missense substitutions.

Analyses of the ACE Polymorphism

The ACE insertion/deletion (I/D) polymorphism was analyzed by PCR as described previously (20). The amplified product was 190 bp (D allele) or 490 bp (I allele). An additional insertion-specific amplification was performed in all apparent D/D individuals to prevent mistyping (20). PCR products were visualized by ethidium bromide staining after electrophoresis in 1.75% agarose gels.

Statistical Analyses

Time from birth to ESRD (or death) was computed by using the Kaplan-Meier method. The effects of covariates on time to ESRD was tested by using the univariate Cox proportional hazards model or the log-rank test for continuous or categorical variables, respectively (31). Potential correlation within members of the same family was accounted for and significance tested using a robust variance estimator (32). The functional form of the relationship between nucleotide location and risk of ESRD was investigated by plotting the Martingale residuals (from a Cox model with no covariates) with a local regression (LOESS) smoother, against nucleotide location (33). All tests were two-sided, and a robust P < 0.05 was considered statistically significant. In Figure 2, the running average of age at ESRD is based on the best fitting cubic spline function of the age versus nucleotide relationship.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Plot of age at onset of ESRD (orange) or age of patients with renal function (censored; green) compared with the position of the mutation along the transcript. A running average of age at onset of ESRD is plotted using only the patients with ESRD (orange).

The PKD1 Population

Information about renal status was collected on 324 PKD1 patients from 80 different pedigrees in which the PKD1 mutation was characterized (28). The age at the onset of ESRD was known for 152 patients, and the age at death, not due to ESRD or for unknown reasons, was documented for an additional 36 patients. The remaining 136 patients had adequate renal function, and their age and SC were recorded. Of the total population, 177 (54.6%) were women and 147 (45.4%) were men. There were 207 patients whose BP status was known, of whom 147 (71.0%) were hypertensive.

To characterize the severity of disease in the PKD1 population and to compare it with other PKD1 populations defined by linkage, survival curves from birth until ESRD, or ESRD or death, were computed by using the Kaplan-Meier method. The median age at the onset of ESRD was 54 yr, and ESRD or death was 53 yr (Figure 1A). As we wished to determine what factors influenced the severity of renal disease, the age at onset of ESRD was used as the endpoint in all other survival plots. Comparison of genders showed no significant difference (Figure 1B and Table 1). To see if the ACE polymorphism was associated with disease severity in this mutation-defined PKD1 population, this marker was typed in 216 patients from whom DNA was available. Of these, 64 were DD (29.6%), 108 ID (50.0%), and 44 II (20.4%), giving allele frequencies of 54.6% (D) and 45.7% (I) and a distribution in Hardy-Weinberg equilibrium. Renal survival analysis showed no significant difference among the three haplotypes (Figure 1C and Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. (A) Cumulative probabilities of survival to end-stage renal disease (ESRD) or ESRD or death. Neither gender (B) nor the angiotensin-converting enzyme (ACE) insertion/deletion (I/D) polymorphism (C) have a significant influence on renal survival.

Intrafamilial and Interfamilial Variability

To analyze the extent to which the severity of renal disease varied within families, the age at onset of ESRD, or the age of patients with renal function (censored data), was plotted against the position of the mutation within the transcript (Figure 2). This figure illustrates that the severity of disease varies considerably between patients with mutations in similar locations, including those within the same family (which are plotted at the same nucleotide position). This intrafamilial variability probably reflects a combination of genetic modification and environmental factors. To determine if the renal outcomes among families were significantly different, a Kaplan-Meier renal survival plot was generated for the four largest families, with 17 or more patients available for study (Figure 3A; Table 1). The log-rank test showed a significant difference in time to ESRD in these families (P = 0.013), with one more severe than the average and three with milder disease. These data indicate that characteristics of the mutation itself may influence the disease presentation. Genetic modification factors may, however, account for some of the interfamilial differences, although a single factor would not be expected to have a strong influence in these large, multi-generation families.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Renal survival plots showing (A) significant differences in renal survival among four different PKD1 families with greater than 17 patients (mutations: P1; IVS7+1G→A; P13, C2229X; P17/P17A, Q2243X, and P125, E3631D) and (B) no significant influence of mutation type (truncating, in-frame, and missense).

Analyses of the PKD1 Mutation

To test whether characteristics of the mutation were associated with the severity of disease, renal survival plots were compared among different mutation types: (1) truncating mutations; (2) in-frame changes; and (3) missense events; however, no significant differences were found (Figure 3B and Table 1). Second, the importance of the mutation’s location was tested in terms of the nucleotide position along the gene with the patients separated into two groups at the median position (nucleotide 7812; exon 19). Renal survival plots comparing the two populations showed that the 5′ group (0 to 7812 nt) had significantly more severe disease (P = 0.025; Figure 4A). Analysis of the plot of age at ESRD compared with mutation position demonstrated the same result (without taking censoring into account), with the mean age at ESRD lower at the 5′ than at the 3′ end (Figure 2). This difference was also reflected in the median age to ESRD and probabilities of renal survival at various ages (see Table 1 for details). Of six patients with ESRD by 35 yr, all were in the 5′ group (Figure 2). To determine the significance of mutation location more clearly, data for each of the three mutation types was divided at the median position and analyzed separately (Figure 4B and Table 1). This analysis showed that location was associated with severity of renal disease in each case, with the three 5′ populations associated with more severe disease than their 3′ counterparts, although statistical significance was only achieved for the in-frame mutations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Renal survival plots comparing (A) patients with mutations in the 5′ region of the gene (0 to 7812 nt) compared with the 3′ group and (B) each mutation type (truncating, in-frame, and missense) divided at nucleotide position 7812. (C) Plot of Martingale residuals of the Cox model against nucleotide position showing a running average (red line) and zero position (black line). The overall average residual is zero with positive values, as seen in the 5′ area, indicating an excess of ESRD relative to expected. All 324 patients are included, with only ESRD patients having positive residuals. Patients with location less than the median (7812) were found to have an earlier onset of ESRD (Table 1). Hence, the purpose of this exploratory analysis is to use the running average to estimate more precisely the location of a change in ESRD risk. The figure suggests a change in risk around 7 to 8 kb, near the median location. (Four points with values less than −2 are shown on the −2 axis).

To assess whether the change in disease severity along the gene was a gradual gradient or associated with a specific cut point, the Martingale residuals from a Cox model with no covariates were plotted against nucleotide position (Figure 4C). In this analysis, the zero score represents an average likelihood of developing ESRD, with patients plotted above the line developing ESRD earlier than the average and those below the line having less severe disease. The running average shows a greater likelihood of ESRD in the 5′ population and a cut point to less ESRD at about 7 to 8 kb, indicating that the median position (7812 nt) is a reasonable estimate of the point of change.

Families with Atypical Renal Presentation

The penetrance of the PKD1 phenotype, when ESRD is considered an end point, is very high (Figure 1). However, two families in whom comprehensive histories were available had consistently mild cystic disease (see Materials and Methods). Although mutation position is an indicator of disease severity in the whole population, this does not seem to explain these extreme cases, as one had a 5′ mutation and one had a 3′ change. It is possible in these cases that unusual characteristics of the mutation, such as an ability to generate a partially functional polycystin-1 from PKD1 with a missense mutation (P228), may explain the mild disease. In the other family (P499), the mutation is a nonsense change early in the gene and mosaicism is possible in the father, as he has an apparent negative family history. This was, however, not evident in leukocyte DNA, and his son also had mild disease.

This study contained six pedigrees including one early onset case. Two of these pedigrees had mutations in the 5′ region, and four had mutations in the 3′ area and were of various types (see Materials and Methods) and so the position or type of these changes was not a good indicator of early onset disease. Two of these changes were also found in families with more typical disease (28), and there is clear intrafamilial phenotypic variability in the early-onset families (see Materials and Methods for clinical details).