Vps34 PI 3-kinase inactivation enhances insulin sensitivity through reprogramming of mitochondrial metabolism

Vps34 PI3K is thought to be the main producer of phosphatidylinositol-3-monophosphate, a lipid that controls intracellular vesicular trafficking. The organismal impact of systemic inhibition of Vps34 kinase activity is not completely understood. Here we show that heterozygous Vps34 kinase-dead mice are healthy and display a robustly enhanced insulin sensitivity and glucose tolerance, phenotypes mimicked by a selective Vps34 inhibitor in wild-type mice. The underlying mechanism of insulin sensitization is multifactorial and not through the canonical insulin/Akt pathway. Vps34 inhibition alters cellular energy metabolism, activating the AMPK pathway in liver and muscle. In liver, Vps34 inactivation mildly dampens autophagy, limiting substrate availability for mitochondrial respiration and reducing gluconeogenesis. In muscle, Vps34 inactivation triggers a metabolic switch from oxidative phosphorylation towards glycolysis and enhanced glucose uptake. Our study identifies Vps34 as a new drug target for insulin resistance in Type-2 diabetes, in which the unmet therapeutic need remains substantial.


2.The authors show that AMPK phosphorylation is increased in myotubes/hepatocytes from Vps34
Het KI mice/ Vps34 inhibitor-treated cells, and they claim this would be the key mechanism by which Vps34-deficiency/inhibition causes enhanced insulin sensitivity and protection from high fat diet induced embolic abnormalities. However, there are several questionable observations/interpretations that need to be addressed. a.What is the mechanism by which AMPK is activated via Vps34-deficiency/inhibition? The authors show there is no change in ATP content in muscle/liver of Vps34 KI mice. Would this be different if authors measure ATP (AMP) levels in myotubes/hepatocytes (due to reduced mitochondria respiratory capacity) from Vps34 KI mice? If so, how do the authors reconcile in vivo and in vitro data?
b.The authors suggest that enhanced glucose uptake in Vps34-deficient muscle cells is mediated via AMPK-TBC1D1 axis and show increased p-S660 TBC1D1 signal. There is no evidence that this site has any functional significance. Other site S237/1 has been better characterized with some genetic evidence associated (O'Neill et al, PNAS, 2011, Frosig, J Physiol, 2010, Chen et al, Diabetologia, 2017. c. Fig 9c and d: Vps34-deficient muscle cells have higher basal AMPK activity and insulinindependent 2DG uptake. They also show that Vps34 inhibitor causes an increase in 2DG in the absence of insulin. Based on these data with enhanced AMPK signaling coupled, they conclude that enhanced insulin-independent glucose uptake in muscle is likely AMPK-dependent via Vps34 inhibition. If that is the case, Vps34 inhibitor-mediated increase in 2DG should be abolished in Vps34-deficient muscle cells. It might be the case that Vps34 is simply a mitochondrial poison which indirectly activates AMPK independently of Vps34. Moreover, to verify that insulinindependent increase in glucose uptake in Vps34-deficient muscle cells is due to AMPK (Fig. 9c), please include AICAR/A769662 and show there is no or blunted increase by these activators (alternatively AMPKa1/a2 can be knocked down by RNAi if feasible in the authors lab).
Reviewer #2 (Remarks to the Author): The study evaluates a whole body heterozygous loss-of-function Vps34 mutant regarding a number of metabolic parameters. Mutants presented increased glucose tolerance and reduced predisposition for diet-induced hepatic steatosis. The phenotype was mostly due to reduced hepatic autophagy, which limited substrate availability of mitochondrial respiration and gluconeogenesis. Some of the metabolic outcomes were reproduced with a pharmacological inhibitor of Vsp34. In general this is a fine work providing important advance in the field and identifying a potentially interesting target for the treatment of diabetes. Because of the potential impact of these findings, some additional experiments and controls are required.
Major comments 1. One major concern about the outcomes of partially inhibiting Vsp34 is that mitochondria function is considerably modified. This could have a short term beneficial effect; however, in the long run this could severely impact on metabolic fitness. It is very important that ageing mutants are analyzed for: adiposity, blood lipids, fat deposition in blood vessels and muscle and markers of oxidative stress. Also, longevity should be determined (a simple Kaplan-Meier graph would be fine). 2. As repeatedly recommended in the series of guidelines for methods in autophagy research, transmission electron microscopy (TEM) is the gold-standard method for evaluation of autophagosome (Klionski et al Autophagy 2016 12:1-222). Critical experiments evaluating the impact of loss-of-function mutation of Vps34 should include TEM. 3. The kinase activity shown in figure 1b should include a positive control, another PI3K isoform. 4. Why in figure 2b fasting glucose is about 80 mg/dl for both WT and mutant, whereas in the other experiments (Fig 2a and 2d) glucose is different between WT and mutant? 5. Authors should calculate kITT for the experiment shown in figure 2d. for stating that mutants are more sensitive to insulin glucose decay during ITT is more important than actual glucose levels.
Minor 1. In abstract (line 51) and discussion (line 370) authors propose that drugs that inhibit Vps34 could be potentially useful in patients that are intolerant to metformin. They say that about 10 % of patients are intolerant to metformin but no reference is presented. In fact, most patients (in my experience much more than 90%) tolerate metformin and the only real major problem associated with the use of this drug is lactic acidosis, which is an extremely rare event. Please consider reviewing this information. 2. It seems there is a splice in the bottom immunoblots for LC3 in Supplementary Figure 4. If that is the case, the splice should be clearly labeled or the experiment repeated in order to obtain a satisfactory image in a single gel.
Reviewer #3 (Remarks to the Author): In this study Bilanges and coworkers addresses the impact of Vps34 (Vps34 phosphoinositide 3kinase (PI3K)), a central player of autophagy, on energy homeostasis. The authors report that systemic Vsp34 inactivation in mice leads to enhanced insulin sensitivity and improved glucose tolerance, with reduced hepatic glucose production and increased glucose uptake in muscle. At the mechanistic level, Vps34 inactivation dampened hepatic autophagy, limiting substrate availability for mitochondrial respiration and gluconeogenesis. In muscle, Vps34 inactivation triggers a metabolic switch from oxidative phosphorylation towards glycolysis and enhanced glucose uptake. Interestingly, Vps34 inactivation did not affect Akt signaling but activated the AMPK pathway, suggesting potential clinical utility for Vps34 inhibitors in Type-2 diabetic patients. This a very well performed study with potential therapeutic outcomes. Major points 1. In Figure 2, the ITT between the two gentotypes seems rather similar. How would it look expressed in percent of base line? The area under the curve should be measured. 2. Figure 3. It is surprising that insulin concentrations were not increased in response to HDF (i.e fasting hyperinsulinemia ?). In the same line, it is surprising not to have any increase in liver weight uponHFD treatment.
3. The observation that circulating levels of adiponectin is interesting. What is the 4. Glucose uptake is increased in skeletal muscle, with a only a tendency in white and brown adipose tissue. To exclude the participation of adipose tissues to the phenotype, insulin signaling experiments (such as Akt, IRS1/2 phopshorylation assays etc..) should be performed in white and brown adipose tissues.,and compared to muscle and liver. 5. Key genes of the b-oxidation pathway should be measured (PPARa, LCPT1 etc..). Does increased in b-oxidation rates explain the lack of steatosis in response to HFD diet ? NCOMMS-16-30432 -"Vps34 PI 3-kinase inactivation enhances insulin sensitivity through reprogramming of mitochondrial metabolism"

Reviewer #1 (Remarks to the Author):
To investigate the role that Vps34 PI3K, a key lipid kinase known as autophagy initiator, plays on organismal metabolism at systemic levels, the authors generated a kinase-inactive (D761A) knockin (KI) mouse model. Because they found that homozygous Vps34 D761A/D761A KI mice are lethal at embryonic stage, they instead characterized heterozygous Vps34 D761A/+ KI mice. Assuming that Vps34 D761A/+ would cause 50% loss of kinase activity, the authors anticipated that this may better mimic the physiological/pharmacological effects of Vps34 inhibitor (at systemic levels). In the current manuscript, they report that Vps34 D761A/+ KI mice display enhanced glucose tolerance and insulin sensitivity, which was associated with partially dampened hepatic autophagy, altered metabolic switching and increased glucose uptake in muscle. Mechanistically, they found that even though Vps34-deficiency does not affect Akt signaling, it promoted activation of AMPK pathway (leading to insulin-independent glucose uptake in muscle). Collectively, they propose that systemic inhibition of Vps34 (by pharmacological means) would be beneficial as potential treatment of metabolic disorders.

General comments:
The authors used an elegant and appropriate model to ask a key biological/physiological question. As clearly described, Vps34 exists as multi-protein complex and knocking out Vps34 is expected to alter expression of its binding proteins, which may cause unexpected secondary effects. Given that here the authors' intention is to inhibit Vps34 activity like pharmacological intervention with complex intact, kinase-inactive KI is the right approach to take. The authors report some interesting metabolic phenotype of the Vps34 Het KI mice, but there are multiple fundamental issues (as described below) and data not compelling enough to support their conclusion.
We are pleased with the overall assessment of our study by this Reviewer who clearly captured the importance of using gene 'knock-in' over gene 'knock-out' approaches for Vps34 studies, especially for modelling pharmacological kinase-inhibition. He/she further raised the following comments: Major comments:

The authors claim that Vps34 Het KI mice show enhanced insulin sensitivity, but (as described below) there is no clear evidence supporting this notion. Please clarify what is the definition of insulin sensitivity in the authors' point of view.
Our definition of increased insulin-sensitivity is an improved insulin-stimulated clearance of blood glucose. Fig. 2D, the authors concluded that KI mice are more insulin sensitive compared to control animals. However, if you actually look at the data, KI mice have lower basal glucose, but upon inulin injection, kinetic (time-course) and magnitude of changes in blood glucose levels look nearly identical between control and KI mice. Please show the data as delta (%) changes from the baseline of respective genotype. If the glucose curve is comparable between control and KI mice, do the authors still claim KI mice are insulin sensitive? If so, please justify.

a. Based on the ITT results shown in
As suggested by the Referee, we have now presented the data (Fig. 2d, 3d) as delta (% of baseline) of the respective genotype. The revised figures (included below) clearly show an enhanced insulin responsiveness in Vps34 KI mice compared to WT mice, both under normal chow diet (Fig. 2d) and high-fat diet (Fig 3d). We trust that this addresses this key concern. Updated Fig. 2d: Updated Fig. 3d: In Response to Referee 3, we have now also calculated the glucose disappearance rate from the ITT tests (kITT; %/min) as follows: kITT = (0.693x100)/t1/2 where t1/2 is the "half-life" calculated from the slope of the plasma glucose concentration. We found that the WT and Vps34 KI mice had a k ITT of 1.35 and 2.77%/min, respectively, indicating a faster glucose clearance in the Vps34 mutant mice, consistent with a better insulin sensitivity.
b. Assuming that insulin sensitivity is similar between control and KI (based on delta % change data), I will need to question the authors' interpretation regarding GTT. What is the potential reason that the KI animals are more glucose tolerant (assuming insulin sensitivity is similar between genotypes as argued above)?
As we have now shown, insulin sensitivity is improved in Vps34 KI mice.
Please show that data 1st phase insulin secretion (plasma insulin levels following 5-10 min glucose injection) and/or show muscle/liver p-Akt blots.
The muscle/liver p-Akt blots were presented in the original version of the manuscript (muscle and liver: Supplementary Fig. 6e,f; primary myotubes and hepatocytes: Fig. 9a,b). These data show a similar insulin-mediated Akt signalling in Vps34 KI and WT cells/tissues.

If the authors claim that basal AMPK levels might be increased and this explains higher glucose disposal per se, please show p-AMPK/p-ACC blots.
We have now added p-AMPK/p-ACC blots for muscle and liver (inserted in original blot in Fig.  S6e, f -highlighted in the blots below), documenting enhanced basal AMPK signalling in Vps34 KI tissues compared to WT controls, confirming the in vitro data on primary cells shown in the original Fig. 9a and b.

It would be also informative if the authors perform AICAR tolerance test and to see if the KI mice show blunted response (blood glucose kinetic).
This is an excellent suggestion, we have now performed an AICAR tolerance test, shown in the figure below. These data demonstrate a clearly blunted response to AICAR in Vps34 D761A/+ mice compared to WT mice. This suggests that the increased basal AMPK levels in Vps34 D761A/+ tissues indeed underlie the higher glucose disposal in these mice. We have now included these data as a new panel of Fig. 9g and inserted the following text in the results section (page 6) of the manuscript: "To assess the involvement of AMPK in the increased insulin-independent increase in glucose disposal in Vps34-deficient muscle cells, blood glucose was assayed after an intraperitoneal injection of the cell-permeable AMPK activator 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR). As shown in Fig. 9g, AICAR caused a significant reduction in blood glucose levels in WT mice, likely due to the combined effects of enhanced muscle glucose uptake activation and inhibition of hepatic glucose production. Interestingly, the hypoglycaemic effect of AICAR was blunted in Vps34 D761A/+ mice, indicating that the increased basal AMPK levels in Vps34 D761A/+ tissues are functionally related to the higher glucose disposal in these mice." Legend: AICAR tolerance test upon intraperitoneal injection of 0.15 g/kg of AICAR in mice after overnight starvation. ≥5-10 mice/genotype were used. Data represent mean  SEM (non-parametric Mann-Whitney t-test) *p<0.05, **p<0.01, ***p<0.001.
c. Fig. 4: Results of hyperinsulinemic-euglycemic clamp are not interpretable. As the authors do not show validation of "clamp" and also other associated key parameters.
A IC A R T o le r a n c e T e s t We now show the requested parameters in a new Supplementary Table 3 (shown below).
The key parameters presented in the Table show that the basal and hyperinsulinaemic glucose levels are clamped at euglycemic levels and there is no difference between the genotypes.

The authors show that AMPK phosphorylation is increased in myotubes/hepatocytes from Vps34 Het KI mice/ Vps34 inhibitor-treated cells, and they claim this would be the key mechanism by which Vps34-deficiency/inhibition causes enhanced insulin sensitivity and protection from high fat diet induced embolic abnormalities. However, there are several questionable observations/ interpretations that need to be addressed. a. What is the mechanism by which AMPK is activated via Vps34-deficiency/inhibition?
Our data suggest that Vps34 inactivation alters cellular energy metabolism by modulating the availability of key substrates for mitochondrial respiration, leading to a cellular stress. We propose that altogether, the reduced mitochondrial respiration and reduced ATP levels in Vps34 KI cells lead to AMPK activation. In the previous version of our manuscript, we had included a summary scheme of the overall mechanism as a supplementary figure, we have now moved this scheme to the main Figure 10d. This should help the Reader in capturing the data from the manuscript.

The authors show there is no change in ATP content in muscle/liver of Vps34 KI mice. Would this be different if authors measure ATP (AMP) levels in myotubes/hepatocytes (due to reduced mitochondria respiratory capacity) from Vps34 KI mice? If so, how do the authors reconcile in vivo and in vitro data?
Fig. 6e in the original manuscript documented a ~25% reduction in ATP levels in Vps34 KI liver compared to WT, with no significant changes in ATP levels in the muscle. Possible explanations for this were mentioned in the text -we stated that this could be due to compensatory mechanisms in the muscle and/or the notion that muscle, unlike the liver, can produce ATP from sources other than mitochondria (glycolysis and possibly phosphocreatinine) to cope with its high energy demands, possibly masking the reduction in mitochondrial ATP production.
To circumvent compensatory mechanisms induced by long-term inactivation of Vps34, we next tested whether acute Vps34 inhibition alters ATP levels in the murine C2C12 myoblast and the Hepa1.6 hepatoma cell lines. As shown in the figure below, inhibition of Vps34 dramatically reduced the ATP levels in these cells. This was also observed by Seahorse analysis in oligomycintreated myotubes (original Fig. 6b,c). This indicates that Vps34 inhibition can reduce ATP levels in cells of muscle origin. We have now inserted these new data and relevant text in the revised manuscript ( Supplementary Fig. 5e and text page 4) as follows: 'To circumvent compensatory mechanisms that could be induced by long-term inactivation of Vps34, we next tested whether acute pharmacological inhibition of Vps34 altered ATP levels in the murine C2C12 myoblast and Hepa1.6 hepatoma cell lines. As shown in Supplementary Fig. 5e, inhibition of Vps34 dramatically reduced the ATP levels in these cells, further indicating that Vps34 activity can control ATP levels in cells of muscle origin.'

New Supplementary Fig. 5e
Legend: Decreased ATP levels in Hepa.16 hepatoma and C2C12 myoblast cells upon Vps34 inactivation. Cells were cultured in presence or absence of 1 µM Vps34-IN1 overnight in Complete media (CM) or starvation media. ATP levels were determined using ATP bioluminescence kit (Roche).  Fig 9c and d: Vps34-deficient muscle cells have higher basal AMPK activity and insulinindependent 2DG uptake. They also show that Vps34 inhibitor causes an increase in 2DG in the absence of insulin. Based on these data with enhanced AMPK signaling coupled, they conclude that enhanced insulin-independent glucose uptake in muscle is likely AMPK-dependent via Vps34 inhibition. If that is the case, Vps34 inhibitor-mediated increase in 2DG should be abolished in Vps34-deficient muscle cells. It might be the case that Vps34 is simply a mitochondrial poison which indirectly activates AMPK independently of Vps34. Moreover, to verify that insulinindependent increase in glucose uptake in Vps34-deficient muscle cells is due to AMPK (Fig. 9c)

, please include AICAR/A769662 and show there is no or blunted increase by these activators (alternatively AMPKa1/a2 can be knocked down by RNAi if feasible in the authors lab).
We agree with the Reviewer that the proposed experiment would be an elegant way to demonstrate that insulin-independent increase in glucose uptake in Vps34-deficient muscle cells is due to AMPK. We agree that Vps34 inhibitor-mediated increase in 2DG should be abolished in Vps34-deficient muscle cells. However, this would only be the case upon homozygous Vps34 kinase-inactivation but not upon heterozygous inactivation as in our experiments, in which the remaining WT Vps34 allele will still respond to Vps34 inhibitor. We would therefore expect a Vps34 inhibitor to still have an impact on 2DG uptake in our cells.
However, we believe that the new data of the AICAR tolerance test (see comment #1b and new Fig. 9e) which reveal a blunted AICAR response, provide valid alternative support for the involvement of higher basal AMPK activity in glucose uptake.

Reviewer #2 (Remarks to the Author):
The study evaluates a whole body heterozygous loss-of-function Vps34 mutant regarding a number of metabolic parameters. Mutants presented increased glucose tolerance and reduced predisposition for diet-induced hepatic steatosis. The phenotype was mostly due to reduced hepatic autophagy, which limited substrate availability of mitochondrial respiration and gluconeogenesis. Some of the metabolic outcomes were reproduced with a pharmacological inhibitor of Vsp34. In general this is a fine work providing important advance in the field and identifying a potentially interesting target for the treatment of diabetes. Because of the potential impact of these findings, some additional experiments and controls are required.
Major comments 1. One major concern about the outcomes of partially inhibiting Vsp34 is that mitochondria function is considerably modified. This could have a short term beneficial effect; however, in the long run this could severely impact on metabolic fitness. It is very important that ageing mutants are analyzed for: adiposity, blood lipids, fat deposition in blood vessels and muscle and markers of oxidative stress. Also, longevity should be determined (a simple Kaplan-Meier graph would be fine).
We agree that this type of study would be very interesting but believe -in line with the Editor's opinion -that this experiment is out of scope for the present study and within the given timeframe. This is a fair comment, but we would add that TEM evaluation can be subjective, and thus requires expert analysis, and quantitative approaches for subtle phenotypes.

As repeatedly recommended in
As part of our study, we had in fact performed multiple experiments using serial block-face scanning electron microscopy (or SBF SEM) on liver tissues to generate high resolution 3-dimensional images as well as transmission electron microscopy (TEM) on liver and various muscle types and primary hepatocytes, in both starved and fed conditions. However, we did not detect distinctive differences (including morphologically abnormal mitochondria and/or abnormal intracellular compartments) from qualitative analysis of images between the Vps34 heterozygotes and WT samples. We have therefore decided not to include any EM data in our manuscript.
For the perusal of the Referee and Editor, we here provide the results of our quantitative in-depth analysis of primary hepatocytes in fed and starved (EBSS 1h) conditions. We trust that the very labour-intensive nature of these experiments will be appreciated -this analysis has taken us several months to perform, using the most advanced EM techniques.
Given that the results obtained were very subtle, and in our view would not add to the manuscript, we would therefore prefer not to include these data in the revised manuscript, unless advised otherwise by the Referee/Editor. However, we have now mentioned our observations in a statement in the results section of the manuscript (page 4) as follows: 'Qualitative and quantitative electron microscopy analysis did not reveal any robust quantitative differences between WT and Vps34D761A/+ primary hepatocytes under starved conditions, confirming our data  that 50% inactivation of Vps34 kinase activity does not fully abolish the formation of autophagosomes and autophagolysosomes under starvation (data not shown)'.
Our results are summarised below: We took images of all autophagosome (AP) and AP-like structures in 5 cells for each condition, sampling from 3 separate experiments on WT and Vps34 KI hepatocytes. The cells were chosen by placing a 20x20 grid over a low magnification TEM image of a single section, and using a random number generator to select individual grid squares. The entirety of a cell intersecting the chosen square was imaged.
From our existing data, we reasoned that there might be subtle changes in the number of forming (open) autophagosomes and therefore we developed an approach to quantify open autophagosomes and the length of the autophagosome membrane. We considered that APs containing material inside similar in granular appearance to the cytoplasm and lacking electron-dense material as "open" AP-like structures. In contrast, the APs containing electron-dense material were considered as "closed" (Fig. a below). Qualitative and quantitative EM analysis did not reveal any robust quantitative differences between WT and Vps34 D761A/+ primary hepatocytes under starved conditions, confirming our data that 50% inactivation of Vps34 does not abolish the formation of autophagosomes and autophagolysosomes under starvation ( Fig. a and b below). We also found that the length of the AP membrane structures in Vps34 D761A/+ primary hepatocytes was comparable to that in control cells. However, we found a significant increase in the number of autophagosomes which were not closed. Our quantitative analysis revealed a ~5-fold increase of open AP-like structures that do not contain mitochondria in Vps34 D761A/+ primary hepatocytes compared to WT under basal condition (Fed, complete medium) (Fig. b). In EBSS there is an increased tendency towards a similar phenotype which is not significantly different. This correlates with our data reporting a decrease of WIPI2 staining in Vps34 D761A/+ primary hepatocytes compared to WT (Fig. 5e). We speculate that the mild reduction in PI3P (Fig. 5d) seen upon heterozygous inactivation of Vps34 may affect the closure of the AP and LC3 lipidation and recruitment.
ferricyanide/1% osmium tetroxide for 1 h, prior to further washing in PB and incubation in 1% tannic acid in 0.05 M PB for 45 min to enhance membrane contrast. After a brief incubation in 1% sodium sulphate in 0.05 M PB, the coverslips were washed twice in distilled water, and dehydrated through an ascending series of Ethanol (30%, 50%, 2x70%, 2x90%, 2x100%), and propylene oxide before infiltration with Epoxy resin and polymerisation overnight at 60°C. The coverslips were removed from the resin blocks by plunging briefly into liquid nitrogen.
The polymerised blocks were trimmed by hand using a single edged razor blade to form a trapezoid block face for serial ultrathin sectioning. Using a diamond knife, serial ultrathin sections of approximately 70 nm thickness were collected on 1% formvar coated single slot grids. The sections were counterstained with lead citrate to further enhance contrast prior to viewing in the electron microscope (FEI Tecnai G2 Spirit BioTWIN with Gatan Orius CCD camera). For quantification of 'open' and 'closed' autophagosome-like structures, cells were chosen at random by placing a 20x20 grid over a low magnification image of a single section, and using a random number generator to select individual grid squares. The entirety of a cell intersecting the chosen square was imaged. The imaging conditions and depth of section (with respect to distance from the coverslip surface) were kept consistent between samples. All autophagosome-like structures in each of 5 cells from WT and Vps34 D761A/+ cultures (n=3 for each) were examined. The autophagosome-like structures were classified as 'open' when the contents of the structure were similar in appearance to the surrounding cellular cytoplasm; 'closed' structures displayed a distinct difference in the density and/or granularity of their contents when compared to the surrounding cytoplasm. The grid overlay for cell selection and viewing was done using ImageJ software.