Periodic Reporting for period 1 - SPHERES (Lipid droplet hypertrophy : the link between adipocyte dysfunction and cardiometabolic diseases)
Berichtszeitraum: 2020-09-01 bis 2022-02-28
Increased adipocyte size is the single adipose characteristic most strongly linked to cardiometabolic diseases, the most prevalent chronic non-communicable disease worldwide. Surprisingly, almost nothing is known about the underlying mechanisms and whether it can be therapeutically targeted. Deploying approaches ranging from biophysical to cellular and whole-body investigations, SPHERES addresses these fundamental biologically and clinically relevant issues. The goal of SPHERES is to understand the dynamics and consequences of adipocyte hypertrophy (enlargement) through investigation of its large lipid droplet (LD).
The initial part of the project was chiefly devoted to WP1 that forms the basis for the two subsequent WPs.
WP1. Beyond-state-of-the-art models/methods
The Rydén (Couchet) lab with Lauschke (Shen) and Langin (Dufau) groups have set up human and mouse 3D cell models referred as to adipocyte spheroids (Advanced Sciences, 2021). Kovacs and Araujo from the Antonny’s group went to Rydén to transfer the technique to IPMC. They have established its suitability for cell imaging and lipid analysis at the single spheroid level. The Rydén lab has been joined by Assoc Prof Niklas Mejhert, an expert in LD biology after 5 years at Harvard. Together with postdoc Couchet, they have generated an immortalized human adipocyte cell model highly suitable for adipocyte spheroids (MS in preparation).
The Langin and Rydén’s lab have developed human immortalized adipose stem cells (hiASCs) that differentiate into adipocytes. Using CRISPR/Cas9 gene editing in hiASCs, Langin’s lab produced adipocytes lacking perilipin 1 (Plin1) and hormone-sensitive lipase (HSL), two targets essential for adipocyte LD phenotype (Bottin, Griseti). In the Langin lab, CRISPR/Cas9 technology was used by Tavernier and Mouisel to produce mice lacking Plin1 and HSL (JCI Insight, 2022) . The Rydén/Mejhert lab have set up CRISPR/Cas9 editing in hiASCs for both deletions and endogenous tagging of proteins.
Using the lipid transport protein OSBP and HSL as pilot proteins, Kovacs in Antonny lab and Laurell/Griseti/Bottin in Langin lab have set up protocols for the efficient determination of the proximity landscape of membrane-bound and LD-associated proteins. The Rydén and Mejhert lab have successfully isolated the LD proteome of white adipose tissue from subjects with or without obesity. Proteomic analyses identify a set of novel hits that will be tested functionally by Drs Hmeadi, Frendo-Cumbo and a new postdoc. They have also set up the LD knowledge portal (https://lipiddroplet.org/) where together with Drs. Farese and Walther, they have integrated data from relevant studies on LDs (Developmental Cell, 2022).
In Langin lab, Flores-Flores used the deep learning-based segmentation method, Cellpose, to measure adipocyte size on mouse adipose tissue histological sections obtained by Mouisel, Denechaud and Marques. The lab bought a powerful computer for matrix calculus and deep learning. Measurement of adipocyte cell size is 10x more rapid than existing tools and is currently performed on more than 1000 adipose tissue images. Couchet in the Rydén and Mejhert lab and Kovacs in the Antonny lab have established imaging approaches for adipocyte spheroids.
Bello and Bigay in the Antonny lab have set up a protocol for the bacterial expression and purification of full-length human Plin1 using affinity tag-based chromatography and gel-filtration under urea conditions. They have determined the urea conditions at which Plin1 can be handled in solution, its intrinsically disordered nature, and its ability to be further incorporated into artificial droplets. Comparison of the biochemical and biophysical properties of Plin1 with that of Plin3 and Plin4 is done with Sabbagh (PhD student of the Copic lab) to define what distinguishes these LD proteins. Araujo (Antonny lab) has set up a protocol by which she can form artificial droplets of defined phospholipid composition, diameter and density. The model amphipathic protein Plin4, studied by Copic, was incorporated into the artificial droplets. This is instrumental for the formation of droplets covered with Plin1, mimicking the white adipocyte’s LD.
WP2. Importance of LD-associated proteins in fat cell hypertrophy and function
In Langin lab, Fournes-Fraresso and Denechaud phenotyped mice with inducible adipocyte-specific deletion of the adipose triglyceride lipase and HSL, the two main lipases controlling lipolysis. The Rydén and Mejhert lab have performed in vivo studies of insulin before and after weight loss. This shows that insulin action in the obese state is selectively impaired where parts of the lipogenic pathway remain intact (Diabetes, 2021). Using spatial transcriptomics of human white adipose tissue from adult humans, the Rydén/Mejhert lab have showed that human fat is composed of three distinct mature adipocyte subtypes. Biopsies obtained before and after insulin stimulation in vivo, showed that only one fat cell subtype responded to insulin (Cell Metabolism, 2021) suggesting that adipocyte heterogeneity impacts on metabolic phenotype. Many of the adipocyte subtype marker genes encode LD-associated proteins.
WP3. Lipid dynamics and interactions within large adipocyte LDs
Copic and d’Ambrosio have defined the molecular determinants that allow Plin4 to form a stable coat surface on LDs (eLife, 2021). They also compared the dynamics of the amphipathic regions of Plin1-4 revealing striking differences between Plins. Together with the evolutionary comparisons, this allows establishment of probes to monitor LD surface reactivity. The Copic lab (Franckhauser, Rezzik, Sabbagh) studies the amphipathic regions that control the balance of Plin4 between LD binding and amyloid fibril formation in vivo and in vitro and have set up cellular systems (Rezzik) for imaging studies. Using homology modelling and molecular dynamics simulations, Bigay and Gautier investigate the LD interaction regions of Plin1 focusing on the helical-bundle like-region. In Plin3, this forms a stable structure poorly available to the LD surface. In Plin1, however, the structure of this region seems looser making it a potential reservoir of additional LD-interacting helices. These molecular analyses allow studies of Plin chimeras to understand the interaction with LDs.
A lipid analysis methodology was established by Araujo (using thin layer chromatography) and by Debayle and Fleuriot (using liquid chromatography-mass spectrometry). Kovacs and Gautier are studying the impact of acyl chain length in triglyceride on the lipolysis stage both using the spheroid system and molecular dynamic simulations using the high-performance computer purchased for the SPHERES project.
WP1. Beyond-state-of-the-art models/methods
The Rydén (Couchet) lab with Lauschke (Shen) and Langin (Dufau) groups have set up human and mouse 3D cell models referred as to adipocyte spheroids (Advanced Sciences, 2021). Kovacs and Araujo from the Antonny’s group went to Rydén to transfer the technique to IPMC. They have established its suitability for cell imaging and lipid analysis at the single spheroid level. The Rydén lab has been joined by Assoc Prof Niklas Mejhert, an expert in LD biology after 5 years at Harvard. Together with postdoc Couchet, they have generated an immortalized human adipocyte cell model highly suitable for adipocyte spheroids (MS in preparation).
The Langin and Rydén’s lab have developed human immortalized adipose stem cells (hiASCs) that differentiate into adipocytes. Using CRISPR/Cas9 gene editing in hiASCs, Langin’s lab produced adipocytes lacking perilipin 1 (Plin1) and hormone-sensitive lipase (HSL), two targets essential for adipocyte LD phenotype (Bottin, Griseti). In the Langin lab, CRISPR/Cas9 technology was used by Tavernier and Mouisel to produce mice lacking Plin1 and HSL (JCI Insight, 2022) . The Rydén/Mejhert lab have set up CRISPR/Cas9 editing in hiASCs for both deletions and endogenous tagging of proteins.
Using the lipid transport protein OSBP and HSL as pilot proteins, Kovacs in Antonny lab and Laurell/Griseti/Bottin in Langin lab have set up protocols for the efficient determination of the proximity landscape of membrane-bound and LD-associated proteins. The Rydén and Mejhert lab have successfully isolated the LD proteome of white adipose tissue from subjects with or without obesity. Proteomic analyses identify a set of novel hits that will be tested functionally by Drs Hmeadi, Frendo-Cumbo and a new postdoc. They have also set up the LD knowledge portal (https://lipiddroplet.org/) where together with Drs. Farese and Walther, they have integrated data from relevant studies on LDs (Developmental Cell, 2022).
In Langin lab, Flores-Flores used the deep learning-based segmentation method, Cellpose, to measure adipocyte size on mouse adipose tissue histological sections obtained by Mouisel, Denechaud and Marques. The lab bought a powerful computer for matrix calculus and deep learning. Measurement of adipocyte cell size is 10x more rapid than existing tools and is currently performed on more than 1000 adipose tissue images. Couchet in the Rydén and Mejhert lab and Kovacs in the Antonny lab have established imaging approaches for adipocyte spheroids.
Bello and Bigay in the Antonny lab have set up a protocol for the bacterial expression and purification of full-length human Plin1 using affinity tag-based chromatography and gel-filtration under urea conditions. They have determined the urea conditions at which Plin1 can be handled in solution, its intrinsically disordered nature, and its ability to be further incorporated into artificial droplets. Comparison of the biochemical and biophysical properties of Plin1 with that of Plin3 and Plin4 is done with Sabbagh (PhD student of the Copic lab) to define what distinguishes these LD proteins. Araujo (Antonny lab) has set up a protocol by which she can form artificial droplets of defined phospholipid composition, diameter and density. The model amphipathic protein Plin4, studied by Copic, was incorporated into the artificial droplets. This is instrumental for the formation of droplets covered with Plin1, mimicking the white adipocyte’s LD.
WP2. Importance of LD-associated proteins in fat cell hypertrophy and function
In Langin lab, Fournes-Fraresso and Denechaud phenotyped mice with inducible adipocyte-specific deletion of the adipose triglyceride lipase and HSL, the two main lipases controlling lipolysis. The Rydén and Mejhert lab have performed in vivo studies of insulin before and after weight loss. This shows that insulin action in the obese state is selectively impaired where parts of the lipogenic pathway remain intact (Diabetes, 2021). Using spatial transcriptomics of human white adipose tissue from adult humans, the Rydén/Mejhert lab have showed that human fat is composed of three distinct mature adipocyte subtypes. Biopsies obtained before and after insulin stimulation in vivo, showed that only one fat cell subtype responded to insulin (Cell Metabolism, 2021) suggesting that adipocyte heterogeneity impacts on metabolic phenotype. Many of the adipocyte subtype marker genes encode LD-associated proteins.
WP3. Lipid dynamics and interactions within large adipocyte LDs
Copic and d’Ambrosio have defined the molecular determinants that allow Plin4 to form a stable coat surface on LDs (eLife, 2021). They also compared the dynamics of the amphipathic regions of Plin1-4 revealing striking differences between Plins. Together with the evolutionary comparisons, this allows establishment of probes to monitor LD surface reactivity. The Copic lab (Franckhauser, Rezzik, Sabbagh) studies the amphipathic regions that control the balance of Plin4 between LD binding and amyloid fibril formation in vivo and in vitro and have set up cellular systems (Rezzik) for imaging studies. Using homology modelling and molecular dynamics simulations, Bigay and Gautier investigate the LD interaction regions of Plin1 focusing on the helical-bundle like-region. In Plin3, this forms a stable structure poorly available to the LD surface. In Plin1, however, the structure of this region seems looser making it a potential reservoir of additional LD-interacting helices. These molecular analyses allow studies of Plin chimeras to understand the interaction with LDs.
A lipid analysis methodology was established by Araujo (using thin layer chromatography) and by Debayle and Fleuriot (using liquid chromatography-mass spectrometry). Kovacs and Gautier are studying the impact of acyl chain length in triglyceride on the lipolysis stage both using the spheroid system and molecular dynamic simulations using the high-performance computer purchased for the SPHERES project.
Development of new methods in WP1 have made good progress despite the Covid-19 pandemic. All tasks will be finalized during the next phase of the project. Using innovative WP1 technologies combined with advanced established techniques/models and unique clinical cohorts, we will determine the role of LD-associated proteins and their protein interactomes in determining LD size and adipocyte function. As described above, the work has been initiated for some of WP2 tasks. Similarly, results have already been obtained in WP3 where lipid dynamics and interactions within adipocyte LD are investigated.