Mesenchymal-specific Alms1 knockout in mice recapitulates metabolic features of Alström syndrome

Objective Alström Syndrome (AS), caused by biallelic ALMS1 mutations, includes obesity with disproportionately severe insulin resistant diabetes, dyslipidemia, and fatty liver. Prior studies suggest that hyperphagia is accounted for by loss of ALMS1 function in hypothalamic neurones, whereas disproportionate metabolic complications may be due to impaired adipose tissue expandability. We tested this by comparing the metabolic effects of global and mesenchymal stem cell (MSC)-specific Alms1 knockout. Methods Global Alms1 knockout (KO) mice were generated by crossing floxed Alms1 and CAG-Cre mice. A Pdgfrα-Cre driver was used to abrogate Alms1 function selectively in MSCs and their descendants, including preadipocytes. We combined metabolic phenotyping of global and Pdgfrα+ Alms1-KO mice on a 45% fat diet with measurements of body composition and food intake, and histological analysis of metabolic tissues. Results Assessed on 45% fat diet to promote adipose expansion, global Alms1 KO caused hyperphagia, obesity, insulin resistance, dyslipidaemia, and fatty liver. Pdgfrα-cre driven KO of Alms1 (MSC KO) recapitulated insulin resistance, fatty liver, and dyslipidaemia in both sexes. Other phenotypes were sexually dimorphic: increased fat mass was only present in female Alms1 MSC KO mice. Hyperphagia was not evident in male Alms1 MSC KO mice, but was found in MSC KO females, despite no neuronal Pdgfrα expression. Conclusions Mesenchymal deletion of Alms1 recapitulates metabolic features of AS, including fatty liver. This confirms a key role for Alms1 in the adipose lineage, where its loss is sufficient to cause systemic metabolic effects and damage to remote organs. Hyperphagia in females may depend on Alms1 deficiency in oligodendrocyte precursor cells rather than neurones. AS should be regarded as a forme fruste of lipodystrophy.


Animal studies
Alms1 Global-and MSC-KO mice were group-housed in individually ventilated cages (IVCs) at the Biological Research Facility at the University of Edinburgh where a 12-hour light/dark cycle (lights on 0700; off 1900) and controlled temperature/humidity (19-21°C/50%) were maintained.Prior to 6 weeks old, mice had ad libitum access to standard chow (CRM, Special Diet Service).
Glucose concentration was determined from the tail vein using the AccuChek Performa Nano [Roche, Switzerland].Blood samples were collected in EDTA-coated capillary tubes and placed on ice before spinning at 4°C for 10 mins at 2000g and separating plasma.Plasma was stored at -80°C and thawed for biochemical analysis.Insulin was assayed by electrochemiluminescence immunoassay at the Medical Research Council Metabolic Diseases Unit (MRC-MDU) Mouse Biochemical Assay Laboratory at the University of Cambridge (Supplementary Methods).
Body length was measured using a digital calliper from nose-anus under anaesthesia at 24 weeks prior to cardiac puncture.At 24 weeks, a terminal bleed was undertaken by cardiac puncture under general anaesthesia and tissues were harvested immediately post mortem, weighed, and sectioned before fixing of tissues in 4% paraformaldehyde or snap freezing in liquid nitrogen.

Plasma biochemistry
Plasma biochemical assays, including cytokine assays, were undertaken by the Core Biochemical Assay Laboratory (CBAL) at the Cambridge University Hospitals NHS Foundation Trust.The Siemens Dimension EXL autoanalyser was used to quantify triglyceride, total cholesterol, HDL-C, alanine transaminase (ALT) and aspartate transaminase (AST).LDL-C is calculated from the triglyceride, HDL-cholesterol and total cholesterol concentrations as described in the Friedwald formula (LDL-C = Total cholesterol -HDL-C -(Triglycerides/2.2).Free fatty acids were quantified by enzymatic colorimetric assay and testosterone was quantified by enzyme-linked immunosorbent assay (ELISA); both reactions were measured on the PerkinElmer Victor-3 Plate Reader.Insulin, leptin, adiponectin and inflammatory cytokines were quantified by electrochemiluminescence immunoassay with measurements read on the MesoScale Discovery Sector s600.The reagent and kit details are detailed in Supplementary Table 1.
Adipocyte area was measured in QuPath after manual adipocyte identification using the wand detection tool.Typically this selected entire adipocytes, but some adjustments were required.At least 100 adipocytes were measured per adipose depot, a number sufficient for accurate representation of adipose depots (5).Adipocyte areas from QuPath were binned in GraphPad Prism into 1000μm 2 bins.Calculation of adipocyte numbers were based on assumed adipocyte sphericity, and an assumed adipose tissue density of 0.96g/ml (6), with depot masses divided by mean adipocyte mass to calculate total adipocyte number per depot.Lipid content and PSR, SBB, and DAB staining in adipose and liver sections was measured with an automated self-defined pixel classifier.
For the detection of mTom and mGFP in Pdgfrα:Cre x mTom/mGFP brains, the brains were dissected, post-fixed in formalin and cryoprotected with 30% sucrose at 4°C before coronal sectioning in 5 series at 25 µm using a freezing microtome (8000, Bright Instruments, UK).Sections were stored in anti-freeze solution at 4°C until processing.Dual immunohistochemistry was then performed using methods adapted from those previously described (7).Briefly, sections were washed with PBS with 0.2% Tween20 for 30 min and then PBS alone before blocking in 1% bovine serum albumin and 5% Donkey Serum for 1 hr at room temperature.Incubation with primary anti-RFP (rabbit, 1:1000, 600-401-379, Rockland Immunochemicals, USA) and anti-GFP (chicken, 1:1000, ab13970, Abcam, UK) antibody was for 24 hours at room temperature in blocking solution before washed with PBS with 0.2% Tween20 followed by PBS alone.Incubation with secondary antibodies (1:500, anti-rabbit AlexaFluor594, anti-chicken AlexaFluor488, Invitrogen, UK) was for 1 hr in blocking solution at room temperature.Of note, the usual permeabilization step with Triton X-100 was not used in any of the steps to avoid disruption of the cytoplasmic membrane.Images were acquired using Axioskope2 microscope and Axiovision software (Zeiss, Germany).All images were converted to 8-bit and merged using ImageJ (Fiji).

Gene Expression Analysis
Total RNA was extracted from 50 mg samples of liver and adipose depots by TRIzol TM according to the manufacturer's protocol.Concentration of eluted RNA was measured using the NanoDrop ONE before dilution to 100ng/μL in nuclease-free water.Reverse transcription was undertaken using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) in the Eppendorf Mastercycler X50s, using 1000ng RNA per reaction.cDNA solutions were diluted 1 in 4. Real-time quantitative PCR (RT-qPCR) was performed using TaqMan reagents including minor groove binder probes (8) on a LightCycler® 480 Instrument II (Roche).Gapdh was evaluated as a housekeeping control gene.Primer efficiency was calculated by standard dilution curve prior to experimental reactions, and reactions were run in triplicate with reverse transcriptase negative and non-template controls run on the same plate.Crossing point (Cp) values were calculated using LightCycler 480 software using the Abs quant/ 2nd derivative max function, and values for Alms1 (probe Mm01189441_m1) normalised to Gapdh values after adjusting for primer efficiency, as described by Pfaffl (9).

Supplementary Methods References
Figure S1.Validation of global and MSC-specific Alms1 KO mice.(A) Schematic of Alms1 tm1c and Alms1 tm1d alleles, both generated from the EUCOMM Alms1 tm1c allele.(B) Liver cDNA gel electrophoresis following PCR amplification across area including exon 7 which is 160bp long.Lane 1: DNA ladder; lanes 2&3: Alms1 global WT cDNA amplicon; lanes 4&5: Alms1 global KO cDNA amplicon; lanes 6&7: Alms1 MSC WT cDNA amplicon; lanes 8&9: Alms1 MSC KO cDNA amplicon; lane 10, non-template control.(C-F) qPCR validation of Alms1 loss in global and MSC-specific Alms1 KO mice using a Taqman probe spanning the exon 6-7 junction of Alms1, normalised to Gapdh.Global WT/KO and MSC WT/KO experiments were performed with identical design at different times, reflected in the dotted line separating the two cohorts.N = 4/group.Data represent individual animals with bars representing mean ± sd.Comparison between WT and KO in (C-F) used an unpaired twotailed Student's t-test with Bonferroni correction.iWAT = inguinal white adipose tissue

Figure S3 .
Figure S3.Histological evaluation of gonadal white adipose tissue of global and mesenchymal stem cell-specific Alms1 knockout mice.(A,B) Representative microscopic images of haematoxylin and eosin-stained gonadal white adipose tissue (gWAT) sections from female and male WT, global or mesenchymal stem cell (MSC) knockout (KO) mice at 24 weeks of age on high fat diet.Scale bars 200μm.(C,D,F,G) Size distribution of cross sectional area of adipocytes in gWAT, represented in bins of 1000μm 2 .(E,H) Calculation of the total number of adipocytes in the gWAT depot of each animal.Each data point represents one animal, with bars representing mean ± sd.Comparison between WT and KO groups in used an unpaired two-tailed Student's t-test with Bonferroni correction for multiple testing.(C-H) For females N = 5, 5, 8 and 8 for global WT, global KO, MSC WT and MSC KO respectively.For males N = 6, 6, 7 and 7 for global WT, global KO, MSC WT and MSC KO respectively.

Figure S4 .
Figure S4.Histological evaluation of liver and interscapular brown adipose tissue of global and MSC-specific Alms1 knockout mice.Representative macroscopic images of liver (A,B) and microscopic images of haematoxylin and eosin stained interscapular brown adipose tissue (iBAT) (C,D) from WT, global and mesenchymal stem cell (MSC) Alms1 knockout (KO) mice at 24 weeks of age.Scale bars are 200μm.

Figure S5 .
Figure S5.Histological evaluation of metabolic tissues from global and mesenchymal stem cell-specific Alms1 knockout mice.Quantification of picrosirius red (PSR) fibrosis staining in liver (A,B), gonadal white adipose tissue (gWAT) (C,D) and inguinal WAT (iWAT) (E,F) in 24 week old male and female global and mesenchymal stem cell (MSC) knockout (KO) mice.(G-N) Quantification of senescence markers in gWAT and iWAT: (G,H) Sudan Black B staining for lipofuscin.(I-N) Immunohistochemical staining for (I,J) p16 (K,L) p21 and (M,N) LaminB1.Each data point represents one animal, with bars representing mean ± sd.