SUPPLEMENTAL DATA Cardioprotective GLP-1

SUPPLEMENTAL DATA

Cardioprotective GLP-1 metabolite prevents ischemic injury by inhibiting mitochondrial trifunctional protein-α.

M. Ahsan Siraj*, Dhanwantee Mundil*, Sanja Beca, Abdul Momen, Eric A. Shikatani, Talat Afroze, Xuetao Sun, Ying Liu, Siavash Ghaffari, Warren Lee, Michael Wheeler, Gordon Keller, Peter Backx, Mansoor Husain SUPPLEMENTAL FIGURES AND FIGURE LEGENDS

A B

Figure S1, related to Figure 1. Pre-treatment with GLP-1(28-36) in a mouse permanent LAD- ligation model of experimental MI. (A) C57BL/6J mice were pre-treated for 14-d by continuous infusion with peptides shown, followed by LAD coronary artery ligation and were sacrificed on d4 post-MI. Non-significant changes in heart/body weight ratios were observed amongst treatment groups measured at d4 post-MI. (B) Non-significant changes were observed in non-fasting blood glucose levels between d0 and d14, measured using glucose-strips from tail vein blood samples. n=7-13/group.

Figure S2, related to Figure 1. Concentration-dependent effect of GLP-1(28-36) on LVDP recovery in isolated hearts undergoing IRI. Isolated hearts from male 10-12 wk old wild-type mice were perfused with varying doses of GLP- 1(28-36). LVDP recovery, expressed as % of LVDP recorded at end of reperfusion divided by LVDP prior to ischemia, was significantly increased at a minimum concentration of 6 nM, n=3-12/group. ****P<0.0001 by One-way ANOVA with Bonferroni post hoc test. A Human cytosolic coronary Ventricular Atrial HL-1 heart CMs CMs cells extract SMC EC

sAC-tr

GAPDH

B +/+ +/- -/- sAC 55 sAC- tr 35

GAPDH 37

kDa

Figure S3, related to Figure 2. Expression of sAC in different cardiac cell types. Monoclonal R21 antibody (CEP Biotech) was used to detect the 50 kDa isoform of sAC, which is known to be the most active form of sAC in mammalian cells. GAPDH was used as loading control (A). Western blot analysis to detect specificity of the R21 antibody. sAC bands were not detected in total heart lysates from sAC-/- mice, but were present in lysates from wild-type (sAC+/+), and heterozygous (sAC-/+) littermates (B). sAC-tr denotes truncated active isoform of sAC protein (50 kDa). A B C

Figure S4, related to Figure 3. GLP-1(28-36) does not stimulate intracellular cAMP accumulation or cytoprotection in cardiomyocytes. Cells were incubated with 450 μM IBMX for 30-min to inhibit cAMP degradation by phosphodiesterases, then treated for 10-min with 100 nM each of GLP-1(28-36), Scram(28-38) and PBS as negative controls and IPE and Forskolin as positive controls (n=3/treatment, each in triplicate), and lysed for measurement of cAMP levels by enzymatic immunoassay. GLP-1(28-36) failed to increase intracellular cAMP accumulation in (A) mouse ventricular CMs or atrial CMs and (B) hESC-CMs, compared to Scram-treated controls. (C) hESC-CM were pre-treated with 100 nM of

GLP-1(28-36) or Scram for 20-min followed by incubation with H2O2 (100 μM) for 48-h to induce oxidative stress (n=3/each in triplicate). Cell culture media was replenished once with fresh peptide

after 24-h incubation with H2O2. At the end of 48-h, LDH release was assayed in duplicates from aliquots of cell culture media by ELISA. Data shown are mean±SE Unlabelled GLP-1(28-36) Biotinylated GLP-1(28-36) 0 mins

A B

Biotinylated GLP-1(28-36) Biotinylated GLP-1(28-36) 5 mins 15 mins

C D

Figure S5, related to Figure 3. Cellular uptake of GLP-1(28-36). Mouse caSMC were stimulated with either (A) 50 µg/ml unlabelled GLP-1(28-36) or 50 µg/ml biotinylated GLP-1 for (B) 0- min, (C) 5-min, (D) 15-min. Cells were washed, fixed, permeabilized and stained with Streptavidin-Alexa Fluor 488 conjugate to detect biotinylated GLP-1(28-36) (green). Hoechst dye (1 mg/ml) was used to counterstain nuclei (blue). Images acquired by an Olympus FluoView confocal microscope, 40X objective. Scale bar 10 µm. FAM- 28-36 A C E FAM- 28-36 Rhodamine Dextran FAM- 28-36

Control siRNA Control significant are taken and detailed and microscope with Dyngo Human Figure

macropinocytosis Mean within

DAPI

Control Control at 4

A S

caEC above 40

6 (

± ,

. 30

X . SE 24 related containing

(C µM)

magnification

. .

- were D) (E of n= .

Cells -

3 transfection F) Human to

incubated independent

inhibitor Figure Human

fluorescent were Cav

and

caEC Dyngo4A -

1 siRNA 1 then

Amiloride cells

Amiloride fluorescent incubated .

with

were

GLP fixed experiments,

mounting were

250

1 in transfected

( (

28 µM 1 4 finally with

mM) intensities % -

paraformaldehyde, media GLP 250 )

and Clathrin incubated enters **** -

µM 1 with

( imaged P

28 and were

< GLP

the 36 clathrin siRNA .

0001

imaged ) quantified with -

1 cells as for ( 28

- by detailed 250

, -

15 washed 36

caveolin

by with student’s min µM )

using for macropinocytosis

GLP

above and spinning

15 3

D X 1 ImageJ min

with 1

dynamin t or ( . - 28

test,

All

and

control - PBS, 36 disk

. images

Data

) N imaged

for .

S mounted inhibitor, confocal -

shown

siRNA 15 =

(A were

min

- not B) as

A 7-36 9-36 28-36 SNAP IPE PBS

P-ENOS

ENOS

GAPDH

B C

Figure S7, related to Figure 3. GLP-1(28-36) increase eNOS phosphorylation in isolated hearts. (A) Representative western blot of p-eNOS and eNOS expression in tissue lysates from isolated hearts perfused with PBS control, 0.6nM GLP-1(28-36), 0.3nM GLP-1(2-36), 0.3nM GLP-1(9-36), Isoproterenol or positive control Nitric oxide donor S-Nitro-N-acetyl-DL-pencillamine (SNAP). (B) Densitometric quantification of p-eNOS expression relative to total eNOS. (C) Densitometric quantification of total eNOS expression normalized to loading control GAPDH. Data shown are mean±SE, ***P<0.001, ****P<0.0001 by One-way ANOVA with Bonferroni post hoc test.

Assay Cntl

A B C

PepTag

Scram 28 IPE PKA No MaxPKA

- + +

Figure S8, related to Figure 3. sAC-dependent actions of GLP-1(28-36) in human caSMC and caEC. (A) GLP-1(28-36) increased cAMP-dependent PKA activity in human caSMC vs. Scram control, as shown by increased amount of phosphorylated Peptag® substrate with densitometric analysis (lower panel, n=3/group) of agarose gel electrophoresis bands (upper panel, representative image of 3 independent experiments). (B) Human caSMC underwent similar inhibitor and peptide treatment as in A above, followed by 48-h incubation with H2O2 (100 μM) to induce oxidative stress (n=3/group, triplicate wells each). Cellular toxicity was measured by LDH release, assayed in duplicates from aliquots of cell culture media, as previously described. Inhibiting sAC with KH7 resulted in increased LDH release vs. inhibiting tmAC with Ddox or PBS controls, indicating loss of cytoprotective effects of GLP-1(28-36) without functional sAC. (C) Human caEC were incubated with 450 μM IBMX for 30-min to inhibit cAMP degradation by phosphodiesterases, then treated with 2 mM Probenecid , 100 nM GLP-1 (28-36) and 10 μM Forskolin (n=2/treatment, each in triplicate), and lysed for measurement of cAMP levels. Probenecid treatment inhibited the ability of GLP-1(28-36) to increase intracellular cAMP levels in caECs. Data shown are mean±SE, *P<0.05, **P<0.01 by One-way ANOVA with Bonferroni post hoc test. Ddox= 2,5 Dideoxyadenosine. Pre-Clear WHE with Figure S9, related to Figure 5. Pulldown protocol of streptavidin magnetic beads GLP-1(28-36). Whole heart protein extracts (WHE) from wild-type mice were pre-cleared with streptavidin Affinity Capture (SA) magnetic beads for 1h at 4oC. Biotin labeled GLP- 1(28-36) (BIOT(28-36)) or scrambled control Incubate WHE with (BIOT(scram)) were added to the WHE and mixed end- BIOT(28-36) or BIOT(Scram) over-end for 4h at 4oC. Streptavidin magnetic beads were added and mixed for 2h at 4oC to immobilize Immobilization biotin-labeled peptides, after which the SA beads were Add streptavidin magnetic isolated using a magnetic stand, washed and the beads proteins associated with BIOT(28-36) and/or BIOT(scram) were eluted from the beads using a low Isolation pH IgG elution buffer. In one set of experiments, the eluents were separated by SDS-PAGE, stained with Wash and Elute with IgG Coumassie blue and significantly visible bands were pH2.0 elution buffer cut, trypsinized and subjected to mass spectrometry analysis. In another group, the eluents underwent in- solution trypsin digestion, and proteins were identified by mass spectrometry analysis.

In-solution SDS-PAGE digestion In-gel digestion

Mass Spectrometry Analysis

28-36 Scram 28-36 Scram 28-36 Scram

44 93 72 79 74 61 105 40 29

Replicate 1 Replicate 2 Replicate 3

Figure S10, related to Figure 5. Venn diagram depicting number of proteins bound to BIOT(28-36) and BIOT(scram). Following affinity pull down and LC-MS/MS, samples were analysed using Sequest and X! Tandem and proteins were identified using Scaffold_4.3.4 to compare interacting partners of BIOT(28-36) vs. BIOT(scram). Data represent three independent pulldown experiments labeled as Replicate 1, 2, and 3. 28-36 denotes BIOT(28-36); Scram denotes BIOT(scram).

Replicate 1

Replicate 2

Replicate 3

Figure S11, related to Figure 5, Scatterplot denoting semi-quantitative analysis of fold-change increase in total spectral count bound to BIOT(28-36) vs. BIOT(scram). Affinity pull-down, LC- MS/MS and protein identification were as previously described. Semi-Quantitative analysis was performed in Scaffold_4.3.4. Each graph represents an independent biological replicate. Each data point in red towards the lower right portion of the graph denotes a protein with significant fold-change increase in total spectral count vs. BIOT(scram). 28-36 denotes BIOT(28-36); Scram denotes BIOT(scram). A

Figure S12, related to Figure 6. Seahorse glucose oxidation assay traces from caSMC cultured in high glucose media, human caEC, and fatty acid oxidation in human caEC. Seahorse XFe24 traces of glucose oxidation measured as Oxygen Consumption Rate (OCR) by sequentially injecting 20 mM glucose, 1 µM Oligomycin, 1 µM FCCP, 1 µM Rotenone and 2 µM Antimycin-A to (A) caSMC cultured in high glucose DMEM and (B) human caEC. (C) Seahorse XFe24 traces of exogenous palmitic acid oxidation measured as OCR by sequentially injecting 200 µM Palmitic acid conjugated to BSA, or BSA control, 1 µM Oligomycin, 1 µM FCCP, 1 µM Rotenone and 2 µM Antimycin A to human caEC. (n=3/treatment in all seahorse assays, each in triplicate). Data shown are mean±SE SUPPLEMENTAL TABLES AND TABLE LEGENDS:

Total Spectral Counts N=1 N=2 N=3 Mw Accession BIOT BIOT BIOT BIOT BIOT BIOT N (Out of Protein kDa No. 28-36 Scram 28-36 Scram 28-36 Scram 3) Trifunctional enzyme subunit ECHA_ 1 alpha, mitochondrial 83 200 25 166 11 94 11 3 MOUSE GN=Hadha

ATP synthase subunit alpha, ATPA_M 2 60 70 29 12 4 16 0 3 mitochondrial GN=Atp5a1 OUSE

Cluster of ADP/ATP ADT1_M 3 33 65 20 20 3 14 4 3 translocase 1 GN=Slc25a4 OUSE Cluster of Sarcoplasmic/endoplasmic AT2A2_ 4 115 55 15 15 3 21 0 3 reticulum calcium ATPase 2 MOUSE GN=Atp2a2

Glyceraldehyde-3-phosphate G3P_MO 5 36 38 14 19 6 20 5 3 dehydrogenase GN=Gapdh USE

Mitochondrial inner IMMT_ 6 84 37 2 25 6 16 0 3 membrane protein GN=Immt MOUSE

Isocitrate dehydrogenase IDHP_M 7 [NADP], mitochondrial 51 31 11 13 2 12 0 3 OUSE GN=Idh2 Tubulin beta-4B chain TBB4B_ 8 50 18 5 9 0 11 0 3 GN=Tubb4b MOUSE Ryanodine receptor 2 RYR2_M 9 565 13 0 2 0 3 0 3 GN=Ryr2 OUSE ATP-dependent 6- PFKAM_ 10 phosphofructokinase, muscle 85 12 0 5 0 6 0 3 MOUSE type GN=Pfkm

Table S1. List of most significant protein bound to BIOT(28-36). Biological Fold Total spectrum count Replicate Change BIOT(28-36) BIOT (Scram) 1 200 25 8 2 166 11 15 3 94 11 8.5 Mean±SE 153.3±31.2 15.7 ± 4.7 10.5±2.2

Table S2. Relative binding affinity of mouse heart MTPα to biotinylated peptides Spectral counts reveal MS-based identification of MTPα in streptavidin magnetic bead pull-down experiments using mouse whole heart lysates incubated with biotinylated-GLP-1(28-36) [BIOT(28- 36)] and -Scram(28-36) [BIOT(Scram)]. From three biological replicates, the relative binding affinity for MTPα was ~11 fold higher with BIOT(28-36) than BIOT(Scram).

Peptide name Amino acid sequence

GLP-1(7-36)NH2 H-HAEGTFTSDVSSYLEGQAAKEFIAWLVKGR-NH2

GLP-1(28-36)NH2 H-FIAWLVKGR-NH2

SCRAM(28-36)NH2 H-AGKFWRILV-NH2

Table S3. Amino acid sequences of synthetic peptides used Amino acid sequences of GLP-1(7-36), GLP-1(28-36) and SCRAM(28-36) used in this study. SUPPLEMENTAL METHODS

Animals: C57BL/6J mice were obtained from Jackson Laboratories (Bar Harbor, ME) and housed for at least 2 weeks (wk) before experimentation. sAC+/- mice, B6;129S5-

Adcy10tm1Lex/Mmnc, identification number 11659-UNC, were obtained from Mutant Mouse

Regional Resource Center (Chapel Hill, NC), a NIH-funded strain repository, and were donated by Lexicon Genetics Incorporated. The mice were bred in our facility with C57BL/6J strain to generate sAC-null (sAC-/-) and genotyped according to the supplier’s protocol. Breeding and genotyping of Glp1r-/- mice have been described (1). All experiments were performed on 10-12 weeks old male mice.

Reagents: C-terminal amidated peptides (GLP-1, GLP-1(28-36) and Scram(28-36) were synthesized by Bachem (Torrance, CA) and supplied in trifluoroacetate salt with >98% purity as confirmed by HPLC. Amino acid sequences of test peptides and scrambled version are listed in

Table S3. Osmotic minipumps with flow moderator (Model #1002) were from Alzet®

(Cupertino, CA). KH7 and 2,5-dideoxyadenosine (Ddox) were a generous gift from Dr. D.H.

Maurice (Queen’s University, Ontario, Canada). All other reagents were from Sigma-Aldrich

(Ontario, Canada).

Experimental MI model, in vivo: Drug infusion and LAD ligation: C57BL/6J mice were anesthetized with intraperitoneal (i.p) ketamine 100 mg/kg (MTC Pharmaceuticals, Cambridge,

ON) and xylazine 10 mg/kg (Bayer Inc., Etobicoke, ON) and implanted subcutaneously (s.c) with osmotic minipumps filled and primed (as per manufacturer’s protocol) with either peptides

1 or vehicle (0.9% saline). All peptides were diluted with 0.9% saline at a targeted delivery rate of

18.5 nmol/kg/day (d) for GLP-1(28-36) and SCRAM(28-36) (2) and 3.5 pmol/kg/min for GLP-1

(3) , in order to achieve an estimated steady-state plasma concentration of 100 pM, as previously reported (3). After 2-wk of continuous peptide infusion, animals were re-anesthesized with i.p ketamine/xylazine and underwent permanent surgical ligation of the proximal left anterior descending (LAD) artery to generate an experimental MI, as previously detailed (4, 5). The surgeon was blinded to treatment groups. At 4-d post-MI, animals were sacrificed and hearts excised for measurement of infarct size using 2,3,5-triphenyl tetrazolium chloride (TTC) staining, as described (6). The 4-d post-MI time point allows accurate measurement of infarct area prior to onset of cardiac rupture events (6).

Infarct size measurements: Freshly excised hearts were washed twice in cold PBS to remove excess blood, and cut into ~2-mm sections perpendicular to the long axis. Sections were incubated at 37oC in 2% TTC for 15-min, after which they were washed with PBS and fixed in

4% paraformaldehyde for 1-h prior to image acquisition using a digital scanner. Image J software

(NIH, USA) was used to measure infarct area (TTC unstained white area) and total left ventricle

(LV) area. For accuracy, measurement was performed and averaged by two separate investigators, with one blinded to treatment groups. Infarct size was expressed as a % of total ventricular area, which in global ischemia, is equal to the area at risk (AAR).

Ischemia reperfusion injury (IRI) model, ex vivo: Isolated heart preparations: 10-12-wk old

C57BL/6J or Glp1r-/- mice were heparinized (10 IU heparin/g body weight) and after 5-min anesthetized with i.p ketamine (100 mg/kg)/xylazine (10 mg/kg) mixture. Upon withdrawal of the pedal reflex, mouse hearts were excised quickly via midline incision and the dissected aorta

2 was immediately cannulated onto the 20 gauge stainless steel cannula to ensure retrograde, non- recirculating, perfusion of the coronary circulation with warmed (37oC) and oxygenated (95%

O2, 5% CO2) Kreb’s Heinseleit solution using the Langendorff apparatus, as described (7). To maintain optimal contractile function of the heart, cannulation time was restricted to less than 1- min, and the heart was allowed to perfuse for 3-min prior to any other intervention. Once the heart was observed to be beating regularly, a fluid-filled plastic balloon (3-4-mm diameter), connected to a pressure transducer, was inserted into the LV via a left atrial incision. The balloon was then inflated to measure heart rate (HR), LV- end systolic pressure (LVESP) and - end diastolic pressure (LVEDP), obtained from computer analysis software (Acknowledge 3.7.1.

Biopack System, CA). Iso-volumetric LVEDP was maintained between 8-10 mmHg by adjusting the volume of the balloon inside the ventricles. Isolated hearts exhibiting a heart rate <300 bpm during first 10-min of perfusion were excluded from the study.

IRI protocol: The heart was allowed to equilibrate for 20-min in order to allow adjustment to the retrograde perfusion conditions. After 20-min of equilibration, hearts were perfused with either

Kreb-Hensleit buffer or specific inhibitors for 20-min, followed by 20-min perfusion with peptide of interest. All inhibitors and peptides (diluted 10-fold to the working concentrations) were administered through a syringe pump (Harvard Apparatus) at one-tenth of the coronary flow rate. The drugs were thoroughly mixed with perfusion buffer to achieve the desired concentration at 37oC, prior to entering the aorta. Global ischemia was then generated by clamping inflow to the heart for 30-min, after which reperfusion was reinstated for 40-min. In the untreated control group, hearts were perfused with Kreb-Henseleit buffer for a total of 40- min prior to ischemia. At the end of IRI protocol, hearts were dismounted, wiped dry and weighed.

Measurement of mechanical function: LV developed pressure (LVDP) was defined as the index of contractile function and measured as the difference between LVESP and isovolumetric

LVEDP. Functional recovery was defined as the measure of cardiac performance and expressed as a % of LVDP at the end of 40-min reperfusion over LVDP just prior to global ischemia.

Functional recovery represents the improvement in myocardial mechanical functions following prolonged global ischemia.

Coronary flow: Coronary venous effluents emerging from the perfused heart every min were collected at time-intervals throughout the I/R protocol, and expressed as flow rate ml/min.

Coronary flow rate was used to (1) adjust inhibitor or peptide pumping rate to achieve the desired final perfusion concentration, and (2) determine extracellular release of lactate dehydrogenase (LDH) or cAMP.

Lactate dehydrogenase release: In addition to cardiac function, the extent of myocardial tissue injury was assessed by measuring LDH release in coronary effluents. LDH is an enzyme released extracellularly when injured or necrotic cells lose membrane integrity. An enzymatic assay kit,

TOX7 (Sigma-Aldrich, ON) was used to measure LDH levels. This assay is based on the reduction of NAD (substrate) by LDH to produce reduced NADH, which stoichiometrically converts tetrazolium (dye) to a colored substance that can be measured spectrophotometrically.

More specifically, the coronary effluents were centrifuged at 250 x g for 4 min to pellet cells. An aliquot of the supernatant was mixed with 2X the volume of LDH assay mixture consisting of

1:1:1 ratios substrate:dye:cofactor in 96-well flat bottom plates. The plates were incubated in the dark for 20-min at room temperature, at the end of which the reaction was terminated by addition of 1/10 volume 1N HCl to each well. Absorbance at 490 nm (Abs490) was measured in a Thermo

Max microplate reader (Molecular Devices, Union City, CA). Background absorbance was

4 measured at 690 nm (Abs690) and LDH activity was expressed as Abs490- Abs690 per coronary

-1 -1 -1 flow rate and standardized to wet heart weight (Abs490- Abs690· ml ·min ·g ).

IRI model, in vivo: Drug injection and IRI protocol: 10-12-wk old sAC-/- and sAC+/+ littermates were anesthetized with i.p. ketamine 100 mg/kg and xylazine 10mg/kg. Mice were place on heating pad at 37oC and the jugular vein was exposed by dissecting skin and superficial neck muscles to inject 3.6 mM/kg GLP-1(28-36) or Scram(28-36) in order to achieve ~6 nM GLP-

1(28-36) or Scram(28-36) plasma concentration. After 30-min, mice were orally intubated and ventilated with a rodent respirator (Harvard Apparatus, Holliston, MA, USA). Surgical thoracotomy was performed and a 7-0 silk suture was passed under the LAD at the lower edge of left atrium and tied with polyethylene (PE-10) tube to produce occlusion. After 45-min of ischemia, the occlusion was released by removing PE-10 tube and leaving a loose knot allowing reperfusion. At 2-d post IRI, mice were re-anesthetized, LAD was re-occluded and 5% Evans blue dye was injected via the jugular vein to assess AAR. Hearts were excised for measurement of infarct size by TTC staining, as described (6).

AAR and infarct size: Freshly excised hearts were washed twice and harvested in cold DMEM without phenol red, and cut into ~2 mm slices perpendicular to the long axis. Images were acquired with stereomicroscope for Evans blue staining. Slices were then incubated in 2% TTC for 30 min at 37oC, after which they were fixed with 10% formalin for 1-h prior to second image acquisition for TTC staining. Image J (NIH, USA) software was used to measure AAR (area not stained by Evans blue) from the first image, and infarct size (white/pale TTC-unstained area) and total LV area from the second image. Measurements were performed by two independent

5 investigators, with one of them blinded to the treatment groups. Infarct size is expressed as % of

AAR and total LV area.

Cell culture and drug treatments: Primary mouse caSMC: Whole coronary arteries were dissected from C57BL/6J mice, sAC-null and wild-type littermate controls and coronary artery smooth muscle cells (caSMC) were isolated as previously described (8). caSMCs were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum

(FBS: Wisent Inc., PQ), 1% Platelet-Derived Growth Factor-BB (PDGF-BB: Sigma-Aldrich,

ON) and 1% penicillin-streptomycin (Wisent Inc., PQ) and passages 3-5 were used for experiments.

Primary mouse neonatal ventricular cardiomyocytes (CM): Ventricular CM were isolated from

1-3-d old C57BL6/J mice as described (6, 9). All cell types were serum-deprived for 24-h prior to treatment with phosphodiesterase inhibitor IBMX (250 µM) for 30-min at 37oC in a humidified incubator, after which the drug of interest and/or vehicle PBS was added to the media to achieve desired final concentrations. After 10-min incubation, cells were lysed for measurement of cAMP, PKA or ATP. All experiments were performed at least 3 times and within each experiment every condition was tested with three replicates.

Primary mouse coronary artery endothelial cells (caEC): Primary mouse caEC were purchased from Cell Biologics, Chicago, IL (Catalog# C57-6093). The cells were cultured on 0.2% gelatin coated plates, maintained in complete EC medium (Cell Biologics Catalog #M1168) and supplemented with 5% FBS, 1mM glutamine, 0.1% growth factors (VEGF, ECGS, EG, heparin, hydrocortisone) and 1% penicillin/streptomycin.

Primary human coronary artery smooth muscle cells (caSMC): Primary human caSMCs were

6 purchased from Cascade Biologics (Gibco, C-017-5C) and maintained in M231 media (Gibco,

M-231) supplemented with Smooth Muscle Differentiation Supplement (Gibco, S-007-25).

HL-1 cell line: Mouse atrial cardiomyocyte cell line (HL-1) was cultured on plates coated with

0.02% gelatin and 12.5 µg/ml fibronectin and maintained in Claycomb Media (Sigma-Aldrich,

ON) supplemented with 10% FBS, 100 µM norepinephrine and 2 mM L-glutamine.

Primary human caEC: Primary human caEC were purchased from ScienCell Research

Laboratories, Carlsbad, CA (Catalog#6020). Cells were cultured on 2 µg/cm2 fibronectin coated plates, maintained in EC medium (ScienCell Catalog# 1001) and supplemented with 5% FBS,

1% EC growth supplement and 1% penicillin/streptomycin.

Human embryonic stem cell-derived cardiomyocytes (hESC-CM): hESC-CM were differentiated as described previously (10). Embryoid bodies (EBs) were dissociated between differentiation d-

20-24 with collagenase type II (Worthington Biochemical Corporation, Lakewood, NJ). Cells were cultured in STEMPRO® 34 SFM medium (Thermo Fisher Scientific) supplemented with

1% penicillin/streptomycin, 1% glutamine, 30 mg/ml transferrin, 5 mg/ml ascorbic acid and 13 mg/ml monothioglycerol (MTG). All experiments were performed in serum free RPMI cell culture medium.

cAMP assays: Isolated mouse hearts were perfused with buffer containing IBMX (100 µM) for

10-min, then switched to buffer mixed with peptide of interest. Coronary effluents were collected at timed intervals and cAMP levels measured using a cAMP enzymatic immunoassay (EIA) kit

(Cayman Chemical, Ann Arbor, MI). For in vitro measurements of cytoplasmic cAMP, cells (~2 x105) were lysed with 0.1N HCL, followed by acetylation of lysates with a mixture of KOH and acetic anhydride, and cAMP concentrations were determined using the Cyclic AMP ELISA kit

(Cayman Chemicals, Ann Arbor, MI). Specific inhibitors KH7 (20 µM), Ddox (50 µM) were added for 3-h, while probenecid (2 mM) was added only 15-min prior to peptide treatments.

Protein kinase A (PKA) activity assay: PKA activity in mouse and human caSMCs were measured using the PepTag® non-radioactive cAMP-dependent protein kinase kit (Promega,

Madison, WI), as previously described (11).

ATP assays: Following drug treatment, mouse caSMCs (~2 x 105 cells) were washed twice with ice-cold PBS, and lysed by addition of 2.5% (w/v) trichloroacetic acid to extract ATP. Cells were scraped from the dish and centrifuged at 10,000 x g for 5-min at 4oC, after which the supernatant was immediately diluted 10 times and neutralized with TRIS-acetate buffer, pH 7.75. ATP was measured using the ENLITEN® ATP assay system bioluminescence detection kit (Promega,

Madison, WI). Briefly 10 µl of sample was added to 100 µl of luciferin/luciferase (rL/L) reagent, and relative light units (RLU) signals were measured with a luminometer. A delay time of 2-s after rL/L addition and a 9-s integration time were used as per manufacturer’s recommendations.

All samples were assayed in duplicates and ATP levels expressed as average RLU.

Cellular injury in vitro: Human or mouse caSMCs were grown to 70% confluence in fully supplemented media. After 24-h serum-deprivation, cells were pre-treated for 20-min with vehicle PBS or 100 nM of GLP-1(28-36) or SCRAM(28-36), followed by 48-h incubation with hydrogen peroxide (H2O2: 100 µM) (9). Cell culture media was replenished once with fresh vehicle, GLP-1(28-36) or SCRAM(28-36) after first 24-h incubation with H2O2. For human caSMCs, specific inhibitors KH7 (20 µM) and Ddox (50 µM) were added for 3-h prior to peptide

8 treatments. At the end of 48-h H2O2 incubation, cellular necrosis was determined from LDH release assayed in duplicates from aliquots of cell culture media using a commercially available cytotoxicity kit (TOX7: Sigma-Aldrich, ON). For determination of apoptosis, mouse caSMCs were treated for 20-min with peptides of interest followed by 7-h incubation with 100 µM H2O2.

Cells were then lysed for protein determination of the apoptotic marker cleaved caspase-3 by

Western blot (12).

Cellular internalization assays: Human caEC were treated with serum-free media and placed at

37°C for 2-h, then incubated with GLP-1 at different doses (125, 250, 500 and 1000 nM) at 37°C for 15-min and washed twice with warm PBS. Cells were then fixed with 4% paraformaldehyde.

Cells were then washed 3X with PBS and mounted on slides with DAPI-containing (1:5000)

Dako Fluorescence Mounting Medium (Agilent, Canada). Cells were imaged by spinning disk confocal microscopy keeping settings constant between groups.

Dynamin inhibition: To inhibit dynamin, caEC were treated with 30 μM dyngo 4a (Abcam,

Cambridge, MA) in HPMI at 37°C for 30 min prior to the addition of GLP-1 (500 nM).

Clathrin and Caveolin-1 siRNA transfection: For siRNA transfections, a master mix was made containing 150 μL Opti-MEM (ThermoFisher, Canada), 4 μL Lipofectamine RNAiMAX

Reagent (ThermoFisher, Canada), and 32 nmol/L siRNA (clathrin or caveolin-1) or non- targeting control siRNA. Human caECs were transfected at 70% confluence in complete serum

EGM-2 and media was changed 24-h after transfection. All experiments on cells transfected with siRNA were performed 36 to 60-h after transfection when cells reached 100% confluence.

Macropinocytosis assay: To inhibit fluid phase internalization by macropinocytosis, serum- starved caECs were treated with 1 mM of amiloride hydrochloride hydrate (Sigma, Cat. No.

A7410) for 20 min at 37°C.

Soluble adenylate cyclase (sAC) siRNA transfection: Human caSMCs were transiently transfected with siRNA to knockdown sAC as previously described, with minor modifications

(13). siRNA duplexes consisted of a pool of four different predesigned sequences targeting the human sAC mRNA sequence (L-006353-00: Thermo Fisher Scientific Biosciences Corp, ON), with a scrambled, non-targeting siRNA pool used as control (D-001810-10). Briefly, the transfection reagent was prepared by incubating anti-sAC siRNA or scrambled siRNA with

Lipofectamine RNAiMax™ (Life Technologies, ON) in Opti-MEM medium (Life Technologies) for 20-min at room temperature. This mixture was then added to human caSMCs (cultured for

24-h at 50% confluence in antibiotic-free supplemented media) in order to achieve a final siRNA concentration of 50 nM. Transfection was carried out at 37oC for 24-h, after which the medium was switched to normal supplemented medium and cells were allowed to grow for another 48-h.

Knockdown of sAC was determined by qRT-PCR using primers against the human isoform of sAC as described (14), which revealed approximately 50% decrease in sAC gene expression.

MTPα siRNA transfection: GE Dharmacon siRNAs directed against mouse MTPα transcripts were purchased with a negative control siRNA that does not bind to any mouse mRNA.

MTPα siRNA-01 = CCAGAACCCAUAUUAAUUA

MTPα siRNA-02 = GUACGAGUCUGCCUAUGGA

MTPα siRNA-03 = GAGCUUGCAUUGACCAAAG

MTPα siRNA-04 = CCAAUACCGUGUGAUAACA

Negative Control siRNA-01 = UAGCGACUAAACACAUCAA

A combination of all four MTPα-specific siRNAs was electroporated into primary mouse caSMC yielded a 76.2±6.9% (N=7) knockdown of mouse MTPα mRNA compared to negative control siRNA (see qRT-PCR Methods).

Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR): Primary mouse caSMCs were electroporated (0.8 million cells per cuvette) with a total concentration of 100nM siRNA (either 25nM of each of four MTPα siRNAs or single negative control siRNA) using electroporation protocol CM-137 and P1 Nucleofection mix in an Amaxa Nucleofector 4D.

These conditions have been validated previously in our lab to give 90% electroporation efficiency using GFP cDNA and confocal microscopy to count electroporated caSMCs.

Electroporated cells were allowed to recover for 24-h in complete growth medium containing

50µg PDGF-BB, 10% FBS and high glucose (20 mmol/L) DMEM. At 24-h post-electroporation, caSMC were used for Seahorse assays, cAMP assays, whole cell protein extracts or RNA extracts (PicoPure kit; ThermoFisher). Total RNA (800 ng per sample) was DNased to remove genomic contamination, spiked with 2pg EGFP complimentary RNA per sample (cDNA synthesis efficiency control template) and converted to cDNA (VILO kit; ThermoFisher). qRT-

PCR was carried out with 20ng cDNA per well in duplicate using SYBR GREEN master mix

(ThermoFisher) and MTPα, β-microglobulin (housekeeping gene) or EGFP-specific primers and using copy number standards for each gene in an ABI HT7900 thermal cycler. Ct values for each sample and for 10, 102, 103, 104, 105 and 106 copies Standards series for each gene were used to calculate copy numbers for MTPα, β-microglobulin and EGFP mRNA in each sample. The mean copy number for two duplicates per sample was normalized for EGFP copies (cDNA synthesis

11 efficiency) and for β-microglobulin copies to compute normalized MTPα per sample.

Normalized MTPα copies in negative control siRNA-treated caSMC were taken as 100% MTPα expression and used to calculate % MTPα expression in MTPα siRNA-electroporated caSMC and to ascertain % knockdown brought about by MTPα-specific siRNA (76.2±6.9% ; N=7).

MTPα Forward = GCAAGGCAGTGACGCTGGTTATCTTGC (Ta = 66°C)

MTPα Reverse = ACCTGGCCGTTATAAAGCCCCATCAGG (Ta = 66°C)

β-microglobulin Forward = GCTATCCAGAAAACCCCTCAA (Ta = 57°C)

β-microglobulin Reverse = CATGTCTCGATCCCAGTAGACGGT (Ta = 57°C)

EGFP Forward = ACACCCTGGTGAACCGCAT (Ta = 57°C)

EGFP Reverse = CCATGATATAGACGTTGTGGCTGTT (Ta = 57°C)

Western blots: Protein extraction and quantification were as described (6, 15). Whole heart protein extracts from C57Bl6/J mice and cell lysates from caSMCs were resolved on SDS-PAGE

(Novex®: Life Technologies), blotted on PVDF membrane, then probed overnight at 4oC for sAC (R21, 1:1000, CEP Biotech Inc.), cleaved caspase-3 (#9661, 1:1500, Cell Signaling

Technology), HADHA (ab54477, 1:10,000, Abcam ), eNOS (#9672, 1:1000, Cell Signaling

Technology), phospho-eNOS (Thr495, #9574, 1:1000, Cell Signaling Technology) and GAPDH

(sc-25778, 1:10000, Santa Cruz Biotechnology) as loading control. Protein bands were detected using corresponding horseradish peroxidase-conjugated goat anti-mouse or anti-rabbit secondary antibodies (1:10000, Biorad), imaged by chemiluminescence with MicroChemi 4.2 imaging system (DNR Bio-Imaging Systems Ltd., Israel), and quantified using Quantity One version 4.6.

Affinity pulldown and mass spectrometry: Preparation of heart extracts: Hearts were dissected from 8-wk old male C57BL/6J mice, washed twice with cold PBS, snap-frozen and stored at -80oC for subsequent analysis. Frozen hearts were pulverized in a mortar and pestle bathed in liquid nitrogen. Frozen powder was dissolved and homogenized with 100 strokes on ice in 1 ml RIPA lysis buffer (Tris pH 7.4 50 mM; EDTA pH 8.0 1 mM; NP-40 1%; NaCl 150 mM; Na orthovanadate 2.5 mM; PMSF 1 mM; Protease inhibitor cocktail (Complete Mini

Roche, Roch, Madison WI); Phosphatase Inhibitor cocktail (Sigma P5726). Cells were allowed to swell on ice for 30-min, and centrifuged at 14,000 rpm for 30-min at 4oC. Lysate proteins were quantified using BCA reagent (Sigma-Aldrich, ON).

Affinity pulldown with magnetic beads: The affinity pulldown was performed at 4oC with 1000

μg pre-cleared heart extracts in the presence of 50 μg/ul biotin-labeled peptides in a total volume of 300 μl binding/wash buffer (Tris – buffered saline, TBS, Pierce Product No. 28379) containing 0.1% Tween™-20 detergent, (TBST). First, to limit non-specific binding, heart protein extracts (1000 μg) were pre-cleared for 1-h with 50 μl of Pierce™ Streptavidin Magnetic

Beads (Pierce Biotechnology, IL) that were prepared and washed according to manufacturer’s instructions. The pre-cleared heart extracts were then collected using a magnetic stand, and the magnetic beads were discarded. Affinity capture was done by end-to-end mixing of pre-cleared heart extracts with either BIOT(28-36) or control BIOT(scram) for 4-h. The captured complexes were then immobilized on 50 μl Pierce™ Streptavidin Magnetic Beads with further end-to-end mixing for 2-h. After immobilization, the beads were collected with a magnetic stand and the supernatant (unbound fraction) was saved at 4oC for further analysis. The beads were washed 3X with 300 μl TBST. Recovery of the bound proteins was achieved by adding 100 μl of IgG

Elution Buffer, pH 2.0 (Pierce Product No. 21028) to the beads. Elution was performed at room

13 temperature with mixing for 5-min. The beads were magnetically separated, and the supernatant containing the target proteins (eluents) were saved at 4oC for further analysis.

Digestion: Eluents underwent tryptic digestion by either (1) in-solution digestion after elution from beads, or (2) resolving on 10% gels with SDS-PAGE, staining with Coomassie blue and visible bands excised and digested. Peptides were extracted with 25 mM ammonium bicarbonate, evaporated to dryness in a speedvac and reconstituted with 5 μl 0.1% TFA in water. LC MS/MS was performed with no further clean up, using a linear ion-trap instrument (Thermofisher LTQ).

Database search: Tandem mass spectra were extracted and charge state deconvolution and de- isotoping were not performed. All MS/MS samples were analyzed using Sequest (Thermo Fisher

Scientific, San Jose, CA, USA; version 1.4.0.288) and X! Tandem (GPM, thegpm.org; version

CYCLONE (2010.12.01.1)). Sequest was set up to search Uniprot_mouse-Sep92014.fasta

(unknown version, 16678 entries) assuming digestion enzyme trypsin. X! Tandem was set up to search Uniprot_mouse-Sep92014 database (unknown version, 16683 entries) also assuming trypsin. Sequest and X! Tandem were searched with a fragment ion mass tolerance of 0.60 Da and a parent ion tolerance of 2.0 Da. Deamidation of asparagine and glutamine, oxidation of methionine, carbamidomethyl of cysteine, and biotinylation of lysine of the N-terminus were specified in Sequest as variable modifications. Glutamic acid to pyro-glutamic acid of the N- terminus, ammonia-loss of the N-terminus, glutamine to pyro-glutamine of the N-terminus, deamidation of asparagine and glutamine, oxidation of methionine, carbamidomethyl of cysteine, and biotinylation of lysine of the N-terminus were specified in X! Tandem as variable modifications.

Criteria for protein identification: Scaffold (version Scaffold_4.3.4, Proteome Software Inc.,

Portland, OR) was used to validate MS/MS based peptide and protein identifications. Peptide

14 identifications were accepted if they could be established at greater than 5.0% probability to achieve an FDR less than 0.1%. Peptide probabilities from X! Tandem were assigned by the

Peptide Prophet algorithm (16) with Scaffold delta-mass correction. Peptide probabilities from

Sequest were assigned by the Scaffold Local FDR algorithm. Protein identifications were accepted if they could be established at greater than 76.0% probability to achieve an FDR less than 1.0% and contained at least 2 identified peptides. Protein probabilities were assigned by

Protein Prophet algorithm(17). Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters.

Seahorse extracellular flux analyzer assays: Assay optimization and drug treatments. Cellular and mitochondrial bioenergetics was measured using a Seahorse XFe24 extracellular flux analyzer in intact caSMC and caEC. We performed glycolysis stress tests and energy substrate

(glucose and palmitate) oxidation with mito stress test as per the manufacturer’s instructions

(18). Mito stress test drugs (Oligomycin-A and FCCP) working concentrations required for caSMC and caEC were determined with dose-titration assays. 20,000 caSMC and 25,000 caEC per well were cultured in V-7 Seahorse plates. Cells were pretreated for 20-min with 100nM

GLP-1(28-36) or Scram(28-36) before starting each experiment.

Glycolysis stress test: Cells were incubated in glucose-free Seahorse assay media supplemented

o with 1mM pyruvate at 37 C in incubator without CO2 for 1-h prior to the assay. Injectors were loaded to add 20mM Glucose, 1μM Oligomycin and 100mM 2 Deoxy-Glucose (2-DG) and

15 glycolysis, glycolytic capacity and glycolytic reserves were calculated as extracellular acidification rate (ECAR).

Glucose oxidation: Cells were cultured in substrate-limited (DMEM, 1% FBS, 0.5mM Glucose,

1mM GlutaMAX and 1mM pyruvate) 24-h prior to the assay. Cells were then incubated in glucose-free Seahorse assay media supplemented with 1mM pyruvate at 37oC in incubator without CO2 for 1-h prior to the assay. Injectors were loaded to add 20mM glucose, 1µM

Oligomycin, 1µM FCCP, 1µM Rotenone and 2µM Antimycin A during mito stress test.

Exogenous glucose oxidation and others mito stress test parameters such as basal respiration, maximal respiration, spare capacity, ATP production, proton leak and non-mitochondrial respiration were calculated as oxygen consumption rate (OCR).

Palmitate oxidation: Palmitate was conjugated with FFA-free BSA at 4:1 molar ratio

Palmitate:BSA as described (18). Cells were cultured in substrate-limited media (DMEM, 1%

FBS, 0.5mM Glucose, 1mM GlutaMAX, 1mM pyruvate and 0.5mM carnitine) 24-h prior to assay. Cells were cultured in Seahorse assay media supplemented with 1mM pyruvate and 2.5

o mM glucose at 37 C in incubator without CO2 for 1-h prior to the assay. Injectors were loaded to add 200µM Palmitate-BSA conjugate or BSA alone, 1µM Oligomycin, 1µM FCCP, 1µM

Rotenone and 2µM Antimycin A during mito stress test. Exogenous palmitate oxidation and others mito stress test parameters such as basal respiration, maximal respiration, spare capacity,

ATP production, proton leak and non-mitochondrial respiration were calculated as oxygen consumption rate (OCR).

Trypan blue exclusion (cell viability): Cells were trypsinized and suspended in serum free medium. Equal volume of 0.4% trypan blue dye (Life Technologies) was added to the cell suspension and incubated for 3-min. A drop of the mixture was applied to hemocytometer and unstained (viable) and stained (nonviable) cells were counted under inverted microscope. The percentage of viable cells was calculated by dividing number of viable cells with total number cells and multiplying by 100. The viability of the cells was measured for 3 days under complete and serum-free culture medium condition. Cell viability by this Trypan-blue exclusion test reveals 89% viability at 24-h, 88% viability at 48-h and 82% viability at 72-h for caSMC cultures. For caEC cultures, Trypan blue exclusion test reveals 95% viability at 24-h, 92% viability at 48-h and 86% viability at 72-h.

Statistical analyses: Data are expressed as mean ± SEM. One-way ANOVA was used to analyze differences between ≥ 3 treatment groups using Prism v4.0 (GraphPad Software Inc.,

San Diego, CA). Two-way ANOVA was used to compare differences between GLP-1(28-36) and Scram(28-36) across multiple doses or different mouse strain . The Bonferroni post-hoc test was used for multiple comparisons. P<0.05 was considered significant.

References for Methods Supplement

1. Scrocchi LA, Brown TJ, MaClusky N, Brubaker PL, Auerbach AB, Joyner AL, and Drucker DJ. Glucose intolerance but normal satiety in mice with a null mutation in the glucagon-like peptide 1 receptor gene. Nat Med. 1996;2(11):1254-8. 2. Tomas E, Wood JA, Stanojevic V, and Habener JF. GLP-1-derived nonapeptide GLP- 1(28-36)amide inhibits weight gain and attenuates diabetes and hepatic steatosis in diet- induced obese mice. Regulatory peptides. 2011;169(1-3):43-8. 3. Perry T, Holloway HW, Weerasuriya A, Mouton PR, Duffy K, Mattison JA, and Greig NH. Evidence of GLP-1-mediated neuroprotection in an animal model of pyridoxine- induced peripheral sensory neuropathy. Experimental neurology. 2007;203(2):293-301.

4. Zaidi SH, Huang Q, Momen A, Riazi A, and Husain M. Growth differentiation factor 5 regulates cardiac repair after myocardial infarction. Journal of the American College of Cardiology. 2010;55(2):135-43. 5. Ohta K, Nakajima T, Cheah AY, Zaidi SH, Kaviani N, Dawood F, You XM, Liu P, Husain M, and Rabinovitch M. Elafin-overexpressing mice have improved cardiac function after myocardial infarction. American journal of physiology Heart and circulatory physiology. 2004;287(1):H286-92. 6. Noyan-Ashraf MH, Momen MA, Ban K, Sadi AM, Zhou YQ, Riazi AM, Baggio LL, Henkelman RM, Husain M, and Drucker DJ. GLP-1R agonist liraglutide activates cytoprotective pathways and improves outcomes after experimental myocardial infarction in mice. Diabetes. 2009;58(4):975-83. 7. Ban K, Noyan-Ashraf MH, Hoefer J, Bolz SS, Drucker DJ, and Husain M. Cardioprotective and vasodilatory actions of glucagon-like peptide 1 receptor are mediated through both glucagon-like peptide 1 receptor-dependent and -independent pathways. Circulation. 2008;117(18):2340-50. 8. Cornwell TL, and Lincoln TM. Regulation of intracellular Ca2+ levels in cultured vascular smooth muscle cells. Reduction of Ca2+ by atriopeptin and 8-bromo-cyclic GMP is mediated by cyclic GMP-dependent protein kinase. The Journal of biological chemistry. 1989;264(2):1146-55. 9. Ban K, Kim KH, Cho CK, Sauve M, Diamandis EP, Backx PH, Drucker DJ, and Husain M. Glucagon-like peptide (GLP)-1(9-36)amide-mediated cytoprotection is blocked by exendin(9-39) yet does not require the known GLP-1 receptor. Endocrinology. 2010;151(4):1520-31. 10. Yang L, Soonpaa MH, Adler ED, Roepke TK, Kattman SJ, Kennedy M, Henckaerts E, Bonham K, Abbott GW, Linden RM, et al. Human cardiovascular progenitor cells develop from a KDR+ embryonic-stem-cell-derived population. Nature. 2008;453(7194):524-8. 11. Shao W, Wang Z, Ip W, Chiang YT, Xiong X, Chai T, Xu C, Wang Q, and Jin T. GLP- 1(28-36) improves beta-cell mass and glucose disposal in streptozotocin-induced diabetic mice and activates cAMP/PKA/beta-catenin signaling in beta-cells in vitro. American journal of physiology Endocrinology and metabolism. 2013;304(12):E1263-72. 12. Sun X, Momen A, Wu J, Noyan H, Li R, von Harsdorf R, and Husain M. p27 protein protects metabolically stressed cardiomyocytes from apoptosis by promoting autophagy. The Journal of biological chemistry. 2014;289(24):16924-35. 13. Flacke JP, Flacke H, Appukuttan A, Palisaar RJ, Noldus J, Robinson BD, Reusch HP, Zippin JH, and Ladilov Y. Type 10 soluble adenylyl cyclase is overexpressed in prostate carcinoma and controls proliferation of prostate cancer cells. The Journal of biological chemistry. 2013;288(5):3126-35. 14. Schmid A, Sutto Z, Nlend MC, Horvath G, Schmid N, Buck J, Levin LR, Conner GE, Fregien N, and Salathe M. Soluble adenylyl cyclase is localized to cilia and contributes to ciliary beat frequency regulation via production of cAMP. The Journal of general physiology. 2007;130(1):99-109. 15. Noyan-Ashraf MH, Shikatani EA, Schuiki I, Mukovozov I, Wu J, Li RK, Volchuk A, Robinson LA, Billia F, Drucker DJ, et al. A glucagon-like peptide-1 analog reverses the molecular pathology and cardiac dysfunction of a mouse model of obesity. Circulation. 2013;127(1):74-85.

16. Keller A, Nesvizhskii AI, Kolker E, and Aebersold R. Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Analytical chemistry. 2002;74(20):5383-92. 17. Nesvizhskii AI, Keller A, Kolker E, and Aebersold R. A statistical model for identifying proteins by tandem mass spectrometry. Analytical chemistry. 2003;75(17):4646-58. 18. Mdaki KS, Larsen TD, Weaver LJ, and Baack ML. Age Related Bioenergetics Profiles in Isolated Rat Cardiomyocytes Using Extracellular Flux Analyses. PloS one. 2016;11(2):e0149002.