Ana Paula Arruda1,2, Benedicte M Pers1,2, Güneş Parlakgül1, Ekin Güney1, Karen Inouye1 &Gökhan SHotamisligil1
Proper function of the endoplasmic reticulum (ER) andmitochondria is crucial for cellular homeostasis, and dysfunction at either sitehas been linked to pathophysiological states, including metabolic diseases.Although the ER and mitochondria play distinct cellular roles, these organellesalso form physical interactions with each other at sites defined asmitochondria-associated ER membranes (MAMs), which are essential for calcium,lipid and metabolite exchange. Here we show that in the liver, obesity leads toa marked reorganization of MAMs resulting in mitochondrial calcium overload,compromised mitochondrial oxidative capacity and augmented oxidative stress.Experimental induction of ER-mitochondria interactions results in oxidativestress and impaired metabolic homeostasis, whereas downregulation of PACS-2 orIP3R1, proteins important for ER-mitochondria tethering or calcium transport,respectively, improves mitochondrial oxidative capacity and glucose metabolismin obese animals. These findings establish excessive ER-mitochondrial couplingas an essential component of organelle dysfunction in obesity that may contributeto the development of metabolic pathologies such as insulin resistance anddiabetes.
Obese individuals are at increased risk fordeveloping insulin resistance and other comorbidities, including diabetes andcardiovascular disease1,2. Although the molecular mechanisms that underliethese associations are not completely defined, dysfunction of cellular organellessuch as the ER and mitochondria has emerged as a key event in the alterationsthat follow nutrient overload3,4. For example, in the liver of obese animals,the ER membrane lipid composition is altered, its capacity to retain Ca2+ isimpaired5, and ER protein degradation machinery is suppressed6. As aconsequence, the unfolded protein response (UPR) is activated, affecting avariety of inflammatory, metabolic and stress-signaling networks directlyinvolved in metabolic diseases3,7,8. ER stress is also detected in obese humans9,10,and interventions that improve ER function have been shown to restore glucosehomeostasis in mouse models as well as in obese individuals and people withdiabetes11–13.
It has also been established in humans andmouse models that obesity results in mitochondrial dysfunction in skeletalmuscle and adipose tissue, featuring altered oxidative function, ultrastructureabnormalities and increased oxidative stress14–20. In the liver, althoughthere is variability between studies, obesity is associated with alteredoxidative capacity and excessive oxidative stress both in humans and in mice21–24. However,the degree of mitochondrial defects, the underlying molecular mechanisms andthe consequences for systemic metabolic control are not well established4,14,17–20.
Because of the distinct roles that ER andmitochondria play in the cell, the metabolic impacts of ER and mitochondrialdysfunction have largely been viewed and studied independently. However, these organellesphysically and functionally interact and are able to regulate each other’s function25.The sites of physical communication between ER and mitochondria, defined asMAMs, are conserved structures found across eukaryotic phyla and are keydeterminants of cell survival and death through the transfer of Ca2+ and othermetabolites25. In addition, this subdomain of the ER is responsible for thebiosynthesis of two abundant phospholipids, phosphatidylcholine andphosphatidylethanolamine25. Recently, it was also shown that MAMs are importantfor autophagy by regulating autophagosome formation26 and for mitochondrialdynamics by marking sites of mitochondrial fission27. Thus, the function ordysfunction of one organelle can profoundly affect the other, but the relevanceof this interaction to obesity-related cellular dysfunction and metabolichomeostasis has not been studied.
Here, we show that obesity drives anabnormal increase in MAM formation, which results in increased Ca2+ flux fromthe ER to mitochondria in the liver. The mitochondrial Ca2+ overload isaccompanied by increased mitochondrial reactive oxygen species (ROS) productionand impairment of metabolic homeostasis. Suppression of two distinct proteinscritical for ER-mitochondrial apposition and Ca2+ flux, inositol1,4,5-trisphosphate receptor, type 1 (IP3R1) and phosphofurin acidic clustersorting protein 2 (PACS-2), resulted in improved cellular homeostasis andglucose metabolism in obese animals, suggesting that this mechanism is crucialfor metabolic health and could represent a new therapeutic target for metabolicdisease.
RESULTS
Higherhepatic ER-mitochondrial contact sites in obesity
In order to investigate ER andmitochondrial morphology and their physical interaction in obesity, we firstused transmission electron microscopy (TEM) to examine liver sections collectedfrom both genetic (ob/ob) and dietary (high-fat diet, HFD) models of obesity. Inboth models this analysis was performed in the context of established obesity:in 8- to 12-week-old ob/ob mice and in mice after 16 weeks of HFD feeding. Weobserved that the ER membrane displayed a disorganized morphology in liversderived from obese animals along with a markedly higher degree of ER appositionto mitochondria (Fig. 1a–d and Supplementary Fig. 1a,c). Detailed quantitative analysis (describedin Supplementary Fig. 1b) of liver sections from each experimental group (tensections per animal, five different animals per group) demonstrated that theproportion of the ER in close contact with mitochondria relative to total ERcontent was significantly greater in the livers of both ob/ob and HFD mice thanin lean controls (Fig. 1e). This finding was further substantiated by TEM imagesof isolated crude mitochondrial pellets that revealed mitochondria of obeseanimals to be more frequently attached to the ER than mitochondria collectedfrom lean controls (Supplementary Fig. 2a,b). We also employedorganelle-targeted fluorescent proteins as an additional strategy toinvestigate the alterations in ER-mitochondria juxtaposition. For this, weexpressed mitochondria-targeted GFP and ER-targeted DsRed proteins in the liversof lean and ob/ob mice through adenoviral gene delivery and performed imageanalysis in isolated primary hepatocytes from these animals 24 h after theisolation. We used in vivo expression of the fluorescent proteins rather than transienttransfection following primary cell isolation because hepatocyte metabolismchanges quickly after isolation. Obese mice showed a greater degree ofcolocalization between ER and mitochondria in liver cells compared to leancontrols (Fig. 1f and Supplementary Fig. 3a,b), as revealed by significantlyincreased Manders’s colocalization coefficient between the organelle-targetedfluorescent markers (Fig. 1f and Supplementary Fig. 3c).
We also observed a lower level of ERmembrane abundance in livers derived from obese animals and a disorganized ERmorphology compared with lean counterparts. We performed an extensivemorphometric analysis to acquire quantitative comparisons (described in detailin Supplementary Fig. 1b,c). ER content was 50% and 35% lower in liver cellsfrom ob/ob and HFD mice, respectively, compared with their lean controls (Fig.1g). This finding is consistent with the measurement of ER area as percentageof total cell area, which was lower in hepatocytes derived from obese mice thanin those derived from lean mice (Supplementary Fig. 1d,e). As an independent readoutof ER content, we also measured the level of calreticulin, a known ER marker,and found that it was lower in both ob/ob mouse– and HFD mouse–derivedlivers compared to control livers (Supplementary Fig. 4a).
The reduction in ER content in obesity wassomewhat unexpected because, in the context of ER stress, the ER is thought tomorphologically rearrange and expand in order to accommodate protein folding demandsand reestablish homeostasis28. In our analysis, this predicted adaptiveresponse was clearly defective in the context of established obesity. Hepaticmitochondrial content was also lower in ob/ob mice than in control mice, asmeasured by TEM (Fig. 1h), mtDNA content (Supplementary Fig. 4b) and COXIVexpression (Supplementary Fig. 4c). In addition, three-dimensional (3D)reconstruction of serialsection TEMs revealed that mitochondria from the liverof obese mice were rounder and swollen compared to the tubular mitochondria observedin the liver of lean animals, in accordance with previous findings19 (Fig. 1i, SupplementaryFig. 2c and Supplementary Movies 1–4). Taken together,our results collected from multiple mouse models through independent means ofquantitative analysis indicate that obesity leads to alterations in hepatic ERand mitochondrial structure and enhanced MAM formation.
Next, we asked whether the alterations inER and mitochondrial morphology and interaction are an early event during thedevelopment of obesity or occur as a consequence of the pathological state triggeredby overnutrition. We first examined the ER and mitochondrial morphology by TEMin livers derived from lean mice and mice fed HFD for just 4 weeks. Short-termHFD was sufficient to induce ER reorganization (Fig. 2a), lower ER content (Fig.2b) and significantly higher ER-mitochondria interactions (Fig. 2c), but it didnot induce significant changes in total mitochondrial content (Fig. 2d).
In order to determine whether changes in ERmorphology and MAM content result from ER stress, dysfunction or both, we also examinedthe effect of acute ER stress in the liver in vivo by administration of 0.5 μgper g body weight of tunicamycin for 6 and 24 h in lean mice. Tunicamycintreatment resulted in a higher degree of phosphorylation of eukaryoticinitiation factor 2α and c-Jun N-terminal kinase (JNK) and higher expression ofATF4 and GRP78, classical markers of UPR activation (Fig. 2e). Interestingly,after 6 h of tunicamycin treatment we observed a markedly higher degree of MAMsin the liver (Fig. 2f). This finding is in line with a previous report showingthat ER stress induced by tunicamycin induced more MAMs in vitro29. After 24 h,however, there were severe structural defects, with notable ER dilation anddisorganization (Fig. 2f). Taken together, these results suggest that greaterMAM formation is a generalized and early response to ER stress.
Alteredhepatic ER-mitochondrial protein expression in obesity
MAMs have a key regulatory role in severalcellular functions, including the transfer of Ca2+ and lipids from ER tomitochondria25. This subdomain of the ER is enriched in specific proteins suchas the Ca2+ channel IP3R25,30, the mitochondrial-ER tether proteins mitofusin-2(Mfn2) (ref. 31) and PACS-2 (ref. 32), chaperones such as sigma 1 receptor(Sig1R)33 and calnexin, and proteins involved in phospholipid metabolism25 (Fig.3a). To explore the possible functional consequences of increasedER-mitochondrial contact in obesity, we examined the expression of MAM-enrichedproteins in liver lysates from lean, DIO and ob/ob mice. There was a markedlyhigher expression of proteins involved in Ca2+ transport, such as IP3R1 andIP3R2 and Sig1R, in samples from obese mice, along with PACS-2 protein levels (Fig.3b). We found no significant differences in Mfn2 levels. In these experiments,we validated the specificity of the antibodies by employing loss-of-functionmodels (Supplementary Fig. 5).
We then performed Percoll-based subcellularfractionation of liver tissues to collect enriched fractions of mitochondria,ER and MAM and verified the preparations by western blot analysis. Cytochrome Cwas enriched in mitochondrial fractions, IP3R1 and sarcoendoplasmic reticulumCa2+ transport ATPase (SERCA) were preferentially found in the bulk ER,calnexin and protein disulfide isomerase (PDI) were equally distributed betweenMAM and bulk ER, whereas Sig1R is found preferentially at the MAM (Fig. 3c).GRP75, the mitochondrial stress protein, was found mainly in the mitochondriabut was also detected in the MAM fraction, as has been described34. We then verifiedthat the levels of IP3R1 and IP3R2 were significantly higher in MAMs of obeseanimals than in those of lean controls (Fig. 3d). The level of PACS-2 was alsohigher in samples derived from HFD-fed mice than in those from lean controls (Fig.3d). Although found in similar levels in total lysates, Mfn2 levels weregreater in the MAM fractions derived from obese animals than in those from leancontrols (Fig. 3d), suggesting a possible redistribution of this protein from themitochondrial membrane to the MAMs. Altogether, these results not only furthersubstantiated our morphological and microscopic observation of a larger numberof ER-mitochondrial contact sites in obesity but also suggested a greaterdensity of anchoring proteins and components involved in Ca2+ transport atthese sites.
MAMenrichment in obesity and mitochondrial calcium overload
In MAMs, IP3R resides in close proximity tovoltage-dependent anion channel, an ion channel located in the outermitochondrial membrane. This structure facilitates the uptake of Ca2+ into the mitochondrialmatrix through the low-affinity mitochondrial Ca2+ uniporter (MCU)25,30,35.Identifying a greater degree of MAMs enriched with Ca2+ regulatory proteins infatty livers led us to postulate that abnormal Ca2+ transport from ER tomitochondria may occur in obesity, with deleterious consequences for mitochondrialfunction. To address this hypothesis, we first assessed mitochondrial Ca2+ dynamicsin obesity by expressing a mitochondria-targeted Ca2+ fluorescence resonanceenergy transfer (FRET) reporter36 in the liver of lean and ob/ob mice throughadeno-associated viral delivery (procedure described in Fig. 4a). Quantitativesingle-cell imaging of mitochondrial Ca2+ levels in isolated hepatocytesrevealed that mitochondria from obese animals contained a significantly (P <0.01) higher Ca2+ concentration ([Ca2+]m) at baseline (Fig. 4b). In addition, wedetected a higher mitochondrial Ca2+ uptake following stimulation of ER Ca2+ releasewith ATP in primary hepatocytes from obese mice relative to those collectedfrom lean counterparts (Fig. 4c,d). In a complementary approach, we treatedprimary hepatocytes from lean and obese mice with carbonylcyanide p-trifluoromethoxyphenylhydrazone(FCCP), which causes disruption of the mitochondrial membrane potentialfollowed by release of mitochondrial Ca2+ to the cytosol. Addition of FCCP tohepatocytes from obese mice resulted in greater cytosolic Ca2+ content comparedto lean cells, as measured by Fura2-AM (Fig. 4e). Notably, the expression ofMCU, the mitochondria Ca2+ uniporter, was also higher in the ob/ob livertissue, although no difference was found in HFD-fed animals at 16 weeks (SupplementaryFig. 4d). Taken together, these results show that in the setting of obesity,hepatic mitochondria are overloaded with Ca2+, suggesting a previouslyunrecognized mechanism of mitochondrial dysfunction in the liver of obeseanimals. To evaluate whether higher [Ca2+]m results from IP3R-mediated ER Ca2+ releaseto the cytosol followed by mitochondrial uptake or by direct transport throughMAM junctions, we analyzed cytosolic Ca2+ dynamics in response to IP3Rstimulation in primary hepatocytes from lean and ob/ob mice using Fura-2AM.Baseline cytosolic Ca2+ concentration was significantly (P < 0.01) higher incells from obese mice (Supplementary Fig. 6a). However, ATP stimulation led toa similar cytosolic Ca2+ rise in hepatocytes from lean and obese animals (SupplementaryFig. 6b) despite the higher [Ca2+]m peak in cells from obese mice (Fig. 4c,d).These observations indicate that in hepatocytes, obesity leads to more Ca2+ transportfrom the ER through the MAM connections, resulting in elevated mitochondrial Ca2+(Supplementary Fig. 6c).
Within the mitochondrial matrix, atransient increase in Ca2+ level activates matrix enzymes and stimulatesoxidative phosphorylation, but sustained exposure to high Ca2+ concentration isoften detrimental for mitochondrial function37. In primary hepatocytes from ob/obmice we observed decreased basal oxygen consumption and diminished response tothe uncoupling agent FCCP, a sign of mitochondrial dysfunction (Fig. 4f,g).Additionally, we observed a moderately but significantly lower mitochondrialmembrane potential (Fig. 4h) and higher ROS levels (Fig. 4i) in hepatocytesfrom obese animals, as measured by tetramethyl rhodamine methyl ester (TMRM)and MitoSOX fluorescence, respectively. These data, together with the alteredmitochondrial morphology detected in liver tissue, reveal that in obeseanimals, high [Ca2+]m is associated with mitochondrial dysfunction.
Experimentalinduction of MAMs accelerates metabolic decline
In an attempt to mechanistically connectincreased MAM content in obesity with mitochondrial Ca2+ overload, dysfunctionand altered cellular and metabolic homeostasis, we asked whether forcing the ER-mitochondriainteraction is sufficient to cause mitochondrial Ca2+ overload and stress inlean, otherwise healthy, animals. To test this hypothesis, we used arecombinant construct encoding a synthetic linker38 that increasesER-mitochondria contact sites (Fig. 5a). In order to validate the efficacy ofthe construct, we first expressed the linker in hepatocyte-derived mouseHepa1-6 cells. Twenty-four hours following transfection, we sorted the cells onthe basis of RFP fluorescence. TEM analysis confirmed that linker expressionwas associated with a higher degree of contacts between the ER and mitochondria(Fig. 5a). These cells remained viable throughout the experiments (48 h) withno gross phenotypic changes (Supplementary Fig. 7a) or signs of cell death, asmeasured by annexin staining followed by FACS analysis (Fig. 5b). Linkerexpression was associated with higher peak of ATP-stimulated [Ca2+]m, in linewith more ER-mitochondrial interactions that occur upon linker expression (Fig.5c). In addition, this experimental enrichment of ER-mitochondrial contacts ledto a lower degree of basal and maximal mitochondrial respiration (Fig. 5d) accompaniedby higher phosphorylation of JNK and expression of CCAAT-enhancer-bindingprotein homologous protein (CHOP) (Supplementary Fig. 7b), both indicators ofcellular stress.
We then used adenovirus-mediated genetransfer to express the synthetic linker in the liver of lean animals. Linkerexpression resulted in more ER-mitochondrial contact sites in the liver comparedto the livers of mice expressing control adenovirus (Supplementary Fig. 7c). However,in 10-week-old lean mice, linker expression did not appear to have any toxiceffects, nor did it cause gross morphological changes within the liver (SupplementaryFig. 7d) or a metabolic phenotype (Supplementary Fig. 7e,f). This suggests thatin a healthy animal, linker expression was well tolerated and not sufficient todisrupt systemic metabolic homeostasis. However, linker expression in micefollowing a short period (4 weeks) of HFD was associated with a considerablyhigher degree of ER-mitochondria contact sites (Fig. 5e and Supplementary Fig.8a) and a markedly higher level of hepatic lipid accumulation compared withcontrols (Fig. 5f and Supplementary Fig. 8b). This was accompanied by lower mitochondrialoxidative function measured following FCCP treatment (Fig. 5g). At themolecular level, we detected that expression of the linker led to activation ofJNK and higher levels of ATF4 and CHOP protein (Fig. 5h) and ATF3 and ATF4mRNA, with no other changes in UPR markers (Supplementary Fig. 8c). We alsoobserved impaired insulin-induced phosphorylation of the insulin receptor, IRS1and AKT in linker-expressing livers (Supplementary Fig. 8d), which correlatedwith a moderate but significant impairment in glucose homeostasis, as measuredby glucose tolerance test (Fig. 5i), and with higher liver glucose output, asmeasured by pyruvate tolerance test (Supplementary Fig. 8e). To explore thewhole-body glucose fluxes with higher precision, we performedhyperinsulinemiceuglycemic clamp studies. Impaired glucose homeostasis wasclearly evident during the clamp experiments in linker-expressing animals, whichrequired significantly lower rates of glucose infusion to sustain euglycemia (Fig.5j and Supplementary Fig. 8f–h), indicating that they were relatively insulin insensitive. Clamp hepaticglucose production was significantly higher (Fig. 5j), also indicating impairedhepatic insulin sensitivity. Taken together, these results demonstrate thatincreasing ER-mitochondrial interaction in vivo in a context of mildovernutrition is sufficient to cause mitochondrial dysfunction, activation ofstress signaling and impaired hepatic control of glucose production,accelerating the decline of glucose homeostasis independent of body weightchanges (Supplementary Fig. 8i).
We then asked whether modulatingmitochondrial Ca2+ uptake by other means, such as by experimentally increasingthe level of MCU, previously demonstrated to affect [Ca2+]m (refs. 39,40) wouldresult in physiological alterations similar to those observed upon increased MAMformation. We found that overexpression of MCU in Hepa1-6 cells led to higherlevels of oxidative stress and JNK activation without inducing cell death (SupplementaryFig. 9a–c). Taken together, these data demonstrate that in vitro, [Ca2+]m canbe affected either by ER-mitochondrial contact or by expression of MCU andplays an important part in cellular stress responses.
SilencingIP3R1 and PACS-2 improves metabolism in obese mice
Lastly, we asked whether we could improvemitochondrial and cellular function and therefore metabolic homeostasis inobese animals by diminishing ER-mitochondrial Ca2+ flux or ER-mitochondrial connection.To this end, we used two different but complementary approaches: we decreasedER-mitochondrial Ca2+ fluxes by suppressing hepatic IP3R1 and decreasedER-mitochondrial physical interaction by suppressing hepatic PACS-2 expression.We achieved a 70% lower level of hepatic IP3R1 expression via the delivery ofan adenoviral shRNA in mice without influencing the expression levels of otherIP3Rs (Fig. 6a). This resulted in a modest but significant improvement inmitochondrial maximal oxidative capacity (Fig. 6b). In addition, IP3R1knockdown coincided with lower cellular stress, as measured by phosphorylationof JNK (Fig. 6c), and an improvement in insulin signaling capacity (Fig. 6d).These local changes in mitochondrial Ca2+ flux in the liver also improvedsystemic glucose tolerance (Fig. 6e). These findings support our postulate thatIP3R1- mediated ER-mitochondrial Ca2+ transport is crucial in the maintenance ofproper glucose metabolism.
To further explore the functionalimportance of ER-mitochondria interactions, we targeted PACS-2, an integral MAMprotein upregulated in obesity (Fig. 3b,e). PACS-2 has previously been shown tobe involved in MAM formation and interorganelle communication32,41. Usingadenoviral shRNA delivery, we achieved nearly 60% knockdown of PACS-2 in ob/ob mice(Fig. 6f). Deletion of PACS-2 in cells has been shown to lead to a significantreduction in ER-mitochondria connection and protects against apoptosis32. Wefound that knockdown of PACS-2 resulted in lower JNK phosphorylation (Fig. 6f) andhigher mitochondrial maximal respiration with no alterations in basalmitochondria respiration (Fig. 6g). Reduction of PACS-2 levels also resulted inimproved insulin sensitivity (Fig. 6h), enhanced systemic glucose tolerance (Fig.6i) and improved liver steatosis (Supplementary Fig. 9d,e). Taken together,these loss-of-function approaches support the concept that targeting thestructural or functional components of MAMs may be an effective strategy toimprove glucose homeostasis in the context of obesity. Lastly, we asked whetherthe changes we observed in the context of obesity were unique to the liver. Inagreement with previous reports15,42,43, we found that in the soleus muscle,mitochondrial morphology was altered and the expression of MAM-enrichedproteins was higher in ob/ob mice compared to lean controls (Supplementary Fig.10), suggesting that the mechanisms described here may also contribute tocellular dysfunction in muscle. However, a more detailed analysis of MAMformation and function in muscle will be required to fully understand the roleof these structural changes in insulin resistance in obesity.
DISCUSSION
We have recently shown that aberrant Ca2+ handlingin the ER is one of the primary causes of hepatic ER dysfunction and stress inobesity5. Here, we demonstrate that altered Ca2+ signaling through the MAMs hasa fundamental role in connecting ER stress to mitochondrial dysfunction inobesity, for which plausible mechanisms have been elusive. Our extensivemorphological, biochemical and physiological analyses lead us to propose amodel in which high nutrient and energy intake leads to increased MAM formationin the liver as an early event in the course of the development of obesity andmetabolic disease. Increased MAM formation drives higher Ca2+ accumulation inthe mitochondria that, in turn, leads to impairment in mitochondrial oxidative capacity,increased ROS generation, cellular stress, impaired insulin action in the liverand abnormal glucose metabolism (Fig. 6j).
We have also observed that induction ofacute ER stress such as that triggered by short-term tunicamycin treatmentcauses a major rearrangement of the ER around the mitochondria, in accordance withearlier findings in vitro29. These data, together with our findings that theincrease in MAM formation is an early event upon HFD feeding, suggest thatincreasing MAM formation may be a short-term adaptive process driven by stressconditions to boost mitochondrial function and lipid synthesis. It is alsopossible that there are dynamic fluctuations in these interactions duringphysiological changes in metabolic exposures and ER function to supportspecific metabolic adaptations. In obesity, however, chronic maintenance ofthese connections leads to undesirable side effects such as mitochondrial Ca2+ overload.Indeed, in our experiments, increasing MAM formation in the absence ofadditional stressors was well tolerated and did not affect metabolichomeostasis. However, in the context of even a brief HFD intervention,experimental enrichment of ER-mitochondria interactions accelerated theprogression of obesity-related pathologies such as hepatic steatosis andglucose intolerance.
The regulation of mitochondrial function byCa2+ is complex, and the outcomes for cellular homeostasis depend on themaintenance of Ca2+ concentration within a very narrow range. Transient fluctuationsin mitochondrial Ca2+ stimulate the tricarboxylic acid cycle and oxidative phosphorylation,maintaining optimal cellular bioenergetics37. Accordingly, cells deficient inall three IP3Rs display compromised mitochondrial function as a result ofdiminished mitochondrial Ca2+ uptake44. Also, interventions that decreaseER-mitochondrial interaction below physiological thresholds, such astissue-specific deletion of Mfn2 in liver45 and hypothalamus46, lead tomitochondrial dysfunction and insulin and leptin resistance, respectively.Although these experiments do not specifically address the regulation ofERmitochondria Ca2+ flux due to the pleiotropic functions of the targeted molecules,they indicate that the proper maintenance of MAMs and Ca2+ flux is essentialfor mitochondrial function and metabolic homeostasis.
We also showed here that the downregulationof key molecules involved in ER-mitochondria connection and Ca2+ flux such asIP3R1 and PACS-2 results in improved mitochondrial respiration, decreased cellularstress and enhanced glucose tolerance in obese animals. Neither IP3R1 norPACS-2 is exclusively expressed in the MAM, and thus targeting these moleculesmay result in alterations to cellular function outside the MAM. For example, itis known that IP3R1 also regulates cytosolic Ca2+ flux; thus, alternativeexplanations for our observation of improved glucose homeostasis following IP3Rknockdown should be considered. Indeed, it was recently reported that in theliver, IP3R regulates systemic glucose metabolism by increasing cytosolic Ca2+ andglucagon signaling in hepatocytes, and reduced IP3R1 levels consequentlydecreased glucose production47,48. These results are consistent with ourfindings. However, the effects of hepatic IP3R1 suppression on oxygenconsumption rate (OCR) observed here strongly suggest that alteredmitochondrial function also contributes to the metabolic phenotype in vivo.Furthermore, producing metabolic improvements by these two independentinterventions modulating different aspects of ER-mitochondria interactionsstrongly supports their importance in organelle and metabolic homeostasis.
The effect of MAM regulation on metabolismis likely to result from the engagement of multiple pathways and multiplestructural components. Interventions that disrupt individual MAM components maythus only partially recapitulate the impact of overall MAM function in cellularand organismal physiology. Also, MAM regulation affects not only Ca2+ flux butalso lipid biosynthesis, autophagy and mitochondrial fission25–27. As allof these phenomena are altered in obesity, it is possible that alteration ofMAMs has a more pleiotropic effect in the obese condition as well as instage-specific metabolic pathologies mediated distinctly by different molecularcomponents, and these processes should be explored in future studies.
During the review of this manuscript, anindependent group reported that HFD feeding leads to diminished ER-mitochondriacontact sites in liver and primary hepatocytes employing a different methodology49.Of note, MAM formation and content in the liver are highly plastic depending onmetabolic status, age and stress condition, and thus the contrastingconclusions may reflect both differing experimental parameters and analyticalmethodology, including in vitro assays. As liver size and lipid contentincrease in obesity, total MAM content should be evaluated relative toalterations in tissue mass and composition. These issues notwithstanding,cumulative evidence supports the concept the MAM regulation has a major impact onglucose homeostasis45,46.
The association between defectivemitochondrial function, insulin resistance and type 2 diabetes has been a topicof intense investigation and debate in the field14–24. Moststudies have been performed in human skeletal muscle, where insulin resistanceis linked to impaired mitochondrial oxidative function and density15,20.However, whether these alterations cause or result from insulin resistance is presentlyunclear. As we found that skeletal muscle mitochondrial morphology is alteredin soleus muscle of ob/ob mice and expression of MAM-enriched proteins isincreased, altered sarcoplasmic reticulum–mitochondria Ca2+ fluxesmay also have a role in obesityrelated mitochondrial dysfunction in muscle.
Our observations may have implicationsbeyond obesity. Enhanced MAM function and increased ER-mitochondrialinteraction have been observed in neurodegenerative diseases such asgangliosidosis50 and Alzheimer’s disease51 and also in senescence52. Notably, Alzheimer’s disease isalso associated with insulin resistance in the brain53 and shares numerousmolecular features with obesity including ER stress, JNK activation andorganelle dysfunction, as well as aging. Therefore, we suggest that therapiesthat restore the proper function of ER-mitochondrial interface may be a usefulapproach in the treatment of multiple metabolic diseases with these mechanisticfeatures.
Figure 1 Obesity induces MAM formation and changes in ER andmitochondrial morphology in the liver. (a,b) RepresentativeTEM images of liver sections derived from wild-type and leptin-deficient (WTand ob/ob, 8–10 weeks of age, a) or lean and obese (16 weeks of HFD, b) mice at ×6,800magnification. Scale bars, 500 nm. Bottom images are magnifications of theboxed areas in the top images. M, mitochondria; N, nucleus; LD, lipid droplet.(c,d) TEM imagesanalyzed using IMOD software to delineate major cellular structures at ×1,400. Scale bars, 2 ìm. (e) Quantitation of ER length adjacent to mitochondrianormalized by total ER length and by mitochondrial perimeter. (f) Confocal images (left) of primary hepatocytes isolated fromlean (WT) and obese (ob/ob) mice coexpressinga mitochondria-targeted GFP (mito-GFP), an ER-targeted DsRed (ER-DsRed) andstained with a lipid droplet–specificdye (monodansylpentane). Scale bars, 10 ìm. Statisticalquantification (right) of the overlapping area (Manders’s coefficient) between mito-GFP and ER-DsRed, n = 10 (WT) n = 11 (ob/ob) from 4 independent experiments. (g,h) Quantitation ofER perimeter (g) and number of mitochondria (h). The morphometric analyses of the TEM pictures (e,g,h) were calculated from average of 50 pictures (10pictures per animal) in each experimental group (from 5 different animals pergroup). (i) 3D reconstruction of serial TEM sections obtained fromindependent samples taken at ×11,000of single mitochondria. Scale bars, 500 nm. All the graphs represent mean ± s.e.m.,*P < 0.05, Student’st-test.
Figure 2 ER structural and functional changes in response to acutestress. (a) Representative TEM images of liver sections from leanmice following 4-week challenge with regular chow or HFD. Scale bar, 500 nm. (b–d) Quantitation of ER perimeter (b) ER length adjacent to mitochondria normalized by totalER length (c) and number of mitochondria (d). The morphometric analyses were calculated from averageof 18 pictures (~5 pictures per animal) in each experimental group (from 3different animals per group) using ImageJ. (e) Evaluation of cell stress signaling activation in liverfollowing 6 or 24 h of tunicamycin (Tm) exposure by western blotting. (f) Representative TEM images of liver sections fromwild-type mice following 6 or 24 h of TM treatment. Insets representenlargements of the boxed areas. Scale bars, 2 ìm (top images), 500nm (bottom images) and 500 nm (insets), n = 3 mice per treatment. M, mitochondria; LD, lipiddroplet. The graphs represent mean ± s.e.m.,*P < 0.05, Student’s t-test.
Figure 3 Regulation of MAM enriched proteins in obesity. (a) Schematic illustration of functional and structural MAMproteins. S1R, Sig1R. (b) Western blot (left) and ImageJ-based quantification analysis (right) ofthe indicated proteins in liver total lysates from WT, ob/ob, lean and HFD-fed (16 weeks) mice. Arrowhead indicatesthe specific Sig1R band, n = 3 representativeof 3 independent experiments. AU, arbitrary units. (c) Western blot analysis of the indicated proteins ofsubcellular fractions from mouse livers. CM, crude mitochondria; PM, puremitochondria. Cytochrome C (Cyto C) was used as a mitochondrial marker; IP3Rand SERCA are ER markers. This panel is representative of 5 differentpreparations. (d) Western blot and quantification analysis of theindicated proteins in MAMs from WT, ob/ob, lean and HFD mice, n = 3 animals per group. PDI served as a loading control.All the graphs represent mean ± s.e.m., *P < 0.05, Student’s t-test.
Figure 4 Obesity alters mitochondrial Ca2+ flux liver. (a) Schematic depicting the details of the experimental procedure used tomeasure mitochondrial Ca2+ levels in primary hepatocytes. (b) Mitochondrial Ca2+ content in primary hepatocytes from wild-type (WT) and ob/ob mice, n = 81 (WT) or 60 (ob/ob) cells, 4 independent experiments. (c) Trace of [Ca2+]m upon 100 ìM ATP stimulation. This is representative of 15 traces. (d) Quantification of the Ca2+ peak after ATP stimulation, n = 15 cells, 4 independent experiments. (e) Cytosolic Ca2+ measured with 4 ìM of Fura2-AM after 1 ìM FCCP treatment, n = 60 (WT) or 40 (ob/ob) cells, 3 independent experiments. (f,g) OCR of primaryhepatocytes from lean and ob/ob mice in basal assaymedium (basal respiration), in the presence of 2 ìM oligomycin (ATPsynthase inhibitor), 1 ìM FCCP (uncoupling agent) and 2 ìM rotenone (Rot)and 2 ìM antimycin (AA) (complex I and III inhibitors, respectively). Maximalrespiration, proton leak and coupled respiration were determined as describedin the Online Methods section, n = 11 (WT) n = 13 (ob/ob) plates, 7 independent preparations. (h) Mitochondrial membrane potential in isolatedhepatocytes from wild type and ob/ob mice, n = 23 (Wt), n = 15 (ob/ob), representative of 3 independent experiments. AU, arbitraryunits. (i) ROS production measured with 3 ìM MitoSOX loading. n = 57 (WT) and 39 (ob/ ob) cells from 3 independent experiments. All the graphsrepresent mean ± s.e.m., *P < 0.05, Student’s t-test.
Figure 5 Experimental induction of ER-mitochondria interactionsincreases mitochondrial Ca2+ flux and impairs glucose homeostasis. (a) Left, schematic illustrating the mitochondrial (M)-ER syntheticlinker. The construct encodes a monomeric RFP fused to the outer mitochondrialmembrane targeting sequence of mouse A kinase anchor protein 1 at the Nterminus and the ER targeting sequence of yeast ubiquitin-conjugating enzyme 6at the C terminus with a length of ~15 nm. Right, TEM images of Hepa 1-6 cellsexpressing control (top) and linker (bottom) plasmids at ×9,300. Scale bars, 500 nm. Representative of 3 differentexperiments (10 pictures each). (b) FACS analysis of annexin V positivity in control andlinkerexpressing cells, n = 3. (c) Trace of mitochondrial Ca2+ dynamics in single Hepa 1-6 cells expressing the controlor linker construct, measured by FRET following treatment with ATP (100 ìM). Representativeof 3 different experiments. Inset, quantification of the peak in Ca2+ after ATPstimulation, 3 independent experiments. (d) Measurement of OCR in control and linker-expressingHepa1-6 cells. n = 15 from 4independent experiments. (e) TEM images ofliver sections from lean mice expressing the control (AdControl) or linker(AdLinker) construct. Scale bars, 500 nm. (f) Liver sections from HFD-fed mice expressing control orlinker constructs and stained with H&E. Scale bars, 50 ìm. (g) OCR of primary hepatocytes from HFD-fed mice expressingcontrol or linker constructs in the presence of 2 ìM oligomycin, 0.4 ìM FCCP and 2 ìM rotenone (Rot)and 2 ìM antimycin (AA). n = 10, from 3independent preparations. (h) Western blot ofthe indicated proteins in liver lysates from HFD-fed mice expressing control orlinker constructs. (i) Glucose tolerancetest (left) and area under the curve (AUC) (right) of mice expressing controlor linker constructs, n = 6 (AdControl-RD(regular diet), AdLinker-RD and AdControl-HFD) n = 7 (AdLinker-HFD). (j) Glucose infusion rate (left) and hepatic glucoseproduction (HGP) (right) during hyperinsulinemic-euglycemic clamp in HFD-fedmice expressing control or linker constructs, n = 8 for AdControl and 11 for AdLinker. All data are mean ± s.e.m., *P < 0.05 (two-way analysis of variance for panel j or Student’st-test for the others).
联系客服
