Perilipin 5 Deletion Unmasks an Endoplasmic ReticulumStress–Fibroblast Growth Factor 21 Axis in SkeletalMuscleMagdalene K. Montgomery,1 Ruzaidi Mokhtar,1,2 Jacqueline Bayliss,1 Helena C. Parkington,1

Victor M. Suturin,1 Clinton R. Bruce,3 and Matthew J. Watt1

Diabetes 2018;67:594–606 | https://doi.org/10.2337/db17-0923

Lipid droplets (LDs) are critical for the regulation of lipidmetabolism, and dysregulated lipid metabolism contrib-utes to the pathogenesis of several diseases, includingtype 2 diabetes. We generated mice with muscle-specificdeletion of the LD-associated protein perilipin 5 (PLIN5,Plin5MKO) and investigated PLIN5’s role in regulating skel-etal muscle lipid metabolism, intracellular signaling, andwhole-body metabolic homeostasis. High-fat feeding in-duced changes in muscle lipid metabolism of Plin5MKO

mice, which included increased fatty acid oxidation andoxidative stress but, surprisingly, a reduction in inflamma-tion and endoplasmic reticulum (ER) stress. Thesemuscle-specific effects were accompanied by whole-body glucoseintolerance, adipose tissue insulin resistance, and reducedcirculating insulin and C-peptide levels in Plin5MKO mice.This coincided with reduced secretion of fibroblast growthfactor 21 (FGF21) from skeletal muscle and liver, resultingin reduced circulating FGF21. Intriguingly, muscle-secretedfactors from Plin5MKO, but not wild-type mice, reduced he-patocyte FGF21 secretion. Exogenous correction of FGF21levels restored glycemic control and insulin secretion inPlin5MKO mice. These results show that changes in lipidmetabolism resulting from PLIN5 deletion reduce ER stressin muscle, decrease FGF21 production by muscle and liver,and impair glycemic control. Further, these studies high-light the importance for muscle-liver cross talk in meta-bolic regulation.

Lipid droplets (LDs) are intracellular organelles comprisinga core of neutral lipids, including triglycerides and

sterols, surrounded by a phospholipid monolayer and pro-teins that concentrate at the LD surface. LDs are highlyconserved across species and in humans are stored in vir-tually every cell type, where they play critical roles in avariety of cellular processes, including the regulation of lipidmetabolism (1).

LDs reside at the intersection of lipid catabolism andlipid storage, and proteins surrounding the LD coordinatethe release of triglyceride-derived fatty acids that are usedfor mitochondrial b-oxidation and energy production, forgenerating cellular membranes and signaling lipids, and asactivators of nuclear receptors that impact transcriptionalregulation of metabolic pathways (2,3). Other LD-associatedproteins are involved in sequestering fatty acids into LDsduring times of energy surplus to protect the cell fromaccumulating excessive lipid intermediates (e.g., ceramidesand diacylglycerol) and the associated cellular dysfunctionsthat are broadly referred to as “lipotoxic” stress (4). Theimportance of LD proteins is highlighted in a variety ofhuman diseases that result from deletion or loss-of-function mutations of LD proteins, including neutral lipidstorage disease with myopathy (adipose triglyceride lipase[ATGL]) (5), Chanarin-Dorfman syndrome (comparativegene identification 58 [CGI-58]) (6), and partial lipodys-trophy, severe dyslipidemia, and insulin-resistant diabetes(perilipin 1 [PLIN1], seipin, and hormone-sensitive lipase[HSL]) (7).

Since the discovery of PLIN1, which is an abundantLD protein and is critical for the regulation of adipocyte

1Metabolism, Diabetes and Obesity Program, Monash Biomedicine DiscoveryInstitute, and Department of Physiology, Monash University, Clayton, Victoria,Australia2Biotechnology Research Institute, Universiti Malaysia Sabah, Kota Kinabalu,Sabah, Malaysia3Institute for Physical Activity and Nutrition, School of Exercise and NutritionSciences, Deakin University, Burwood, Victoria, Australia

Corresponding author: Matthew J. Watt, matthew.watt@monash.edu, orMagdalene K. Montgomery, magda.montgomery@monash.edu.

Received 7 August 2017 and accepted 16 January 2018.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db17-0923/-/DC1.

© 2018 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, and thework is not altered. More information is available at http://www.diabetesjournals.org/content/license.

594 Diabetes Volume 67, April 2018

METABOLISM

metabolism, intense research efforts have attempted toidentify the LD proteome and the proteins’ functions inmany cell types. Prominent among this list is PLIN5, whichis abundantly expressed in tissues with high oxidative ca-pacity, such as the heart, liver, and skeletal muscle (8), andplays an important role in LD dynamics by modulating theflux of fatty acids from LDs to match the energy require-ments of the cell (9). PLIN5 achieves this by interactingwith and coordinating the actions of the key lipolytic pro-teins ATGL, CGI-58, and HSL (10,11). Moreover, PLIN5appears to be important in the adaptation to nutrientand cellular stress, with PLIN5 expression increasing afterhigh-fat feeding and endurance exercise training (9). Fewstudies have focused on understanding how PLIN5 impactscell signaling pathways to modulate this adaptation to nu-trient stress. A recent study showed that skeletal musclePLIN5 has the capacity to stimulate fibroblast growth factor21 (FGF21) expression in muscle and subsequently increasecirculating FGF21, suggesting that muscle PLIN5 contrib-utes to systemic metabolic control (12). The increase inmuscle FGF21 was accompanied by endoplasmic reticulum(ER) stress, particularly Atf4 within the PERK/eIF2a path-way, which is a potent transcriptional activator of FGF21(13). This is of special interest as LDs originate from andreside in close proximity to the ER and have been impli-cated in ER stress and activation of the unfolded proteinresponse (UPR) (14).

Skeletal muscle plays a critical role in maintaining whole-body energy homeostasis because it is an important sitefor lipid disposal and oxidation and is the major tissue re-sponsible for insulin-stimulated glucose uptake. Muscle insulinresistance is a pathological feature of type 2 diabetes, and LDsfigure prominently in pathologies of insulin resistance, in-cluding obesity and type 2 diabetes (15). In this study, wehave generated mice with muscle-specific deletion of Plin5to determine the role of PLIN5 in regulating skeletal musclemetabolism, cellular stress, and FGF21 expression/secretionduring dietary lipid overload and, in turn, to understand theimpact of skeletal muscle PLIN5 on glycemic control.

RESEARCH DESIGN AND METHODS

Ethics and Experimental DesignExperimental procedures were approved by the MonashUniversity Animal Ethics Committee (MARP/2013/050)and conformed to the National Health and Medical Re-search Council of Australia guidelines regarding the careand use of experimental animals. Generation of muscle-specific Plin5 knockout mice (Plin5MKO) on a C57BL/6 back-ground has been previously described (16). Male mice werehoused at 22°C on a 12:12-h light-dark cycle and were fedeither a rodent chow (5% energy from fat) (Specialty Feeds,Glen Forrest, Australia) or a high-fat, high-sucrose diet(HFD; 43% energy from fat) (High Fat Rodent Diet SF04-001; Specialty Feeds) starting at 8 weeks of age for a total of8 weeks. Body weight was measured weekly. Mice werefasted from 0700 to 1100 h before all experiments unlessotherwise stated.

Body Composition and Energy ExpenditureFat and lean body mass were measured using DEXA(Lunar Pixi; PIXImus, Fitchburg, WI) as previously described(17). Oxygen consumption, carbon dioxide production,physical activity, and food intake were measured usinga 12-chamber Oxymax indirect calorimetry system (Colum-bus Instruments, Columbus, OH). Studies were com-menced after 8 h of acclimation to the metabolic chamber.Expired gases were assessed at 30-min intervals for 48 hat 22°C.

Glucose and Insulin Tolerance TestsMice were gavaged orally with glucose (2 g/kg body weight)or injected in the intraperitoneal (i.p.) cavity with insulin(0.75 units/kg body weight). Blood glucose levels weremonitored (Accu-Chek II glucometer; Roche Diagnostics,Castle Hill, Australia), and plasma insulin and C-peptidelevels were determined by ELISA (Ultra-Sensitive MouseInsulin ELISA/Mouse C-peptide ELISA Kit; Crystal Chem,Elk Grove Village, IL).

Tissue-Specific Glucose UptakeMice were fasted for 4 h and gavaged orally with glucose(2 g/kg body weight) and simultaneously injected i.p.with 2-[1-14C]-deoxy-D-glucose (2DG; 10 mCi/mouse)(PerkinElmer, Melbourne, Australia). Blood glucose concen-tration and tracer levels were determined as indicated (Fig.4E). Blood was deproteinized and tissue-specific 2DG up-take determined as described previously (18). It is notedthat the tracer and tracee were administered by differentroutes, which is not consistent with the principles of glu-cose kinetics analysis. While acknowledging this limitation,this approach allows determination of 2DG accumulation intissues under conditions of hyperglycemia and hyperinsuli-nemia. In a separate set of experiments, 2DG uptake wasmeasured in inguinal adipose tissue ex vivo, as describedpreviously (19).

Lipid AnalysesLipids were extracted in chloroform:methanol (2:1 volumefor volume), and the phases were separated with 4 mmol/LMgCl2 as previously described (20). Triacylglycerol contentwas determined by colorimetric assay (Triglycerides GPO-PAP; Roche Diagnostics, Indianapolis, IN). Ceramide anddiacylglycerol were analyzed by mass spectrometry as pre-viously described (21). Plasma nonesterified fatty acids(NEFAs) and total plasma cholesterol were measured bycolorimetric assay (Wako Diagnostics, Osaka, Japan).Thiobarbituric acid reactive substance (TBARS) and lipidhydroperoxides were determined as previously described(22,23).

Plasma MetabolitesPlasma b-hydroxybutyrate levels were measured by colori-metric assay (Sapphire Biosciences, Redfern, Australia).Plasma and medium FGF21 (Mouse/Rat FGF-21 QuantikineELISA; R&D Systems, Minneapolis, MN) and tumor necro-sis factor-a (TNFa) were measured by ELISA (Mouse TNFa

diabetes.diabetesjournals.org Montgomery and Associates 595

ELISA; elisakit.com, Scoresby, Australia). Aspartate amino-transferase (AST) and alanine aminotransferase (ALT) ac-tivities were determined by enzymatic colorimetric assays(24,25).

Fatty Acid MetabolismFatty acid metabolism in isolated soleus muscle was mea-sured as previously described (17).

ImmunoblottingLysates were prepared in radioimmunoprecipitation assaybuffer, lanes were loaded relative to whole-tissue proteincontent (20 mg protein per lane), proteins were resolved bySDS-PAGE electrophoresis, and immunoblot analysis wasconducted as described previously (16). Primary antibodiesare listed in Supplementary Table 1 and have been validated(16,26–29).

Analysis of Gene ExpressionRNA extraction, cDNA synthesis, and real-time PCR wereperformed as previously described in detail (29). The primersequences are listed in Supplementary Table 2.

Heart PerfusionsMice were anesthetized with 5% isoflurane, the chestopened, the aorta clamped, and the heart removed intoice-cold calcium-free Hanks’ balanced salt solution. A 22-gstainless steel tube was introduced into the aorta, with theopening just above the coronary sinus, and secured inplace. The heart was mounted on a Langendorff apparatus(ML870B2; AdInstruments, Bella Vista, Australia) and per-fused through the coronary network with bubbled (95% O2/5% CO2) EX-CELL 325 Protein-Free CHO Serum-Free Me-dium (SAFC Biosciences) at 37°C for 2 h (recirculation of6 mL of medium). The heart was enclosed by a heatingjacket to ensure stable temperature and humidity. Flowwas gradually increased (displayed on LabChart v7 software)until coronary network pressure was 80 mmHg (Gould pres-sure transducer).

Conditioned Media ExperimentsIntact soleus and extensor digitorum longus (EDL) muscle(each 10–12 mg) were excised from HFD-fed mice andplaced in 0.5 mL pregassed (95% O2/5% CO2) EX-CELLmedium at 37°C for 6 h. Liver slices (;400 mm thickness,20 mg) were incubated under the same conditions. FGF21was measured in the “conditioned medium” (CM; i.e., me-dium containing secreted factors). In addition, the CM frommuscles was applied to primary murine hepatocytes (26) for24 h after 4-h adherence of hepatocytes to the cultureplates (50% CM, 50% M199 hepatocyte medium). Theseexperiments were replicated using heart CM. In separateexperiments, 10% serum was added to the hepatocyte me-dium for 24 h. FGF21 secretion from hepatocytes was mea-sured and corrected for FGF21 present in the muscle CM(contributed 1–2% of total medium FGF21) and in theplasma (3–11% of total FGF21). Interassay variation wasminimized by combining hepatocytes from three mice foreach experiment.

FGF21 “Add-Back” ExperimentsPlin5MKO mice were injected i.p. with 5 ng recombinantmurine FGF21 (Sino Biologicals, Beijing, China) 1 h beforean oral glucose tolerance test (OGTT). Blood glucose,plasma FGF21, insulin, and C-peptide were assessed as in-dicated (Fig. 6).

Statistical AnalysisResults are presented as mean 6 SEM. Data were ana-lyzed with an unpaired two-tailed Student t test or a two-way ANOVA with Bonferroni post hoc analysis, whereappropriate. Statistical significance was accepted atP , 0.05.

RESULTS

Metabolic Profiling of Plin5MKO Mice Fed a StandardLow-Fat DietSimilar to our recent publication (16), Plin5 mRNA wasreduced by 73% and 97% in quadriceps muscle and heartof Plin5MKO mice, respectively (Fig. 1A), without changes inother tissues or other LD-associated proteins (PLIN2 andCGI-58) (Supplementary Fig. 1A–C). Whereas body weightwas unchanged in Plin5MKO compared with wild-type (WT)mice (Fig. 1B), there was an increase in fat mass in Plin5MKO

mice (Fig. 1C) and no change in lean mass between geno-types (Fig. 1D). Energy expenditure was not different be-tween genotypes (Fig. 1E), whereas respiratory exchangeratio (RER) was decreased in Plin5MKO mice (Fig. 1F), dem-onstrating increased whole-body fatty acid oxidation. Foodintake was not different between genotypes (WT 3.33 60.07 [n = 9], Plin5MKO 3.03 6 0.13 [n = 8] g/mouse/day).Glucose clearance during an OGTT was indistinguish-able between WT and Plin5MKO mice (Fig. 1G), as wasplasma insulin measured in fasted mice and during theOGTT (Fig. 1H). There was no difference in insulin action(Fig. 1I).

Metabolic Profiling of Plin5MKO Mice Fed an HFDWe next determined whether increasing dietary fat contentwould reveal metabolic alterations in Plin5MKO mice becausePLIN5 is upregulated by high-fat feeding (30). Whereastotal body weight (Fig. 2A) and individual tissue weights(Supplementary Table 3) were similar between genotypes,lean mass was elevated by 15% (Fig. 2B) and whole-bodyadiposity tended to be decreased (P = 0.099) (Fig. 2C) inPlin5MKO mice. Food intake was similar between genotypes(WT 2.40 6 0.05 [n = 5], Plin5MKO 2.35 6 0.16 [n = 5]g/mouse/day).

Energy expenditure was similar (Fig. 2D) and whole-body fatty acid oxidation was increased in Plin5MKO com-pared with WT mice (Fig. 2E). Skeletal muscle fatty acidoxidation was significantly increased in Plin5MKO mice (Fig.2F), whereas triglyceride esterification tended to be reduced(Fig. 2G), pointing toward fatty acids being preferentiallyshuttled into oxidative pathways in skeletal muscle ofPlin5MKO mice. Fatty acid uptake was not different betweengenotypes (Fig. 2H).

596 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018

Molecular Changes Mediating the Differences inSkeletal Muscle Lipid Metabolism in Plin5MKO MiceThe abundance of electron transport chain complexes wassimilar in skeletal muscle of WT and Plin5MKO mice (Fig. 2Iand Supplementary Fig. 1D–H). Interestingly, mRNA and pro-tein contents of acetyl-CoA carboxylase 2 (ACC2), an enzymethat regulates the entry of fatty acids into the mitochondria,was decreased in Plin5MKO mice (Fig. 2J–L). Whereas phos-phorylation at the deactivating Ser212 residue of ACC2 wasmarkedly decreased in Plin5MKO mice (Fig. 2K), there was nodifference between genotypes when expressed per unit ofACC2 protein (Fig. 2M). There was no difference in phos-phorylated (Thr172) or total AMPK (Supplementary Fig. 1Iand J), the upstream regulator of ACC. Hence, the profounddecrease in ACC2 abundance in skeletal muscle is likely tocontribute to the observed increase in fatty acid oxidation.

Increasing fatty acid–derived mitochondrial electron flowcan lead to oxidative stress. Oxidative stress was increasedin the skeletal muscle of Plin5MKO mice compared with WTmice as indicated by increased lipid hydroperoxide (Fig. 2N)and TBARS (Fig. 2O). The increased muscle fatty acid oxi-dation in Plin5MKO mice also coincided with a decrease inintramyocellular triglyceride content (Fig. 2P), indicativeof increased lipolysis and oxidation of triglyceride-derivedfatty acids. No differences were observed in liver triglycer-ide content (WT 17.4 6 1.5 [n = 6], Plin5MKO 20.6 6 1.2[n = 7] mmol/g tissue) or in circulating triglycerides, NEFAs,and total plasma cholesterol content between WT andPlin5MKO mice (Table 1). There were no differences in mus-cle ceramide and diacylglycerol contents (SupplementaryFig. 2A–D) between genotypes.

Figure 1—Metabolic profiling in WT (white) and Plin5MKOmice (black) fed a standard chow diet. A: Plin5mRNA in quadriceps (Quad), soleus, andEDL muscle, heart, liver, and epididymal (Epi) adipose tissue. B: Body weight (WT n = 18, Plin5MKO n = 16). C: Fat mass (WT n = 4, Plin5MKO n =7). D: Lean mass (WT n = 4, Plin5MKO n = 7). E: Whole-body oxygen consumption (WT n = 13, Plin5MKO n = 10). F: RER (WT n = 13, Plin5MKO n =10). G: Glucose tolerance as determined by OGTT (2 g/kg body weight) (WT n = 4, Plin5MKO n = 7). H: Plasma insulin levels during the OGTT (WTn = 4, Plin5MKO n = 7). I: Changes in blood glucose levels during an insulin tolerance test (0.75 units/kg body weight) (WT n = 4, Plin5MKO n = 7).Data are means 6 SEM. *P , 0.05. WAT, white adipose tissue.

diabetes.diabetesjournals.org Montgomery and Associates 597

Figure 2—Metabolic profiling in WT (white) and Plin5MKO mice (black) fed an HFD. A: Body weight (WT n = 21, Plin5MKO n = 19). B: Lean mass(WT n = 5, Plin5MKO n = 3). C: Fat mass (WT n = 5, Plin5MKO n = 3). D: Whole-body oxygen consumption (WT n = 10, Plin5MKO n = 8). E: RER (WTn = 10, Plin5MKO n = 8). F: Muscle fatty acid oxidation (WT n = 10, Plin5MKO n = 10).G: Muscle triglyceride (TAG) esterification (WT n = 6, Plin5MKO

n = 6). H: Total muscle fatty acid uptake (WT n = 5, Plin5MKO n = 5). I: Protein levels of mitochondrial content, biogenesis, and oxidative capacitymarkers in whole-muscle lysates (lanes were loaded relative to whole-muscle protein content). J: Muscle Acacb mRNA (WT n = 8, Plin5MKO n =9). K: ACC2 protein content. L: Immunoblot quantification of total protein content of ACC2 (WT n = 7, Plin5MKO n = 8). M: Immunoblotquantification of pACC2/ACC2 (WT n = 7, Plin5MKO n = 8). N: Muscle lipid peroxidation, measured as lipid hydroperoxide (LOOH) content(WT n = 12, Plin5MKO n = 10). O: Muscle TBARS levels (WT n = 12, Plin5MKO n = 10). P: Muscle triglyceride (TAG) content (WT n = 7, Plin5MKO n =7). Data are means 6 SEM. *P , 0.05. Representative immunoblotting shows n = 2 per group, but respective quantification analysis wasperformed on n = 7–8 per group. Experiments were performed on four independent cohorts of mice, and analyzed data are from all cohortscombined. ETC C., electron transport chain complex.

598 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018

Altered Patterns of Lipid Storage Are Associated WithDecreased ER StressChanges in lipid catabolism and LD formation are likely toimpact ER stress (14), and deletion of Plin5 impacts theseprocesses in skeletal muscle (Fig. 2). Accordingly, we ana-lyzed the three major ER stress–induced signaling pathwaysin skeletal muscle of WT and Plin5MKO mice, which includethe IRE1/XBP1, PERK/eIF2a, and ATF6 pathways. UnderER stress, IRE1a is autophosphorylated at Ser724 andremoves a 26–base pair intron to produce spliced XBP1,which in turn regulates expression of genes involved inthe UPR. Plin5 deletion in skeletal muscle decreased IRE1aSer724 compared with WT muscle (Fig. 3A), the mRNAexpression (data not shown) and protein contents of thespliced isoform of XBP (XBP1s) (Fig. 3B), and mRNA ex-pression of various transcription targets of XBP1s, includ-ing Edem1, Dnajb11 (HEDJ), Dnajc3 (P58IPK), and Hspa5(GRP78/Bip) (Fig. 3C). In accordance with reduced IRE1aactivation, phosphorylation of the downstream stress kinaseJNK was substantially reduced in the muscle of Plin5MKO

compared with WT mice (Fig. 3D). Together, these datademonstrate a marked reduction of the IRE1 pathway inPlin5MKO skeletal muscle.

There was also evidence for mild inhibition of the PERK/eIF2a pathway; whereas eIF2a phosphorylation (Ser51)was unchanged, ATF4 mRNA content was markedly re-duced (Fig. 3E). We were unable to reliably detect PERKin skeletal muscle using several antibodies. Nuclear factorNrf2 regulates ATF4 transcription (31) and was reduced inPlin5MKO muscle (Fig. 3E). XBP1s and ATF4 regulate thetranscription factor C/EBP homologous protein (CHOP).CHOP mRNA was decreased by 77% and CHOP proteincontent was reduced by 30% (not significant) (Fig. 3F). Incontrast, the ATF6 pathway was unaffected by Plin5 dele-tion (Fig. 3G).

In addition to reduced ER stress signaling, the mRNAcontent of the proinflammatory markers TNFa, IL6, and

MCP1 was decreased in the skeletal muscle of Plin5MKO

compared with WT mice (Fig. 3H). Plasma TNFa was notdifferent between genotypes, indicating localized ratherthan systemic effects (Table 1). Finally, PLIN2 has beenassociated with p62-mediated lipophagy in cultured myo-tubes (32), indicating a potential role for PLIN5 in autophagy-regulated clearance of lipids. There was no differencebetween genotypes for the autophagy markers p62 andthe LC3 (II/I) ratio (Fig. 3I).

ER Stress Is Unchanged in the Heart of Plin5MKO MiceMCK-Cre induced a substantial reduction in Plin5 in theheart (Fig. 1A), and, similar to skeletal muscle, Plin5MKO

mice exhibited a 55% reduction in cardiac triglyceride con-tent (Supplementary Fig. 3A). In contrast to skeletal muscle,ACC2 phosphorylation and protein content and markers ofER stress and inflammation were not different betweengenotypes (Supplementary Fig. 3B–N), and lipid hydroper-oxide levels (Supplementary Fig. 3O) and TBARS (Supple-mentary Fig. 3P) were significantly reduced in the hearts ofPlin5MKO mice. These results demonstrate cell-autonomousroles for PLIN5 in skeletal and cardiac muscle.

Effects of Muscle-Specific Plin5 Deletion on GlucoseMetabolismGiven that oxidative and ER stress and inflammation causeglucose intolerance and insulin resistance (33), we next in-vestigated glucose metabolism in mice. Glucose clearancewas impaired in Plin5MKO compared with WT mice afteran oral glucose load (Fig. 4A), and this was accompaniedby a mild reduction in plasma insulin levels (Fig. 4B) andplasma C-peptide levels (Fig. 4C) in Plin5MKO mice. Plin5MKO

mice also exhibited a mild impairment in whole-body in-sulin action (Fig. 4D). To ascertain tissue-specific differen-ces in glucose disposal, mice were administered 2DG duringan OGTT. Blood glucose was higher in Plin5MKO mice30 min after glucose administration (WT 19.1 6 1.0 vs.Plin5MKO 22.36 1.2 mmol/L, P = 0.05), and the radiotracerconcentration in the blood was similar between genotypes(Fig. 4E). Accumulation of 2DG uptake in skeletal muscle(Fig. 4F) and heart (Fig. 4G) was similar between genotypes,whereas 2DG content was significantly reduced in the whiteadipose tissue of Plin5MKO compared with WT mice (Fig.4H). The finding was confirmed in ex vivo 2DG uptakeexperiments in adipose tissue (Fig. 4I).

Plin5 Deficiency in Skeletal Muscle Impacts Local andSystemic FGF21 LevelsGiven that the data demonstrated no role of muscle-specificPlin5 deletion on muscle glucose uptake, but rather in ad-ipose tissue glucose uptake and reduced plasma insulin andC-peptide levels, we hypothesized that an endocrine fac-tor might be responsible for actions in tissues distant toskeletal muscle. XBP1s, ATF4, and CHOP are powerful tran-scriptional activators of FGF21 (13,34), and these proteinswere reduced in the skeletal muscle of Plin5MKO mice (Fig.3B, E, and F). Consistent with this molecular regulation,FGF21 mRNA content was markedly reduced in the skeletal

Table 1—Plasma metabolites in WT and Plin5MKO mice fed anHFD

WT Plin5MKO

Glucose (mmol/L) 10.5 6 0.4 10.7 6 0.2

Insulin (pmol/L) 698 6 110 527 6 88

Triglyceride (mmol/L) 538 6 45 609 6 85

NEFA (mmol/L) 484 6 83 478 6 50

Cholesterol (mmol/L) 3.6 6 0.7 4.0 6 0.8

b-Hydroxybutyrate (mmol/L) 566 6 58 452 6 22

ALT (units/mL) 1.88 6 0.38 0.82 6 0.20*

AST (units/mL) 0.78 6 0.25 0.31 6 0.09

TNFa (pg/mL) 7.2 6 1.2 6.4 6 1.3

Values are mean 6 SEM. Glucose: WT n = 12, Plin5MKO n = 10;insulin: WT n = 11, Plin5MKO n = 10; triglyceride/TNFa: WT n = 6,Plin5MKO n = 6; NEFA/b-hydroxybutyrate: WT n = 6, Plin5MKO n =5; cholesterol: WT n = 5, Plin5MKO n = 5; ALT/AST: WT n = 5,Plin5MKO n = 4. *P , 0.05.

diabetes.diabetesjournals.org Montgomery and Associates 599

Figure 3—Analysis of ER stress pathways in muscle of HFD-fed WT (white) and Plin5MKO (black) mice. A: IRE1 phosphorylation (pIRE1). B:mRNA (WT n = 10, Plin5MKO n = 8) and protein content of the spliced isoform of XBP1. C: Changes in mRNA transcripts of XBP1 target genes,including Hedj, Edem1, Dnajc3 (i.e., P58IPK), and Hspa5 (i.e., GRP78) (WT n = 9, Plin5MKO n = 9). D: JNK phosphorylation (pJNK). E: eIF2aphosphorylation (peIF2a) and mRNA levels of its target ATF4 (WT n = 9, Plin5MKO n = 10) as well as Nfe2l2 mRNA (WT n = 9, Plin5MKO n = 9). F:mRNA (WT n = 9, Plin5MKO n = 10) and protein content of CHOP.G: Protein content of full-length and spliced ATF6, as well as its target Hsp90b1mRNA (WT n = 8, Plin5MKO n = 10). H: mRNA levels of TNFa, IL6, and Ccl2 (WT n = 8, Plin5MKO n = 9). I: Protein content of the autophagymarkers p62 and LC3 (I + II). Data are means 6 SEM. *P , 0.05. Representative immunoblotting shows n = 2 per group, but respectivequantification analysis was performed on n = 8 per group for all protein measures. Experiments were carried out on four independent cohorts ofmice, and analyzed data are from all cohorts combined.

600 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018

muscle of Plin5MKO compared with WT mice fed an HFD(Fig. 5A), and this was accompanied by a 55% reduction inplasma FGF21 (Fig. 5B). We also assessed liver FGF21 be-cause this is thought to be the primary source of circulatingFGF21 (35). Liver FGF21 mRNA was significantly reducedin Plin5MKO compared with WT mice (Fig. 5C), suggestingthat factors secreted from skeletal muscle may impactFGF21 expression in the liver. FGF21 mRNA was not dif-ferent between genotypes in quadriceps muscle or liver ofchow-fed mice, or in the heart and inguinal adipose tissueof HFD mice (Supplementary Fig. 4A–D).

To test the possibility of skeletal muscle–liver tissuecross talk, secreted products were collected from glycolytic(EDL) and oxidative (soleus) skeletal muscle ex vivo and the

CM was added to primary murine hepatocytes. Intriguingly,muscle CM from Plin5MKO mice, but not WT mice, re-duced FGF21 secretion from hepatocytes (Fig. 5D). Se-creted factors from the heart of Plin5MKO mice did notimpact liver FGF21 secretion (Fig. 5D). We also assessedFGF21 secretion from muscle and liver to ascertain thepotential contribution of each tissue toward circulatingFGF21 levels in WT and Plin5MKO mice. Interestingly,FGF21 secretion was per unit mass higher in muscle com-pared with liver (Fig. 5E). Moreover, FGF21 secretion wasreduced in the soleus muscle and liver of Plin5MKO mice,whereas no differences were observed between genotypes inEDL muscle or heart (Fig. 5F). Interestingly, addition of10% serum from Plin5MKO mice to the hepatocyte culture

Figure 4—Glucose metabolism in HFD-fed WT and Plin5MKO mice. A: Glucose tolerance as determined by OGTT (2 g/kg body weight) (WT n =12, Plin5MKO n = 10). P , 0.001 main effect for genotype by two-way ANOVA; *P , 0.05 by Bonferroni post hoc test. B and C: Plasma insulinand plasma C-peptide levels during the OGTT (WT n = 11, Plin5MKO n = 10). D: Changes in blood glucose levels during an i.p. insulin tolerancetest (0.75 units/kg body weight) (WT n = 12, Plin5MKO n = 11). *P , 0.05 by Bonferroni post hoc test. E–H: Mice were injected with 10 mCi 2DGand gavaged with 2 g glucose/kg body weight, and tissues were collected 30 min after injection. E: Radioactivity in the blood. Tissueaccumulation of 2DG after an OGTT in skeletal muscle (F), heart (G), and epididymal adipose tissue (H) (WT n = 12, Plin5MKO n = 10). I: 2DGuptake into inguinal adipose tissue after 100 nmol/L stimulation with insulin for 20 min ex vivo (WT n = 5, Plin5MKO n = 5). Data are means6 SEM.*P, 0.05. Experiments in A–D were carried out on four independent cohorts of mice, experiments in E–H on three independent cohorts of mice,and I on two cohorts of mice, and analyzed data are from all cohorts combined. CPM, counts per minute.

diabetes.diabetesjournals.org Montgomery and Associates 601

medium resulted in a 30% reduction in hepatocyte FGF21secretion, compared with hepatocytes incubated with WTserum (Fig. 5G). Thus, FGF21 appears to be decreased inthe circulation of high fat–fed Plin5MKO mice as a result ofreduced expression and secretion from skeletal muscle andliver, the latter effect being regulated by an unknown mus-cle-secreted factor.

Rescue of Serum FGF21 in Plin5MKO Mice RestoresSystemic Glucose MetabolismTo test if the reduced plasma FGF21 levels in Plin5MKO miceare likely to impact adipose tissue, we examined the expres-sion of known FGF21 target genes in adipose tissue. Ucp1,Cidea, and Cpt1a expression was significantly reduced inPlin5MKO mice (Fig. 6A–C). Expression of FGF21 nontargets(36) Kctd20 and Spata13 was unchanged (data not shown).

We next tested whether correction of the reduced serumFGF21 could “rescue” the metabolic defects observed inPlin5MKO mice. Injection of recombinant murine FGF21 inPlin5MKO mice was sufficient to increase plasma FGF21 tolevels observed in WT mice (Fig. 6D), to rescue glucose in-tolerance (Fig. 6F), and to restore insulin secretion (Fig. 6G).

DISCUSSION

The results reveal an unanticipated link between lipidmetabolism, ER homeostasis, and the regulation of myokineproduction. Our data show that PLIN5 is required tomaintain the appropriate balance between fatty acid storageand fatty acid oxidation during periods of lipid overload andthat failure to do so causes oxidative stress in skeletal

muscle. When comparing the responses of high fat–fed WTand Plin5MKO mice, the data suggest that intracellular lipidaccumulation activates the UPR in skeletal muscle and thatthis drives the production and secretion of the myokineFGF21, which is known to improve adipose tissue insulinsensitivity and pancreatic insulin secretion. This is of in-terest as an activation of the UPR is commonly associatedwith a deterioration of insulin sensitivity (37). Moreover,this study has identified novel interorgan communicationbetween muscle and liver. PLIN5 deletion alters the factorssecreted by skeletal muscle, which in turn reduce the pro-duction and secretion of FGF21 by the liver, suggestinga potential mechanism by which altered intramyocel-lular lipid metabolism can regulate systemic metabolichomeostasis.

The role of LD proteins in the regulation of lipid andglucose metabolism has garnered much attention in recentyears, particularly in skeletal muscle, which is a criticalregulator of whole-body fatty acid oxidation and insulin-stimulated glucose uptake. There is particular interest inPLIN5’s role in these processes because of its high expres-sion in skeletal muscle and its regulation in response tophysiological perturbations that modulate metabolism, in-cluding high-fat feeding, fasting, and exercise (9). Whole-body fat oxidation was increased in Plin5MKO mice and wasaccompanied by a partitioning of fatty acids away fromstorage and toward oxidation, resulting in reduced intra-myocellular triglyceride accumulation in both skeletal andheart muscle. This finding is consistent with previous

Figure 5—FGF21 as a potential modulator of whole-body glucose metabolism in HFD-fed mice. A: FGF21 mRNA levels in quadriceps muscle(WT n = 8, Plin5MKO n = 9; data from three independent cohorts of mice). B: Plasma FGF21 content (WT n = 22, Plin5MKO n = 21; data from fiveindependent cohorts of mice). C: FGF21 mRNA levels in liver (WT n = 12, Plin5MKO n = 13; data from four independent cohorts of mice). D:FGF21 secretion from primary hepatocytes incubated with CM (i.e., media containing secreted factors) from EDL and soleus muscle (WT n = 12,Plin5MKO n = 14) and from heart (WT n = 5, Plin5MKO n = 5). E: Ex vivo FGF21 secretion from isolated intact EDL muscle and soleus muscle, andFGF21 secretion from liver slices (WT n = 12, Plin5MKO n = 14; data from two independent cohorts of mice). F: Ex vivo FGF21 secretion fromintact heart using the Langendorff heart perfusion technique (WT n = 5, Plin5MKO n = 5; one cohort of mice). G: FGF21 secretion from primaryhepatocytes incubated with media containing 20% serum of HFD-fed WT (white) and Plin5MKO (black) mice (WT n = 6, Plin5MKO n = 6; plasmaused from five independent cohorts of mice). Data are means 6 SEM. *P , 0.05.

602 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018

reports (12,38,39) and the premise that LDs are a centralhub for fatty acid trafficking in muscle (3,40) and thatPLIN5’s major function is to reduce lipolysis by bind-ing ATGL and CGI-58 and preventing their interaction(11,41). The increase in fatty acid oxidation was not relatedto molecular changes reflecting altered mitochondrial con-tent or oxidative capacity, which contradicts a recent reportdescribing PLIN5 as a nuclear protein that promotes tran-scriptional regulation of genes that mediate oxidative func-tion (42), but rather was associated with a substantialreduction in ACC2 mRNA and protein content in muscle.ACC2 produces malonyl CoA, which allosterically inhibitsthe mitochondrial fatty acid transporter CPT1 and therebysuppresses fatty acid oxidation. ACC2 transcription is reg-ulated by several transcription factors, including XBP1s (i.e.,spliced isoform of XBP1), which is activated as part ofthe UPR during ER stress (43). XBP1 was substantially de-creased in the muscle of Plin5MKO mice, concomitant with

a drastic reduction in ACC2 mRNA and protein, acceleratedfatty acid oxidation, and oxidative stress. Changes inACC2 were independent of changes in the upstream re-gulator AMPK. Collectively, these data suggest that PLIN5is required to coordinate shuttling of intracellular tri-glyceride lipolysis toward fatty acid oxidation and thatdeleting this function disrupts the normal sensing of ERstress and results in excessive fatty acid oxidation. Al-though speculative, this lipid-sensing mechanism may berequired to protect cells from further cellular stress (e.g.,oxidative stress) in the face of lipid accumulation and ERstress.

ER stress is induced by several metabolic perturbations,and, although it remains unclear how lipids cause ER stressand activate the UPR, accumulating evidence shows thatthis conserved response plays an important role in main-taining metabolic and lipid homeostasis. LDs originate fromand reside in close proximity to the ER and have been

Figure 6—FGF21 as a potential modulator of whole-body glucose metabolism. Decreased mRNA expression of the FGF21 target genes UCP1(A), CideA (B), and CPT1a (C) in inguinal adipose tissue (WT n = 8, Plin5MKO n = 10; data from two independent cohorts of mice). To determine ifthe decrease in plasma FGF21 in Plin5MKO was driving the whole-body metabolic phenotype, Plin5MKO were injected with 5 ng of FGF21recombinant protein (rFGF21), followed by an OGTT and blood collection 60 min postinjection. D: Plasma FGF21 in WT and Plin5MKO micebefore (left) and after (right) rFGF21 injection (WT n = 8, Plin5MKO n = 7). E: Glucose tolerance after rFGF21 injection (WT n = 8, Plin5MKO n = 9). FandG: Plasma insulin and plasma C-peptide after rFGF21 injection (WT n = 8, Plin5MKO n = 7).H: Schematic highlighting the conceptual advanceof this work. The accumulation of intracellular lipids is associated with ER stress/activation of the UPR, which in turn drives production of ACC2to prevent excessive fatty acid (FA) oxidation and oxidative stress. In addition, the UPR activates transcription of FGF21 and secretion into thecirculation but also the release of muscle-secreted factors that are important for liver FGF21 production and secretion. FGF21 enhances adiposetissue glucose uptake and insulin secretion from the pancreas. Data are means 6 SEM. *P , 0.05.

diabetes.diabetesjournals.org Montgomery and Associates 603

implicated in the UPR (14). In this regard, LD accumulationis closely associated with ER function in yeast (44) and inthe livers of mammals, where the accumulation of LDs isassociated with ER stress (45) and resolution of ER stressprevents lipid accumulation (46). Skeletal muscle containsan extensive network of ER, called the sarcoplasmic reticu-lum, and high-fat feeding induces the UPR in skeletalmuscle of mice (37,47). We show that deletion of Plin5 inskeletal muscle increases fatty acid oxidation, reduces in-tracellular lipid accumulation, and blunts activation of theUPR, most notably the IRE1-XBP1s and PERK-eIF2a path-ways. Such regulation is consistent with mice overexpress-ing Plin5 in skeletal muscle that shows excessive muscletriglyceride accumulation (12,39) and an induction of ERstress pathways (12), and it is consistent with the notionthat PLIN5 is required to regulate healthy intramyocellularlipid flux to buffer lipid loading in myocytes (38,48). We didnot observe any differences in ER stress in the heart de-spite a complete deletion of Plin5, pointing toward skeletalmuscle–specific regulation.

ER stress and activation of the UPR is associated withthe development of insulin resistance (37). We conclusivelyshow that the IRE1-XBP1 and the PERK-eIF2a arms of theUPR are downregulated in the skeletal muscle of Plin5MKO

mice. This coincided with decreased insulin and C-peptidelevels during an oral glucose load and a modest impairmentin whole-body insulin action, which was associated withreduced adipose tissue insulin sensitivity and, surprisingly,no changes in muscle 2DG after an oral glucose challenge.XBP1s, ATF4, and CHOP are transcriptional regulators ofFGF21 (13,34,49), and our analysis demonstrated an;50%decrease in muscle and liver FGF21 expression and circu-lating FGF21 levels in Plin5MKO mice. FGF21 is a regulatorof b-cell insulin secretion (50) and adipose tissue insulinsensitivity (51,52), and the marked downregulation ofFGF21 provides a plausible explanation for this metabolicphenotype in Plin5MKO mice. The substantial increase inmuscle and plasma FGF21 in mice overexpressing Plin5 inskeletal muscle supports this notion (12). Our data alsosupport the premise that upregulation of FGF21 by periph-eral tissues might be a more general response to cellularstresses (53,54).

Whether muscle-derived FGF21 impacts whole-bodyphysiology remains controversial, with some reports sug-gesting that FGF21 is not typically expressed in muscle andis only induced in situations of muscle stress (55,56),whereas others describe FGF21 as an important myokine(57,58). Our results linking ER stress with FGF21 produc-tion and our finding under ex vivo conditions that FGF21secretion was greater from muscle than from liver supportsboth interpretations. Extending on these data, we have alsodemonstrated that muscle-secreted factors in Plin5MKO micesignificantly reduce hepatocyte FGF21 secretion, whereasheart-secreted factors had no impact, highlighting the likelyimportance of skeletal muscle–liver communication in reg-ulating systemic FGF21 levels. This is further highlightedby plasma of Plin5MKO mice having the capacity to reduce

hepatocyte FGF21 secretion. Identifying the specific muscle-secreted factors driving the inhibition of FGF21 secretionin Plin5MKO mice is beyond the scope of this study, butsuch information will provide insight in understandingmyokine regulation of metabolism. Our results supportthe recent findings in mice overexpressing Plin5 in skele-tal muscle, which reported increased skeletal muscle andserum FGF21 levels that were proposed to induce the“browning” of adipose tissue and improved systemic metab-olism (12). In this respect, we show that correction ofplasma FGF21 levels in Plin5MKO mice by recombinantFGF21 injection can rescue their glucose intolerance andrestore insulin secretion to rates of WT mice. In addition,the present studies extend on these previous observationsby elucidating the likely mechanism mediating this effect;that is, muscle-specific Plin5 deletion leads to remodeling ofLD dynamics and lipid accumulation, a decrease in UPR/ERstress, and downregulation of essential transcription factorsfor FGF21 expression and secretion in muscle.

In conclusion, these findings demonstrate that PLIN5deficiency in skeletal muscle reduces intramyocellular tri-glyceride accumulation, increases fatty acid oxidation, andcauses oxidative stress. Our studies comparing WT andPlin5MKO mice have unmasked the likely importance of lipidsensing by the ER for regulating systemic glucose metabo-lism, showing that under conditions of dietary lipid over-supply, PLIN5 is required for the adaptive maintenance ofUPR-driven transcription of FGF21 in skeletal muscle,which leads to secretion of FGF21 that can positively im-pact adipose tissue insulin sensitivity and systemic glycemiccontrol. Moreover, skeletal muscle secretes as yet unidenti-fied factors to promote FGF21 production and secretion bythe liver, highlighting the growing appreciation of muscle-liver cross talk for systemic metabolic control.

Acknowledgments. The authors thank Maria Matzaris (Monash University)for technical assistance.Funding. This work was funded by the Australian National Health and MedicalResearch Council (NHMRC) (1047138). M.K.M. is supported by a Peter DohertyFellowship and M.J.W. is supported by a Senior Research Fellowship from theNHMRC (1077703). C.R.B. is supported by an Australian Research Council FutureFellowship (FT160100017).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. M.K.M. and M.J.W. planned and conducted theexperiments, analyzed the data, and wrote the manuscript. R.M. planned andconducted the experiments and analyzed the data. J.B. and V.M.S. conducted theexperiments. H.C.P. and C.R.B. planned and conducted the experiments. All authorsedited the manuscript. M.J.W. is the guarantor of this work and, as such, had fullaccess to all the data in the study and takes responsibility for the integrity of the dataand the accuracy of the data analysis.

References1. Walther TC, Farese RV Jr. Lipid droplets and cellular lipid metabolism. AnnuRev Biochem 2012;81:687–7142. Haemmerle G, Moustafa T, Woelkart G, et al. ATGL-mediated fat catabolismregulates cardiac mitochondrial function via PPAR-a and PGC-1. Nat Med 2011;17:1076–1085

604 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018

3. Meex RC, Hoy AJ, Mason RM, et al. ATGL-mediated triglyceride turnover andthe regulation of mitochondrial capacity in skeletal muscle. Am J Physiol EndocrinolMetab 2015;308:E960–E9704. Unger RH. Minireview: weapons of lean body mass destruction: the role ofectopic lipids in the metabolic syndrome. Endocrinology 2003;144:5159–51655. Fischer J, Lefèvre C, Morava E, et al. The gene encoding adipose triglyceridelipase (PNPLA2) is mutated in neutral lipid storage disease with myopathy. Nat Genet2007;39:28–306. Lefèvre C, Jobard F, Caux F, et al. Mutations in CGI-58, the gene encodinga new protein of the esterase/lipase/thioesterase subfamily, in Chanarin-Dorfmansyndrome. Am J Hum Genet 2001;69:1002–10127. Gandotra S, Le Dour C, Bottomley W, et al. Perilipin deficiency and autosomaldominant partial lipodystrophy. N Engl J Med 2011;364:740–7488. Wolins NE, Quaynor BK, Skinner JR, et al. OXPAT/PAT-1 is a PPAR-induced lipiddroplet protein that promotes fatty acid utilization. Diabetes 2006;55:3418–34289. Mason RR, Watt MJ. Unraveling the roles of PLIN5: linking cell biology tophysiology. Trends Endocrinol Metab 2015;26:144–15210. Granneman JG, Moore HP, Mottillo EP, Zhu Z. Functional interactions betweenMldp (LSDP5) and Abhd5 in the control of intracellular lipid accumulation. J BiolChem 2009;284:3049–305711. Wang H, Bell M, Sreenivasan U, et al. Unique regulation of adipose triglyceridelipase (ATGL) by perilipin 5, a lipid droplet-associated protein. J Biol Chem 2011;286:15707–1571512. Harris L-ALS, Skinner JR, Shew TM, Pietka TA, Abumrad NA, Wolins NE.Perilipin 5-driven lipid droplet accumulation in skeletal muscle stimulates the ex-pression of fibroblast growth factor 21. Diabetes 2015;64:2757–276813. Wan X-s, Lu X-h, Xiao Y-c, et al. ATF4- and CHOP-dependent induction ofFGF21 through endoplasmic reticulum stress. Biomed Res Int 2014;2014:80787414. Velázquez AP, Tatsuta T, Ghillebert R, Drescher I, Graef M. Lipid droplet-mediated ER homeostasis regulates autophagy and cell survival during starvation. JCell Biol 2016;212:621–63115. Coen PM, Dubé JJ, Amati F, et al. Insulin resistance is associated with higherintramyocellular triglycerides in type I but not type II myocytes concomitant withhigher ceramide content. Diabetes 2010;59:80–8816. Mohktar RAM, Montgomery MK, Murphy RM, Watt MJ. Perilipin 5 is dis-pensable for normal substrate metabolism and in the adaptation of skeletal muscle toexercise training. Am J Physiol Endocrinol Metab 2016;311:E128–E13717. Huijsman E, van de Par C, Economou C, et al. Adipose triacylglycerol lipasedeletion alters whole body energy metabolism and impairs exercise performance inmice. Am J Physiol Endocrinol Metab 2009;297:E505–E51318. Spry C, Saliba KJ, Strauss E. A miniaturized assay for measuring small mol-ecule phosphorylation in the presence of complex matrices. Anal Biochem 2014;451:76–7819. Crowe S, Wu LE, Economou C, et al. Pigment epithelium-derived factor con-tributes to insulin resistance in obesity. Cell Metab 2009;10:40–4720. Watt MJ, Heigenhauser GJF, Dyck DJ, Spriet LL. Intramuscular triacylglycerol,glycogen and acetyl group metabolism during 4 h of moderate exercise in man. JPhysiol 2002;541:969–97821. Borg ML, Andrews ZB, Duh EJ, Zechner R, Meikle PJ, Watt MJ. Pigmentepithelium-derived factor regulates lipid metabolism via adipose triglyceride lipase.Diabetes 2011;60:1458–146622. Bou R, Codony R, Tres A, Decker EA, Guardiola F. Determination of hydro-peroxides in foods and biological samples by the ferrous oxidation-xylenol orangemethod: a review of the factors that influence the method’s performance. AnalBiochem 2008;377:1–1523. Buege JA, Aust SD. Microsomal lipid peroxidation. Methods Enzymol 1978;52:302–31024. Arnold PM, Parslow GR. Designing a coupled assay system for aspartateaminotransferase. Biochem Educ 1995;23:40–4125. Bergmeyer HU. IFCC methods for the measurement of catalytic concentrationsof enzymes: part 3. IFCC method for alanine aminotransferase (L-alanine: 2-oxoglutarate aminotransferase, EC 2.6.1.2). Clin Chim Acta 1980;105:147–154

26. Montgomery MK, Fiveash CE, Braude JP, et al. Disparate metabolic response tofructose feeding between different mouse strains. Sci Rep 2015;5:1847427. Liu M, Montgomery MK, Fiveash CE, et al. PPARa-independent actions ofomega-3 PUFAs contribute to their beneficial effects on adiposity and glucose ho-meostasis. Sci Rep 2014;4:553828. Montgomery MK, Brown SHJ, Lim XY, et al. Regulation of glucose homeostasisand insulin action by ceramide acyl-chain length: a beneficial role for very long-chainsphingolipid species. Biochim Biophys Acta 2016;1861:1828–183929. Montgomery MK, Hallahan NL, Brown SH, et al. Mouse strain-dependentvariation in obesity and glucose homeostasis in response to high-fat feeding. Dia-betologia 2013;56:1129–113930. Badin P-M, Vila IK, Louche K, et al. High-fat diet-mediated lipotoxicity andinsulin resistance is related to impaired lipase expression in mouse skeletal muscle.Endocrinology 2013;154:1444–145331. Wang H, Sreenivasan U, Gong D-W, et al. Cardiomyocyte-specific perilipin5 overexpression leads to myocardial steatosis and modest cardiac dysfunction. JLipid Res 2013;54:953–96532. Lam T, Harmancey R, Vasquez H, et al. Reversal of intramyocellular lipid ac-cumulation by lipophagy and a p62-mediated pathway. Cell Death Discov 2016;2:1606133. Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threadsand missing links. Cell 2012;148:852–87134. Jiang S, Yan C, Fang QC, et al. Fibroblast growth factor 21 is regulated bythe IRE1a-XBP1 branch of the unfolded protein response and counteracts endo-plasmic reticulum stress-induced hepatic steatosis. J Biol Chem 2014;289:29751–2976535. Itoh N. FGF21 as a hepatokine, adipokine, and myokine in metabolism anddiseases. Front Endocrinol (Lausanne) 2014;5:10736. Muise ES, Souza S, Chi A, et al. Downstream signaling pathways in mouseadipose tissues following acute in vivo administration of fibroblast growth factor 21.PLoS One 2013;8:e7301137. Koh HJ, Toyoda T, Didesch MM, et al. Tribbles 3 mediates endoplasmic re-ticulum stress-induced insulin resistance in skeletal muscle. Nat Commun 2013;4:187138. Mason RR, Mokhtar R, Matzaris M, et al. PLIN5 deletion remodels intracellularlipid composition and causes insulin resistance in muscle. Mol Metab 2014;3:652–66339. Bosma M, Sparks LM, Hooiveld GJ, et al. Overexpression of PLIN5 in skeletalmuscle promotes oxidative gene expression and intramyocellular lipid contentwithout compromising insulin sensitivity. Biochim Biophys Acta 2013;1831:844–85240. Kanaley JA, Shadid S, Sheehan MT, Guo Z, Jensen MD. Relationship betweenplasma free fatty acid, intramyocellular triglycerides and long-chain acylcarnitines inresting humans. J Physiol 2009;587:5939–595041. Wang C, Zhao Y, Gao X, et al. Perilipin 5 improves hepatic lipotoxicity by in-hibiting lipolysis. Hepatology 2015;61:870–88242. Gallardo-Montejano VI, Saxena G, Kusminski CM, et al. Nuclear perilipin 5 in-tegrates lipid droplet lipolysis with PGC-1a/SIRT1-dependent transcriptional regu-lation of mitochondrial function. Nat Commun 2016;7:1272343. Piperi C, Adamopoulos C, Papavassiliou AG. XBP1: a pivotal transcriptionalregulator of glucose and lipid metabolism. Trends Endocrinol Metab 2016;27:119–12244. Fei W, Shui G, Gaeta B, et al. Fld1p, a functional homologue of human seipin,regulates the size of lipid droplets in yeast. J Cell Biol 2008;180:473–48245. Fei W, Wang H, Fu X, Bielby C, Yang H. Conditions of endoplasmic reticulumstress stimulate lipid droplet formation in Saccharomyces cerevisiae. Biochem J2009;424:61–6746. Özcan U, Yilmaz E, Özcan L, et al. Chemical chaperones reduce ER stress andrestore glucose homeostasis in a mouse model of type 2 diabetes. Science 2006;313:1137–114047. Deldicque L, Cani PD, Philp A, et al. The unfolded protein response is activatedin skeletal muscle by high-fat feeding: potential role in the downregulation of proteinsynthesis. Am J Physiol Endocrinol Metab 2010;299:E695–E705

diabetes.diabetesjournals.org Montgomery and Associates 605

48. Laurens C, Bourlier V, Mairal A, et al. Perilipin 5 fine-tunes lipid oxidation tometabolic demand and protects against lipotoxicity in skeletal muscle. Sci Rep 2016;6:3831049. De Sousa-Coelho AL, Marrero PF, Haro D. Activating transcription factor4-dependent induction of FGF21 during amino acid deprivation. Biochem J 2012;443:165–17150. Wente W, Efanov AM, Brenner M, et al. Fibroblast growth factor-21improves pancreatic b-cell function and survival by activation of extracellularsignal-regulated kinase 1/2 and Akt signaling pathways. Diabetes 2006;55:2470–247851. Luo Y, McKeehan WL. Stressed liver and muscle call on adipocytes with FGF21.Front Endocrinol (Lausanne) 2013;4:19452. Arner P, Pettersson A, Mitchell PJ, Dunbar JD, Kharitonenkov A, Rydén M.FGF21 attenuates lipolysis in human adipocytes - a possible link to improved insulinsensitivity. FEBS Lett 2008;582:1725–1730

53. Kim KH, Jeong YT, Oh H, et al. Autophagy deficiency leads to protection fromobesity and insulin resistance by inducing Fgf21 as a mitokine. Nat Med 2013;19:83–9254. Keipert S, Ost M, Johann K, et al. Skeletal muscle mitochondrial uncouplingdrives endocrine cross-talk through the induction of FGF21 as a myokine. Am JPhysiol Endocrinol Metab 2014;306:E469–E48255. Fisher FM, Maratos-Flier E. Understanding the physiology of FGF21. Annu RevPhysiol 2016;78:223–24156. Markan KR, Naber MC, Ameka MK, et al. Circulating FGF21 is liver derived andenhances glucose uptake during refeeding and overfeeding. Diabetes 2014;63:4057–406357. Izumiya Y, Bina HA, Ouchi N, Akasaki Y, Kharitonenkov A, Walsh K. FGF21 is anAkt-regulated myokine. FEBS Lett 2008;582:3805–381058. Liu X, Wang Y, Hou L, Xiong Y, Zhao S. Fibroblast growth factor 21 (FGF21)promotes formation of aerobic myofibers via the FGF21-SIRT1-AMPK-PGC1apathway. J Cell Physiol 2017:232:1893–1906

606 ER Stress and FGF21 in Plin5MKO Mice Diabetes Volume 67, April 2018