Biedl Syndrome

CLINICAL RESEARCH www.jasn.org

Risk Factors for Severe Renal Disease in Bardet–Biedl Syndrome

† † ‡ † Elizabeth Forsythe,* Kathryn Sparks, Sunayna Best,* Sarah Borrows, Bethan Hoskins, | †† Ataf Sabir,§ Timothy Barrett, Denise Williams,¶ Shehla Mohammed,** David Goldsmith, ‡‡ ‡ † David V. Milford,§ Detlef Bockenhauer, Lukas Foggensteiner, and Philip L. Beales*

*Genetics and Genomic Medicine Program, University College London Great Ormond Street Institute of Child Health, †National Bardet–Biedl Syndrome Service, Department of Clinical Genetics, Great Ormond Street Hospital, London, United Kingdom; ‡‡University College Hospital Center for Nephrology, Great Ormond Street Hospital, London, United Kingdom; ‡Nephrology Unit, Queen Elizabeth II Hospital, Birmingham, United Kingdom; §Nephrology Unit and |Department of Endocrinology, Birmingham Children’s Hospital, Birmingham, United Kingdom; ¶Clinical Genetics Unit, Birmingham Women’s Hospital, Birmingham, United Kingdom; and **Clinical Genetics Unit and ††Nephrology Unit, Guy’s Hospital, London, United Kingdom

ABSTRACT Bardet–Biedl syndrome is a rare autosomal recessive, multisystem disease characterized by retinal dystrophy, renal malformation, obesity, intellectual disability, polydactyly, and hypogonadism. Nineteen disease-causing genes (BBS1–19)havebeenidentified, of which mutations in BBS1 are most common in North America and Europe. A hallmark of the disease, renal malformation is heterogeneous and is a cause of morbidity and mortality through the development of CKD. We studied the prevalence and severity of CKD in 350 patients with Bardet–Biedl syndrome–related renal disease attending the United Kingdom national Bardet–Biedl syndrome clinics to further elucidate the phenotype and identify risk indicators of CKD. Overall, 31% of children and 42% of adults had CKD; 6% of children and 8% of adults had stage 4–5 CKD. In children, renal disease was often detected within the first year of life. Analysis of the most commonly mutated disease-associated genes revealed that, compared with two truncating mutations, two missense mutations associated with less severe CKD in adults. Moreover, compared with mutations in BBS10,mutationsinBBS1 associated with less severe CKD or lack of CKD in adults. Finally, 51% of patients with available ultrasounds had structural renal abnormalities, and 35% of adults were hypertensive. The presence of structural abnormalities or antihypertensive medication also correlated statistically with stage 3b–5 CKD. This study describes the largest reported cohort of patients with renal disease in Bardet–Biedl syndrome and identifies risk factors to be considered in genetic counseling.

J Am Soc Nephrol 28: 963–970, 2017. doi: 10.1681/ASN.2015091029

Bardet–Biedl syndrome (BBS) is a rare autosomal inter- and intrafamilial phenotypic variation is recessive ciliopathy characterized by rod-cone dys- observed.2 trophy, renal malformations, learning difficulties, Structural renal and urinary tract anomalies obesity, postaxial polydactyly, and hypogonad- and renal dysfunction is a cause of considerable ism.1 Nineteen disease-causing genes have been identified (BBS1–BBS19) in the last two decades. BBSgenescode for proteins that localize to the cilia Received September 16, 2015. Accepted August 7, 2016. or the basal body and are thought to be involved in Published online ahead of print. Publication date available at cilia development and maintenance.2 Mutations in www.jasn.org. BBS genes lead to defective cilia. Sequencing of Correspondence: Dr. Elizabeth Forsythe, Genetics and Genomic known disease-causing genes confirms a clinical Medicine Programme, Institute of Child Health, 30 Guilford Street, diagnosis of BBS in around 80% of patients.2 Vari- London, WC1N 1EH, UK. Email: [email protected] able expressivity is a hallmark of BBS and both Copyright © 2017 by the American Society of Nephrology

J Am Soc Nephrol 28: 963–970, 2017 ISSN : 1046-6673/2803-963 963 CLINICAL RESEARCH www.jasn.org morbidity and reported to affect 53%–82% of patients with Diagnosis was on the basis of clinical phenotyping. All BBS.2–6 The primary renal phenotype is highly variable, rang- patients underwent genetic testing. On sequencing 13 disease- ing from cystic tubular disease, dysplastic renal disease, and related genes, molecular confirmation of the diagnosis was FSGS to concentrating defects.3–7 Lower urinary tract dys- achieved in 80% of all pedigrees (216 of 270) and in 77% of all function is observed in many patients and may have upper patients (265 of 350). A full list of genotypes can be found in renal tract sequelae.8 It is thought that ciliary dysfunction Supplemental Table 1. The sex distribution was 54% male and leads to disturbance of the noncanonical Wnt-signaling path- 46% female. Sixty-nine percent of patients were white, 28% way which may contribute to the development of cystic kidney were Southeast Asian, and the remaining 3% had a mixture of disease classically associated with BBS.2 Secondary renal dis- other backgrounds (black, Chinese, and Ashkenazi Jewish). ease may occur as a consequence of hypertension and diabetes, The distribution of genotypes is demonstrated in Figure 1. both of which are frequently observed in this population. For the purpose of statistical analysis, mutation type was clas- The high frequency of renal disease in BBS is a cause of great sified according to severity. One hundred twenty-five patients anxiety among patients because of the devastating effect this had two missense mutations, 82 had two truncating mutations can have on quality of life, morbidity, and mortality.4–6 This (nonsense, frameshift, splice site, or a combination), 39 had a study examines renal disease in the largest reported cohort of combination of missense and truncating mutations, and the patients with BBS and identifies indicators of disease that remaining 19 patients had other mutation combinations, in- are directly relevant to patient management and clinical cluding exon deletions and start-codon aberrations. stratification. Age of Onset of Renal Disease One hundred fifty-six pediatric patients were seen in the RESULTS pediatric clinics. Of these, we were able to retrospectively obtain the earliest recorded age of onset of stage9 2–5 CKD Overview of the BBS Population (CKD2–5) in 49 pediatric patients attending the BBS clinic Three hundred fifty patients attended the adult and pediatric (Figure 2). All pediatric patients with CKD4–5werediagnosed national BBS clinics in Birmingham and London over a 4-year before the age of 5 years. The majority of patients with any period (2010–2014). The patient population ranged in age stage of CKD presented before the age of 10 years. Because the from birth to 60 years old, with a peak frequency in the 6–10 peak referral age to the clinic is late childhood, a later-recorded years-of-age category, and with few older adults. onset of CKD is likely to reflect significant ascertainment bias

Figure 1. The BBS1 genotype predominates in the UK population of patients with Bardet-Biedl syndrome followed by mutations in BBS10 and BBS2. Distribution of genotypes.

964 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 963–970, 2017 www.jasn.org CLINICAL RESEARCH

Ninety percent of children and 96% of adults were able to comply. Urinalysis was normal in all those considered free of renal disease. Table 1 demonstrates the number of patients seen in the adult and pediatric clinics and the prevalence of CKD indicators. Table 2 demonstrates the prevalence of CKD2–5 in the adult and pediatric populations presenting to the BBS clinics. The age at which patients reached CKD5 is known for 20 patients in this cohort. Of these patients, 70% (n=14) had reached CKD5 by the age of 20 years. The age of onset of CKD5 was delineated retrospectively and Kaplan–Meier survival to CKD5 is demonstrated in Figure 4. Figure 2. CKD4-5 was predominantly diagnosed within the first Genotype and mutation-type analyses in adults with BBS year of life in pediatric patients with BBS. Age at which renal revealed statistically significant correlations with the presence disease (CKD2–5) was first noted in pediatric patients with BBS of severe renal disease defined as CKD3b–5(eGFR,45 ml/min (n=49). per 1.73 m2). Univariable logistical regression analysis indicated that mutations in BBS2, BBS10,andBBS12 were more as patients may have asymptomatic renal disease and no pre- likely to be associated with severe renal disease than mutations vious renal sonography. in BBS1 (P values of 0.02, ,0.001, and 0.03, respectively) Observations from the national BBS clinics suggest that (Table 3). Univariable logistical regression analysis of muta- patients either develop CKD4–5 in childhood or maintain tion type revealed that truncating mutations and truncating/ normal or near-normal renal function into adulthood. Figure missense mutations were statistically associated with severe 3 demonstrates that the frequency of CKD4–5 remains similar renal disease in comparison to two missense mutations in adults and children (8% and 6% respectively). (P values of ,0.001 and 0.01, respectively) (Table 3). Proportional frequencies of stage 2, 3,4, and 5CKDin adults Prevalence of CKD were compared for the commonest genotypes, BBS1 and eGFR results were available for 189 adults and 133 pediatric BBS10, as outlined in Figure 5. Patients with BBS1 mutations patients. Seventy percent of adult patients had at least two are more likely to be disease-free or have early stage CKD. eGFR readings. The prevalence of each stage of CKD in adults Mutations in BBS10 were more frequently represented with and children is demonstrated in Figure 3, where CKD2–5is increasing stage of CKD. Of note, patients attending our clin- present in 42% and 31%, respectively. In the adult population ics with mutations in BBS1 are statistically significantly older (n=194), 107 patients were considered free of renal disease or than those with mutations in BBS10 (P=0.001). had CKD1 and five did not have an eGFR or a renal ultrasound Previous research indicates that the recurring missense scan. Forty-three adult patients had a normal renal ultrasound mutation M390R in BBS1 may be hypomorphic.10,11 We as- scan and 47 had a structural abnormality. In the pediatric sessed this by comparing adult patients who were homozygous population (n=156), 84 patients were considered free of renal for M390R to patients with other mutation types in BBS1 disease or had CKD1, and 23 patients did not have an eGFR. (Supplemental Figure 1). This did not reveal an obvious hy- Eighty-seven pediatric patients had a renal ultrasound scan, 43 pomorphic effect of homozygous BBS1 M390R mutations. It of which revealed a structural abnormality. Nine patients had may reflect the relatively small group of patients, or that the ma- neither a documented eGFR nor an ultrasound scan. All pa- jority of patients with mutations in BBS1 who are not homozy- tients were requested to provide a sample for urinalysis. gous for M390R are heterozygous for this mutation. Because BBS

Figure 3. The majority of patients with BBS have normal eGFR. Distribution of CKD stages in (A) adults (n=194) and (B) children (n=156).

J Am Soc Nephrol 28: 963–970, 2017 Bardet–Biedl Syndrome 965 CLINICAL RESEARCH www.jasn.org

Table 1. CKD assessment in the adult and pediatric populations Althoughunder-reported,6%oftheadult n/Total (% of Total) BBS population reportedurologicalcomplica- CKD Marker Pediatric Patients Adult Patients tions requiring specialist management. Urolog- ical abnormalities included neuropathic bladder, Total no. of patients 156 (100%) 194 (100%) vesico-ureteric reflux, urinary incontinence, and eGFR fl Total no. of patients who had eGFR 133/156 (86%) 189/194 (97%) bladder out ow obstructions. Normal or CKD1 84/133 (63%) 107/189 (57%) CKD2–5(,90 ml/min per 1.73 m2) 49/133 (37%) 82/189 (43%) Presence of Structural Abnormalities CKD5 (,15 ml/min per 1.73 m2) 8/133 (5%) 12/189 (8%) One hundred seventy-seven ultrasound reports Renal ultrasound scan 87/156 (55%) 90/194 (46%) from the entire cohort were available for our Normal 44/87 (51%) 43/90 (48%) assessment.Ofthese,87(49%)wereunremark- Abnormal 43/87 (49%) 47/90 (52%) able and 90 (51%) revealed structural defects. No ultrasound scan and no eGFR 9/156 (6%) 5/194 (3%) Abnormalities were categorized as atrophic/ Urinalysis 140/156 (90%) 186/194 (96%) scarring, echogenic or loss of corticomedullary differentiation, cystic or dysplastic, other de- is an autosomal recessive disease, it is possible that the hypomor- velopmental abnormality, or hydronephrosis, as seen in Figure 6. phic effect of M390R predominates even in those who are hetero- No consistent pattern was evident in the type of renal structural zygous. Of note, only one patient homozygous for BBS1 M390R aberrations observed; cystic disease in patients ranged from uni- had progressed beyond CKD3. The patient had renal transplant lateral single cysts to multiple bilateral cystic disease. Develop- aged 23 and was subsequently referred to the Bardet-Biedl syn- mental abnormalities included horseshoe, ectopic, duplex, and drome service. absent kidneys. Where several abnormalities were present, the Weassessedfor thepresenceofmicro-andmacroalbuminuria predominant structural defect is reported. On assessing genotype by analyzing urinary albumin-to-creatinine ratios as a proxy correlations (BBS1 versus BBS10), no association with the pres- for glomerular injury. Urinary albumin-to-creatinine ratios ence of structural abnormality was identified (P=0.19). were available for 139 adult and pediatric patients. Seven (5%) Correlating the presence of all causes of structural abnormality in had proteinuria (defined as urinary albumin-to-creatinine adults with CKD staging revealed a strong correlationwith CKD3b–5 ratio .30 mg/mmol); three of whom were diabetic. Thirty- (P=0.04). All patients with a reported renal ultrasound scan and two patients (28%) had microalbuminuria (defined as urinary severe renal disease had a detectable structural abnormality (n=7). albumin-to-creatinine ratio .3.5 mg/mmol12), and two of these Ultrasound reports from 39 pediatric patients with known patients were diabetic. This could be matched to an eGFR in 119 renal structural abnormalities who had both antenatal and patients. There was a statistically significant correlation between postnatal sonography failed to identify the anatomic aberra- severe renal disease and urinary albumin-to-creatinine ratios tion prenatally in 14 patients (36%). (P=0.01). The sample size was inadequate for correlation with Five pediatric patients had abnormal antenatal renal genotype and mutation type. ultrasound reports and normal postnatal sonography. In all patients, nonspecific echogenicity was reported antenatally and no specific struc- Table 2. Prevalence of CKD in adults and children, according to age group tural abnormalities were detected. Normal/CKD1/ In this cohort, 30 patients presented with Age Group, yr CKD2 CKD3 CKD4 CKD5 Total No eGFR sonographic evidence of cystic kidney disease Pediatric patients, n which is classically associated with BBS. 0–522270536Twenty-four of these patients had molecular 6–10 42 7 5 2 1 57 confirmation of BBS. Patients with mutations 11–15 29 6 2 0 1 38 in BBS1 and BBS10 accounted for the major- – 16 18 14 9 1 0 1 25 ity of genotypes represented (42% and 21%, Total (% of total) 107 (69%) 24 (15%) 15 (10%) 2 (1%) 8 (5%) 156 respectively). Figure 7 demonstrates some of Adult patients, n the structural renal aberrations commonly 16–20 27 0 3 0 2 32 21–25 29 5 1 1 3 39 detected on ultrasonography. 26–30 17 5 1 1 1 25 31–35 14 4 1 1 3 23 Hypertension and Diabetes 36–40 6 7 5 0 1 19 Thirty-five percent of adult patients (n=67) 41–45 5 8 2 1 1 17 in this cohort were hypertensive. There 46–50 6 5 2 0 1 14 was a statistically significant correlation be- 51–55 6 8 3 0 0 17 tween CKD3b–5 and the presence of anti- 56–60+ 2 5 1 0 0 8 hypertensive medication (P,0.01) (Table Total (% of total) 112 (58%) 47 (24%) 19 (10%) 4 (2%) 12 (6%) 194 3). There was a significant association

966 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 963–970, 2017 www.jasn.org CLINICAL RESEARCH

after the onset of visual decline, which typically occurs toward the end of the first decade of life, accounting for the high frequency of children aged 6–10 years old. Many children presenting in the first year of life are siblings of patients known to the service. The clinical service has been in operation since 2010 and patients .60 years old may not have a known diagnosis of BBS, as familiarity with the syndrome has only increased in the last two decades. The variation of genotypes presented in this study reflects the United Kingdom BBS population and is similar to that observed by others in Europe and North America.13,14 Patients with mutations in BBS1 generally present later to the BBS clinics than patients with mutations in BBS10. This may relate to a milder phenotype and later onset of retinal degeneration. Our study suggests that the onset of primary renal disease in Figure 4. End stage renal disease is not a common feature in the UK children predominantly occurs in infancy. For many adult BBS population. Kaplan–Meier survival curve: age at which patients patients it was difficult to accurately determine the age of onset were first diagnosed with CKD5 (n=350). of renal disease because patients with CKD are managed locally with an annual specialist BBS review. A striking observation is Table 3. Univariable logistical regression analysis of risk factors for severe renal the relatively modest difference in preva- , 2 disease (eGFR 45 ml/min per 1.73 m ) in adults with known common genotypes lence of CKD4–5 between adults (8%) and Confidence Interval children (6%) (Figure 3). This supports Risk Factor Odds Ratio P Value 2.5% 97.5% our hypothesis that patients with BBS pri- – Genetic factors marily either develop CKD4 5inchild- Genotype (n=154) hood or remain entirely or relatively free BBS1 mutation (n=90) (Reference) of severe renal disease. The small propor- BBS2 mutation (n=22) 4.4 1.28 15.19 0.02a tion of adult-onset severe renal disease may BBS9 mutation (n=6) 2.4 0.12 17.74 0.46 relate to comorbidities associated with a BBS10 mutation (n=26) 7.4 2.49 23.32 ,0.01 BBS, such as urological complications, hy- a BBS12 mutation (n=10) 5.9 1.08 28.39 0.03 pertension, obesity, and diabetes. These are Mutation type (n=149) potentially modifiable risk factors which Missense/missense (n=76) (Reference) , a should be managed appropriately. The Truncating/truncating (n=40) 11.4 3.9 41.8 0.01 . Missense/truncating (n=33) 6.3 1.5 28.6 0.01a number of patients 30yearsislimited Diabetes (n=137) 0.62 0.14 0.99 0.47 in this study and it is therefore not possible Hypertension (n=137) 5.43 2.21 14.29 ,0.01a to draw conclusions about the risk of de- Body mass index (n=93) 1.04 0.96 1.10 0.32 veloping renal disease in older patients Age (n=154) 1.02 0.99 1.96 0.15 with BBS. aStatistically significant result. The prevalence of CKD in this cohort is lower than anticipated on the basis of between the presence of hypertension and albuminuria previous estimates.3 Forty-two percent of adults have (P,0.001). The most commonly prescribed antihypertensive CKD2–5 and only 8% of adult patients develop CKD3b–5. medications were angiotensin-converting enzyme inhibitors There appears to be both genotype and mutation-type corre- (52%), followed by diuretics (21%), calcium channel blockers lations with CKD, with increased risk of developing CKD3b–5 (15%), beta blockers (8%), and angiotensin-receptor blockers (3%). for those patients who have truncating mutations and muta- Fifteen percent of adult patients (n=28) were on hypogly- tions in BBS10. This is in keeping with a previous study in- cemic medication. There was no statistically significant asso- dicating similar findings for cardiovascular risk factors in this ciation with CKD3b–5(P=0.47). group.10 However, it is difficult to ascertain if genotype and mutation type represent independent risk factors, or if the statistical significance of mutation types reflects the high prev- DISCUSSION alence of missense mutations in BBS1 and truncating mutations in other genotypes. Only one of the patients in this study Toour knowledge, this is the largest reported studycharacterizing with homozygous BBS1 M390R mutations developed CKD5. the renal phenotype in BBS. The age distribution of our patient BBS1 M390R is the most common mutation observed in the population is most likely a reflection of a number of factors BBS population in Europe and Northern America, and pa- affecting patientreferral.Childrenareoftenreferredtotheservice tients homozygous for M390R make up a significant

J Am Soc Nephrol 28: 963–970, 2017 Bardet–Biedl Syndrome 967 CLINICAL RESEARCH www.jasn.org

abnormalities and CKD. This study validates the requirement for a baseline postnatal renal ultrasound scan after a diagnosis of BBS. The presence of a structural abnormality should alert clinicians that close monitoring is required to identify any deterioration in renal function. Anecdotal reports of structural renal abnormalities detected antentally but absent on postnatal sonography could not be confirmed in this study. One report16 suggests that mutations in BBS10 are associated with antenatal severe cystic kidney disease, which is incompati- ble with life. In this study, the prevalence of BBS10 mutations among patients presenting with postnatal sonographic evidence of cystic kidney disease (21%; n=5) was consistent with that of the overall study popu- Figure 5. Mutations in BBS10 area associated with more severe renal disease as lation (20%). The authors are not aware compared to mutations in BBS1. Percentage distribution of CKD stages in adults. of a higher rate of pregnancy losses in fam- Mutations in BBS1 versus BBS10; absolute numbers are indicated above each column. ilies with BBS10 mutations. It has previously been suggested that proteinuria is consistently absent in BBS- associated renal disease.17 This study demonstrates the coexistence of proteinuria and CKD and the correlation between CKD3b–5 and albuminuria, and is likely the result of renal mass reduction caused by structural abnormalities. We have previously reported on the high prevalence of antihypertensive medication in this group.10 It is not clear if the statistical correlation between CKD and antihypertensive medication is related to antihypertensive agents prescribed as a part of CKD management to decrease the rate of progression or if hypertension represents an independent statistical risk factor in this cohort. In summary, this study maps the prevalence of renal disease in BBS and charac- terizes the highly variable renal phenotype. Figure 6. Structural renal abnormalities are highly variable in BBS. Prevalence of We have identified risk indicators as well as structural abnormalities detected on sonography. potentially protective factors for renal disease. Adults who harbor missense muta- proportion of the patient population (45% of the United tions in BBS1 and have normal renal ultrasound scans in Kingdom population are homozygous or heterozygous for adulthood are less likely to develop CKD3b–5. Patients with this mutation [B. Hoskins, P.L. Beales, unpublished data]). truncating mutations in BBS10, hypertension, and abnormal This significant subgroup of patients could be counseled re- renalultrasoundscansareatsignificantly increased risk of garding their lower risk of progressing to CKD4–5. CKD3b–5 compared with the general BBS population. Pri- Guidelines for the management of BBS recommend that mary renal disease as a consequence of BBS appears to present every patient should have a baseline renal ultrasound exam- in early childhood. ination to assess for the presence of any structural abnormal- The evidence presented here could have a direct clinical ities.15 Although some patients who have abnormal renal implication for patients with BBS. The presence or absence of ultrasound scans do not go on to develop CKD, there is a risk factors should be considered when counseling patients statistically significant correlation between structural and may be used to stratify the clinical service. Previous

968 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 963–970, 2017 www.jasn.org CLINICAL RESEARCH

Figure 7. Renal ultrasonography demonstrates common structural abnormalities associated with Bardet-Biedl syndrome. (A) Dem- onstrates the typical cystic dysplastic appearance associated with BBS: a small subcortical cyst, one large cyst, and loss of corticomedullary differentiation. (B) Subcapsular cysts and increased echogenicity. (C) Nephrocalcinosis. (D) Renal pelvic dilation. LT, left; RT, right. recommendations advise that patients should be reviewed by a calculated according the Schwarz–Haycock formula (height [cm] 3 nephrologist annually unless CKD is present, in which case 31/creatinine [mmol/L]). Adults were categorized as hypertensive if closer monitoring is required. On the basis of this study, the they were on antihypertensive medication, or if they fulﬁlled the authors suggest that adults with the lowest risk of renal disease criteria for antihypertensive treatment in patients with diabetic ac- could receive community nephrology follow-up and low-risk cording to the Kidney Disease Improving Global Outcomes study children could be seen less frequently in specialist clinics. All guidelines.19 Referrals were made primarily from the British national patients with ESRD require frequent specialist follow-up and patient support group, clinical geneticists, and ophthalmologists in those with identiﬁable risk factors or early stable renal disease the United Kingdom. (CKD1–3) warrant annual specialist follow-up. A multina- tional study could facilitate the development of a statistical Mutation Analysis renal risk calculator for this unique population. Mutation analysis was undertaken through the United Kingdom national BBS gene panel, which encompasses 11 BBS genes including BBS1–BB10 and BBS12, as well as two BBS-associated genes, MKS1 CONCISE METHODS and ALMS1.

Patients Statistical Analyses The following renal parameters were ascertained retrospectively for all Genotype-phenotype analysis was targeted to patients with mutations 350 patients attending the national BBS clinics: known history of renal in the two most commonly affected genes, BBS1 and BBS10,aswellas disease; stage of CKD (if present); any abnormalities noted on renal the less common genotypes BBS2 and BBS12, where adequate sample ultrasound scanning; eGFR; renal function tests; and relevant sizes were available. Correlation with mutation type was also assessed. concomitant factors, including presence of hypertension, diabetes, Patients with two known missense mutations were compared with andobesity.Allpatientsgaveinformedconsent orassent.Patientswere two known truncating (nonsense or frameshift) mutations and a seen in the pediatric clinics if they were 16 years of age or under, or 18 combination of missense/truncating mutations. The nonparametric years of age or under and in full-time education. All other patients Mann–Whitney U test was performed to assess differences in median were seen in the adult clinics. All blood and urinary tests were age for genotypes BBS1 and BBS10. For the purpose of genotype and completed after a 6-hour starvation period. The Modiﬁcation of Diet mutation type analysis children were not included because renal fail- in Renal Disease formula was applied to estimate GFR in all adults, in ure appears to occur and progress primarily in childhood, hence CKD keeping with its common use for patients with obesity and diabetes in stage was not considered to be stable until adulthood. Multivariable the general population.18 eGFR for the pediatric population was regression analysis was applied to evaluate genotype-phenotype

J Am Soc Nephrol 28: 963–970, 2017 Bardet–Biedl Syndrome 969 CLINICAL RESEARCH www.jasn.org analysis and assess the effect of confounders on chronic renal disease. 6. Gourdol O, David L, Colon S, Bouvier R, Ayral A, Aguercif M, François R: The relative burden of each risk factor was described in odds ratios. Renal involvement in the Laurence-Moon-Bardet-Biedl syndrome. – Statistical analyses were conducted in R (R Foundation for Statistical Apropos of 3 cases. Pediatrie 39: 175 181, 1984 fi 7. Marion V, Schlicht D, Mockel A, Caillard S, Imhoff O, Stoetzel C, van Dijk Computing, Vienna, Austria). A 5% con dence level was considered P, Brandt C, Moulin B, Dollfus H: Bardet-Biedl syndrome highlights the statistically significant. All tests were two-tailed. major role of the primary cilium in efficient water reabsorption. Kidney Int 79: 1013–1025, 2011 8. Harnett JD, Green JS, Cramer BC, Johnson G, Chafe L, McManamon P, Farid NR, Pryse-Phillips W, Parfrey PS: The spectrum of renal disease in ACKNOWLEDGMENTS Laurence-Moon-Biedl syndrome. N Engl J Med 319: 615–618, 1988 9. Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, Hogg RJ, Perrone RD, Lau J, Eknoyan G; National Kidney Foundation: National The authors wish to thank patients and colleagues and in particular the Kidney Foundation practice guidelines for chronic kidney disease: evalu- – Laurence-Moon-Bardet Biedl Syndrome patient group for their ation, classification, and stratification. Ann Intern Med 139: 137–147, 2003 ongoing support and E. Bagkeris (University College London) for 10. Forsythe E, Sparks K, Hoskins BE, Bagkeris E, McGowan BM, Carroll PV, statistical support and advice. Huda MS, Mujahid S, Peters C, Barrett T, Mohammed S, Beales PL: The national Bardet–Biedl Syndrome clinic is funded by the Genetic predictors of cardiovascular morbidity in Bardet-Biedl syndrome. Clin Genet 87: 343–349, 2015 National Health Service (NHS) Highly Specialised Services. The 11. Castro-Sánchez S, Álvarez-Satta M, Cortón M, Guillén E, Ayuso C, National Institute for Health Research Wellcome Clinical Research Valverde D: Exploring genotype-phenotype relationships in Bardet- Facility at Birmingham Children’s Hospital was used to host some of Biedl syndrome families. JMedGenet52: 503–513, 2015 the data collection via the European Wolfram, Alstrom and Bardet- 12. Amin R, Widmer B, Prevost AT, Schwarze P, Cooper J, Edge J, Biedl syndrome Registry Project. Marcovecchio L, Neil A, Dalton RN, Dunger DB: Risk of microalbuminuria and progression to macroalbuminuria in a cohort with E.F. is funded by the Medical Research Council. E.F., D.B., and childhood onset type 1 diabetes: prospective observational study. BMJ P.L.B. are supported by the National Institute for Health Research 336: 697–701, 2008 Biomedical Research Centre at Great Ormond Street Hospital for 13. Deveault C, Billingsley G, Duncan JL, Bin J, Theal R, Vincent A, Fieggen Children NHS Foundation Trust and University College London. D.B. KJ, Gerth C, Noordeh N, Traboulsi EI, Fishman GA, Chitayat D, receives support from the European Union, FP7 (grant agreement Knueppel T, Millán JM, Munier FL, Kennedy D, Jacobson SG, Innes AM, Mitchell GA, Boycott K, Héon E: BBS genotype-phenotype assessment 2012-305608), and the European Consortium for High-Throughput of a multiethnic patient cohort calls for a revision of the disease defi- Research in Rare Kidney Diseases. P.L.B.was supported by a Wellcome nition. Hum Mutat 32: 610–619, 2011 Trust Senior Fellowship and is a NIHR Senior Investigator 14. Billingsley G, Deveault C, Héon E: BBS mutational analysis: a strategic approach. Ophthalmic Genet 32: 181–187, 2011 15. Forsythe E, Beales PL: Bardet-Biedl Syndrome. GeneReviews at GeneTests: Medical Genetics Information Resource 1993–2016, 2015. DISCLOSURES Available at: http://www.genetests.org. Accessed August 11, 2015 None. 16. Putoux A, Mougou-Zerelli S, Thomas S, Elkhartoufi N, Audollent S, Le Merrer M, Lachmeijer A, Sigaudy S, Buenerd A, Fernandez C, Delezoide AL, Gubler MC, Salomon R, Saad A, Cordier MP, Vekemans ‘ ’ REFERENCES M, Bouvier R, Attie-Bitach T: BBS10 mutations are common in Meckel - type cystic kidneys. JMedGenet47: 848–852, 2010 17. Putoux A, Attie-Bitach T, Martinovic J, Gubler MC: Phenotypic vari- 1. Beales PL, Elcioglu N, Woolf AS, Parker D, Flinter FA: New criteria for ability of Bardet-Biedl syndrome: focusing on the kidney. Pediatr improved diagnosis of Bardet-Biedl syndrome: results of a population Nephrol 27: 7–15, 2012 survey. J Med Genet 36: 437–446, 1999 18. Goderis G, Van Pottelbergh G, Truyers C, Van Casteren V, De Clercq E, 2. Forsythe E, Beales PL: Bardet-Biedl syndrome. Eur J Hum Genet 21: Van Den Broeke C, Buntinx F: Long-term evolution of renal function in 8–13, 2013 patients with type 2 diabetes mellitus: a registry-based retrospective 3. Imhoff O, Marion V, Stoetzel C, Durand M, Holder M, Sigaudy S, Sarda cohort study. BMJ Open 3: e004029, 2013 P, Hamel CP, Brandt C, Dollfus H, Moulin B: Bardet-Biedl syndrome: a 19. Wheeler DC, Becker GJ: Summary of KDIGO guideline. What do we study of the renal and cardiovascular phenotypes in a French cohort. really know about management of blood pressure in patients with Clin J Am Soc Nephrol 6: 22–29, 2011 chronic kidney disease? Kidney Int 83: 377–383, 2013 4. Tieder M, Levy M, Gubler MC, Gagnadoux MF, Broyer M: Renal abnormalities in the Bardet-Biedl syndrome. IntJPediatrNephrol3: 199–203, 1982 5. O’Dea D, Parfrey PS, Harnett JD, Hefferton D, Cramer BC, Green J: The importance of renal impairment in the natural history of Bardet-Biedl This article contains supplemental material online at http://jasn.asnjournals. syndrome. Am J Kidney Dis 27: 776–783, 1996 org/lookup/suppl/doi:10.1681/ASN.2015091029/-/DCSupplemental.

970 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 963–970, 2017 MUTATIONS TYPES TYPES (N=36) MUTATIONS STAGES1. CKDSUPPLEMENTALFIGURE OF ADULTS.IN PERCENTAGEDISTRIBUTION

Percentage of affected patients in each mutation group 10 20 30 40 50 60 70 0 renal disease No knownNo 31 24 CKD2 23 6 CKD3 1 3 CKD4 1 CKD5 BBS1 1 M390R/M390R M390R/M390R VERSUS(N=55)OTHER 2 BBS1 other BBS1 other combinationsmutation BBS1 M390R/M390R BBS1 Supplemental table 1: List of BBS genotypes, mutation types and protein changes

Gene DNA mutation 1 DNA mutation 2 Protein change 1 Protein change 2 No of patients

BBS1 c.1169T>G c.1040delT p.Met390Arg p.Met347Argfs*27 1

BBS1 c.1169T>G c.1169T>G p.Met390Arg p.Met390Arg 95

BBS1 c.1169T>G c.1339G>A p.Met390Arg Splicing 1

BBS1 c.1169T>G c.1508_1509delinsAC p.Met390Arg p.Leu503His 2

BBS1 c.1169T>G c.1514_1515del p.Met390Arg p.Leu505Profs*52 1

BBS1 c.1169T>G c.1570_1572del p.Met390Arg p.Asn524del 1

BBS1 c.1169T>G c.1614del p.Met390Arg p.Leu539Cysfs*40 1

BBS1 c.1169T>G c.217G>T p.Met390Arg p.Gly73* 3

BBS1 c.1169T>G c.607del p.Met390Arg p.Met203* 1

BBS1 c.1169T>G c.851del p.Met390Arg p.Tyr284Serfs*5 1

BBS1 c.1169T>G c.890G>A p.Met390Arg p.Arg297Gln 1

BBS1 c.1169T>G Exon deletion p.Met390Arg Exon deletion 1

BBS1 c.1169T>G c.599_604del p.Met390Arg p.Ile200_Thr201del 1

BBS1 c.145del c.145del p.Asp49Metfs*8 p.Asp49Metfs*8 1

BBS1 c.145del c.951+1G>A p.Asp49Metfs*8 Splicing 1

BBS1 c.1584T>A c.1169T>G p.Tyr528* p.Met390Arg 1 BBS1 c.1708C>T c.1708C>T p.Arg570* p.Arg570* 3

BBS1 c.345dupT c.1169T>G p.Lys116* p.Met390Arg 1

BBS1 c.47+8C>T c.1169T>G Splicing p.Met390Arg 1

BBS1 c.479G>A c.479G>A p.Asp145Glyfs*21 p.Asp145Glyfs*21 4

BBS1 c.48-1G>A c.1169T>G Splicing p.Met390Arg 1

BBS1 c.559_604del c.1708C>T p.lle200_Thr201del p.Arg570* 1

BBS1 c.683T>C c.1169T>G p.Leu228Pro p.Met390Arg 2

BBS1 c.723+1G>A c.1169T>G Splicing p.Met390Arg 1

BBS1 c.871C>T c.1169T>G p.Gln291* p.Met390Arg 1

BBS1 c.871C>T c.36C>A p.Gln291* p.Cys12* 1

BBS1 c.915+1G>A c.1169T>G Splicing p.Met390Arg 1

BBS1 c.951+1G>A c.951+1G>A Splicing Splicing 1

BBS1 c.98dup c.1169T>G p.Asn33Lysfs*66 p.Met390Arg 1

BBS1 exon 8 + 9 deletion exon 8 + 9 deletion Deletion Deletion 2

BBS2 c.1148_1149dup c.2107C>T p.His384Serfs*34 p.Arg703* 1

BBS2 c.1197del c.2060-1G>T p.His399Glnfs18 Splicing 1

BBS2 c.1237C>T c.1237C>T p.Arg413* p.Arg413* 1

BBS2 c.1683_1700del c.1683_1700del p.Asp562_Ile567del p.Asp562_Ile567del 1 BBS2 c.175C>T c.416G>A p.Gln59* p.Gly139Asp 1

BBS2 c.175C>T c.613-1G>C p.Gln59* Splicing 1

BBS2 c.1814C>G c.1814C>G p.Ser605* p.Ser605* 1

BBS2 c.311A>C c.1895G>C p.Asp104Ala p.Arg632Pro 4

BBS2 c.522T>A c.823C>T p.Asp174Glu p.Arg275* 1

BBS2 c.565C>T c.565C>T p.Arg189* p.Arg189* 1

BBS2 c.612+1G>A c.823C>T Splicing p.Arg275* 1

BBS2 c.635T>G c.823C>T p.Met212Arg p.Arg275* 2

BBS2 c.72C>G c.1559C>T p.Tyr24* p.Leu520* 2

BBS2 c.72C>G c.504del p.Tyr24* p.Leu168Phefs*33 1

BBS2 c.72C>G c.72C>G p.Tyr24* p.Tyr24* 3

BBS2 c.72C>G c.823C>T p.Tyr24* p.Arg275* 1

BBS2 c.780_782dup c.1572_1575del p.Leu260_lle261insMet p.His525Phefs*21 1

BBS2 c.823C>T c.1286_1287del p.Arg275* p.Val429Glyfs*44 1

BBS2 c.823C>T c.823C>T p.Arg275* p.Arg275* 3

BBS4 c.1106+3A>G c.1106+3A>G Splicing Splicing 1

BBS4 exon 7 + 8 deletion exon 7 + 8 deletion Exon 7+8 deletion Exon 7+8 deletion 1

BBS5 c.258+1G>A c.258+1G>A Splicing Splicing 1 BBS5 c.2T>A c.2T>A p.Met1? p.Met1? 4

BBS6 c.47del c.47del p.Pro16Hisfs*2 p.Pro16Hisfs*2 1

BBS6 c.705del c.705del p.Lys235Asnfs*45 p.Lys235Asnfs*45 1

BBS6 c.775del c.775del p.Thr259Leufs*21 p.Thr259Leufs*21 3

BBS6 c.830T>C c.830T>C p.Leu277Pro p.Leu277Pro 1

BBS7 c.1306-1_1308del c.1306-1_1308del Splicing Splicing 1

BBS7 c.187G>A c.878A>C p.Gly63Arg p.Gln293Pro 1

BBS7 c.580_582del c.580_582del p.Ala194del p.Ala194del 2

BBS7 c.712_715del c.712_715del p.Arg238Glufs*59 p.Arg238Glufs*59 1

BBS7 exon 7 + 8 deletion exon 7 + 8 deletion Deletion Deletion 2

BBS8 c.775_778del c.775_778del p.Asp259Hisfs*16 p.Asp259Hisfs*16 1

BBS8 c.1049+2_1049+4del c.1049+2_1049+4del Splicing Splicing 3

BBS9 c.1028G>A c.1028G>A p.Gly343Glu p.Gly343Glu 1

BBS9 c.113-1G>A c.113-1G>A Splicing Splicing 2

BBS9 c.122T>G c.557A>G p.Ile41Ser p.Tyr186Cys 4

BBS9 c.214del c.214del p.Val72Trpfs*12 p.Val72Trpfs*12 1

BBS9 c.223C>T c.1146del p.Arg75* p.Asp382Glufs*2 1

BBS9 c.442+1G>C c.2249c>G Splicing p.Ala750Gly 1 BBS9 Exon 5 deletion Exon 5 deletion Exon 5 deletion Exon 5 deletion 1

BBS10 c.118A>T c.290T>A p.Lys40* p.Leu97His 1

BBS10 c.145C>T c.687delT p.Arg49Trp p.Val230Phefs*7 1

BBS10 c.1676dup c.1676dup p.Tyr559* p.Tyr559* 1

BBS10 c.2119_2120del c.964_966del p.Val707X p.Tyr322del 1

BBS10 c.235dup c.235dup p.Thr79Asnfs*17 p.Thr79Asnfs*17 1

BBS10 c.235dup c.989T>C p.Thr79Asnfs*17 p.Val330Ala 1

BBS10 c.271dup c.1244delA p.Cys91Leufs*5 p.His415Leufs*16 1

BBS10 c.271dup c.2119_2120del p.Cys91Leufs*5 p.Val707* 2

BBS10 c.271dup c.271dup p.Cys91Leufs*5 p.Cys91Leufs*5 21

BBS10 c.271dup c.515insTA p.Cys91Leufs*5 p.Leu172Phefs*4 1

BBS10 c.271dup c.687del p.Cys91Leufs*5 p.Val230Phefs*7 2

BBS10 c.271dup c.699_674dup p.Cys91Leufs*5 p.Gly224_Val225dup 1

BBS10 c.271dupT c.1091del p.Cys91Leufs*5 p.Asn364Thrfs*5 1

BBS10 c.271dupT c.118A>T p.Lys40* p.Cys91Leufs*5 1

BBS10 c.271dupT c.235dup p.Thr79Asnfs*17 p.Thr79Asnfs*17 2

BBS10 c.271dupT c.646dupG p.Cys91Leufs*5 p.Asp216Glyfs*39 1

BBS10 c.285A>T c.2119_2120del p.Arg95Ser p.Val707* 1 BBS10 c.306_309del c.899A>C p.Asp102Glufs*6 p.His300Pro 1

BBS10 c.365del c.365del p.Asn122Ilefs*13 p.Asn122Ilefs*13 2

BBS10 c.515_516insTA c.2119_2120del p.Leu172Phefs*4 p.Val707* 1

BBS10 c.530A>G c.530A>G p.Tyr177Cys p.Tyr177Cys 7

BBS10 c.776_777del c.776_777del p.Ile259Serfs*44 p.Ile259Serfs*44 1

BBS12 c.1063C>T c.1063C>T p.Arg355* p.Arg355* 1

BBS12 c.1092del c.1277G>A p.Glu365Argfs*18 p.Cys426Tyr 2

BBS12 c.1115_1116del c.1996_1998del c.1115_1116del p.Val666del 2

BBS12 c.1397G>A c.1397G>A p.Gly466Asp p.Gly466Asp 1

BBS12 c.1418_1420del c.1418_1420del p.Phe473del p.Phe473del 1

BBS12 c.1483_1484del c.1483_1484del p.Glu495Argfs*3 p.Glu495Argfs*3 1

BBS12 c.1589T>C c.1589T>C p.Leu530Pro p.Leu530Pro 3

BBS12 c.1616C>T c.1616C>T p.Gly539Val p.Gly539Val 1

BBS12 c.530A>G c.530A>G p.Tyr177Cys p.Tyr177Cys 1

BBS12 c.172G>T c.1092del p.glu58* p.Glu365Argfs*18 1

BBS12 c.62C>A c.62C>A p.Ser21* p.Ser21* 1

BBS12 c.1502C>T c.2047G>T p.Thr501Met p.Val683Leu 1