Target for diabetes treatment and prevention

Abstract

The present invention relates to the identification of ALMS1 as the missing player involved in the regulation of the insulin-mediated glucose uptake through GLUT4 sorting vesicles, and to the down-regulation of ALMS1 by αPKC. Accordingly, the present invention relates to a molecule capable of preventing the binding of αPKC on ALMS1 for use for treating or preventing diabetes, in particular type 2 diabetes. In addition, the present invention relates to a method for identifying molecule capable of preventing the binding of αPKC on ALMS1.

Claims

1. A method of treating or delaying the progression or onset of diabetes mellitus, diabetic retinopathy, diabetic neuropathy, diabetic nephropathy, insulin resistance, hyperglycemia, obesity, and/or hyperinsulinaemia comprising administering a molecule inhibiting the binding of αPKC (Protein Kinase C alpha type) to ALMS1 (Alstrom syndrome protein 1) to a subject in need of treatment, wherein the molecule is a peptide, the peptide comprises a fragment of SEQ ID NO: 14, and said fragment is between 5 and 50 amino acids in length and spans one or more amino acid residue(s) selected from the group consisting of S227, S232, F262, D266, W267, K269, E271, G285, T296, V329, and I332 of SEQ ID NO: 14.

2. The method according to claim 1, wherein said method treats or delays the progression or onset of type 2 diabetes mellitus in said subject.

3. The method according to claim 1, wherein the molecule does not interfere with the binding of TBC1D4 to ALMS1.

4. The method according to claim 1, wherein the peptide is between 5 and 20 amino acids in length.

5. The method according to claim 1, wherein the peptide is a stapled peptide.

6. The method according to claim 1, wherein the peptide is between 5 and 30 amino acids in length.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Metabolic characterization of the Alms.sup.foz/foz mice

(2) (A) Mean body weight of WT and Alms.sup.foz/foz male mice (n=6-8 mice per genotype). (B) Photograph of visceral adipose tissue from WT and Alms.sup.foz/foz Scale bar: 25 μM. (C) Insulin tolerance test (I.T.T.) performed on WT and Alms.sup.foz/foz mice and the corresponding histogram showing the Area under the curve (A.U.C.) for each genotype (n=6-8 mice per group). p<0.001). (D) Mean body weight of WT and Alms.sup.foz/foz male mice (n=6-8 mice per genotype). (E) Photograph of visceral adipose tissue from corresponding WT and Alms.sup.foz/foz Scale bar: 25 μM. (F) Insulin tolerance test performed on WT and Alms.sup.foz/foz mice and the corresponding histogram showing the A.U.C. for each genotype (n=6-8 mice per group). *** stands for p-value<0.001. (G) Immunoblots for the indicated proteins in insulin sensitive tissues from nonobese WT and Alms.sup.foz/foz mice. (H) Results of radioactive counts in different target tissues after injection of radioactive deoxyglucose to WT and Alms.sup.foz/foz mice (n=5 mice per genotype). * stands for p-value=0.05.

(3) FIG. 2. ALMS1 silencing effect in human mature adipocytes

(4) (A) Photographs showing the lack of absorption of 2-NBDG (green) in control (shCTRL shRNA) or ALMS1-deprived adipocytes (ALMS1 shRNA) silenced mature adipocytes in absence of INS. (B) Photographs depicting lack of absorption of 2-NBDG in ALMS1 shRNA compared to CTRLshRNA. Nuclei were counterstained with DAPI, DIC: Differential Interference Contrast pictures. (C) 3D images of CTRLshRNA or ALMS1shRNA mature adipocytes stained for intracellular Triglycerides (TG), plasma membrane in red (PM) and nuclei in blue (DAPI). (D) Measurements of fluorescent levels correlating with amounts of intracellular TG in mature adipocytes (n=16 wells per condition measured) *p-value=0.05. (E-F) Immunodetection of AKT and pS473-AKT in CTRLshRNA and ALMS1shRNA treated mature adipocytes in presence and absence of INS. (G) 3D images of CTRLshRNA and ALMS1shRNA mature adipocytes showing cellular localization of Insulin receptor (IR in red) and GLUT4 (in green) in absence of Ins. Cut-view images displaying the dynamics of GLUT4 localization in absence of Ins. (H), after 30 min. INS. stimulation (I) and with 30 min INS. stimulation followed by 2 hours of absence of INS. (J) in CTRLshRNA and ALMS1shRNA mature adipocytes. Scale bars: 25 μm in A, B, C and 5 μm in G-J.

(5) FIG. 3. Predicted interaction sites on ALMS1 protein and modelling of its partner TBC1D4.

(6) (A) Predicted 3D structure of the ALMS1 protein with helices and loops. (B) Predicted 3D structure of the ALMS1 protein with the potential interacting sites represented by red dots. (C) Primary sequence of TBC1D4 protein with indicated localization of binding sites or interacting domains. (D) Predicted 3D structure of the TBC1D4 protein.

(7) FIG. 4. ALMS1 is required for TBC1D4 cellular trafficking

(8) (A) In silico predicted 3D structure showing spatial interaction between ALMS1 and TBC1D4 with an enlarged view of the interaction site highlighting the predicted interacting amino acid residues (L66, Y61 and S2879) of the ALMS1 protein. (B) 3D image from immunostained mature adipocytes depicting co-localization of TBC1D4 (green) and ALMS1 (red). Nuclei were counterstained with DAPI (blue). (C-D) Immunoblots for the indicated proteins on cell lysates (50 μg total protein loaded per lane) for CTRLshRNA and ALMS1shRNA mature adipocytes treated with or without insulin. 3D images of immunofluorescence experiments performed on either CTRLshRNA or ALMS1shRNA or TBC1D4shRNA mature adipocytes depicting cellular localization of GLUT4 in absence of Insulin (−INS) (E) or in the presence of INS. (F). PM: Plasma membrane and nuclei counterstained with DAPI. 3D images of immunofluorescence experiments performed on either CTRLshRNA or ALMS adipocytes showing cellular localization of GLUT4 (green) and TBC1D4 in absence of INS (G) or when treated 30 min. with INS. (H). Scale bars: 10 μm.

(9) FIG. 5. TBC1D4 is not the sole interacting partner of ALMS1 playing a role in the adipocyte biology

(10) (A-C) Photographs showing absorption of 2-NBDG in either CTRLshRNA or ALMS1shRNA or TBC1D4shRNA deprived adipocytes after 30 min Ins. stimulation. (D-F) 3D images obtained using non-permeablized fixated mature adipocytes stimulated with INS. following immunodetection of GLUT4 membrane bound (green). Plasma membrane (PM) was stained with Image-iT (red) and nuclei were counterstained with DAPI. (G) Immunoblots of 2 proton pumps subunits (ATP6V0D1 and ATP6V1A) identified by mass spectrometry in the IP experiments using ALMS1 as bait, αPKC, GLUT4 and β-Tubulin in cellular extracts from white adipose tissue (WAT) and kidney. 50 μg total protein loaded per lane. (H) Photograph of Duolink positive signal detected in adipocytes using antibodies against ALMS1 and ATP6. (I) Immunofluorescence pictures showing cellular localizations of ATP6V0D1 and ALMS1 and merged in mature adipocytes upon INS. stimulation. (J) In silico predicted binding sites of TBC1D4 (red) and PKC (yellow) which are only 20 Angstroms away from each other in the ALMS1 3D structure. (K-L) Immunodetections of αPKC, TBC1D4 and α-Actinin in immunoprecipitates using ALMS1 as bait in adipocytes cultured in absence or presence of INS.

(11) FIG. 6. Restoring acidification in ALMS1-deprived adipocytes reinstate glucose absorption

(12) (A-B) Time lapse pictures were performed on either control or ALMS1-deprived Acridine orange stained adipocytes stimulated with INS. (C-D) Time lapse pictures were performed on either control or ALMS1-deprived Acridine orange stained adipocytes stimulated with an electroneutral K.sup.+/H.sup.+ exchange ionophore, Nigericin (NIG.). (E) Top to bottom: Scanning electron microscopy (SEM) pictures of control adipocytes stimulated either without Ins. or with INS. or with NIG. White arrows show swelled vesicles. (F) Corresponding Transmitted Electron microscopy (TEM) pictures shown in (E) showing vesicles fusion with the plasma membrane in presence of INS. and NIG. (G) Top to bottom: SEM pictures of ALMS1-deprived adipocytes stimulated either without INS or with INS. or with NIG. (H) Corresponding TEM pictures shown in (G) showing vesicles fusion with the plasma membrane only in presence of NIG. (I) Photographs showing the intracellular content of 2-NBDG (green) in control mature adipocytes either in absence of INS. (top panel) or after 30 minutes INS. stimulation (middle panel) or after 30 min. NIG. Stimulation (bottom panel). (J) Photographs showing the intracellular content of 2-NBDG (green) in ALMS1-deprived mature adipocytes either in absence of INS. (top panel) or after 30 minutes INS. stimulation (middle panel) or after 30 min. NIG. stimulation (bottom panel). Scale bars: 20 μm except for F and H: 500 nm.

(13) FIG. 7. GLUT4 trafficking requires ALMSome protein complex

(14) (A) 3D images obtained using non-permeabilized fixated mature adipocytes stimulated with NIG. following immunodetection of GLUT4 membrane bound (green). Plasma membrane (PM) was stained with Image-iT (red) and nuclei were counterstained with DAPI. (B) Photographs showing intracellular TG content 24 hrs. after NIG. treatment. (C) Schematic representation of ALMS1 cellular localization and protein partner in absence of INS. stimulation in mature adipocyte. (D) Schematic representation of ALMS1 dynamics and protein partners after INS. stimulation in mature adipocyte.

(15) FIGS. 8A-8B. Glucose absorption is triggered in absence of INS through specific interference of αPKC binding site in the ALMSome

(16) (A) Photographs showing absorption of 2-NBDG in presence or absence of INS in adipocytes infected with either CTRL lentiviral particles or αPKC domain carrying lentiviral particles. (B) Quantification of intracellular glucose analogue 2-NB in presence or absence of INS in adipocytes infected with either CTRL lentiviral particles or αPKC domain carrying lentiviral particles. (n=8 per group).

(17) FIG. 9. min-αPKC-FLAG construct characterization in adipocytes

(18) Top panel: Immunodetection of min-αPKC-FLAG using an anti-FLAG antibody in mature adipocytes 48 hours post lentiviral infection. 2.sup.nd and 3.sup.rd panels: 3D image of the adipocyte showing the perinuclear localization of min-αPKC-FLAG. Last panel: Schematic representation of the experimental approaches used to assess the effect of min-αPKC-FLAG on glucose absorption.

EXAMPLES

(19) Alström syndrome (ALMS) is a rare autosomal recessive disorder characterized by several clinical features including obesity and early-onset diabetes. It originates due to mutations in the ALMS1 gene coding for a protein of 460 kDa.

(20) The function of the ALMS1 gene and how it causes the Alström syndrome phenotype has hitherto been unknown, with studies into its function being impeded by the extremely large size of the encoded protein and its low levels of expression.

(21) Alström syndrome (ALMS) is a rare monogenic childhood obesity syndrome for which there is only one causative mutated gene identified to date, the ALMS1 gene. ALMS is classified as a member of the ciliopathy disorders that includes Bardet Biedl syndrome, a group of syndromic disorders originating from mutations in the large number of different proteins that together play a critical role in primary cilium function. Alms1 encodes the 461 kDa ALMS1 protein that was originally described to bear a purely centriolar localization, although more recent data has also suggested a cytoplasmic localization of ALMS1.

(22) ALMS is clinically identified by collective multisystem phenotype thought to reflect the ubiquitous tissue expression of ALMS1, closely mimicking many of the phenotypic features of BBS. Common clinical features of ALMS include retinal degeneration, hearing loss, childhood obesity, early-onset type 2 diabetes mellitus (T2DM) dilated cardiomyopathy, renal and hepatic dysfunction, hypothyroidism, short stature, hyperlipidemia, and organ fibrosis. Children with ALMS develop obesity in early childhood that is associated with early onset of T2DM at around 16 years of age with a much higher overall prevalence of early onset T2DM in ALMS than seen with other childhood obesity syndromes resulting in a similar body mass index (BMI) including BBS. The reason for this predilection for T2DM in children with ALMS that is out of proportion to their degree of obesity has remained elusive.

(23) The inventors investigated the role of the ALMS1 protein during the adipogenic differentiation process and found that the ALMS1 protein expression levels increased during adipogenesis. ALMS1 suppression, in adipogenic differentiating mesenchymal stem cells, inhibited the anti-adipogenic cascades but surprisingly was not favoring adipogenesis.

(24) In addition, the inventors showed the ALMS1 protein complex is also required in mature adipocytes for efficient GLUT4 retention in its insulin-responsive compartment and its ability to fuse with the plasma membrane in response to insulin stimulation. Inactivation of ALMS1 decreased the amount of glucose able to be absorbed by mature adipocytes upon insulin stimulation, therefore contributing to hyperglycaemia and the onset of diabetes.

(25) Previous studies in the spontaneous mutant Alms.sup.foz/foz and genetrapped Alms1knockout murine ALMS models confirmed that these mice, similarly to affected human children, develop obesity in early adolescence due to hyperphagia, and also exhibit impaired glucose tolerance, hyperinsulinemia and islet hypertrophy, consistent with severe insulin resistance, although the tissue origin or mechanism for this insulin resistance has previously not been characterised. Previously published studies of in vitro studies on the murine 3T3-L1 fibroblast cell line showed that inhibition of ALMS1 gene expression resulted in mild impairment of adipogenesis but was reported to have no effect on the insulin signaling pathway in the resulting mature adipocytes as measured by insulin-mediated AKT phosphorylation. This data led directly away from the invention presented here that Alms1 does indeed play a critical hitherto unrecognized role in the insulin signaling pathway and in GLUT4 mediated glucose transport.

(26) Indeed, despite the previously published contrary data, the inventors when carefully studying the phenotype of the Fat Aussie murine ALMS model (Alms1.sup.foz/foz) identified that insulin resistance in this model preceded rather than followed the development of obesity. They further identified the adipose tissue as the specific site driving the insulin resistance and subsequent development of glucose intolerance and T2DM in ALMS. They confirmed that insulin signaling in Alms1.sup.foz/foz adipocytes was intact all the way down to phosphorylation of TBC1D4, the last known member of the insulin-mediated glucose uptake pathway but then through a subsequent series of investigations identified a protein complex they termed the Almsome, composed of several key proteins that associate with ALMS1 and which together are required for the tethering and fusion of the GLUT4 vesicles to the adipocyte plasma membrane (PM) in response to insulin signaling. The Almsome thereby represents the hitherto unidentified ultimate step in insulin-mediated glucose uptake into adipocytes, with insulin resistance in ALMS due to disruption of Almsome function leading to failure of GLUT4 membrane fusion and thereby a block to adipocyte glucose transport.

Example 1

(27) Alms1.sup.foz/foz Mice Display Severe Specific Adipose Tissue Insulin Resistance Even in the Absence of Obesity

(28) Animal Husbandry

(29) Alms1.sup.foz/foz mice and Alms1.sup.+/+ (WT) littermates were maintained on a C57BL/6J background in the animal facility on a 12 hourly light/dark cycle. Mice had free access ad libitum to water and either normal chow containing 5.4% fat, energy content 12 MJ/kg (Gordon's rat and mouse maintenance pellets, Gordon's specialty stockfeeds, Australia) or high fat diet (HFD) containing 23% Fat, High Simple carbohydrate, 0.19% cholesterol, energy content 20 MJ/kg (SF03-020, Specialty feeds, Australia). Primers flanking the foz mutation were used for PCR genotyping: forward ACA ACT TTT CAT GGC TCC AGT (SEQ ID NO:12); reverse TTG GCT CAG AGA CAG TTG AAA (SEQ ID NO: 13).

(30) Six month old obese and young (<60 days old) nonobese Alms1.sup.foz/foz mice and wildtype (WT) littermates were used to investigate what primary metabolic impairment leads Alms1.sup.foz/foz mice to develop T2DM. Six month old Alms1.sup.foz/foz mice were obese with an average body weight of 45.5 g±1.7 g compared to 26.4 g±1.3 g for WT littermates (FIG. 1A) and as previously shown had fasting hyperglycaemia and impaired glucose tolerance with elevated HOMA scores. An insulin tolerance test (ITT) showed that unlike WT (FIG. 1B) and heterozygous littermates, glycaemia in obese Alms1.sup.foz/foz mice was unresponsive to insulin administration (FIG. 1B), even when doses of insulin as high as 20 U/kg were administered. (FIG. 1C). Obesity of Alms1.sup.foz/foz mice was due to severe adipocyte hypertrophy (FIG. 1B) rather than the adipocyte hyperplasia more typically seen in obese BBS mice. To determine what the primary defect was that was causing the glucose intolerance in Alms1.sup.foz/foz mice, young lean Alms1.sup.foz/foz mice were studied to remove the confounding effect of obesity on insulin responsiveness. At 2 months of age, WT and Alms1.sup.foz/foz males had a similar average body weight of −24 g (FIG. 1D). ITT in these mice showed that nevertheless the young nonobese Alms1.sup.foz/foz males already exhibited significantly reduced insulin responsiveness (FIG. 1E), consistent with insulin resistance preceding obesity in this model. Immunodetection of TRAP, Akt, p-AKT, GLUT4, C/EBP-α and GAPDH performed on insulin sensitive tissues namely, heart, liver, skeletal muscles and white adipose tissue (WAT) of 6-month-old non-fasted Alms1.sup.foz/foz and WT showed no major differences in protein levels except for a consistent increase in the p-AKT to total AKT ratio in WAT, consistent with a paradoxical increase rather than reduction in activation of upstream members of the insulin signaling pathway in glucose intolerant Alms1.sup.foz/foz mice (FIG. 1G). To identify which tissues alone or together might be the primary source of the insulin resistance observed in Alms1.sup.foz/foz mice, the tissue distribution of insulin-mediated deoxyglucose (DOG) uptake was compared in WT and Alms1.sup.foz/foz mice. This confirmed that severely impaired DOG uptake was limited to the WAT of Alms1.sup.foz/foz mice with a compensatory increase in DOG uptake into the insulin-responsive gastrocnemius and soleus muscles when compared to WT mice.

(31) These studies demonstrate that although Alms1.sup.foz/foz mice become obese and develop progressive T2DM with age, the major initial defect contributing to insulin resistance and hyperglycaemia is a failure in the absence of functional ALMS1 of adipose tissue glucose uptake in response to insulin signaling, with this defect predating the development of obesity.

Example 2

(32) Silencing of Alms1 in Human Adipocytes Blocks Glucose Uptake Through Impaired GLUT4 Cellular Sorting

(33) Materials. From Molecular Probes, Invitrogen: Acridine Orange, Image-iT® LIVE Plasma Membrane and Nuclear Staining Labeling Kit, 2-NBDG (2-(N-7-nitrobenz-2-oxa-1, 3-diazol-4-yl) amino)-2-deoxyglucose), Hoechst 33258 and Cell Light™ Early Endosomes-RFP* BacMam 2.0*; Catalog #: A3568, 134406, N13195, H3569 and C10587. From Lonza: AdipoRed™ Assay Reagent (Catalog #: PT-7009). Lentiviral particles from Santa Cruz Biotechnology, INC.: ALMS1 shRNA (h) Lentiviral Particles, TBC1D4 shRNA (h) Lentiviral Particles and Control shRNA Lentiviral Particles-A; Catalog #: sc-72345-V, sc-61654-V and sc-108080 respectively. From Tocris Biosciences: Nigericin Sodium Salt (Catalog #: 4312).

(34) Biochemical tests. Mice were tested for insulin resistance by the insulin tolerance test (ITT) and intraperitoneal glucose tolerance test (IPGTT). For the ITT, mice were fasted 4 hours with no access to food but free access to water. Mice were weighed and insulin (Humulin R, Eli Lilly, USA) was injected ip at 0.75 U/kg body weight in 0.9% saline for injection (Pfizer, USA). Tail blood was obtained and the plasma glucose was determined for each mouse using a glucometer (Optium Xceed, Abbott, USA) and blood glucose test strips (Optium point of care, Abbott, USA) at 0, 15, 30 and 60 min after insulin injection. For the IPGTT, mice were fasted 18 hours and injected at 2 mg/g body weight with D-glucose (Analar, VWR, USA) in 0.9% saline for injection. Plasma glucose was determined for each mouse using a glucometer with sampling via tail vein at 0, 15, 30, 60 and 120 min after glucose injection. For plasma insulin measurement, blood was collected on conscious animals via cheek bleeding. After collection, blood samples were kept on ice and spun at 17000 g, 10 min at 4° C. Insulin levels were assayed using a commercial ultrasensitive mouse insulin ELISA kit (Crystal Chem Inc., USA). The homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated using individual mouse fasting insulin and fasting glucose levels. The following formula was used:
HOMA-IR=[fasting glucose(mg/dL)×fasting insulin(μU/mL)]/405.

(35) Cell culture. Human white visceral preadipocytes (Catalog #: C-12732; PromoCell) and human mesenchymal stem cells (Catalog #: C-12974; PromoCell) derived from healthy bone marrow were purchased. The preadipocytes were seeded according to manufacturer's protocol and cultured in the Preadipocyte growth medium (Catalog #: C-27410; PromoCell) to confluence. One day before inducing terminal adipogenesis, the cells were infected with specific lentiviral particles consisted of a pool of 3 shRNAs target-specific constructs purchased from Santa Cruz Biotechnology and on the next day, adipogenic differentiation was induced by changing the medium to the Preadipocyte Differentiation Medium (Catalog #: C-27436; PromoCell) for 2 days. After the differentiation phase, the medium was finally changed to the Adipocyte Nutrition medium (Catalog #: C-27438; PromoCell). For the culture without insulin, Adipocyte Basal Medium (Catalog #: C-2431; PromoCell) without insulin was complemented with 5 g/L of deoxyglucose, 8 μg/mL d-Biotin, 400 ng/mL Dexamethasone. For the hMSCs, they were cultured in Mesenchymal Stem Cell Growth Medium (Catalog #: C-28010; PromoCell) to confluence. hMSCs were transfected with specific siRNAs as described above and on the next day adipogenic differentiation was induced by changing the medium to the MSC Adipogenic Differentiation Medium (Catalog #: C28011; Promocell).

(36) RNA extraction, cDNA synthesis, q-PCR and Taqman. Total RNA was prepared from the different tissues and cells using a RiboPure™ kit (Catalog #: AM1924; Ambion) followed by a DNAse treatment with the TURBO DNA-Free™ (Catalog #: AM1907; Ambion). RNA integrity was assessed by gel electrophoresis and RNA concentration by Eppendorf Biophotometer Plus with the Hellma® Tray Cell (Catalog #: 105.810-uvs; Hellma). Reverse transcription of 1 μg total RNA to cDNA was performed using the BioRad iScript™ cDNA synthesis kit (Catalog #: 170-8891; BioRad). Real-time quantitative polymerase chain reaction amplification was performed in a BioRad CFX96 ™ Real-Time System using the iQ™ SYBR® Green Supermix (Catalog #: 170-8886; BioRAd) and primer sets optimized for tested targets for SYBR Green-based real-time PCR for the real-time PCR. Taqman analysis was carried out with the specific gene assay with the Taqman® Fast Advanced Master Mix (Catalog #: 4444557; Applied Biosystems). The normalized fold expression of the target gene was calculated using the comparative cycle threshold (CO method by normalizing target mRNA C.sub.t to those for GAPDH using the CFX Manager Software Version 1.5 and was verified using the Lin-Reg program.

(37) Western blots and immunofluorescence microscopy. Male Alms1.sup.foz/foz and WT littermates were anaesthetized. The following insulin sensitive tissues: liver, heart, muscle and adipose tissue were harvested and directly placed in RIPA buffer (Tris 50 mM, NaCl 150 mM, 0.1% SDS, 1% Triton-X100) supplemented with Complete mini protease inhibitor cocktail and PhosSTOP phosphatase inhibitor cocktail (Roche, Switzerland). Samples were sonicated and centrifuged 30 min at 17 000 g, 4° C. 30 min. Protein concentration assayed with BCA assay (Thermo Fisher Scientific, USA). Cellular proteins from cells were obtained by trichloroacetic acid precipitation and immunoblot analyses were performed using 30-50 μg total protein. Specific antibody binding was visualized using the SuperSignal® West Femto Maximum Sensitivity Substrate (catalog #: Lf145954, Pierce) on a BioRad Versadoc™ Imaging System or ImageQuant LAS 4000 imager (GE Healthcare, UK). Nonspecific proteins stained with Ponceau S were used as loading controls to normalize the signal obtained after specific immunodetection of the protein of interest using the Bio-Rad Quantity One program. For immunofluorescence experiments, the cells were seeded on permanox 8-wells Lab-Tek II Chamber Slide (Catalog #: 177445; NUNC). Cells were treated as indicated. Then both cells and tissues cryosections were processed for protein detection after methanol fixation and permeabilized with 0.1% Triton X-100. The microscopy slides were mounted for detection with Vectashield Mounting Medium (Catalog #: H-1200; Vector Laboratories). To view membrane-associated proteins, cells were formalin fixated for 15 min and were directly blocked, followed by immunostaining and acquisition using an upright Zeiss Axiolmager Z2 microscope. Image analysis, 3D reconstitution and Time Lapse experiments and endosomes tracking experiments were performed using either the Zeiss AxioVision program with the corresponding 3D and Tracking Zeiss modules or the Zeiss Zen 2012 imaging platform.

(38) Fluorescence measurement. The preadipocytes were cultured in a 96 well plate and 12 wells infected with the either ALMS1 shRNA lentiviral particles or CTRL shRNA lentiviral particles and differentiated the next day into mature adipocytes. 3 weeks later, the intracellular triglycerides were stained with AdipoRed staining following the manufacturer's procedure and the fluorescence was measured on a Tecan Infinite 200 quad4 monochromator (Tecan, Lyon, France) at a wavelength of 520 nm. The generated data were then analyzed using the Tecan Magellan Data Analysis software using as blank unstained adipocytes.

(39) Co-Immunoprecipitation experiments. For the co-immunoprecipitation experiments, we used the Dynabeads® Antibody Coupling kit (Catalog #: 143.11D, Invitrogen) in combination with the Dynabeads® co-immunoprecipitation kit (Catalog #: 143.21D, Invitrogen). The hMSCs were cultured to confluence and adipogenic differentiation was triggered by medium change. 7 days after adipogenic differentiation was initiated by medium change, the adipocytes, cultured with our without Ins. 24 hours prior to lysis, were lysed under native conditions and used according to the kit. After immunoprecipitation and release from the beads, the samples were loaded on a NuPage 3-8% TrisAcetate Gel (Catalog #: EA0375BOX, Invitrogen) with a Hi Mark™ Prestained HMW Protein Standard (Catalog #: LC5699, Invitrogen).

(40) Protein preparation and identification by mass spectrometry. In gel digestion: The gel digestion procedure was carried out as described by Rabilloud et al. (ref). Preparation of the gel pieces before trypsin digestion was performed by a liquid handler robot (QuadZ215, Gilson International, France). Briefly, gel bands were washed alternately with 100 μl of 25 mM NH.sub.4HCO.sub.3 and then 100 μl of 50% acetonitrile (ACN) (3 min wash under shaking and the liquid was discarded before addition of the next solvent). This hydrating/dehydrating cycle was repeated twice and the pieces of gel were dried for 20 min before reduction (10 mM DTT/25 mM NH.sub.4HCO.sub.3 buffer at 56° C. for 45 min) and alkylation (25 mM iodoacetamide/25 mM NH.sub.4HCO.sub.3 buffer for 45 min, room temperature). Afterwards, gel spots were again washed with 3 cycles of 25 mM NH.sub.4HCO.sub.3/ACN alternately. Following 20 min drying step, the gel pieces were rehydrated by three volumes of trypsin (Promega, V5111), 12.5 ng/μ1/in 25 mM NH.sub.4HCO.sub.3 buffer (freshly diluted) and incubated overnight at room temperature. Tryptic peptides were extracted from gel by vigorous shaking for 30 min in adapted volume of 35% H.sub.2O/60% ACN/5% HCOOH and a 15 min sonication step.

(41) MALDI-TOF (/TOF) mass spectrometry and database search. MALDI mass measurement was carried out on an Autoflex III Smartbeam (Bruker-Daltonik GmbH, Bremen, Germany) matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (MALDI-TOF TOF) used in reflector positive mode. A prespotted anchorchip target (PAC system from Bruker Daltonik, technical note TN011) with HCCA matrix was used to analyse tryptic digests. The resulting peptide mass fingerprinting data (PMF) and peptide fragment fingerprinting data (PFF) were combined by Biotools 3.2 software (Bruker Daltonik) and transferred to an intranet version of the search engine MASCOT (Matrix Science, London, UK). Variable modifications (N-term protein acetylation, methionine oxidation and cysteine carbamidomethylation) and one tryptic missed cleavage were taken into account and the peptide mass error was limited to 50 ppm. Proteins were identified by searching data against an NCBI non-redundant protein sequence database and then submit to the human restricted database. In all results, the probability scores were greater than the score fixed as significant with a p-value of 0.05. NanoLC-MSMS mass spectrometry and database search: For nanoLC-MS/MS analysis, peptides were transferred in glass inserts, compatible with the LC autosampler system (nanoLC-U3000, Dionex, US). The LC system was coupled to an ESI-Q-TOF mass spectrometer (MicroTOFQ-II, Bruker, Germany). The method consisted in a 60 min run at a flow rate of 300 nL/min using a gradient from two solvents: A (99.9% water: 0.1% formic acid) and B (99.92% acetonitrile: 0.08% formic acid). The system includes: a 300 μm×5 mm PepMap C18 used for peptides preconcentration and a 75 μm×150 mm C18 column used for peptides elution. The TOF analyzer was calibrated each day: data were acquired and processed automatically using Hystar 2.8 and DataAnalysis 2.6 softwares. Consecutive searches against the NCBInr database first and then against the human sub-database were performed for each sample using local versions of Mascot 2.2 (MatrixScience, UK) and Proteinscape 2.0 (Bruker, Germany). False-positive rate (FPR) for protein identification was estimated using a reverse decoy database: protein validation was done using a FPR below 1%. Moreover, proteins identified by only 1 peptide were checked manually: MS/MS spectra were inspected according to conventional fragmentation rules.

(42) In situ Proximity ligation assay (PLA). Duolink in situ PLA kit with antimouse PLUS probe and anti-rabbit MINUS probe (catalog #: 90701 and 90602; OLINK Bioscience) were used in combination with the appropriate primary antibodies according to the manufacturer's procedure. Human primary preadipocytes and mature adipocytes were cultured on 8-well Lab-Tek II chamber slide (Nunc) and treated as for immunofluorescence microscopy until the primary antibody incubation step. After washing, cells were decorated with PLA PLUS and MINUS probes (1:20 dilution) for 2 hrs at 37° C. Hybridization and ligation of probes, amplification, and final SSC washing were performed according to the manufacturer's procedure. Fluorescence transfer based on protein-protein interaction was visualized using the Duolink Detection kit 613 (OLINK Bioscience) and images were acquired.

(43) Statistics. Statistical analyses were performed using GraphPad Prism 5 software (GraphPad Software, Inc., USA). Results are shown as means±standard deviation. Significance of the results was determined by paired t tests or the non-parametric Mann-Whitney U test was used for statistical comparison of BMI and AUC data. A value of P<0.05 was considered to denote statistical significance and was marked with an asterisk.

(44) Using primary human preadipocytes as an in vitro model, the inventors localized ALMS1 primarily in a cytoplasmic rather than the previously reported centrosomal pool. ALMS1 was silenced during adipogenesis and although a significant decrease in the anti-adipogenic factor Pref-1 was observed, no major differences could be detected in expression levels of pro-adipogenic transcription factors such as the cEBPs and PPARγ.

(45) Following ALMS1 silencing in 2-week-old mature adipocytes, glucose absorption capacity was assessed using labelled glucose analogue (2-NBDG). In the absence of insulin stimulation, no 2-NBDG uptake could be detected in ALMS1-silenced and control mature adipocytes (FIG. 2A). On the other hand, insulin stimulation resulted in increased 2-NBDG uptake in the control human mature adipocytes (FIG. 2B, top panel) but not ALMS1-silenced cells (FIG. 2B, bottom panel). Further to reduced glucose absorption in ALMS1-silenced adipocytes, the inventors observed a reduction in intracellular triglycerides (TG) in these cells a week later (FIGS. 2C-D). Of note, this reduced glucose absorption in ALMS1-deficient adipocytes was not associated with decreased phosphorylation of AKT, the downstream signaling target of insulin, as pS473-AKT levels after 30 minutes incubation with insulin were similar in both control and ALMS1-silenced human adipocytes (FIGS. 2E-F), consistent with the normal to increased levels of AKT phosphorylation previously observed in Alms1.sup.foz/foz murine adipocytes (FIG. 1G).

(46) The inventors next investigated the dynamics of GLUT4 in human adipocytes in the absence of ALMS1. Insulin receptor (IR) cellular localization to the plasma membrane was not impaired following ALMS1 silencing being detected in the vicinity of the plasma membrane (PM) in the absence of insulin. (FIG. 2G, top panel). By contrast, in ALMS1-deficient adipocytes in the absence of insulin GLUT4 lost its perinuclear localization and was detected dispersed throughout the cell cytoplasm rather than assuming its usual perinuclear localisation. (FIG. 2G, middle and bottom panels and 2H). Upon insulin stimulation, GLUT4 was observed to move to the PM within the actin mesh (FIG. 21) in both control and ALMS1-silenced adipocytes. Two hours post insulin stimulation in the absence of insulin, GLUT4 was still detected dispersed throughout the cytoplasm of the ALMS1-silenced adipocytes whereas control adipocytes had their GLUT4 appropriately re-localized to the perinuclear region (FIG. 2J). As there is an equilibrium between exocytosis and endocytosis of GLUT4 vesicles to and from the PM, the inventors checked to exclude that the impaired GLUT4 sorting in Alms1-silenced adipocytes was not due to defective GLUT4 endocytosis. Examination of dynamin, a key molecule in endocytosis, demonstrated no difference in protein levels nor cellular localization following ALMS1 silencing in adipocytes. Furthermore, the mean velocities of labelled endosomes were similar between ALMS1-silenced and control adipocytes, arguing against a defect in endocytosis being the cause of reduced GLUT4 presence in the PM in response to insulin signaling.

Example 3

(47) ALMS1 is Required for TBC1D4 Targeting to the PM in Response to Insulin Signaling

(48) To understand the molecular mechanism underlying the effect of ALMS1 inactivation on GLUT4 localisation, the inventors identified interacting partners of ALMS1 in human adipocytes. Immunoprecipitation (IP) using ALMS1 as the bait was performed using young mature human adipocytes (4 days after differentiation trigger) followed by identification of ALMS1 interacting partners by mass spectrometry. Amongst proteins were immunoprecipitated with ALMS1, was TBC1D4, a known AKT substrate GTPase required for proper retention of GLUT4 in the GLUT4 sorting vesicles (GSVs) and for the translocation of GLUT4 to the cell membrane for intracellular glucose uptake.

Example 4

(49) Development of Structural Homology Models of ALMS1, TBC1D4 and αPKC.

(50) As the crystal structure of Alms1 has not yet been solved, in silico structural homology modeling was used to predict the 3D structure of ALMS1 and identify structural motifs that could bind potential interacting ligands (FIGS. 3A-C).

(51) Structural Model of ALMS1. The model of ALMS1 was constructed using fragment modeling method with the homology modelling program, Modeller. The amino acid sequence for each exon of ALMS1 was submitted to profile-based threading algorithm available at PISRED server and suitable templates were identified. Then those identified template proteins were aligned with the respective exon sequences and each exon was modeled separately using Modeller. The energy optimization and selection of models were conducted based on discrete optimized protein energy score. Finally, models were assembled to construct the structure of full length ALMS1 and the full-length protein was relaxed and minimized using the molecular dynamics simulation program NAMD.

(52) Structural Model of the PTP binding domain of TBC1D4. The PTP binding domain of TBC1D4 is located within the first 160 residues. No reliable homologues structure was identified to model the structure in between the PTP binding domain and the Rab binding domain. Crystal structure of the PTP domain of murine Disabled-1(Dab-1), 1NU2 (E-value=5.2e-17), which was identified by HMM based template search at Swiss model was used as the template for constructing the PTP binding domain of TBC1D4.

(53) The PTP domain of TBC1D4 interacting with ALMS-1. The macromolecular docking was performed by using the ClusPro 2.0 algorithm. Residues located in the interaction surface with >=0.4 angstrom overlap were considered as interacting residues. Interproscan revealed that the ALMS-1 contained a WD40-like domain within the first 3871 residues. WD40 domain containing proteins are a family of proteins functioning as scaffolds for macro-molecular interactions.

(54) The PTP binding domain of TBC1D4 interacting with ALMS1. Initially, the PTP binding domain and the RabGTP binding domain of TBC1D4 were docked to the ALMS1 model using the Cluspro 2 server to determine the most probable site of interaction. Then both domains were docked to their respective interacting sites on ALMS1 using Autodock 4.2 and their binding affinities were calculated. Based on the affinities, the PTP binding domain of TBC1D4 binds ALMS1 with ˜100 fold higher affinity compared to the RabGTP binding domain. Hence, the inventors predict that the PTP binding domain may have a higher probability to interact with the ALMS1 molecule compared to the RabGTP binding domain.

(55) Modelling the PTP domain of TBC1D4. The phospho-tyrosine binding domain of TBC1D4 was modeled after identifying a suitable template from the Swiss model template identification algorithm.

(56) Docking TBC1D4 PTP domain and RabGTP binding domain to ALMS1. Initially, the PTP binding domain and the RabGTP binding domain of TBC1D4 were docked to ALMS1 using the Cluspro 2 server and the site of interaction was identified. Then both domains were docked to their respective interacting sites using Autodock 4.2 and their binding affinities were calculated.

(57) TABLE-US-00002 Predicted ALSM-1 residue 65, 66, 69, 72, 73, 74, 75, 76, 77, 78, 80, 87, 2875, 2876, 2877, 2878, 2879, 2880, 2881, 2882, numbers, with the potential to 2883, 2884, 2885, 2887, 2888, 2889, 2890, 2892, 2893, 2894, 2895, 2897, 2909, 2910, interact with another ligand 2912, 2929, 2931, 2932, 2933, 2934, 2935, 3557, 3558, 4131, 144, 145, 146, 147, 148, 149, 150, 151, 193, 194, 195, 198, 199, 200, 201, 205, 208, 211, 214, 226, 227, 229, 233, 234, 235, 236, 239, 242, 243, 246, 248, 249, 250, 251, 252, 314, 319, 321, 986, 1341, 1344, 2269, 113, 114, 115, 116, 123, 126, 127, 128, 1340, 1438, 1439, 1440, 1441, 1442, 1443, 1444, 1446, 1447, 1448, 1449,1450,1451,1452,1453, 1454, 1457, 1458, 1459, 1478, 1915, 1918, 1919, 1920, 1922, 1923, 1930, 2041, 2042, 2043, 2257, 2267, 2483, 2484, 3866, 218, 219, 220, 221, 222, 223, 224, 277, 278, 279, 282, 285, 286, 287, 288, 686, 688, 689, 690, 691, 699, 1856, 1858, 1859, 1861, 1862, 1863, 1864, 1865, 1866, 1867, 1868, 1869, 1870, 1871, 1872, 1949, 1968, 1969, 1971, 1974, 1979, 1980, 1981, 1982, 1983, 1984, 2104, 2107, 2111, 2870, 2872, 2874, 2915, 3285, 3286, 3287, 793, 795, 796, 797, 1285, 1314, 1408, 1409, 1422, 1423, 1425, 1426, 1427, 1430, 1431, 1671, 1672, 1794, 1797, 2538, 2539, 2540, 2555, 2556, 2557, 2563, 2564, 2565, 2567, 2568, 2588, 2591, 2599, 2603, 2699, 2701, 2702, 3108 Predicted residues from ALMS1 E17, D58, S59, G62, H65, L66, Q736, T737, E738, D828, S829, T1088, D1089, mediating the interaction with A1169, Q1170, F2882, L2883, E2884 aPKC Predicted residues from aPKC F114, D116, C118, L121, N138, Q142, I145, P148, G433, E545, S562, S566, mediating the interaction with F597, D601, W602, K604, E606, G620, T631, V664, I667 ALMS1 Predicted residues from TBC1D4 G75, A76, P77, A78, R80, E81, V82, I83 mediating the interaction with ALMS1 Predicted residues from ALMS1 H65, L66, S2879 mediating the interaction with TBC1D4

(58) The homology model revealed that Alms1 assumes an apple core type structure with a large number of bindings sites of potential ligands centered around the core. The TBC1D4 crystal structure was similarly not solved and hence the inventors used a homology modeling approach to predict the structure of the PTP binding domain of TBC1D4 (FIGS. 3C-D). Subsequently, in silico docking studies were performed which predicted high affinity binding of TBC1D4 with ALMS1 through hydrogen bonding of TBC1D4 residues G75, A76, P77, A78, R80, E81, V82, 183 with interacting residues H65, L66, S2879 on Alms1 (FIG. 4A). Co-localization of ALMS1 and TBC1D4 was then confirmed in human adipocytes by immunofluorescence studies (FIG. 4B). The expression levels of GLUT4, TBC1D4, and TRAP were next tested in ALMS1-silenced adipocyte with or without insulin stimulation but no significant differences were found (FIG. 4C). Upon phosphorylation by activated AKT, phosphorylated TBC1D4 (p-TBC1D4) in adipocytes targets RAB proteins such as RAB14 and RAB10 prior to GSVs being targeted to the PM. However, upon insulin stimulation of Alms1-silenced adipocytes, no difference could be detected in the levels of TBC1D4, p-TBC1D4, RAB14 and RAB10 (FIG. 4D). The inventors next focused on GLUT4 cellular localization. In the absence of insulin stimulation, TBC1D4 silencing reproduced the ALMS1 silencing effect seen in mature adipocytes, i.e. a mislocalization of GLUT4 throughout the cytoplasm (FIG. 4E). In response to insulin stimulation, GLUT4 was released from the perinuclear region in control adipocytes spreading-out throughout the adipocyte cytoplasm (FIG. 4F) thereby reproducing the GLUT4 distribution pattern seen in in ALMS1 and TBC1D4-silenced adipocytes in the absence (FIG. 4E) and presence (FIG. 4F), of insulin. The inventors subsequently investigated the effect of ALMS1 silencing on the cellular dynamics of TBC1D4 in response to insulin. In the absence of insulin, TBC1D4 was localized to the perinuclear region in both control and ALMS1-silenced adipocytes (FIG. 4G) but notably, in response to insulin, TBC1D4 was only transported to the PM in control but not ALMS1-silenced adipocytes (FIG. 4H).

Example 5

(59) ALMS1 Forms a Dynamic Protein Complex, the ALMSome, Required for Insulin-Stimulated Glucose Transport in Human Mature Adipocytes

(60) Although the inventors showed that ALMS1 silencing prevented TBC1D4 targeting to the PM, it remained to be seen whether this impairment on its own explained the major reduction in glucose uptake observed in ALMS1-deficient adipocytes. The inventors therefore compared the cellular uptake of 2-NBDG upon insulin stimulation in ALMS1 or TBC1D4-silenced or control adipocytes and found almost no 2-NBDG absorbed in the ALMS1-silenced adipocytes compared to control adipocytes (FIGS. 5A-B), whereas whilst reduced compared to controls a substantial amount of 2-NBDG was still absorbed by TBC1D4-silenced adipocytes (FIG. 5C). Subsequent GLUT4-antibody binding assays on either ALMS1 or TBC1D4-silenced or control adipocytes following 30 minutes insulin stimulation showed a high proportion of GLUT4 in the PM in control and TBC1D4-silenced adipocytes but not in ALMS1-silenced cells (FIGS. 5D-F), indicating that the secondary defect in TBC1D4 targeting to the PM in Alms1-silenced cells did not, in itself, explain the very severe defect in glucose transport and GLUT4 PM expression in ALMS1-deficient cells. A further examination of the Alms1 IP data revealed several subunits of V type ATPase proton (H.sup.+) pumps (A, B, D1 and G2) that the inventors then showed to be expressed in mature adipocytes (FIG. 5G) together with αPKC, the activating kinase of the H.sup.+ pumps under insulin control. The inventors confirmed that ALMS1 was in close vicinity with the V-ATPase H.sup.+ pumps in mature adipocytes in the presence of insulin both by an in situ PLA Duolink approach targeting ALMS1 and the proton pumps subunits A1 and D1 (FIG. 5H) and also by co-immunostaining ALMS1, VATPase A1 and D1 and αPKC (FIGS. 5H, I). ALMS1 co-localized with the proton pump subunit V0D1 (FIG. 5I) that is integrated into the GSV membrane indicating that ALMS1 is transported in the adipocyte together with the proton pumps localized within the GSVs. Using their in silico-based structural model of ALMS1 interacting partners the inventors identified a binding motif for PKC on Alms1. The binding sites for TBC1D4 and αPKC on Alms1 were in such close proximity that the model showed that simultaneous docking of both proteins to Alms1 was not possible due to steric hindrance (FIG. 5J). The inventors thus hypothesized. To test their hypothesis that ALMS1 binding of αPKC or alternatively TBC1D4 was under the reciprocal control of insulin signaling in the adipocytes the inventors performed further IPs again using ALMS1 as bait but this time using human mature adipocytes cultured in the presence or absence of insulin with IPs being immunoblotted for both αPKC and TBC1D4. The results revealed that αPKC could only be pulled down by Alms1 and detected by immunoblotting in extracts of adipocytes incubated in the absence of insulin (FIG. 5K) whereas TBC1D4 was only pulled down by ALMS1 and detected in extracts of adipocytes incubated in the presence of insulin, consistent with the inventors model of reciprocal insulin-regulated Alms1 binding (FIG. 5L).

Example 6

(61) The ALMSome is Required for the Acidification of GSVs Prior to GLUT4 Delivery to the Plasma Membrane

(62) While AKT-phosphorylation of TBC1D4 has been known to in some way lead to GLUT4 trafficking, the ultimate GSV-PM fusion step is an insulin regulated non-AKT dependent event that requires osmotic swelling of the GSVs under the action of the vATPase H.sup.+ pump. However, knowledge of the actual signal and mechanism for activation of the H.sup.+ pump by insulin was missing. The inventors tested if ALMS1 inactivation could prevent the acidification of the GSVs and therefore the chemo-osmotic-mediated release of GLUT4 to the PM using the acidotrophic dye, acridine orange, which emits a green fluorescence at low concentration and an orange-red fluorescence at high concentrations in the lysosomes in which acridine orange is protonated and sequestered. In absence of insulin, no orange-red fluorescence was detected in the adipocytes. By contrast, insulin induced a rapid appearance of red color in control human mature adipocytes (FIG. 6A) but not in ALMS1-silenced adipocytes (FIG. 6B) indicating loss of insulin-mediated acidification of lysosomes in ALMS1-silenced adipocytes.

(63) The inventors next tested whether acidifying ALMS1-silenced adipocytes using Nigericin (NIG.), an electroneutral K.sup.+/H.sup.+ exchange ionophore known to cause osmotic swelling of the GSVs would bypass the Alms1-associated defect in GLUT4 fusion and glucose absorption. NIG. treatment resulted in a rapid acidification of both control and ALMS1-silenced adipocytes (FIGS. 6C-D), thereby activating the swelling and fusion of the intracellular vesicles. In parallel, electron microscopy analysis showed vesicles sitting next to the PM without fusion in absence of insulin in both control and ALMS1-silenced adipocytes (FIGS. 6E-F, top panels). Insulin treatment caused a swelling of the vesicles (FIG. 6E, middle panels) associated with fusion of vesicles with the PM for glucose absorption only in the control adipocytes (FIG. 6E, middle panels) but not in ALMS1-silenced adipocytes (FIG. 6F, middle panels). However, NIG. induced vesicular swelling and fusion with the PM in both control and ALMS1-silenced adipocytes (FIGS. 6E-F, bottom panels). The NIG treatment restored glucose absorption in ALMS1-silenced adipocytes. While insulin had little effect in inducing 2-NBDG absorption in ALMS1-silenced adipocytes (FIGS. 6G-H, top and middle panels), NIG not only restored vesicle fusion but could also be shown to restore 2-NBDG absorption in the ALMS1-silenced adipocytes to levels seen in control cells (FIGS. 6G-H, bottom panels) This restored glucose transport in NIG-treated ALMS1-silenced adipocytes correlated with restored GLUT4 fusion with the PM (FIG. 7A) but not with TBC1D4 targeting to the PM (FIG. 7B); and led to restoration of TG-filling of ALMS1-silenced adipocytes 24-hours post NIG treatment.

Example 7

(64) Identification of Peptide Inhibitors of PKC Binding to Alms1

(65) Once the site of binding interaction between two proteins is known, as known in the art it is possible using knowledge of the conformation and amino acids of each protein involved in mediating the interaction, to use computational models to design peptides or small molecule drugs which by binding in the region of the interaction site are able to sterically or otherwise hinder the binding interaction. The inventors therefore sought to identify peptides that would inhibit the interaction of ALMS1 and αPKC or TBC1D4 using their previously described ALMS1, TBC1D4 and αPKC structural models described in Example 4. Peptides predicted using this method to block the interaction between αPKC and ALMS1 included the sequences

(66) TABLE-US-00003 (SEQ ID NO: 5) LDSDSHYGPQHLESIDD, (SEQ ID NO: 6) DSHQTEETL, (SEQ ID NO: 7) QQTLPESHLP, (SEQ ID NO: 8) QALLDSHLPE. (SEQ ID NO: 9) PADQMTDTP, (SEQ ID NO: 10) HIPEEAQKVSAV or (SEQ ID NO: 11) SCIFLEQ.
A peptide identified using this method to block the interaction between TBC1D4 and ALMS1 was the sequence GCGAPAAREVILVL (SEQ ID NO: 12).

Example 8

(67) Expression of the Specific ALMS1-Interacting αPKC Interacting Domain in Mature Adipocytes Triggers Glucose Absorption in Absence of Insulin.

(68) Next, the inventors verified the hypothesis that insulin mediates the release of αPKC from the ALMSome complex in order to induce glucose absorption. For that, they cloned the interacting domain of αPKC (SEQ ID NOs: 14 and 15) in a lentiviral vector together with a Flag-TAG. The selected sequence was the minimum sequence of αPKC (min-αPKC-FLAG) so as to prevent sterical hindrance with the TBC1D4 interaction site on ALMS1. The expressed min-αPKC-FLAG in the adipocytes competes with the endogenous αPKC to prevent it from binding Almsome and hence favor the insulin-mediated TBC1D4 binding to Almsome. Mature adipocytes were then infected with either control or min-αPKC lentiviral particles to assess the impact of min-αPKC-FLAG on glucose absorption. 48 hours post-infection, min-αPKC-FLAG was immunodetected using an antibody against the FLAG-Tag (FIG. 9). For the in vitro proof of concept, we treated mature adipocytes as described (FIG. 9) and then incubated the treated mature adipocytes with 2-NBDG to assess the effect of min-αPKC-FLAG on glucose absorption. Of interest, 2-NBDG was absorbed in min-αPKC-FLAG treated adipocytes in absence of INS (FIG. 8A, left column) which corresponded to a 3.5 times increase compare to control (FIG. 8B). On the other hand, no significant difference was observed in presence of INS (FIGS. 8A, right column and 8B). These data demonstrate that targeting the interaction of ALMS1 and αPKC is sufficient to trigger glucose absorption in the adipocytes irrespective of the presence of INS.

(69) Production of Lentiviral Vector Carrying the αPKC Domain

(70) The ALMS1-interacting domain of human PKCα was amplified from human HEK293 cell cDNA with N-terminal FLAG tag using Forward 5′-gtacGAATTCGCCACCATGGATTACAAGGATGACGACGATAAGCTCACGGACTTCAAT TTCCTC-3′ (SEQ ID NO: 16) and Reverse 5′-tagcGGATCCTCATACTGCACTCTGTAAG ATGGG-3′ (SEQ ID NO: 17) primers and cloned into lentiviral vector pCDH-EF1-MCS-IRES-puro (System Biosciences). For virus production, PKCα lentiviral vectors were transfected into 293TN cells (System Biosciences) along with packaging plasmids psPAX2 and pMD2.G (Addgene) with the weight ratio of 3:2:1 respectively by using Lipofectamine 2000 (Life Technologies). Forty-eight hours after transfection, the culture supernatant was harvested by centrifugation at 500×g for 10 min, followed by filtration through 0.45 μm syringe filter with PES membrane (Sartorius). The virus solution was then concentrated by adding ½ volume of cold 30% (wt/vol) PEG6000 dissolved in 0.5M NaCl and incubated overnight at 4° C. with occasional mixing. The mixture was then centrifuged at 3000×g for 15 min at 4° C. Then the pellet containing lentiviral particles was resuspended in 1 mL DMEM medium and stored at −80° C. before infection of target cells.