CLINICAL RESEARCH www.jasn.org

Polycystic Disease with Hyperinsulinemic Hypoglycemia Caused by a Promoter in Phosphomannomutase 2

† ‡ Oscar Rubio Cabezas,* Sarah E. Flanagan, Horia Stanescu, Elena García-Martínez,§ † † | Richard Caswell, Hana Lango-Allen, Montserrat Antón-Gamero,§ Jesús Argente,* ¶** † †† ‡ ‡‡ Anna-Marie Bussell, Andre Brandli, Chris Cheshire, Elizabeth Crowne, ‡ ‡‡ || Simona Dumitriu, Robert Drynda,§§ Julian P Hamilton-Shield, Wesley Hayes, ‡ ‡ Alexis Hofherr,¶¶ Daniela Iancu, Naomi Issler, Craig Jefferies,*** Peter Jones,§§ † ‡ ‡ Matthew Johnson, Anne Kesselheim, Enriko Klootwijk, Michael Koettgen,¶¶ ††† ‡‡‡ ‡ ‡ ‡ Wendy Lewis, José María Martos, Monika Mozere, Jill Norman, Vaksha Patel, † ||| || ||| Andrew Parrish, Celia Pérez-Cerdá,§§§ Jesús Pozo,* Sofia A Rahman, Neil Sebire, ‡ || ‡ Mehmet Tekman, Peter D. Turnpenny,¶¶¶ William van’t Hoff, Daan H.H.M. Viering, † ‡ ‡|| ||| Michael N. Weedon, Patricia Wilson, Lisa Guay-Woodford,**** Robert Kleta, ||†††† † ‡|| ||| Khalid Hussain, Sian Ellard, and Detlef Bockenhauer

Due to the number of contributing authors, the affiliations are listed at the end of this article.

ABSTRACT Hyperinsulinemic hypoglycemia (HI) and congenital polycystic kidney disease (PKD) are rare, genetically heterogeneous disorders. The co-occurrence of these disorders (HIPKD) in 17 children from 11 unrelated families suggested an unrecognized genetic disorder. Whole-genome linkage analysis in five informative families identified a single significant on 16p13.2 (logarithm of odds score 6.5). Se- quencing of the coding regions of all linked failed to identify biallelic . Instead, we found in all patients a promoter mutation (c.-167G.T) in the phosphomannomutase 2 (PMM2), either homo- zygous or in trans with PMM2 coding mutations. PMM2 encodes a key enzyme in N-glycosylation. Abnor- mal glycosylation has been associated with PKD, and we found that deglycosylation in cultured pancreatic b cells altered insulin secretion. Recessive coding mutations in PMM2 cause congenital disorder of glyco- sylation type 1a (CDG1A), a devastating multisystem disorder with prominent neurologic involvement. Yet our patients did not exhibit the typical clinical or diagnostic features of CDG1A. In vitro, the PMM2 pro- moter mutation associated with decreased transcriptional activity in patient kidney cells and impaired binding of the transcription factor ZNF143. In silico analysis suggested an important role of ZNF143 for the formation of a chromatin loop including PMM2. We propose that the PMM2 promoter mutation alters tissue-specific chromatin loop formation, with consequent organ-specificdeficiency of PMM2 leading to the restricted phenotype of HIPKD. Our findings extend the spectrum of genetic causes for both HI and PKD and provide insights into gene regulation and PMM2 pleiotropy.

J Am Soc Nephrol 28: 2529–2539, 2017. doi: https://doi.org/10.1681/ASN.2016121312

Received December 14, 2016. Accepted February 22, 2017. Correspondence: Dr. Detlef Bockenhauer, UCL Centre for Nephrology, Rowland Hill Street, London NW3 2PF, UK. O.R.C., S.E.F., H.S., R.K., K.H., S.E., and D.B. contributed equally Email: [email protected] to this work. Copyright © 2017 by the American Society of Nephrology Published online ahead of print. Publication date available at www.jasn.org.

J Am Soc Nephrol 28: 2529–2539, 2017 ISSN : 1046-6673/2808-2529 2529 CLINICAL RESEARCH www.jasn.org

Table 1. Clinical details Patient ID Variable 1.1 1.2 1.3 1.4 2.1 2.2 3.1 4.1 Sex Male Female Female Female Female Male Male Male Mutation Nucleotide c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. T/c.-167G.T T/c.-167G.T T/c.-167G.T T/c.-167G.T T/c.422G.A T/c.422G.A T/c.470T.C T/c.422G.A Predicted p.?/p.? p.?/p.? p.?/p.? p.?/p.? p.?/p.Arg141 p.?/p.Arg141 p.?/Phe157 p.?/p.Arg141 His His Ser His Hyperinsulinism Birth weight, g 2880 3400 3000 3750 3850 3400 3325 2960 Age at diagnosis, 610481008210 mo Treatment None None None None Diazoxide Diazoxide Diazoxide Diazoxide

Renal disease Antenatal Cysts/ No No No No No No No abnormalities polyhydramnios Age at renal 0 6 48 11 48 0 21 0 diagnosis, mo eGFR 70 100 85 115 95 90 90 50 at diagnosis, ml/min per 1.73 m2 Kidney size, .95th .95th .95th .95th .95th .95th .95th .95th percentile Morphology Cysts Hyperecho- Hyperecho- Hyperecho- Cysts Cysts Cysts Cysts genicity/ genicity/ genicity/ cysts cysts cysts Age at last f/u, yr 10 15 20 14 11 4 1 2 eGFR at last f/u, 70 90 ,10 (18) 75 80 40 90 50 ml/min per 1.73m2 Liver disease Morphology Normal Normal Normal Normal Cysts Cysts Normal Normal (ultrasound) Liver function Normal Normal Normal Normal Normal Normal Normal Normal tests Shown are pertinent clinical data for the 17 patients. Note the presence of the PMM2 promoter mutation c.-167G.T in all patients. ID, identifier; N/A, not available; f/u, follow-up. aUltrasound imaging of the liver was normal, yet a biopsy revealed ductal plate malformation (see Figure 3).

Autosomal recessive polycystic kidney disease (ARPKD), a had been identified in these genes in any of the patients rare disorder with an estimated incidence of 1:20,000 live described here. births, is characterized by the combination of polycystic The co-occurrence of these two rare disorders in multiple kidneys and hepatic fibrosis.1 It is caused by mutations in patients and without identified mutations in known disease PKHD1, although mutations in other ciliary disease genes genes strongly suggested the existence of a yet undescribed may phenocopy the disorder.2,3 Causative mutations are Mendelian disorder with recessive inheritance. identified in approximately 85% of cases.4 Here, we inves- tigated patients with an ARPKD-like clinical presentation that were genetically unsolved and prominently character- RESULTS ized by a concurrent clinical diagnosis of hyperinsulinemic hypoglycemia (HI). HI in itself is also a rare disorder, with Patients an estimated incidence of 1:50,000 and most commonly We noted the co-occurrence of the clinical diagnoses of HI associated with mutations in the ABCC8 and KCNJ11 and polycystic kidney disease (HIPKD) in 17 patients from 11 genes involved in the regulation of insulin release from families of European descent (Supplemental Figure 1, Table 1), pancreatic b cells.5,6 However, no causative mutations including a consanguineous family with four affected

2530 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2529–2539, 2017 www.jasn.org CLINICAL RESEARCH

Table 1. Continued Patient ID 5.1 6.1 7.1 8.1 9.1 9.2 10.1 10.2 11.1 Female Male Female Female Male Male Female Male Female

c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. c.-167G. T/c.422G.A T/c.422G.A T/c.422G.A T/c.255+1G.A T/c.620T.C T/c.620T.C T/c.470T.C T/c.470T.C T/c.470T.C p.?/p. p.?/p. p.?/p. p.?/p.? p.?/p. p.?/p. p.?/p. p.?/p. p.?/p. Arg141His Arg141His Arg141His Phe207Ser Phe207Ser Phe157Ser Phe157Ser Phe157Ser

N/A 4200 4252 4500 3440 3450 N/A 3720 N/A 4 0 N/A 12 14 0 18 11 14

N/A Diazoxide Diazoxide/ N/A Diazoxide Diazoxide N/A N/A Diazoxide cornstarch

N/A Cysts N/A N/A Cysts Cysts N/A N/A No

36 0 36 15 0 0 N/A N/A 20

95 20 120 75 30 40 N/A N/A 105

.95th .95th .95th .95th N/A N/A N/A N/A .95th

Cysts Cysts Cysts Cysts Cysts Cysts Cysts Cysts Cysts

14 6 11 7 4 0.3 N/A N/A 11 85 ,10 (2) 141 40 65 70 N/A N/A 94

Coarse Cysts Cysts Normala Normal Normal Cysts N/A Cysts echotexture N/A Normal Normal Normal Normal Normal Normal Normal N/A

individuals. Multiple siblings were affected in three further were discovered incidentally during childhood. Blood tests pedigrees, consistent with autosomal recessive inheritance were consistent with CKD stage 1 or 2 in nine patients, (Figure 1, Supplemental Figure 1). whereas two patients (1.1 and 6.1, see Table 1) progressed to Eight patients from seven families had been referred for ESRD at ages 20 and 2 years, respectively, with need for dialysis genetic testing for HI through an international study of 1250 or transplantation. Native kidneys were removed at the time families. All patients had been tested for ABCC8 and KCNJ11 of transplant in patient 6.1 due to their massive size (20 cm mutations, which are the most common cause of HI, yet none longitudinal diameter, normal ,7.7), and histology was were found. Mutations in PKHD1, the causative gene for consistent with a predominantly glomerulocystic disorder ARPKD, had been excluded by linkage and/or sequencing (Figure 2). analysis in 14 patients. Liver Disease Kidney Disease In addition, liver cysts were seen in eight patients. Other pa- All patients were found to have enlarged kidneys with multiple tients had no abnormalities on liver imaging (Table 1), includ- cysts on imaging, prompting an initial clinical diagnosis of ing patient 8.1, whose liver biopsy (age 1 year) showed ductal ARPKD (Figure 2). Severity of kidney disease was variable: plate malformation (Figure 3). Liver function tests were normal three patients presented antenatally; in others kidney cysts or only borderline elevated.

J Am Soc Nephrol 28: 2529–2539, 2017 Polycystic Kidney Disease with Hyperinsulinemic Hypoglycemia 2531 CLINICAL RESEARCH www.jasn.org

Figure 1. Genetic studies identify a shared mutation. (A–E) Pedigrees of informative families used for linkage analysis. (F) The combined parametric multipoint linkage analysis reveals a single significant peak with a maximum LOD score of 6.5 on chromosome 16. (G) Sequencing of the linked region reveals the presence of the mutation c.-167G.T in the promoter of PMM2 in all patients. Shown are electropherogram traces of the region c.-173_–c.161 from a patient homozygous for the mutation, a control (wild-type) subject, and a heterozygote parent.

HI metric linkage analysis in this and four other informative HI was diagnosed at a median age of 10 months of life, often families confirmed and refined this to a single significant presenting with hypoglycemic seizures. The spectrum of se- locus of 2.3 Mb, including 14 annotated genes on chromo- verity varied and in one patient (1.3) diagnosis of HI was made some 16p.13.2 with a combined LOD score of 6.5 (Figure 1). only at 4 years of age. Most were treated with diazoxide and Using next-generation sequencing, we identified a noncod- responded to this therapy, although some patients did not ing variant in the promoter region for PMM2 (encoding receive any treatment. Phosphomannomutase 2, a key enzyme in N-glycosylation), c.-167G.T, not previously documented in databases (Figure Other Organs 1). Sanger sequencing confirmed the presence of this mutation No manifestations in other organs or dysmorphologies were in all patients. It was found in homozygous state in the four noted. Specifically, there was no evidence for neurologic patients from the consanguineous family and in compound impairment. Patients 2.1, 2.2, and 6.1 underwent detailed audio- heterozygote state with previously described deleterious cod- logic and ophthalmologic testing including electroretinogram, ing mutations in PMM2 in all others (Table 1). Sequencing of which revealed no abnormalities. the parents from families 2–11 confirmed that these mutations Serum transferrin glycosylation was assessed by isoelectric were inherited in trans. focusing in 11 patients and was normal in all, excluding a sub- stantial global defect in glycosylation. Functional Studies of the Promoter Variant The identified c.-167G.Tmutation lies in a region previously Genetic Analysis described as a promoter of PMM2.7 We assessed the effect of Autozygosity analysis of SNP-chip genotype data in the con- the promoter variant on PMM2 transcription in vitro,using sanguineousfamily revealedahomozygous2.5 Mbregionon human kidney and pancreatic b cell lines. Cells were trans- chromosome 16p.13.2. Whole-genome multipoint para- fected with luciferase under the control of either wild-type

2532 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2529–2539, 2017 www.jasn.org CLINICAL RESEARCH

Figure 2. Renal imaging and histology in children with HIPKD. (A) Ultrasound images of the kidney from patient 2.1, left kidney, age 11 years; (B) patient 2.2, right kidney, age 4 years; and (C) patient 6.1, right kidney, age 2 years. Note the presence of cysts of various sizes. Kidney length was .95th percentile for age in all patients. (D) Axial MRI image of the kidney from patient 2.1, age 11 years; (E and F) axial and coronal MRI images (without contrast) from patient 6.1, age 2 years. Note the massively enlarged kidneys with cysts of various sizes. (G) Macroscopic appearance of the nephrectomy specimen from patient 6.1, age 2 years, removed at the time of transplant. Note the large kidney size (20 cm longitudinal, normal ,7.7 cm) and numerous macroscopic cysts. (H and I) Histology of the same kidney demonstrating multiple cysts lined by simple attenuated epithelium, with glomeruli noted in some (glomerulocystic disease; hematoxylin and eosin; original magnifications, 320 and 3100, respectively). Dist, distance; Lt, left; Rt, right. or mutant promoter. In both cell models, the promoter To assess the consequences of decreased PMM2 activity on variant caused a significant reduction in luciferase activity insulin secretion, we investigated the effect of deglycosylation in a (Figure 4, A and B). We also assessed PMM2 transcription murine pancreatic b cell line.8 We noted a significant increase in kidney cells derived from the nephrectomy from pa- in insulin secretion after stimulation with the protein kinase C tient 6.1, compound heterozygous for the promoter muta- activator, phorbol-12 myristate-13 acetate (Figure 4F), establishing tion c.-167G.T and the missense mutation c.422G.A; a direct link between glycosylation and insulin secretion. p.R141H. Digital quantitative PCR revealed reduced expression of the allele containing the promoter variant compared with the other allele (Figure 4C). The promoter mutation lies in a rec- DISCUSSION ognized binding site for the transcription factor ZNF143.7 We therefore assessed binding of ZNF143 to the promoter using an HIPKD: A Disorder with Pancreatic, Kidney, and Liver Electrophoretic Mobility Shift Assay (EMSA), which demon- Manifestations strated significantly reduced binding to the mutant promoter We describe a previously unreported disorder presenting with compared with wild type (Figure 4, D and E). ARPKD-like kidney disease and HI in childhood.

J Am Soc Nephrol 28: 2529–2539, 2017 Polycystic Kidney Disease with Hyperinsulinemic Hypoglycemia 2533 CLINICAL RESEARCH www.jasn.org

underwent a liver biopsy to establish a di- agnosis, which showed ductal plate malfor- mation (Figure 3). This is the only patient in our cohort who had a liver biopsy and it suggests that the absence of abnormalities on imaging or blood tests does not exclude the possibility of microscopic structural changes in the liver of patients with HIPKD. None of the patients reported here had ev- idence of portal hypertension, a complica- tion seen in about a quarter of patients with ARPKD.9

HIPKD and the Promoter Mutation The identification of a promoter mutation in PMM2 was unexpected, because there wasnoevidenceforasystemicdisorderof glycosylation in our patients. Autosomal recessive mutations in PMM2 are the cause of a devastating multisystemic disease Figure 3. Liver involvement in HIPKD. (A) Shown is an axial T2-weighted (STIR se- called congenital disorder of glycosylation quence) MRI image of patient 2.1 showing numerous small liver cysts. (B) Ultrasound 1a (CDG1A). Neurologic problems usually image of the liver of patient 2.2 showing a cyst. (C and D) Photomicrographs of the liver predominate the phenotype of CDG1A, biopsy specimen from patient 8.1, demonstrating abnormal and expanded portal tracts with affected patients presenting with se- consistent with ductal plate malformation (hematoxylin and eosin; original magnifi- vere developmental delay, strabism, and 3 3 cations, 40 and 100, respectively). muscular hypotonia, none of which was noted in our patients. Moreover, typical dysmorphic features of CDG1A, such as Most patients presented with hypoglycemia in the first year protruding ears, inverted nipples, and abnormal subcutane- of life, consistent with congenital HI. Hypoglycemia was the ous fat pads, were not seen in any of our patients with HIPKD. first recognized manifestation of HIPKD in most patients, Infantile demise is common.10 Later on, in surviving patients, typically presenting with seizures. However, severity varied other complications, such as thrombotic events, skeletal de- with delayed diagnosis in some. formities, , retinitis pigmentosa, hypogonadism, and All patients had enlarged cystic kidneys, which typically peripheral neuropathy become apparent.11 None of our pa- led to an initial diagnosis of ARPKD. Yet, kidney manifes- tients presented any evidence for such involvement, clearly tations of HIPKD differ from ARPKD. Histology of the distinguishing HIPKD from CDG1A. kidneys from patient 6.1 shows predominantly, if not ex- AkeydiagnosticfeatureofCDG1Aisthepresenceof clusively, glomerular cysts, whereas in ARPKD cysts arise specific abnormalities seen with isoelectric focusing of trans- from fusiform dilatation of the collecting duct.3 Sev- ferrin due to abnormal glycosylation. Importantly, this test eral of the patients had renal cysts seen on antenatal was normal in all patients with HIPKD tested. Although ultrasound, yet none had severe manifestations with oli- normal, or only slightly abnormal, transferrin isoelectric gohydramnios and consequent complications, such as patterns have been reported in a few patients with a mild pulmonary hypoplasia. Instead, two patients (1.1 form of CDG1Awith isolated mild neurologic and novisceral and 6.1), including the one with the earliest progression involvement, this is clearly clinically distinct from patients to ESRD (age 2 years), had antenatal documentation of with HIPKD.12,13 polyhydramnios. In contrast, about 30%–40% of patients The phenotype of HIPKD appears to be restricted to kidneys with ARPKD die in the neonatal period from complications and pancreatic b cells with additional liver involvement, which of oligohydramnios.4 is distinct from CDG1A, where visceral involvement is only Liver involvement is an obligate aspect of ARPKD and the noted in the severe early-onset form with neurologic and presence of liver manifestations initially seemed to support a dysmorphic manifestations. A potential explanation for diagnosis of ARPKD in our patients. Liver abnormalities on this organ-specific involvement would be a unique sensitiv- imagingwerenotedineightpatients,mostlyshowingscattered ity of kidney, pancreas, and liver to impaired PMM2 func- cysts and/or a heterogeneous echotexture. Liver enzyme levels tion. Under this hypothesis, the promoter mutation would in plasma were normal or only borderline abnormal. Inter- affect PMM2 function to just such a degree that only kidney, estingly, liver imaging was normal in patient 8.1, who pancreatic b cells, and liver are affected, whereas PMM2

2534 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2529–2539, 2017 www.jasn.org CLINICAL RESEARCH

Figure 4. The c.-167G.T mutation impairs promoter function. Luciferase activity is significantly reduced with mutant compared with wild-type promoter in (A) human pancreatic b cell line 1.1B4 and (B) human kidney cell line RPTEC/TERT1. Light gray bars show lu- ciferase activity with wild-type promoter, black bars with the c.-167G.T mutation, and the crosshatched bars show activity with de- letion of the ZNF143 binding site c.-163_-180del. Asterisk indicates significance (wild type versus respective mutation; P,0.01). (C) In kidney cells derived from patient 6.1, expression of the allele with the mutant promoter (c.-167G.T) is significantly reduced compared with the other allele. Asterisk indicates significance (P,0.01). (D) EMSA demonstrating reduced affinity of mutant promoter to ZNF143. Note the impaired binding of ZNF143 (indicated by arrow) to the mutant (lane 3) compared with wild-type (lane 1) PMM2 promoter. Binding is specific, because addition of excess unlabeled wild-type (lane 2) or mutant promoter probe (lane 4) abolishes binding to the labeled probes. Moreover, no specific binding is seen in the negative control (no ZNF143) to either wild-type (lane 5) or mutant (lane 6) promoter probe. (E) Densitometry confirms significantly reduced affinity of ZNF143 to mutant promoter (P,0.01). Bands from lane 1 and 3 indicated by arrow in (D) were measured by densitometry and results pooled from five independent experiments. (F) De- glycosylation alters insulin secretion in a murine pancreatic b cell line (MIN6). There is significant increase in insulin secretion after stimulation with high glucose (P,0.01), as expected. Note the significant (P,0.01) increase in insulin secretion with deglycosylation after further stimulation with phorbol-12 myristate-13 acetate (PMA). activity was still sufficient to avoid manifestations in other with visceral manifestations as seen in HIPKD also had as- organs.However,thisisnotconsistent with observations sociated debilitating neurologic and dysmorphic features, in patients with CDG1A, as those with mild forms show which are absent in HIPKD. isolated neurologic involvement only.12,13 This strongly sug- Mutations affecting PMM2 function or transcription can gests that the brain is most sensitive to loss of PMM2 func- therefore cause either CDG1A or HIPKD and such pleiotropy tion and mild impairment thus primarily causes neurologic is well recognized: mutations in TRPV4, for instance, are as- problems. The severity of manifestations seen in patients sociated with at least eight different disorders, comprising with HIPKD in the three key organ systems involved, yet such divergent clinical phenotypes as skeletal dysplasias or with no evidence of systemic involvement, clearly argues motor and sensory neuropathies.14 Mutations in OCRL cause for a tissue-specific effect of the promoter mutation. Al- either the systemic disorder Lowe syndrome or the kidney- though specific neurologic investigations, such as an MRI specific Dent disease.15 Yet, although the pleiotropic effects of scan of the brain, have not been performed due to a lack of most of these genes are poorly understood, our data here clinical indication, previous reports of patients with CDG1A clearly point to tissue-specific dysregulation of PMM2 by the

J Am Soc Nephrol 28: 2529–2539, 2017 Polycystic Kidney Disease with Hyperinsulinemic Hypoglycemia 2535 CLINICAL RESEARCH www.jasn.org promoter mutation as the key disease mechanism in HIPKD. Figures 2 and 3). HNF4A could affect PMM2 transcription Indeed, it has been suggested that altered chromatin looping either positively in the setting of the wild-type promoter, be- allows for tissue-specific interactions between enhancers and pro- cause it would require binding of ZNF143 to establish the moters, thus constituting a key mechanism for genetic pleiot- functional loop, or negatively, by destabilizing a weakened in- ropy.16 Our promoter mutation appears to confirm this concept. teraction between mutant PMM2 promoter and ZNF143/ CTCF. Under both scenarios, there would be decreased Promoter Mutation and Chromatin Loops PMM2 transcription from the mutant promoter in HNF4A- can be modulated by regulatory regions expressing cells, compatible with tissue-specificdiseaseas many megabases away from the coding region and it is the seen in HIPKD. three-dimensional architecture of chromatin that is critical to allow this interaction.17 This involves the formation of struc- HIPKD and Glycosylation tural chromatin “loops,” typically between 40 kb and 3 Mb in Defective glycosylation has been previously linked to the de- size and flanked by binding sites for CCCTC–binding factor velopmentofHIorcystickidneydiseaseandbothphenotypes (CTCF) in convergent orientation.17 These loops, also called have been described as part of the spectrum of clinical man- “topologically associated domains” (TAD), serve as funda- ifestations in the multivisceral form of CDG1A.25–29 Yet, mental regulatory units of transcription.18 Inspection of given the predominance of neurologic problems, only few publicly available data for transcription-factor binding and data on the renal involvement in CDG1A are available, chromatin interactions shows that the region around PMM2 mainly that cysts can be seen on imaging, which on antenatal contains a number of CTCF sites and functional promoters ultrasound may manifest as increased echogenicity (“antenatal (Supplemental Figure 2). Interaction between pairs of CTCF bright kidneys”), and in cases with postmortem examination tu- sites has been observed in a variety of combinations, poten- bular microcysts were noted.25,29,30 Interestingly, the latter is differ- tially dynamic and tissue-specific in nature, to form localized ent to the only available histology in our HIPKD cohort, which chromatin loops. Moreover, these data suggest an interaction shows predominantly, if not exclusively, glomerular cysts (Figure 2). between ZNF143 and specific CTCF sites, consistent with pre- The consequences of impaired glycosylation for cystogenesis vious observations of functional interactions between these are evident from the recent identification of mutations in GANAB two DNA binding , which affect the three-dimensional as a cause of kidney and liver cysts.31 Several animal studies pro- structure of such chromatin loops by linking CTCF with distal vide further links: mice deleted for Aqp11 develop cystic kidney regulatory elements, facilitating specific gene regulation.19–21 disease due to altered glycosylation of polycystin1,32 and defective Here, ZNF143 binds to the PMM2 promoter, as confirmed by glycosylation of polycystin2 leads to cyst formation.33 Thus, im- our EMSA data (Figure 4). By simultaneously binding CTCF paired glycosylation of key proteins involved in cystic kidney dis- and the PMM2 promoter, ZNF143 could establish a func- ease may be responsible for the development of renal cysts in tional chromatin loop enabling interaction between the HIPKD. There is also evidence for the importance of glycosylation promoter and a number of regulatory binding sites. This for targeted membrane trafficking of the sulfonyl urea receptor would allow specific spatiotemporal regulation of PMM2, (SUR), a key protein for regulated insulin secretion.34 This com- depending on the expression profile of respective regula- plements our own data, showing altered insulin secretion after tory factors. Although the expression of all genes contained deglycosylation (Figure 4E). Similarly, polycystic liver disease is in the chromatin loop could be affected by the promoter associated with mutations in genes that cause abnormal glycosyl- mutation, the fact that it is found in trans with coding mu- ation or processing of glycoproteins.35 tations in PMM2 clearly implicates specificimpairmentof In conclusion, wereportapreviouslyundescribeddiseasewith PMM2 regulation as the key disease mechanism in HIPKD. the key manifestations of HI and polycystic kidney disease. All Although other genes in the loop and especially TMEM167, patients share a noncoding mutation, which disrupts binding of which is regulated by the same promoter as PMM2,may the transcription factor ZNF143 and impairs specificregulationof also be affected, this effect is expected to be the same as in PMM2 expression. Our findings are consistent with a critical role the asymptomatic heterozygous carriers of the promoter of protein glycosylation for normal insulin secretion and kidney mutation without a PMM2 coding mutation in trans.The morphology. Selective organ involvement, as seen in the patients key role of PMM2 is further supported by the decreased presented here, has to our knowledge not previously been asso- PMM2 expression with the promoter mutation (Figure ciated with a promoter mutation and thus provides important 4). Although PMM2 is expressed ubiquitously, there is ev- insight into gene regulation. Identification of a disease-causing idence for additional tissue-specificregulation.22 Apoten- promoter mutation emphasizes the importance of assessing non- tial candidate involved in such tissue-specificregulationis coding variants in the genetic analysis of disease. the transcription factor HNF4A, which is expressed pre- dominantly in the three organ systems involved in HIPKD: liver, pancreas, and kidney.23 Indeed, we previously linked CONCISE METHODS mutations in HNF4A with HI and kidney disease.24 Clusters of HNF4A binding site are present in the loop (Supplemental Full details of all methods can be found in the Supplemental Material.

2536 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2529–2539, 2017 www.jasn.org CLINICAL RESEARCH

Genetic Studies Charity; Kids Kidney Research; St Peter’s Trust for Kidney, Bladder All subjects and/or their parents gave informed consent for genetic and Prostate Research; The David and Elaine Potter Foundation; testing and all studies were approved by the respective institutional and the European Union, FP7 (grant agreement 2012-305608 research ethics review boards. Linkage analysis was performed as de- “European Consortium for High-Throughput Research in Rare tailed previously.36 Kidney Diseases [EURenOmics]”). S.E.F. has a Sir Henry Dale Next-generation sequencing was performed through Perkin Elmer Fellowship jointly funded by the Wellcome Trust and the Royal (www.perkinelmer.com) and Illumina (www.illumina.com). Sequencing Society (grant number: 105636/Z/14/Z). S.E. is a Wellcome Trust data were analyzed with an in-house pipeline, as well as Ingenuity variant Senior Investigator. analysis software (www.ingenuity.com). Identified mutations in PMM2 were confirmed with Sanger sequencing. DISCLOSURES EMSA None. EMSA was performed as described.37 ZNF143 was expressed in vitro and bound to biotinylated probes specificforeitherwild-typeor fi mutant promoter. Speci city was assessed by competition with excess REFERENCES of unlabeled probes. 1. Zerres K, Mücher G, Becker J, Steinkamm C, Rudnik-Schöneborn S, Insulin Secretion Heikkilä P, Rapola J, Salonen R, Germino GG, Onuchic L, Somlo S, Insulin-secreting MIN6 cells were maintained and insulin secretion Avner ED, Harman LA, Stockwin JM, Guay-Woodford LM: Prenatal di- measured as described previously.38,39 To assess the effect of degly- agnosis of autosomal recessive polycystic kidney disease (ARPKD): Molecular genetics, clinical experience, and fetal morphology. Am J cosylation, the cells were incubated in the absence or presence of Med Genet 76: 137–144, 1998 Peptide N-Glycosidase F and Endoglycosidase H. 2. Gunay-Aygun M, Parisi MA, Doherty D, Tuchman M, Tsilou E, Kleiner DE, Huizing M, Turkbey B, Choyke P, Guay-Woodford L, Heller T, Szymanska K, Johnson CA, Glass I, Gahl WA: MKS3-related ciliopathy Luciferase Assay with features of autosomal recessive polycystic kidney disease, neph- Promoter activity was assessed by placing three different promoter ronophthisis, and Joubert Syndrome. J Pediatr 155: 386–392.e1, 2009 9 1 2 constructs 5 to the renilla luciferase gene: ( )wildtype,() 3. Hartung EA, Guay-Woodford LM: Autosomal recessive polycystic kid- c.-167G.T, and (3) c.-180_-163del, the latter deleting the entire ney disease: A hepatorenal fibrocystic disorder with pleiotropic effects. predicted ZNF143 binding site.7 Pediatrics 134: e833–e845, 2014 4. Guay-Woodford LM, Bissler JJ, Braun MC, Bockenhauer D, Cadnapaphornchai MA, Dell KM, Kerecuk L, Liebau MC, Alonso-Peclet Digital PCR MH, Shneider B, Emre S, Heller T, Kamath BM, Murray KF, Moise K, fi Analysis of allele-speci c PMM2 expression was performed in pri- Eichenwald EE, Evans J, Keller RL, Wilkins-Haug L, Bergmann C, Gunay- mary renal cells from patient 6.1 obtained after nephrectomy. Cells Aygun M, Hooper SR, Hardy KK, Hartung EA, Streisand R, Perrone R, were prepared and cultured as described previously.40 RNA was iso- Moxey-Mims M: Consensus expert recommendations for the diagnosis and lated and cDNA generated as described and used as template for management of autosomal recessive polycystic kidney disease: Report of an international conference. JPediatr165: 611–617, 2014 digital PCR with primers specific for the alleles containing either 5. Flanagan SE, Kapoor RR, Hussain K: Genetics of congenital hyper- 41 the promoter or missense mutation. insulinemic hypoglycemia. Semin Pediatr Surg 20: 13–17, 2011 6. Mohamed Z, Arya VB, Hussain K: Hyperinsulinaemic hypoglycaemia: Bioinformatics Genetic mechanisms, diagnosis and management. J Clin Res Pediatr – The genomic region around the PMM2 promoter was assessed for Endocrinol 4: 169 181, 2012 7. Anno YN, Myslinski E, Ngondo-Mbongo RP, Krol A, Poch O, Lecompte chromosomal segmentation, transcription factor, and CTCF binding O, Carbon P: Genome-wide evidence for an essential role of the human sites using the University of California in Santa Cruz Staf/ZNF143 transcription factor in bidirectional transcription. Nucleic browser with publicly available tracks. Chromatin loop formation Acids Res 39: 3116–3127, 2011 was assessed using available Hi-C data.42 8. Ishihara H, Asano T, Tsukuda K, Katagiri H, Inukai K, Anai M, Kikuchi M, Yazaki Y, Miyazaki JI, Oka Y: Pancreatic beta cell line MIN6 exhibits characteristics of glucose metabolism and glucose-stimulated insulin se- Statistical Analyses cretion similar to those of normal islets. Diabetologia 36: 1139–1145, 1993 Appropriate statistical assays were used as indicated in the Supple- 9. Shneider BL, Magid MS: Liver disease in autosomal recessive polycystic P, mental Material for the respective experimental data. A 0.05 was kidney disease. Pediatr Transplant 9: 634–639, 2005 considered significant. 10. Grünewald S: The clinical spectrum of phosphomannomutase 2 de- ficiency (CDG-Ia). Biochim Biophys Acta 1792: 827–834, 2009 11. Monin ML, Mignot C, De Lonlay P, Héron B, Masurel A, Mathieu- Dramard M, Lenaerts C, Thauvin C, Gérard M, Roze E, Jacquette A, ACKNOWLEDGMENTS Charles P, de Baracé C, Drouin-Garraud V, Khau Van Kien P, Cormier- Daire V, Mayer M, Ogier H, Brice A, Seta N, Héron D: 29 French adult patients with PMM2-congenital disorder of glycosylation: Outcome This work was supported by the Medical Research Council (grant of the classical pediatric phenotype and depiction of a late-onset number 98144); the Great Ormond Street Hospital Children’s phenotype. Orphanet J Rare Dis 9: 207, 2014

J Am Soc Nephrol 28: 2529–2539, 2017 Polycystic Kidney Disease with Hyperinsulinemic Hypoglycemia 2537 CLINICAL RESEARCH www.jasn.org

12. Casado M, O’Callaghan MM, Montero R, Pérez-Cerda C, Pérez B, kidneys in infants and children due to early microcystic changes. Pe- Briones P, Quintana E, Muchart J, Aracil A, Pineda M, Artuch R: Mild diatr Radiol 36: 108–114, 2006 clinical and biochemical phenotype in two patients with PMM2-CDG 30. Pérez-Dueñas B, García-Cazorla A, Pineda M, Poo P, Campistol J, Cusí (congenital disorder of glycosylation Ia). Cerebellum 11: 557–563, 2012 V, Schollen E, Matthijs G, Grunewald S, Briones P, Pérez-Cerdá C, 13. Vermeer S, Kremer HP, Leijten QH, Scheffer H, Matthijs G, Wevers RA, Artuch R, Vilaseca MA: Long-term evolution of eight Spanish patients Knoers NA, Morava E, Lefeber DJ: Cerebellar ataxia and congenital with CDG type Ia: Typical and atypical manifestations. Eur J Paediatr disorder of glycosylation Ia (CDG-Ia) with normal routine CDG Neurol 13: 444–451, 2009 screening. JNeurol254: 1356–1358, 2007 31. Porath B, Gainullin VG, Cornec-Le Gall E, Dillinger EK, Heyer CM, Hopp K, 14. Nilius B, Voets T: The puzzle of TRPV4 . EMBO Rep 14: Edwards ME, Madsen CD, Mauritz SR, Banks CJ, Baheti S, Reddy B, Herrero 152–163, 2013 JI,BanalesJM,HoganMC,TasicV,WatnickTJ,ChapmanAB,VigneauC, 15. Hoopes RR Jr., Shrimpton AE, Knohl SJ, Hueber P, Hoppe B, Matyus J, Lavainne F, Audrezet MP, Ferec C, Le Meur Y, Torres VE; Genkyst Study Simckes A, Tasic V, Toenshoff B, Suchy SF, Nussbaum RL, Scheinman Group, HPoPKDG, Consortium for Radiologic Imaging Studies of Polycystic SJ: Dent disease with mutations in OCRL1. Am J Hum Genet 76: 260– Kidney Disease; Harris PC: Mutations in GANAB, encoding the glucosidase 267, 2005 IIalpha subunit, cause autosomal-dominant polycystic kidney and liver dis- 16. Lonfat N, Montavon T, Darbellay F, Gitto S, Duboule D: Convergent ease. Am J Hum Genet 98: 1193–1207, 2016 evolution of complex regulatory landscapes and pleiotropy at Hox loci. 32. Inoue Y, Sohara E, Kobayashi K, Chiga M, Rai T, Ishibashi K, Horie S, Su Science 346: 1004–1006, 2014 X, Zhou J, Sasaki S, Uchida S: Aberrant glycosylation and localization of 17. Dekker J, Mirny L: The 3D genome as moderator of chromosomal polycystin-1 cause polycystic kidney in an AQP11 knockout model. J communication. Cell 164: 1110–1121, 2016 Am Soc Nephrol 25: 2789–2799, 2014 18. Dixon JR, Selvaraj S, Yue F, Kim A, Li Y, Shen Y, Hu M, Liu JS, Ren B: 33. Hofherr A, Wagner C, Fedeles S, Somlo S, Köttgen M: N-glycosylation Topological domains in mammalian genomes identified by analysis of determines the abundance of the transient receptor potential channel chromatin interactions. Nature 485: 376–380, 2012 TRPP2. J Biol Chem 289: 14854–14867, 2014 19. Bailey SD, Zhang X, Desai K, Aid M, Corradin O, Cowper-Sal Lari R, 34. Conti LR, Radeke CM, Vandenberg CA: Membrane targeting of ATP- Akhtar-Zaidi B, Scacheri PC, Haibe-Kains B, Lupien M: ZNF143 pro- sensitive . Effects of glycosylation on surface ex- vides sequence specificity to secure chromatin interactions at gene pression. JBiolChem277: 25416–25422, 2002 promoters. Nat Commun 2: 6186, 2015 35. Janssen MJ, Waanders E, Woudenberg J, Lefeber DJ, Drenth JP: 20. Heidari N, Phanstiel DH, He C, Grubert F, Jahanbani F, Kasowski M, Congenital disorders of glycosylation in hepatology: The example of Zhang MQ, Snyder MP: Genome-wide map of regulatory interactions in polycystic liver disease. JHepatol52: 432–440, 2010 the human genome. Genome Res 24: 1905–1917, 2014 36. Bockenhauer D, Feather S, Stanescu HC, Bandulik S, Zdebik AA, 21. Phanstiel DH, Boyle AP, Heidari N, Snyder MP: Mango: A bias- Reichold M, Tobin J, Lieberer E, Sterner C, Landoure G, Arora R, correcting ChIA-PET analysis pipeline. Bioinformatics 31: 3092–3098, Sirimanna T, Thompson D, Cross JH, van’t Hoff W, Al Masri O, Tullus K, 2015 Yeung S, Anikster Y, Klootwijk E, Hubank M, Dillon MJ, Heitzmann D, 22. Grünewald S, Schollen E, Van Schaftingen E, Jaeken J, Matthijs G: High Arcos-Burgos M, Knepper MA, Dobbie A, Gahl WA, Warth R, Sheridan residual activity of PMM2 in patients’ fibroblasts: Possible pitfall in the E, Kleta R: Epilepsy, ataxia, sensorineural deafness, tubulopathy, and diagnosis of CDG-Ia (phosphomannomutase deficiency). Am J Hum KCNJ10 mutations. NEnglJMed360: 1960–1970, 2009 Genet 68: 347–354, 2001 37. Garner MM, Revzin A: A gel electrophoresis method for quantifying the 23. Drewes T, Senkel S, Holewa B, Ryffel GU: Human hepatocyte nuclear binding of proteins to speci fic DNA regions: Application to compo- factor 4 isoforms are encoded by distinct and differentially expressed nents of the Escherichia coli lactose operon regulatory system. Nucleic genes. Mol Cell Biol 16: 925–931, 1996 Acids Res 9: 3047–3060, 1981 24. Hamilton AJ, Bingham C, McDonald TJ, Cook PR, Caswell RC, Weedon 38. Jones PM, Salmon DM, Howell SL: Protein phosphorylation in electri- MN, Oram RA, Shields BM, Shepherd M, Inward CD, Hamilton-Shield cally permeabilized islets of Langerhans. Effects of Ca2+, cyclic AMP, a JP, Kohlhase J, Ellard S, Hattersley AT: The HNF4A R76W mutation phorbol ester and noradrenaline. Biochem J 254: 397–403, 1988 causes atypical dominant Fanconi syndrome in addition to a b cell 39. Liu B, Barbosa-Sampaio H, Jones PM, Persaud SJ, Muller DS: The phenotype. J Med Genet 51: 165–169, 2014 CaMK4/CREB/IRS-2 cascade stimulates proliferation and inhibits apo- 25. Strøm EH, Strømme P, Westvik J, PedersenSJ:Renalcystsinthecarbohydrate- ptosis of b-cells. PLoS One 7: e45711, 2012 de ficient glycoprotein syndrome. Pediatr Nephrol 7: 253–255, 1993 40. Kuo NT, Norman JT, Wilson PD: Acidic FGF regulation of hyper- 26. Shanti B, Silink M, Bhattacharya K, Howard NJ, Carpenter K, Fietz M, proliferation of fibroblasts in human autosomal dominant polycystic Clayton P, Christodoulou J: Congenital disorder of glycosylation type kidney disease. Biochem Mol Med 61: 178–191, 1997 Ia: Heterogeneity in the clinical presentation from multivisceral failure 41. Klootwijk ED, Reichold M, Helip-Wooley A, Tolaymat A, Broeker C, to hyperinsulinaemic hypoglycaemia as leading symptoms in three Robinette SL, Reinders J, Peindl D, Renner K, Eberhart K, Assmann N, infants with phosphomannomutase deficiency. J Inherit Metab Dis 32 Oefner PJ, Dettmer K, Sterner C, Schroeder J, Zorger N, Witzgall R, [Suppl 1]: S241–S251, 2009 Reinhold SW, Stanescu HC, Bockenhauer D, Jaureguiberry G, 27. Chang Y, Twiss JL, Horoupian DS, Caldwell SA, Johnston KM: In- Courtneidge H, Hall AM, Wijeyesekera AD, Holmes E, Nicholson herited syndrome of infantile olivopontocerebellar atrophy, micro- JK, O’Brien K, Bernardini I, Krasnewich DM, Arcos-Burgos M, Izumi Y, nodular cirrhosis, and renal tubular microcysts: Review of the NonoguchiH,JiaY,ReddyJK,IlyasM,UnwinRJ,GahlWA,WarthR, literature and a report of an additional case. Acta Neuropathol 86: Kleta R: Mistargeting of peroxisomal EHHADH and inherited renal 399–404, 1993 Fanconi’ssyndrome.NEnglJMed370: 129–138, 2014 28. de Lonlay P, Seta N, Barrot S, Chabrol B, Drouin V, Gabriel BM, Journel 42. Rao SS, Huntley MH, Durand NC, Stamenova EK, Bochkov ID, Robinson H, Kretz M, Laurent J, Le Merrer M, Leroy A, Pedespan D, Sarda P, JT, Sanborn AL, Machol I, Omer AD, Lander ES, Aiden EL: A 3D map of Villeneuve N, Schmitz J, van Schaftingen E, Matthijs G, Jaeken J, Korner the human genome at kilobase resolution reveals principles of chro- C, Munnich A, Saudubray JM, Cormier-Daire V: A broad spectrum of matin looping. Cell 159: 1665–1680, 2014 clinical presentations in congenital disorders of glycosylation I: A series of 26 cases. JMedGenet38: 14–19, 2001 29. Hertz-Pannier L, Déchaux M, Sinico M, Emond S, Cormier-Daire V, Saudubray JM, Brunelle F, Niaudet P, Seta N, de Lonlay P: Congenital This article contains supplemental material online at http://jasn.asnjournals. disorders of glycosylation type I: A rare but new cause of hyperechoic org/lookup/suppl/doi:10.1681/ASN.2016121312/-/DCSupplemental.

2538 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 2529–2539, 2017 www.jasn.org CLINICAL RESEARCH

AFFILIATIONS

† *Pediatric Endocrinology, Hospital Infantil Universitario Niño Jesús, Madrid, Spain; University of Exeter Medical School, Institute of ‡ Biomedical and Clinical Science, Exeter, United Kingdom; University College London Centre for Nephrology, University College London, | London, United Kingdom; §Pediatric Nephrology, Hospital Universitario Reina Sofía, Córdoba, Spain; Instituto de Investigación La Princesa, Universidad Autónoma de Madrid, Madrid, Spain; ¶Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición (CIBEROBN), Instituto de Salud Carlos III, Madrid, Spain; **Madrid Institute for Advanced Studies on Food, Comité de Ética de la Investigación de la Universidad Autónoma de Madrid, and Centro Superior de Investigaciones Científicas, Carretera de Cantoblanco 8.28049, Madrid, †† ‡‡ Spain; Walter-Brendel-Center of Experimental Medicine, Ludwig-Maximilians-University Munich, Munich, Germany; University of Bristol and Bristol Royal Hospital for Children, Bristol, United Kingdom; §§Diabetes Research Group, King’s College, London, United Kingdom; || Great Ormond Street Hospital for Children NHS Foundation Trust, London, United Kingdom; ¶¶Renal Division, Department of Medicine, Faculty of Medicine, University of Freiburg, Freiburg, Germany; ***Starship Children’s Hospital, Liggins Institute, University of Auckland, ††† ‡‡‡ Auckland, New Zealand; East of Scotland Genetic Service, Dundee, United Kingdom; Pediatric Endocrinology, Hospital Clínico Universitario Virgen de la Arrixaca, Murcia, Spain; §§§Centro de Diagnóstico de Enfermedades Moleculares, Universidad Autónoma de ||| Madrid, Center for Biomedical Research in Rare diseases, Instituto de Investigacion Hospital Universitario La Paz, Madrid, Spain; University College London Institute of Child Health, London, United Kingdom; ¶¶¶Clinical Genetics, Royal Devon and Exeter NHS Foundation Trust, †††† Exeter, United Kingdom; ****Children’s National Health System, Washington, DC; and Department of Pediatric Medicine, Sidra Medical and Research Center, Doha, Qatar

J Am Soc Nephrol 28: 2529–2539, 2017 Polycystic Kidney Disease with Hyperinsulinemic Hypoglycemia 2539 Supplementary Appendix

Table of Contents

DETAILED METHODS 2 Next generation sequencing 2 Luciferase promoter studies 3 Assessment of allele specific expression of PMM2 in primary renal cells 4 Allelic discrimination of PMM2 in patient derived cells 6 Digital PCR sample preparation 7 Data analysis from digital PCR 8 Electrophoretic Mobility Shift Assay (EMSA) 9 Insulin secretion studies 11 Bioinformatic analysis of the promoter mutation 12 Bioinformatic analysis of HNF4A binding sites 13

SUPPLEMENTAL RESULTS 14 Chromatin loops 14 Specificity of HNF4A binding sites 15

SUPPLEMENTAL FIGURE LEGENDS 16 Supplemental Figure 1: Pedigrees of the 11 Families 16 Supplemental Figure 2: Bioinformatic analysis of the PMM2 genomic region suggests an organ- specific disease mechanism 17 Supplemental Figure 3: Cartoon of a potential disease mechanism 18

SUPPLEMENTAL REFERENCES 19

1 Detailed Methods

Next generation sequencing DNA from the patients in families 2 (2.1, 2.2), 6 (6.1), and 8 (8.1) underwent commercially available exome capture (Agilent SureSelect V4) and next generation sequencing (Perkin Elmer, Branford, CT, USA). Data analysis was performed with Ingenuity Variant Analysis (Version 4.1.20160602; Qiagen,

Redwood City, CA, USA). In these four individuals 715,799 variants were noted.

457,942 variants remained after quality control (call quality ≥ 20). Excluding common variants (allele frequency ≥ 1 % in either of the 1000 Genomes project

(phase 3v5b), ExAc (0.3), or NHLBI ESP exomes (ESP6500SI-V2) left 74,187 variants. Of those 24 variants were within the region of linkage (i.e. rs11077126

(NC_000016.10:g.7719191) – rs117691315 (NC_000016.10:g.9925462).

Applying a recessive model of inheritance showed three variants in a

(compound) heterozygote state in a single gene only (i.e. a heterozygote promoter mutation c.-167G>T in PMM2 in all patients (2.1, 2. 2, 6.1, 8.1) and another heterozygote mutation in PMM2 in each patient (c.422G>A (2.1, 2.2,

6.1); c.255+1G>A (8.1)).

Exome sequencing was performed separately also for patients 9 and 10 after exome capture with Agilent SureSelect v4. We aligned reads to hg19 using BWA and called variants using GATK as previously described.1 Ninety-one percent of target bases were covered at ≥20X. Excluding common variants as above we identified 10 variants in the region of linkage across the two patients. The only promoter or coding variants were c.-167G>T in PMM2 (heterozygous in both patient 9 and 10) and p.F157S (heterozygous in patient 9) and p.F207S

2 (heterozygous in patient 10). We performed whole genome sequencing on DNA from patient 7 at Complete Genomics as previously described.2 (Variants were aligned to NCBI build 37. 96.2% of bases were covered at ≥20X and 97.1% of the genome was fully called. A total of 3921750 SNPs and indels were called across the genome. In the linkage region there were a total 115 variants called after filtering out any variant at ≥1% in the 1000 Genomes project or present in

69 publically available genomes from Complete Genomics that were not flagged as clinically associated in dbSNP.2 Of these, only two variants were coding

(p.R141H in PMM2) or occurred in the promoter region of a gene (c.-167G>T in

PMM2).

Luciferase promoter studies The LightSwitch Promoter Reporter construct (S714930, Activemotif, UK) having the promoter region of human PMM2 gene cloned 5’ of the Renilla luciferase gene (RenSP) was used to assess promoter activity for different mutations. This construct was generated through cloning of the 960 bp 5’ to the PMM2 startcodon into the 3656 bp vector pLightswitch_Prom. The fragment was cloned between the MluI and BglII restriction sites. Two different mutations were generated: the patients’ mutation, c.-167G>T, and the c.-163_-180del, which deletes the ZNF143 binding site, detailed previously as SBS1.3 The mutated constructs were generated using the Quickchange II XL Site-directed mutagenesis kit (200522, Agilent Technologies, California, USA) according to the manufacturer’s protocol. The following oligos were used for the mutagenesis:

5’-CTGCGTTGCACCCTTGGAGTTGCGGTC-3’,

5’-GACCGCAACTCCAAGGGTGCAACGCAG-3’,

3 5’-GGCGGGCGTGATCTTGCGGTCCAGGA-3’ and

5’-TCCTGGACCGCAAGATCACGCCCGCC-3’.

All reporter constructs were sequence verified. The empty vector (Active Motif,

S790005) was used as negative control.

Two different cell lines were transfected with all reporter constructs. The human renal proximal tubular cell line RPTEC/TERT1 (LGC standards, ATCC-CRL-

4031, UK) was transfected in Lipofectamine® 2000 (Invitrogen, Thermo Fisher

Scientific), 24 hours after seeding (30,000 cells/well in 96 well plates) with 200 ng of each experimental construct (wild type or mutated promoter or empty vector). The human pancreatic Beta-cell line 1.1B4 (Sigma Aldrich/ECACC,

10012801-1VL, UK) was transfected in Lipofectamine® 2000, 24 hours after seeding (7,500 cells/well in 96 well plates) with 50 ng luciferase construct. After

24 hours, cell lysates were prepared for the Renilla luciferase component of the

LightSwitch™ Dual Assay Kit (DA010, Active Motif, UK) according to the manufacturer’s protocol and the luminescence was measured on a multimode

DTX880 plate reader (Beckman Coulter, UK) for 2s. Each transfection experiment and luciferase assay was done in triplicate. The empty reporter vector was used as negative control and luminescence readouts were plotted relative to the negative control.

Assessment of allele specific expression of PMM2 in primary renal cells Primary renal cells

Primary renal cell cultures were obtained from the renal cortex of the nephrectomy specimen from patient 6.1, compound heterozygous for the PMM2 promoter mutation (c.-167G>T) and a missense mutation (c.422G>A). Cortical tissue was separated from medulla and minced into explants <1 mm3. Explants

4 were plated into pre-wetted flasks (Corning) with multiple explants in each flask and cultured in a minimal volume of 4.5 g/L D-Glucose DMEM GlutaMax™-I (Life technologies, 31966-047) supplemented with 10% fetal calf serum (Sera

Laboratories International Ltd), 1% antibiotic-antimycotic (Sigma,A5955-100ML) at 37°C, 5% CO2. Explants were left undisturbed for at least 1 week to maximise attachment and the initiation of cellular outgrowth. Thereafter the medium was changed every 4-5 days taking care not to dislodge the explants from the substrate. At confluence cellular outgrowth and explants were trypsinised

(Trypsin/EDTA Life Technologies 253200-062) explants sedimented by gravity and cell suspension plated in new flasks, cells were cultured under standard conditions (37°C, 5% CO2) with the medium changed (every 4-5 days). Cells were passaged at confluence and split 1:3. Cells were further maintained in 4.5 g/L D-Glucose DMEM GlutaMax™-I (Life Technologies, 31966-047) with 10% fetal bovine serum (Life Technologies, 10270-106), 2.5 µg/mL Amphotericin B

(PAA laboratories, P11-001) and penicillin/streptomycin (1000 U/mL) (Life

Technologies, 15140-122) at 37°C, 5% CO2 and grown until about 80% confluence with a regular medium change every three to four days.

RNA isolation

RNA was isolated from primary renal cell cultures at different passages and days

(passage 3, 2 extractions; passage 4, 1 extraction). RNA Isolation was performed using TRIzol® (Life Technologies, 15596018), according to the manufacturer’s protocol for adherent cells. RNA was dissolved in UltraPure

DEPC-treated Water (Sigma, D5758-100ML). The concentration of RNA was determined using the Nanodrop ND1000 spectrophotometer (Thermo Fisher)

5 and concentrations were in the range of 150 – 200 ng/µL per sample. RNA was stored at -80°C. cDNA synthesis

Each of the three RNA isolates was transcribed into cDNA by a one-step reverse transcription reaction using the iScript™ Reverse Transcription Supermix for RT- qPCR Kit (Bio-Rad, 1708840). A total amount of 1000 ng per RNA sample was set into the cDNA synthesis step. As suggested by the manufacturers’ protocol a no-reverse-transcriptase control for each RNA template and a no-template control were used as negative controls. cDNA samples were stored at -20°C.

Allelic discrimination of PMM2 in patient derived cells Absolute quantification of allelic discrimination was implemented and analysed using the chip-based QuantStudio™ 3D Digital PCR system (Thermo Fisher) and the QuantStudio™ 3D AnalysisSuite Cloud Software (Thermo Fisher) respectively. A custom allele specific assay, containing a primer/probe mix was designed using the TaqMan® Assay Design Tool (custom TaqMan® SNP

Genotyping Assay 40x, Thermo Fisher, 4331349, Assay-ID: AHGJ9ED).

We used the missense mutation c.422G>A to discriminate the two patient alleles, as c.422G is linked with the mutant and c.422A with the wild type promoter.

The probes were differently labelled with a fluorescent dye (FAM™ or VIC®) for discrimination of the expression levels of each PMM2 allele. Both probes were designed to span the nucleotide position 422 (Ref.-Seq.: NM_000303.2) of

PMM2 within the cDNA of patient and control. Therefore, the FAM™-labelled probe (FAM™-5’-CCAAGAAGAACGCATTGA-3’-MGB) was used to detect the

6 allele with the mutant promoter and the VIC®-labelled probe (VIC®-5’-

CCAAGAAGAACACATTGA-3’-MGB) was specific for the allele with the wildtype promoter. For amplification, the assay also consisted of the following forward and reverse primers: 5’-GTGTCCCCTATTGGAAGAAGCT-3’ and

5’-CTTTTGTCTTATATTTTCTTTTTTATCGAGTTCGTAGA-3’, respectively.

Control constructs (PMM2 wt/mut)

As positive controls, two linearized plasmid constructs (pcDNA™3.1(-

)_PMM2_wt.dna; pcDNA™3.1(-)_PMM2_mut.dna) were used in a 1:1 ratio, mimicking the two PMM2 alleles of the patient with respect to the c.422G>A mutation. For construct design, the mammalian expression vector pcDNA™3.1(-

) (5.4 kb; Invitrogen, V79520) was used and linearised with BAM HI and Xho I for ligation of the insert. The insert for both, wild type and mutant construct, was generated via PCR resulting in a 349 bp fragment of the PMM2 gene. The primers used, also introduced the restriction sites for BAM HI and Xho I on both ends of the amplicon to make ligation in correct orientation possible. Thereafter, the constructs were transformed into E.coli for recombination. After selection

(ampicillin resistance) of positive clones the desired plasmid containing the

PMM2 wt or mutant fragment was purified via plasmid preparation and linearized through BAMHI and XhoI digestion.

Digital PCR sample preparation Twenty-five nanogram per patient cDNA sample (based on previous RNA determination) and a total of 200 fg of the two constructs (control) in a 1:1 ratio

(100 fg each) were used to ensure an optimal concentration of approximately

7 200 to 2000 copies/µL of target sequence in the final reaction. In addition, three different negative controls were measured: A no RT patient control (pool of all three RNA isolates set into RT PCR, Bio-Rad, 1708840) with a total RNA amount of 25 ng, a no template control (Bio-Rad, 1708840) and the two constructs individually (200 fg each).

All samples were made to a volume of 6 uL using nuclease free water (Sigma,

W1754-1VL). Each (n=3) cDNA sample was measured once and for the control, the 1:1 construct mix was prepared three times independently (n=3) and each was assayed once. For the final reactions, master mix, custom TaqMan® SNP

Genotyping Assay (1:2 dilution), nuclease-free water and sample were combined and a total volume of 14.5 µL was loaded on each chip according to the manufacturer’s recommendations (Thermo Fisher, MAN0007720). All dPCR reactions were carried out using the GeneAmp PCR system 9700 (Thermo

Fisher) and samples were amplified with the following conditions: Initial denaturation at 96°C for 10 mins followed by 39 cycles of 60°C for 2 mins; 98°C for 30 sec and 60°C of 2 mins for extension (see also Thermo Fisher,

MAN0007720).

Data analysis from digital PCR All chips (20,000 partitions of 755 pL) were read on the QuantStudio™ 3D Digital

PCR Instrument (chip reader). The data were then imported into a web-based

QuantStudio™ 3D AnalysisSuite™ Cloud Software for further analysis.

Quality threshold was determined by measuring the amount of ROX-signal of the master mix. For each chip, the quality threshold was set to a default value of

0.5 and the target number of copies/µL did never exceed 2,000. The clusters in the scatterplot of the AnalysisSuite software were examined to judge both, adequacy of separation and outliers, and adjusted manually. Only chips that met

8 the aforementioned criteria were incorporated in the pooled statistical analysis that was subsequently done using Microsoft Office Excel. To assess significance of expression levels of the PMM2 alleles in the patient a two-tailed and paired t- test was done. For comparison of patient versus control a two-tailed and unpaired t-test was done for each allele separately.

Electrophoretic Mobility Shift Assay (EMSA)

Oligonucleotides

All oligonucleotides were obtained from Integrated DNA Technologies; two 29- mer 5’-biotinalyted double stranded oligonucleotide probe containing either the wild type or mutant PMM2 promoter sequence c.-150_-178:

5’-GATCTGCGTTGCACCCTGGGAGTTGCGGT-3’ (wild type) and

5’-GATCTGCGTTGCACCCTTGGAGTTGCGGT-3’ (mutant).

The same two sequences were also purchased unlabeled for the competition experiments.

ZNF143 expression

ZNF143 expression was performed using the TNT T7 Quick Coupled

Transcription/Translation System (Promega, L1170) and confirmed by Western analysis using an antibody (Origene, 4C5 anti DDK-Ab, TA50011) that detects the DDK tag. Expression of this transcription factor was achieved using a

PrecisionShuttle pCMV6-Entry vector consisting of a T7 promotor, c terminal

(myc and DDK) tag and the open reading frame (ORF) for ZNF143 (Origene,

RC205850). The same vector without the ORF for ZNF143 was used as a negative control.

9

EMSA

EMSA was performed essentially as described by Garner and Revzin,4 but with slight modifications; binding reactions and detection of the 5’-biotinylated double stranded oligonucleotide probes corresponding to the c.-156_-184 fragment of the PMM2 and TMEM186 bidirectional promotor containing the predicted binding site for ZNF143 were performed using the LightShift Chemiluminescent

EMSA Kit (Thermoscientific, 20148) as described by the manufacturer’s instructions.

Briefly, binding reactions were performed using 1 X binding buffer (10 mM Tris,

50 mM KCl, 1 mM DTT; pH 7.5), 2.5 % glycerol, 5 mM MgCl2, 50 ng/µL of poly(dI- dC), 0.05% NP40, 2 µL of the respective TnT reaction (vector only or vector containing the ORF of ZNF143) reaction and 2 pmol of biotinylated probes in a total volume of 20 µL for 20 minutes at room temperature. Competition were performed by adding 200-fold excess of unlabeled probes to the binding reaction prior to addition of expressed transcription factor. Five microliter of loading dye was added to each of the samples and all were subjected to gel electrophoresis at 4oC using a pre-electrophorese 4 % native polyacrylamide gel in 0.5 X TBE buffer at a constant 100 V. DNA and protein were then electroblotted onto a pre- soaked Biodyne B Nylon Membrane in 0.5 % TBE. Transfer was performed at

4oC in the Xcell SureLock Mini-Cell system at a constant 380 mA for 30 minutes using pre-cooled 0.5 % TBE. Once transferred, the DNA was crosslinked to the membrane by exposing it to a transilluminator for 15 minutes. The membrane was immunoblotted with a 1:300 dilution of streptavidin-horseradish peroxidase conjugate antibody in block buffer. Detection of the biotin-labelled DNA was

10 achieved by chemiluminescence. Blots were imaged using the Syngene

Dyversity imaging system and images were analyzed using the Syngene software. EMSA was performed in 5 independent experiments.

Densitometry

ImageJ version 1.42q was used to select and determine the mean grey value in a given area of the bands of interest and the same area representing the background noise. The value obtained from the background noise was used to subtract the value obtained from each of the bands. Results from 5 independent experiments were assessed. A two-tailed and unpaired t-Test was performed on these results.

Insulin secretion studies Insulin-secreting MIN6 cells5 were kindly provided by Prof J I Miyazaki

(University of Tokyo, Japan) and maintained at 37°C (95% air/ 5% CO2) in

DMEM supplemented with 10% fetal calf serum (FCS), 2mM L-glutamine, 100

U/ml penicillin and 0.1 mg/ml streptomycin. For measurements of insulin secretion MIN6 cells were seeded into 96-well plates at a density of 30,000 cells/well and maintained as adherent monolayers under standard tissue culture conditions. To deglycosylate cell surface proteins the cells were incubated

(37°C, 120 min) in a physiological salt solution6 supplemented with 2mM glucose, 2mM CaCl2 and 0.5 mg/ml BSA in the absence or presence of Peptide

N-Glycosidase F (PNGase F, 4000U/ml, New England Biolabs) and

Endoglycosidase H (Endo H, 4000U/ml, New England Biolabs). Immediately following deglycosylation the cells were washed and incubated (37°C, 60 min,

200µl/well) in the physiological salt solution supplemented with 2mM glucose, or

11 20mM glucose, or 20mM glucose plus the protein kinase C activator, phorbol-

12 myristate-13 acetate (PMA, 500nM). At the end of the incubation period supernatant samples were removed and insulin was measured by radioimmunoassay as described previously.7 Differences in insulin secretion between treatment groups were assessed by two-way ANOVA.

Bioinformatic analysis of the promoter mutation The genomic region on chromosome 16 including surrounding PMM2 was analysed with the UCSC Genome browser8 using human genome data (HG19).

Transcription factor binding was assessed with the UCSC Genome Browser transcription factor binding site track (TFBS) using ENCODE ChiP-Seq data from 67 cell lines (March 2012 Freeze). Chromatin segmentation was assessed using all available ENCODE data from human cell lines (GM12878, H1-hESC,

K562, HeLa-S3, HepG2, HUVEC).

Prediction of chromatin looping in the genomic region on chromosome 16 was performed by analyzing ENCODE CHiP-Seq data for binding of CTCF and markers of Cohesin (RAD21 and SMC3) and assessing of the orientation of the

CTCF sites, so that loops would be delimited by CTCF in convergent orientation.9

We further assessed evidence for loop formation in this region by analyzing Hi-

C data from Rao et al.10 Data from this project had been deposited in Gene

Expression Omnibus (GEO) under the accession number GSE63525. Data on loops involving PMM2 were available for 3 cell types: K562, IMR90 and

GM12878. We further analyzed ENCODE ChIA-PET data from the cell lines with available CTCF data: K562 and MCF-7.

12 We further assessed whether chromatin formation may be different in tissues affected by HIPKD compared to those that are not. Cell lines with available

CHiA-PET or Hi-C data are either derived from various forms of cancer (K562,

GM12878, MCF-7) or fetal lung (IMR90) and thus may not be representative of tissues affected by HIPKD. We therefore looked at indirect evidence for altered chromatin architecture by assessing ENCODE data for accessibility to formaldehyde (FAIRE mapping), DNase I or transcription factor binding,

Bioinformatic analysis of HNF4A binding sites HNF4A binding sites identified from ChIP-seq (Supplemental Figure 2) were analyzed for HNF4A specificity using a positional weight matrix based on the data from Fang et al11, as well as their web-based tool

(http://nrmotif.ucr.edu/NRBSScan/H4SBM.htm).

13

Supplemental Results

Chromatin loops Analysis of CTCF and Cohesin binding sites and CTCF orientation identified a potential loop including PMM2 spanning approximately 220 kb (chr 16: 8754981 to 8974276), as depicted in Supplemental Figure 2. Direct evidence for the existence of this predicted loop was provided from Hi-C data from the K562 and

IMR90 cell lines and ChIA-PET data from MCF-7 cells. (Supplemental Figure 2).

Analysis of chromatin segmentation, FAIRE mapping, DNase I and transcription factor binding shows a distinct and unique pattern for HepG2 cells compared to other cell types (data not shown, except for chromatin segmentation in

Supplemental Figure 2), suggesting most notably increased chromatin accessibility over the ABAT gene. HepG2 is a liver-derived cell line and thus likely representative of HIPKD affected tissue. The altered chromatin accessibility coincides with the presence of HNF4A in HepG2 cells, which binds to a number of sites in ABAT and expression of ABAT in this cell line; moreover, novel peaks of DNase I hypersensitivity in HepG2 cells predominantly occur at

HNF4A binding sites, consistent with altered chromatin structure in HNF4A expressing cells, which may interfere with PMM2 transcription from the mutant promoter.

14 Specificity of HNF4A binding sites HNF4A belongs to a nuclear receptor family including RXR and COUPTF2 that can share a classical DNA binding motif (AGGTCA). However, a similar motif

(CAAAGTCCA) is highly specific for HNF4A binding.11 We analyzed the 13

HNF4A binding sites identified in the PMM2 containing loop (Supplemental

Figure 2) and noted that the binding sites at the 5’ and 3’ end of the loop

(assessed as strong binding sites by ENCODE analysis of ChIP-seq data and detailed in Supplemental Figure 3), located in the respective ABAT and

CARHSP1 promoter best fit the HNF4A specific binding motif, consistent with the hypothesis of HNF4A being the key transcription factor affecting tissue- specific PMM2 transcription.

15 Supplemental Figure legends

Supplemental Figure 1: Pedigrees of the 11 Families Shown are the pedigrees of the 17 patients with HIPKD described here. Dark symbols denote affected, empty symbols unaffected individuals. A small black circle in an empty symbol indicates a parent in whom the carrier state for the respective mutation of the affected child has been confirmed. Note that inheritance is consistent with being autosomal recessive.

*: Families used for linkage analysis (Figure 1)

M: carrier of the promoter mutation c.-167G>T

M: carrier of a coding mutation in PMM2

16 Family 1* c.-167G>T/c.-167G>T Family 4 Family 2* Family 3 c.-167G>T/p.R141H c.-167G>T/p.R141H c.-167G>T/p.F157S

M/N M/N M/N M/N

M/N M/N M/N M/N M/N M/N 4.1 3.1 2.1 2.2 M/M M/M M/M M/M 1.1 1.2 1.3 1.4 M/M M/M M/M M/M

Family 8* Family 7 Family 5 Family 6* c.-167G>T/c.255+1G>A c.-167G>T/p.R141H c.-167G>T/p.R141H c.-167G>T/p.R141H M/N M/N M/N M/N M/N M/N

8.1 7.1 5.1 6.1 M/M M/M M/M M/M

Family 11* Family 9 Family 10 c.-167G>T/p.F157S c.-167G>T/p.F207S c.-167G>T/p.F157S

M/N M/N M/N M/N M/N

9.1 9.2 11.1 10.1 10.2 M/M M/M M/M M/M M/M Supplemental Figure 2: Bioinformatic analysis of the PMM2 genomic region suggests an organ-specific disease mechanism A) Bioinformatic analysis. Shown is the genomic region on chromosome

16p13.2. Chromatin interaction analysis (ChIA-PET, Hi-C) identifies a loop of chromatin (size approximately 220 kb) that includes PMM2 (HG19). The upper part of the picture shows the genes in the region. Below are three tracks indicating DNA binding sites based on ENCODE analysis of ChiP-seq data, thus reflecting transcription factor binding in experimental cells. Grayscale according to strength.

1) CTCF: Note the binding sites (yellow arrows), which are in convergent

orientation, as expected for CTCF sites limiting chromatin loops.

2) ZNF143: Note the strong ZNF143 binding site immediately 5’ of PMM2

(green arrow), where the c.-167G>T mutation is located.

3) HNF4A: Note the clusters of HNF4A binding sites at the flanks of the

predicted loop, the closest of which is <2 kb from the CTCF site.

The lower 6 tracks show chromatin segmentation from the six available cell lines

(ENCODE). Color-coding as follows: gray: repressed or low activity region, blue:

CTCF-enriched elements (insulator regions), red: promoter region, green: transcribed region, orange: enhancer, yellow: weak enhancer.

Note that there is evidence for general expression (all 6 cell lines) on the centromeric side of the PMM2 promoter (to the right of the green arrow) and specific expression (only in the hepatic cell line, left arrowhead) on the telomeric side of the promoter (to the left of the green arrow) consistent with organ-specific chromatin regulation.

17 > Supplemental Figure 3: Cartoon of a potential disease mechanism Simplified cartoon (not to scale) depicting the proposed disease mechanism. A)

Wild type promoter: Note the establishment of a chromatin loop delimited by binding sites for CTCF (yellow). Binding of ZNF143 to the wild type PMM2 promoter (green) and to CTCF changes the 3-dimensional structure of the loop

(the functional loop) and thus can allow interaction between the PMM2 promoter and clusters of HNF4A binding sites (black). B) Mutant promoter: there is impaired binding of ZNF143 to the mutant PMM2 promoter (red) with altered loop formation in HNF4A expressing cells and thus impaired PMM2 transcription. HNF4A is expressed predominantly in the organ systems involved in HIPKD: pancreatic beta cells, kidney and liver.

18 A

B X X X

Wildtype Promoter ZNF143 CTCF HNF4A sites Mutant Promoter Supplemental References

1. Flanagan, SE, Haapaniemi, E, Russell, MA, Caswell, R, Lango Allen, H, De Franco,

E, McDonald, TJ, Rajala, H, Ramelius, A, Barton, J, Heiskanen, K, Heiskanen-

Kosma, T, Kajosaari, M, Murphy, NP, Milenkovic, T, Seppanen, M, Lernmark, A,

Mustjoki, S, Otonkoski, T, Kere, J, Morgan, NG, Ellard, S, Hattersley, AT:

Activating germline mutations in STAT3 cause early-onset multi-organ

autoimmune disease. Nature genetics, 46: 812-814, 2014.

2. Drmanac, R, Sparks, AB, Callow, MJ, Halpern, AL, Burns, NL, Kermani, BG,

Carnevali, P, Nazarenko, I, Nilsen, GB, Yeung, G, Dahl, F, Fernandez, A, Staker, B,

Pant, KP, Baccash, J, Borcherding, AP, Brownley, A, Cedeno, R, Chen, L,

Chernikoff, D, Cheung, A, Chirita, R, Curson, B, Ebert, JC, Hacker, CR, Hartlage, R,

Hauser, B, Huang, S, Jiang, Y, Karpinchyk, V, Koenig, M, Kong, C, Landers, T, Le,

C, Liu, J, McBride, CE, Morenzoni, M, Morey, RE, Mutch, K, Perazich, H, Perry, K,

Peters, BA, Peterson, J, Pethiyagoda, CL, Pothuraju, K, Richter, C, Rosenbaum,

AM, Roy, S, Shafto, J, Sharanhovich, U, Shannon, KW, Sheppy, CG, Sun, M,

Thakuria, JV, Tran, A, Vu, D, Zaranek, AW, Wu, X, Drmanac, S, Oliphant, AR,

Banyai, WC, Martin, B, Ballinger, DG, Church, GM, Reid, CA: Human genome

sequencing using unchained base reads on self-assembling DNA nanoarrays.

Science, 327: 78-81, 2010.

3. Anno, YN, Myslinski, E, Ngondo-Mbongo, RP, Krol, A, Poch, O, Lecompte, O,

Carbon, P: Genome-wide evidence for an essential role of the human

Staf/ZNF143 transcription factor in bidirectional transcription. Nucleic acids

research, 39: 3116-3127, 2011.

19 4. Garner, MM, Revzin, A: A gel electrophoresis method for quantifying the binding

of proteins to specific DNA regions: application to components of the

Escherichia coli lactose operon regulatory system. Nucleic acids research, 9:

3047-3060, 1981.

5. Ishihara, H, Asano, T, Tsukuda, K, Katagiri, H, Inukai, K, Anai, M, Kikuchi, M,

Yazaki, Y, Miyazaki, JI, Oka, Y: Pancreatic beta cell line MIN6 exhibits

characteristics of glucose metabolism and glucose-stimulated insulin secretion

similar to those of normal islets. Diabetologia, 36: 1139-1145, 1993.

6. Gey, GO, Gey, MK: Maintenance of human normal cells in continuous culture:

preliminary report. In Cultivation of mesoblastic tumors and normal cells and

notes on methods of cultivation. American Journal of Cancer, 27: 45-76, 1936.

7. Jones, PM, Salmon, DM, Howell, SL: Protein phosphorylation in electrically

permeabilized islets of Langerhans. Effects of Ca2+, cyclic AMP, a phorbol ester

and noradrenaline. The Biochemical journal, 254: 397-403, 1988.

8. Kent, WJ, Sugnet, CW, Furey, TS, Roskin, KM, Pringle, TH, Zahler, AM, Haussler,

D: The human genome browser at UCSC. Genome Res, 12: 996-1006, 2002.

9. Dekker, J, Mirny, L: The 3D Genome as Moderator of Chromosomal

Communication. Cell, 164: 1110-1121, 2016.

10. Rao, SS, Huntley, MH, Durand, NC, Stamenova, EK, Bochkov, ID, Robinson, JT,

Sanborn, AL, Machol, I, Omer, AD, Lander, ES, Aiden, EL: A 3D map of the

human genome at kilobase resolution reveals principles of chromatin looping.

Cell, 159: 1665-1680, 2014.

11. Fang, B, Mane-Padros, D, Bolotin, E, Jiang, T, Sladek, FM: Identification of a

binding motif specific to HNF4 by comparative analysis of multiple nuclear

receptors. Nucleic acids research, 40: 5343-5356, 2012.

20