Family enriched for diabetes and end-stage renal disease

Individuals from the family examined in this study were enrolled in the Joslin Study on the Genetics of Type 2 Diabetes between 1993 and 200315. Briefly, between 1993 and 2003, families with an apparent autosomal dominant mode of inheritance of type 2 diabetes (T2D), irrespective of their nephropathy status, were recruited to the Joslin Study of Genetics of Nephropathy in Type 2 Diabetes Family Collection through T2D probands receiving medical care at the Joslin Clinic. The protocols and written informed consent procedures used in this study were approved by the Committee on Human Subjects of the Joslin Diabetes Center.

All participating family members provided written informed consent prior to participating in this study. All family members were of European ancestry. After obtaining informed written consent, trained recruiters administered previously described study protocols that included a structured interview, seated blood pressure measurements, and the collections of blood and urine samples. All blood and urine specimens have been stored at −80 °C since the time of their collection. ESRD status for members of this family was updated through the United States Renal Data System.

Genetic analyses

Whole genome sequencing of DNA samples from six individuals with ESRD was performed on the Illumina HiSeq2500 platform available at Macrogen (Rockville, MD). FASTQ files from paired-end sequencing were subsequently aligned to the Hg19/GRCh37 reference genome using the Sentieon Genomics DNASeq software pipeline per the developer’s guidelines34. The DNASeq pipeline incorporates the Broad Institute’s BWA-GATK Best Practices Workflow but performs alignments and variant calling in a computationally efficient process. The six individual samples were jointly genotyped with 524 ethnically matched background controls comprised of 291 individuals from the 1000 Genomes Project and 233 healthy samples from the Utah Genome Project35. The pipeline created individual genomic variant call files and a final jointly-called variant call file (VCF). Sample qualities in final VCF were evaluated using Peddy36 to determine and confirm sex, relatedness, heterozygosity and ancestry of each individual to identify any potential sample quality issues. The VCF was functionally annotated using ANNOVAR for downstream analyses37.

A unified linkage analysis and rare variant association testing using pedigree Variant Annotation, Analysis, and Search Tool (pVAAST), version 2.2.0, was performed on six individual sample target genomes and 524 ethnically matched background controls35. pVAAST incorporates pedigree information into disease-gene prioritization procedures that incorporates amino acid substitution and calculates burden of each gene candidate. pVAAST was performed using 1 × 10−7 permutations and configured to evaluate penetrance values in the range of 0.5–0.995, with LOD and CLRT filters removed. In keeping with best practices from the developers of this tool, we considered any gene with a variant achieving a pVAAST p value < 0.005 as a potential positive signal. Additional reranking of pVAAST outputs occurred using Phenotype Driven Variant Ontological Re-ranking tool (PHEVOR) using Human Phenotype Ontology (HPO) terms and connections to Gene Ontology (GO) terms to prioritize potentially damaging alleles, using the HPO terms a) for kidney disease: HP:0000077 (Abnormality of the kidney), HP:0000112 (Nephropathy), and HP:0003774 (Stage 5 chronic kidney disease), and b) for diabetes: HP:0000819 (Diabetes) and HP:0005978 (Type 2 diabetes mellitus) and a combination of the pVAAST p value and Phevor score38.

Sanger sequencing confirmation of the ADIPOQ variant was performed in all 6 individuals with ESRD in this family as well as 8 additional family members for whom DNA was available. The genomic region covering the variant was PCR amplified using the Qiagen Taq PCR amplification kit (Qiagen, Valencia, CA) per the manufacturer’s protocol and the following primer pair: forward 5′–GGCTGTAACCAACCTAGGCAGG–3′ and reverse 5′–AGGCAAAGTAGTACAGCCCAGG–3′. Sanger sequencing was performed using BigDye Terminator v3.1 Cycle Sequencing chemistry (Life Technologies, Carlsbad, CA). Capillary electrophoresis was performed by the DNA Sequencing Core Facility at the Health Science Center at the University of Utah using a 3730 DNA Analyzer (Applied Biosystems, Foster City, CA). The resulting chromatograms were analyzed using Sequencher version 5.4.6 (Gene Codes Corporation, Ann Arbor, MI).

All 14 family members were genotyped on the HumanCore BeadChip (Illumina, San Diego, CA, USA), which contains 250,000 genome-wide tag SNPs (and other variants) and over 200,000 exome-focused variants, by the DNA Sequencing Core Facility at the Health Science Center at the University of Utah. All samples were passed through a stringent quality control protocol that included filtering for low-quality variants (e.g., call rates <95% and excessive deviation from Hardy–Weinberg equilibrium).

HLA alleles DR3 (haplotype DQA1*05:01-DQB1*02:01) and DR4 (haplotype DQA1*03-DQB1*03:02) were predicted for all 14 family members using genome-wide genotyping data, the SNP2HLA software39, and genome-wide genotyping data from the HLA-region and the Type 1 Diabetes Genome Consortium’s reference genotype-HLA-allele panel.

Circulating adiponectin

Circulating adiponectin levels were measured in serum from 14 family members (including 4 noncarriers and 10 carriers of the ADIPOQ mutation) using an enzyme-linked immunosorbent assay (ELISA) (Millipore EZHADP-61K). Protein concentration was assessed using a BCA kit (Pierce) prior to load equivalent protein levels from all patients. The ELISA was measured using a ThermoFisher Varioskan Lux Microplate Reader.

Fast protein liquid chromatography (FPLC)

FPLC was performed as previously described40. Serum (50 µL) from two representative noncarrier or carrier siblings was injected into an ÄKTA Go FPLC (GE Healthcare). A Superdex 200 10/300 GL column (GE Healthcare) was used to separate adiponectin complexes in HEPES/Ca2+ buffer (25 mM HEPES; 150 mM NaCl; and 1 mM CaCl2, pH 7.4). 250 µL fractions were collected over a 20 mL retention volume. The retention volumes found to contain adiponectin were then used to run Western blots to determine HMW, LMW and trimeric adiponectin. Samples were run on an SDS-Page gel (BioRad Criterion TGX) after being reduced in Laemmli and 355 mM 2-mercaptoethanol and boiled for 10 min. The gel was transferred to PVDF membrane (BioRad). The membranes for each patient were blocked in 5% BSA then probed for adiponectin overnight with rabbit polyclonal anti-adiponectin at a 1:1000 dilution (Abcam, ab75989). There were washed then stained with goat-anti-rabbit AlexaFluor Plus 680 (Invitrogen) at a 1:4000 dilution. Then washed again and imaged on a ThermoFisher iBright system. Western blots derive from the same experiment and were processed in parallel.

Lipidomic analysis

Ceramide levels were measured using targeted lipidomic profiling and liquid chromatography with tandem mass spectrometry (LC-MS/MS) at the University of Utah’s Metabolomics Core. Analysis of lipid species, including C16.0 ceramides, was performed as previously been described16. Briefly, these determinations were obtained from serum specimens from 14 family members as well as 25 unrelated nondiabetic controls using targeted LC-MS/MS platforms at the University of Utah’s Metabolomics Core with each lipid normalized to an internal standard using an established and previously described method developed by the Summers lab16. Statistical analysis was performed using Student’s t test in SAS for Windows, version 9.2 (SAS Institute, Cary, NC) following log10 transformation of these data and results are considered significant at the P < 0.05 level.

Gene expression vectors

Gene sequences for human wild-type and mutant ADIPOQ were cloned into the pVitro2 vector (InvivoGen) expressing blasticidin resistance (Fig. S6). The vector is designed for dual gene expression with both genes driven by two human ferritin composite promoters. Following the addition of wild-type ADIPOQ is an IRES followed by the bsr gene allowing for blasticidin resistance. For cells that express only wild-type or mutant ADIPOQ, the gene sequences were disrupted with a premature stop codon so that only one variant of adiponectin is produced. The vectors were transfected into HEK293T cells (ATCC, CRL-11268) using JetPrime (Genesee Scientific). Twenty-four hours after transfection, 20 µg/mL blasticiden was added to the cell culture media. The cells were grown and split three times to ensure stable expression of the genes. The cells were maintained in 10 cm cell culture dishes in DMEM with 10% FBS, pen/strep, and blasticidin. For cells treated with the proteasome inhibitor MG132 (Sigma), healthy cells were treated for 12–18 h with a concentration of 20 µM in FBS free media.

Western blotting

Stable expressing cells were grown to 80–90% confluency. Then media was removed and supplemented with DMEM with no FBS. Twelve to eighteen hours later the media was collected with the addition of protease inhibitors (Roche) and put-on ice. The cells were lysed using 200 µL of ice cold radioimmunoprecipitation assay buffer containing protease inhibitors (Roche). The cells were vortexed and lightly sonicated to form a homogeneous solution. Protein concentration was assessed using a BCA kit (Pierce). For traditional Western blotting, 2 µg of protein was reduced using Bolt Reducing Agents (ThermoFisher) and heated for 5 min. Protein was run on Invitrogen Bolt 4–12% Bis-Tris gels (ThermoFisher) and then transferred to a nitrocellulose membrane using the Invitrogen Power Blotter. The membranes were then processed for immunodetection using the iBind Western System. Primary antibodies were selected to target adiponectin (Rabbit polyclonal anti-adiponectin from Abcam, ab75989) and beta-actin as a loading control (Mouse monoclonal anti-beta-actin from Cell Signaling, 3700S). Both were used at a 1:1000 dilution. The adiponectin antibody selected for use was made using a peptide that falls within both the wild-type and mutant adiponectin protein. Antibodies targeting the N-terminus and the C-terminus of adiponectin were also used to confirm that the mutant protein is not detected when using a C-terminus antibody (data not shown). The secondary antibodies used were Alexa-fluor Plus 488 Anti-Mouse (ThermoFisher, A32723) and Alexa-fluor Plus 800 Anti-Rabbit (ThermoFisher, A32735) at a 1:2000 dilution. The blots were imaged on a ThermoFisher iBright system. All blots derive from the same experiment and were processed in parallel.

FRET constructs and analysis

To assess if wild-type adiponectin and the mutant adiponectin protein could interact, we designed constructs with each protein tagged with a FRET sensor (Fig. S7). We chose a green fluorescent protein derivative called Clover and a red fluorescent protein derivative called mRuby2 to act as our donor/acceptor pair. Both Clover and mRuby2 were designed and proven to be excellent FRET sensors41. Either Clover or mRuby2 were attached to the protein via a linker sequence at the N-termina (since the same C-terminus is not present in the mutant protein). Individual wild-type adiponectin or mutant adiponectin were labeled with Clover or mRuby2 as single expressing constructs to gather spectral data from individual fluorophores. A double expressing construct with wild-type adiponectin labeled with Clover and wild-type adiponectin labeled with mRuby served as a FRET positive control since it is known that adiponectin forms multimers with itself. A double expressing construct with wild-type adiponectin labeled with Clover and mutant adiponectin labeled with mRuby2 served as the experimental group.

FRET readings were recorded via two methods, acceptor bleaching and sensitized emission.

Acceptor bleaching model

In this model, the acceptor fluorophore is photobleached and donor fluorophore intensities are taken prior and after the photobleaching has occurred. If FRET is occurring, the energy that would have gone toward exciting the acceptor remains with the donor, meaning that the donor intensity is increased after photobleaching the acceptor. The stable expressing cells were grown in coverslip bottom chamber slides using FluoroBrite DMEM (Gibson) with 2% FBS. The slides were mounted into a Pathology Devices, Inc stage heater kept at 37 °C and imaged on a Nikon A1R confocal microscope with a 40× oil immersion objective. Cells were quickly identified and focused using epifluorescence. A pre-bleach image was obtained using a 488 nm argon gas laser to excite Clover and a 561 nm sapphire diode laser to excite mRuby2 using predetermined laser settings for adequate imaging. After a pre-bleach image was obtained the 488 nm laser was turned off and the 561 nm laser was set to 100% power and allowed to excite the cells for 10 min to bleach the mRuby2 signal. Afterwards, a post-bleach image was acquired using the same laser and acquisition settings as the pre-bleach image. Cells containing the empty vector that produce neither Clover nor mRuby2 were imaged to provide background intensities of the cells. The intensity for Clover and mRuby2 in each cell was measured in FIJI imaging software in both the pre- and post-bleach images. The average background intensity from non-expressing cells was subtracted from the intensity measurements of the pre- and post-bleach images. Only cells that had a ≥10% reduction in mRuby2 intensity were used for quantification. FRET efficiency was calculated by \(1 - \frac}}}\), where Fda is the intensity of the donor fluorophore before bleaching and Fd is the intensity of the donor fluorophore after the 10-min bleach. The FRET efficiency percent was graphed using Prism. Intensity modulated ratiometric images were created using RatioImage in MatLab.

Sensitized emission model

In this model, the intensity of the acceptor fluorophore is measured to determine if it gains intensity when the donor fluorophore is physically close. The stable expressing cells were grown in a Greiner µClear bottom black-walled 96-well plate using DMEM with 10% FBS, pen/strep, and blasticidin ensuring that each FRET construct (individual expressing cells and dual Clover/mRuby2 expressing cells) were in triplicate. Once the cells reached 70–80% confluency, the media was replaced with FluoroBrite DMEM (Gibson) with no FBS. The cells were treated with MG132 at this time and incubated for 12–18 h. The plate was then analyzed using a ThermoFisher Varioskan Lux Microplate Reader with the chamber heated to 37 °C and maintaining 5% CO2. The Varioskan was set to read 29 readings from the bottom of each well. To detect Clover, intensity excitation was set to 505 nm and emission detection was at 515 nm. To detect mRuby2, intensity excitation was set to 559 nm and emission detection was at 600 nm. FRET readings for dual Clover/mRuby2 expressing cells were taken with excitation set at 505 nm, to excite Clover and emission detection was set at 600 nm to read mRuby2 fluorescence. Spectral data with each excitation setting were also obtained to verify that the fluorophores emitted at their ideal peak emission curve (data not shown). Fluorescent readings were background subtracted from readings from cells with the empty vector. This FRET method requires subtracting out the spectral bleed-through derived from the donor emission spectrum overlap with the acceptor and the direct excitation of the acceptor by the donor’s excitation light to verify actual FRET. The corrected FRET ratio (FC) was obtained by:

where DSBT is the donor spectral bleed-through and the value obtained from the wild-type adiponectin tagged with Clover in the 505–600 nm reading. ASBT is the acceptor spectral bleed-through and is obtained from the mutant adiponectin tagged with mRuby2 in the 505–600 nm reading. Acceptor intensity is the measure from the 559 nm excitation and 600 nm emission reading. The corrected FRET ratio was graphed using Prism. The experiment was repeated two other times with similar results.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.