|Home | About | Journals | Submit | Contact Us | Français|
Decreases in alpha myosin heavy chain (α-MHC) is a common feature of human heart failure (HF), while α-MHC overexpression in transgenic (TG) rabbits is cardioprotective against tachycardia-induced cardiomyopathy (TIC). Hypothesizing that MHC isoform content alterations would impact sarcomere and mitochondrial energetics protein complement, we investigated the impact of α-MHC overexpression on global cardiac protein expression.
Protein expression was assessed by two-dimensional gel electrophoresis and mass spectrometry on extracts from TG and non-transgenic (NTG) rabbits under TIC or sham-operated conditions.
We observed significant changes in the levels of actin, myosin light chain 2 and desmin between the LV tissue of TG and NTG animals. The proteome was broadly impacted, with significant changes in mitochondrial energetics and chaperone protein families. No changes were observed in total cellular MHC or in myofibril-associated MHC. In myofibrils isolated from TGSham animals, only actin levels were altered in TGsham compared to NTGsham animals, suggesting careful myofibril assembly regulation.
These data suggest myofibril protein composition may protect against TIC, emphasizing protein interconnectivity and demonstrating the need for broad-based proteomic studies in understanding targeted genetic manipulations. This study identifies targets for future development of cardioprotective agents and elucidates tachycardia-induced HF pathways.
Heart failure (HF) is a pathophysiological state characterized by insufficient cardiac output partially due to decreased contractility. Cardiac contraction results from protein interactions between the thin and thick filaments. Thick myofilaments consist of the myosin complex, comprised of a pair of myosin heavy chains (MHC) and two pairs each of myosin light chain 1 and 2 (MLC-1, MLC-2), and associated proteins such as myosin binding protein C. The thin filaments are comprised of α-actin, α-tropomyosin (Tm) and the troponin complex, comprised of troponin I, troponin T (TnT) and troponin C. The sarcomere also contains associated proteins, including desmin and α-actinin. The MHC isoforms α-MHC or β-MHC are classified as either V1 myosin or V3 myosin, respectively. V1 myosin has a high rate of ATP hydrolysis[5, 6], whereas V3 myosin has a slower rate of ATP hydrolysis, but generates more force per unit of ATP. V1 myosin predominates in the rodent ventricle, whereas V3 myosin is the predominant isoform in human and rabbit ventricles.
Although α-MHC levels are low in the human and rabbit ventricle, decreases in α-MHC is a common hallmark of human [9, 10], rabbit , and rodent  heart failure. Transgenic (TG) rabbits with increased α-MHC expression displayed cardioprotection against tachycardia-induced cardiomyopathy (TIC) based on echocardiographic and hemodynamic parameters. These changes included increased pressure development, relaxation, and shortening fraction. Whether protective effects are attributable solely to α-MHC or to broader myocyte alterations resultant from transgene expression is not clear. We hypothesized that alterations to a single MHC isoform would result in general sarcomeric protein complement alterations and mitochondrial energetic changes in response to the increased enzymatic activity of α-MHC. To address this hypothesis we performed 1- and 2-dimensional gel electrophoresis (1DE and 2DE, respectively) studies on cardiac tissue, from TG and non-transgenic (NTG) rabbits that were either sham-operated or underwent TIC. Since functional thick and thin filament proteins must be incorporated into myofibrils, we sought to determine whether increased α-MHC in TG animals was fully incorporated into myofibrils, and whether overexpression of α-MHC altered expression of other myofibrillar proteins. We also examined whether expression of a recently described MHC protein chaperone, UNC-45B, was impacted in TG animals.
The experimental cohort consisted of 28 rabbits, 14 TG and 14 NTG (7 TIC and 7 sham, each group). Epicardial pacing lead placement and pacemaker implantation with sequential increases in LV pacing was performed over a 30-day period to induce cardiomyopathy with severe ventricular dilation as previously described. Sham animals were treated identically, but the pacemaker generator was not activated. All animals were treated humanely and in accordance with the Animal Welfare Act regulations of the U.S. Department of Agriculture and with the Guide for the Care and Use of Laboratory Animals.
From each animal, 100 mg of tissue was homogenized in 10 volumes per mass of tissue in whole tissue extraction buffer (20 mM Tris-HCl (pH 6.8), 7 M urea, 2 M thiourea, 4% ASB-14 (tetradecanoylamidopropyl-dimethylammonio-butanesulfonate), 50 mM NaF, 1μM leupeptin, 1 μM pepstatin, 2.5 mM EDTA, 0.36 μM aprotinin and 0.2 mM Na3VO4) using a dounce homogenizer. Extracts were centrifuged and supernatant stored at -80°C. For the analysis of the myofibril subproteome, myofibrils were purified from both failing NTG rabbits (n=3) and failing TG rabbits (n=3) in a protocol modified from Chandra et al.. Tissue (100 mg) was rinsed in 10 volumes per mass of Relax buffer (75 mM KCl, 10 mM imidazole (pH 7.2), 2 mM MgCl2, 2 mM EGTA, 1 mM NaN3, 4 mM phosphocreatine, 1 mM ATP, 50 mM BDM, 1 mM DTT, 1 mM benzamidine-HCl, 0.1 mM PMSF, 1 μg/mL leupeptin, 1 μg/mL pepstatin, 1% TX-100) and cut into small pieces. The buffer was then altered to include 10 mM EDTA and the tissue was homogenized using a Polytron (Kimatica AG, Littau-Lucerne, Switzerland) homogenizer at 55% of maximum speed for 12 seconds, five times. The homogenized sample was centrifuged at 3,000 × g at 4°C, the supernatant was discarded while the pellet was resuspended in 10 volumes / weight of Standard buffer (5 mM KCl, 10 mM imidazole (pH 7.2), 2 mM MgCl2, 2 mM EGTA, 1 mM NaN3, 1% TX-100). The suspended pellet was homogenized with a Dual glass-on-glass homogenizer and then centrifuged at 3,000 × g for 8 minutes at 4°C. The supernatant was discarded and the pellet was suspended again in Standard buffer (10 volumes / weight) and allowed to sit on ice for 5 minutes to solubilize membranes and non-myofilament proteins. The suspension was centrifuged at 3,000 × g at 4°C, and the pellet suspended in Standard buffer as before, which was repeated 4 times. The last two times, the pellet was suspended in Standard buffer without TX-100. Following the final wash, the pellet was highly enriched for myofibrils (myofilament subproteome). The myofibrils were suspended in K-60 buffer (60 mM KCl, 20 mM MOPS, 2 mM MgCl2) and stored at 4°C for ATPase assays or solubilized into whole tissue extraction buffer for 1-DE. Protein concentrations were determined using a BCA protein assay (Pierce Biotech, Rockford, IL)
For 1DE analysis, ~1 μg of myofibril extracts from NTGTIC (n=3) and TGTIC rabbits (n=3) were labeled with cyanine dye (Cy3) (as previously described) and applied to 4-12% bis-Tris acrylamide gels (Invitrogen) with MOPS running buffer (Invitrogen). A higher load (10 μg protein) preparative equivalent gradient gel was run and silver stained. For 2DE analysis, 15 μg of tissue from each rabbit whole tissue extract was first labeled with cyanine dye (Cy3 or Cy5). Proteins were isoelectric focused in 7 cm pH 4-7 Drystrips (GE Healthcare) in IEF buffer (7 M urea, 2 M thiourea, 4% CHAPS, 1% DTT, 1.5% HED). The second dimension separation (SDS-PAGE) was performed using 4-12% bis-Tris gels (Invitrogen) and MES running buffer (Invitrogen). IPG strips with a pH gradient of 4-7 were used as we were primarily interested in the changes in the myofilament proteins located in this pH range. A preparative 2D gel for protein identification (250 μg unlabeled protein extract from a NTG sham rabbit) was made by isoelectric focusing in 18 cm pH 4-7 Drystrips (GE Healthcare) and a second dimension separation was performed using 20 cm × 20 cm bis-Tris SDS-PAGE gels with a resolving gel containing 10% acrylamide and a stacking gel containing 4% acrylamide. Proteins were separated in MOPS running buffer. Western blotting was performed using the NuPAGE western blotting system and apparatus as described by the manufacturer (Invitrogen), with chicken polyclonal IgY antibodies to UNC-45 (1:10,000 dilution, Genway), donkey anti-chicken alkaline phosphatase secondary antibodies (1:10,000), incubated with substrate (Pierce, Lumiphos) and exposed to film.
For Cy3/5 labeled proteins, gels were scanned using a Typhoon 9400 Imager (GE Healthcare) and analyzed by Ludesi Inc. (Lund, Sweden) for spot matching and quantification. For 1DE, individual protein bands were quantitated using Progenesis Workstation 2005 (Nonlinear Dynamics). For 2DE, the average integrated intensity for each spot was determined for each group of rabbits and four comparisons were performed between the experimental groups. Statistical significance was determined between groups using both a 1-way and 2-way ANOVA to examine differences between groups as well as interaction between transgene incorporation and tachycardia. In order to visualize protein spots that would be analyzed by MS, 2DE gels were silver stained using a non-glutaraldehyde protocol adapted from the protocol of Shevchenko et al as previously described. Proteins observed as multiple protein spots, possibly due to post-translational modification, were further analyzed by MS to confirm their identity as altered forms of the original protein. When multiple isoelectric forms of a protein were identified, the abundance of a particular isolectric form of a protein was compared to the total abundance of that protein.
For MS identification, protein spots (2DE) or bands (1DE) were excised from silver stained gels, destained and washed multiple times with 50% (v/v) ACN in water (supplemental methods), dried via speedvac (Savant) and analyzed by MS. For MALDI-TOF MS analysis, dried peptides were reconstituted in 5 μL of 50% (v/v) ACN, 0.1% (v/v) TFA. Equal volumes (0.5 μL) of peptide solution and matrix (10 mg/ml α-cyano-4-hydroxy-trans-cinnamic acid in 50% (v/v) ACN and 0.1% (v/v) TFA) were dried on a target plate. Dried plates were analyzed using an ABI-4800 TOF-TOF MS (ABI Technologies) and searched against the NCBInr database using Mascot software (Matrix Science). For LC-MS analysis, dried peptides were solubilized in 0.1% formic acid and loaded onto a C18 column connected to an Eksignent nano-2DLC pump as previously described. Eluted peptides were analyzed with a LTQ-LC/MS/MS (ThermoElectron) and queried against the NCBInr database using Mascot software (Matrix Science) searching against mammals.
Myofibrillar ATPase activity was determined for myofibrils purified from NTGTIC (n=3) and TGTIC (n=3) rabbits as previously described . ATPase rates under positive Ca2+ conditions (7 mM EGTA, 7 mM CaCl2) and negative Ca2+ (7 mM EGTA) were examined.
Myocardial tissue extract (50 ug) from NTG rabbits was diluted into 50 mM Tris-HCl (pH 7.9), 100 mM NaCl, 10 mM Mg Cl2, 1 mM DTT and incubated overnight at 37°C with or without the addition of 20 units of calf intestinal alkaline phosphatase (AP, New England Biolabs). AP treated samples were labeled with Cy3 whereas untreated sample was labeled with Cy5. Samples were combined and run on a single 2D gel and scanned as previously described.
Total protein extracts from LV tissue obtained from sham-operated or TIC NTG and TG animals were subjected to 2D-DIGE (n = 7/condition) (Figure S1). A comparison was carried out using both pooled samples (n=3/gel) and also individual animals. Figure 1A shows the quantitative data for the myofilament proteins Tm, TnT, MLC-1 and MLC-2, α-actin and desmin. Significant changes were observed for 55 spots representing 17 proteins, and direct comparisons between all experimental groups are presented in Table 1. The thick filament protein MHC, due to its high molecular mass, is not amenable to quantitative analysis by 2DE. There were no significant changes observed in abundance or ratio of the thin filament associated proteins Tm or TnT following transgene incorporation or TIC. A common effect of TIC in both TG and NTG animals was an increase in MLC-1 (P=0.04) and MLC-2 (P<0.01, Figure 1A). Although the absolute ratios of MLC-1 and MLC-2 differed between the TG and NTG animals, TG rabbits possess a higher MLC-2:MLC-1 ratio compared to NTG rabbits, regardless of TIC status. MHC isoform composition largely modulates catalytic activity of cardiac myosin; the data therefore suggest changes in force generation. Although there were no significant differences in protein abundance of MLC-2 between NTGTIC and TGTIC animals, there was a change in the most basic form of MLC-2 (Spot 4, pI=4.9). As confirmed by treatment with calf alkaline phosphatase, this was due to differential phosphorylation (spot 4; pI = 4.9, unphosphorylated form of MLC-2) (Figure 2). Actin underwent a differential response depending upon transgene expression. Under sham conditions, NTG rabbits possessed 1.2-fold more actin than TG rabbits (P=0.03). TGTIC rabbits showed increased actin, whereas NTGTIC rabbits showed a reduction (Figure 1A, Table 1).
The intermediate filament protein desmin, which is associated with the sarcomere through its interaction at the Z-disk,  showed increased abundance in both NTGTIC and TGTIC rabbits (Figure 1), although desmin increases in TGTIC animals did not reach the same level as NTGTIC animals (P=0.01). Desmin levels were ~35% lower in TGTIC rabbits compared to the NTGTIC rabbits (Figure 1A, P=0.03). We also observed a reduction in the abundance of the most basic form of desmin in TGsham and TGTIC animals (Table 1, Figure S2).
Using 2DE-DIGE it was not possible to determine whether the changes in the myofilament proteins reflected changes in intact myofilaments or in the soluble, non-incorporated protein pool. We hypothesized that the effects of altered α-MHC would be most pronounced in myofilaments, so myofibrils were isolated and compared to total cellular extracts of the myofilament proteins from the LV-free wall tissue of NTGTIC and TGTIC rabbits (Figure 3A and B). Absolute and relative levels of MHC, MLC-1, and MLC-2 were unchanged in intact myofibrils between NTGTIC and TGTIC rabbits. Thus, the increase in MLC-2 observed in the 2DE experiments for the TG rabbits is most likely due to increases in the soluble protein pool. As α-MHC levels are maintained in TGTIC animals (but undetectable in NTGTIC animals), increases in the soluble pool of α-MHC due to TG expression suggests that proteins involved in chaperoning myosin assembly might be elevated during TIC in TG animals. We therefore focused on a newly identified co-chaperone for α-MHC, UNC-45. UNC-45 is essential for formation of thick filament in C elegans.[22-24]. Among vertebrates, there are two UNC-45 homologues: UNC-45A and UNC-45B. UNC-45A is ubiquitous, while UNC-45B is found only in cardiac and skeletal muscle, and has a role in sarcomere formation. Western blot analysis indicates that UNC-45B was upregulated >5-fold in TGTIC animals compared with TGSham animals (Figure S3). Similar to the findings from total cell extracts, there was also higher actin levels in TGTIC rabbits compared to NTGTIC rabbits (20%; P=0.01) from isolated myofibrils. TnT levels were elevated in myofibrils (24%, P=0.04) but not whole cell extracts.
These data suggest that there is a selective increase in at least some of the thin filament proteins relative to the thick filament (e.g. MHC) in the myofibrils of TG rabbits following pacing. Since several key myofilament proteins are differentially regulated, we investigated whether this affected myofibrillar ATPase acvitivty. Ca2+-activated myosin ATPase was increased in TGTIC myofilaments compared to NTGTIC myofilaments (P<0.005; Figure 3C), consistent with increased incorporation of α-MHC, as described by James et al.
We observed changes in proteins involved in cellular energetics and stress responses among TGTIC animals (Figures 4 and S2). The majority of these proteins were from electron transport chain Complex I or Complex III. In Complex I, there was 37% less of the 75 kDa subunit, but 21% more of the 30 kDa subunit in TGSham rabbits compared to NTGSham rabbits. Following TIC, the level of the 75 kDa subunit decreased in NTGTIC rabbits but not in TGTIC rabbits, whereas the 30 kDa subunit increased in NTGTIC rabbits, but decreased in TGTIC rabbits. The 24 kDa subunit of Complex I was approximately 25% lower in NTGSham animals, but significantly reduced in TGTIC animals, whereas levels were unchanged in NTGTIC animals. There was 39% more core protein I in NTGSham rabbits compared to TGSham rabbits (P<0.005), but following TIC there was 45% more in TGTIC rabbits (P=0.04). We also observed changes in oxidoreductase protein 3-hydroxyisobutyrate dehydrogenase (3-HIBD) and the TCA cycle protein ATP-specific succinyl-CoA ligase (SCL) following TIC. There was an increase of 110% for 3-HIBD in TGSham rabbits compared to NTGSham, but there was no significant difference between TIC groups. While there was not a significant difference in ATP-specific SCL between TGSham and NTGSham rabbits, there was a 24% increase in TGTIC rabbits. The only significant change for GTP-specific SCL was in NTGTIC rabbits, with a 22% reduction after TIC.
The ER stress response proteins glucose regulated proteins 78 and 94 (GRP78/GRP94) also showed changes. There were higher levels of GRP 78 (20%; P<0.005) in TGsham rabbits (Figure 4) than NTGSham rabbits. There were higher levels of GRP 78 (49%; P=0.002), GRP 94 (58%; P=0.02) and HSP 71 (21%; P=0.02) in NTGTIC versus TGTIC rabbits. HSP60 levels were only significantly different in NTGTIC animals by 1-way ANOVA and underwent a 26% reduction (P<0.001). Although not different between NTG and TG rabbits, a common increase in transglutaminase 2 was also observed following TIC. The cytosolic antioxidant protein peroxiredoxin 6 (Prx6) was higher (24%; P<0.005) in the TGSham rabbits compared to NTGSham rabbits, with no detectable difference after TIC.
This study investigated the effects of MHC isoform changes in a transgenic rabbit model with cardiac isoform content reflective of the human heart. These rabbits possess increased α-MHC content from 5% (NTG levels) to 40% of the total MHC pool, but have similar hemodynamics to NTG rabbits under normal physiological conditions. TG rabbits showed significant cardioprotection against TIC, as they have less ventricular dilation and better systolic performance than NTG animals. We hypothesized that this was due to differential protein changes resultant from altered protein expression. Our observations demonstrate that TG rabbits possess a distinct protein expression pattern under basal conditions and furthermore exhibit a differential response to TIC.
The study design utilized 7cm 4-7 IPG strips enabling use of the large numbers of samples required to support our statistical interpretation of the data. This pI range is very robust for isoelectrical focusing and aided in data analysis, many proteins, including a majority of mitochondrial proteins are underrepresented.
Following TIC, MLC-1, MLC-2, and desmin levels increased in both NTG and TG animals compared to sham-operated animals within each group. TG animals had higher MLC-2 in comparison to NTG animals under either sham or TIC conditions, whereas no significant differences in MLC-1 were found when comparing TG to NTG animals in either sham or TIC conditions. Furthermore, differences in MLC-2 phosphorylation were observed between TG and NTG groups. Compared to NTGSHAM MLC-2 was less phosphorylated in TGSham rabbits; following TIC, TGTIC animals exhibited increased phosphorylation compared to NTGTIC animals. Phosphorylation of MLC-2 can increase myofilament force production  and this could be a mechanism by which TG rabbits are able to maintain myocardial contractile force. In contrast, dephosphorylated MLC-2 has been previously associated with human heart failure, and was shown to have a higher calcium sensitivity in terms of force production (pCA(50)). We expect that the effects of transgene expression on MLC-2 are divorced from MLC-1, as the latter was neither increased in abundance nor was its phosphorylation status altered with either TIC or transgene incorporation. Since the functional myosin complex is known to contain 2:2:2 molar ratios of MHC and the two light chains, these data imply that either there was a reduced quantity of MLC-1, which limited the amount of myosin complex formed, or that the non myofilament associated pool of MLC-2 is increased. We believe the latter, as purified myofibrils contained an equal stoichiometry of both MLC-1 and MLC-2 with respect to MHC (Figure 2). This suggests that in TIC, MLC2 has either increased steady state mRNA levels or decreased protein turnover rates. Careful regulation of UNC-45 abundance is likely important as over- or under-expression of UNC-45 results in a decrease in myosin levels.[21, 29] Landsverk and colleagues hypothesize that increased or decreased levels of UNC-45 prevent proper folding of myosin, resulting in increased myosin degradation via ubiquitinylation and degradation via the proteosome. We observed increases in UNC-45B levels in TGTIC animals compared with TGSham controls, suggesting an association with MHC expression. Given the protective effect of α-MHC among TGTIC animals, and the reported need for optimal UNC-45 expression for the proper function of myosin, it is not surprising that increased UNC-45 protein expression was observed in these animals.
Actin abundance changes were prominent in NTGTIC or TGTIC animals, with increased actin in TGTIC animals versus a decrease in NTGTIC animals. In 2DE we cannot differentiate between myofilament, soluble or cytoskeletal actin; however, these results were also observed for myofilament-associated actin. Regardless, there were no derangements in sarcomeric regularity (electron micrograph data not shown), so the impact of this change is difficult to interpret.
The intermediate filament protein desmin was increased in the total protein extracts from both NTGTIC and TGTIC rabbits, although the extent differed. Both human dilated cardiomyopathy (DCM) patients and a number of HF animal models have demonstrated increased desmin compared to appropriate controls.[30-32] Desmin increases are most likely a protective rather than a pathogenic response, as TG mice that overexpress desmin specifically in the heart are unaffected and have preserved cardiac function. Post-translational modifications of desmin may also be involved in this response, as TG rabbits show greater amounts of dephosphorylated desmin in total protein extracts (Figure S2). In contrast, equal amounts of desmin were present in isolated myofibrils from both NTGTIC and TGTIC rabbits (Figure 3A and B), suggesting an increased pool of non-sarcomere associated protein with maintenance of the myofilament-associated desmin. In TG mice that over-express desmin, the protein shows normal incorporation and is localized to the z-disc, whereas in DCM patients, desmin is disorganized and distributed amorphously in the cytoplasm.[31, 34] We also observed a reduction in the abundance of the dephosphorylated form of desmin in TGSham and TGTIC animals. This form of desimin has recently been reported by Agnetti and colleagues to be protective in a canine HF model and in human HF.
As expected, rates of ATP hydrolysis were dramatically increased in myofibrils isolated from TG animals versus NTG animals; therefore, it was not surprising that there was differential expression of proteins involved in energy generation between these groups. Complex I 75 kDa subunit and CP1 levels decreased in NTGTIC animals, whereas the 30 kDa and the 24 kDa Complex 1 subunits decreased in the TGTIC rabbits. The oxidoreductase protein 3-HIBD was increased 2.1 fold higher in TGSham animals as compared to NTGSham animals. Thus, while proteins involved in maintaining the ATP pool may be reduced at a time when higher levels of ATP should be necessary, the base increase in 3-HIBD may offset the decreases under pacing conditions. This could result from the communication/integral response between the myofilament requirement for ATP and the mitochondria. Since 3-HIBD is involved in the degradation of aliphatic amino acids, it is also possible that protein catabolism may be an additional source of energy for the TCA cycle by acting as a source of succinyl CoA in TG rabbits. It is also interesting to note that the differential abundances of Complex I and Complex III that occurred in the TGSham rabbits and the change in ATP-specific SCL under TIC have also been observed in rabbit cardiomyocytes preconditioned with adenosine.
The stress response proteins GRP 78, GRP 94 and HSP 71 responded differentially between NTG and TG rabbits. All increased during TIC in NTG animals but decreased in TIC in TG animals. GRP 78 and 94, molecular chaperones induced by glucose starvation, are involved in the unfolded protein response and may be involved in apoptosis regulation. Notably, increases in GRP78 are associated with HF.
Collectively these results suggest that modifications in the relative α-MHC levels results in a broad cellular impact on other members of the cellular proteome. These changes, which range from myofilament proteins to proteins in energetics and redox handling, demonstrate that despite extreme specificity in genetic manipulations, the final outcome on phenotype can be impacted by changes in several protein networks and systems. Therefore, it is very difficult to determine the singular effect of increased α-MHC on the function of the rabbit myocardium either at rest or following TIC outside the scope of other protein changes.
In vitro motility analysis of myosin isolated from end-stage HF patients demonstrated that incorporation of physiological levels of V1 myosin versus V3 myosin had a broad impact on other proteins. This is consistent with our results, in that many of the changes of increased V1 myosin in the cell may not be attributed to myosin expression alone, but rather reflects compensatory and/or auxiliary protein changes occurring in different areas of the cell. Our results also highlight the importance of determining the target sub-proteome of interest. For example, by 2DE there was discordance in the changes of the contractile proteins MHC, MLC-1, and MLC-2 in tissue lysates versus these same proteins in isolated myofibrils, suggesting that protein abundance changes for these proteins in the soluble pool did not carry over into assembled sarcomeres. Taken together, these results suggest that α-MHC interacts with multiple networks, including energetic proteins, contractile proteins and protein chaperones, which collectively confer a protective phenotype against tachycardia-induced heart failure.
The authors would like to acknowledge that LC-MS protein identification was performed in the Technical Implementation and Coordination Core (TICC) of the Johns Hopkins School of Medicine Mass Spectrometry and Proteomics Facility with support from the National Hearth Lung and Blood Institute Proteomics Center grant 1 N01-HV-28180 and NIH NHLBI Proteomics contract NO1-HV00240 (Van Eyk). The authors would like to thank Dr. Robert Cole and Roberto Diez for assistance with the 2DE-DIGE experiments and Bob O'Meally for help with protein identifications. These studies were also supported by NIH grants P01HL059408 (Robbins) and P02 HL69779 (Robbins) and by the Johns Hopkins University Dean's Office (startup funds to DG). Michael Linde (www.lindemedicalwriting.com) contributed to the organization and editing of this manuscript.
Conflict of Interest: The authors have no conflicts of interest to disclose.