Am J Physiol Heart Circ Physiol 304: H1382 H1396, First published March 11, 2013; doi: /ajpheart

Transcription

1 m J Physiol Heart Circ Physiol 34: H38 H396, 3. First published March, 3; doi:.5/ajpheart.63.. Low molecular weight fibroblast growth factor- signals via protein kinase C and myofibrillar proteins to protect against postischemic cardiac dysfunction Janet R. Manning, Sarah O. Perkins, Elizabeth. Sinclair, Xiaoqian Gao, Yu Zhang, Gilbert Newman, W. Glen Pyle, and Jo El J. Schultz Department of Pharmacology and Cell iophysics, University of Cincinnati College of Medicine, Cincinnati, Ohio; and Department of iomedical Sciences, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada Submitted 4 ugust ; accepted in final form 7 February 3 Manning JR, Perkins SO, Sinclair E, Gao X, Zhang Y, Newman G, Pyle WG, Schultz JJ. Low molecular weight fibroblast growth factor- signals via protein kinase C and myofibrillar proteins to protect against postischemic cardiac dysfunction. m J Physiol Heart Circ Physiol 34: H38 H396, 3. First published March, 3; doi:.5/ajpheart.63.. mong its many biological roles, fibroblast growth factor- (FGF) acutely protects the heart from dysfunction associated with ischemia/reperfusion (I/R) injury. Our laboratory has demonstrated that this is due to the activity of the low molecular weight (LMW) isoform of FGF and that FGF-mediated cardioprotection relies on the activity of protein kinase C (); however, which isoforms are responsible for LMW FGF-mediated cardioprotection, and their downstream targets, remain to be elucidated. To identify the pathway(s) that contributes to postischemic cardiac recovery by LMW FGF, mouse hearts expressing only LMW FGF () were bred to mouse hearts not expressing ( KO) or subjected to a selective ε inhibitor (εv ) before and during I/R. Hearts only expressing LMW FGF showed significantly improved postischemic recovery of cardiac function following I/R (P.5), which was significantly abrogated in the absence of (P.5) or presence of ε inhibition (P.5). Hearts only expressing LMW FGF demonstrated differences in actomyosin TPase activity as well as increases in the phosphorylation of troponin I and T during I/R compared with wild-type hearts; several of these effects were dependent on activity. This evidence indicates that both and ε play a role in LMW FGF-mediated protection from cardiac dysfunction and that signaling to the contractile apparatus is a key step in the mechanism of LMW FGF-mediated protection against myocardial dysfunction. low molecular weight FGF; and ε; troponin; actomyosin TPase; postischemic cardiac function CRDIOVSCULR DISESE IS the number one killer in the U.S., and every year over eight million people in the U.S. suffer from myocardial infarction (43). Often, myocardial ischemia is characterized by lowered postischemic left ventricular function, which is associated with increased morbidity and mortality (7). There is a profound and immediate need to address the problem of myocardial dysfunction and infarction associated with cardiac ischemia and reperfusion (I/R). One promising therapeutic strategy to alleviate I/R-induced cardiac injury is fibroblast growth factor- (FGF). FGF, found at elevated levels in the serum of I/R patients (63), has both angiogenic and acute cardioprotective effects (6, 6, 37, 54, 64). This makes FGF an outstanding candidate for ddress for reprint requests and other correspondence: J. J. Schultz, Dept. of Pharmacology and Cell iophysics, Univ. of Cincinnati, College of Medicine, 3 lbert Sabin Way, ML 575, Cincinnati, OH 4567 ( the treatment of cardiac patients; however, while in recent years, there have been a number of noteworthy developments characterizing the mechanism of action of FGF in the ischemic heart, there are several unanswered questions that must be addressed before FGF becomes a therapeutic molecule. Yet to be determined is the distinct role each FGF isoform plays in the ischemic heart. FGF includes two classes of protein isoforms, the low molecular weight (LMW) and the high molecular weight (HMW; Refs. 8, ). LMW FGF is 8 kda in both mouse and human and is found primarily in the cytosol of the cell and extracellular fluid as well as nucleus (, 39, 4). HMW FGF differs from LMW in that it includes a nuclear localization signal and encompasses two isoforms in mouse ( kda) and four in humans ( 34 kda) (, 4). Our laboratory, using gain-of-function and loss-of-function genetic mouse models, has demonstrated that all endogenous FGF isoforms are necessary for protection against myocardial infarction (4, 4), whereas HMW and LMW FGF isoforms play opposing roles in the recovery of the heart from I/R injury (4, 4). Studies using genetically modified mice with only LMW FGF expression (), or wild-type mouse hearts with LMW FGF exogenously added, have shown that this isoform is protective against postischemic dysfunction (3, 4, 4), while hearts with only the HMW isoforms (LMWKO), or with HMW FGF overexpressed, have significantly lower postischemic cardiac function (4). However, the pathways involved in signal transduction for each isoform subtype during I/R-induced cardiac dysfunction have not yet been elucidated; this information is critical for understanding the actions of FGF in the postischemic heart. Our laboratory has previously identified protein kinase C () as necessary for mediating the protection against cardiac dysfunction seen in mouse hearts overexpressing all LMW and HMW isoforms of FGF (7), which suggests that is critical to the signaling mechanisms of LMW and HMW FGF effects on the heart during I/R. mong the isoforms that are activated by overexpression of FGF in the mouse heart that may also phosphorylate targets modulating cardiac function are ε and (, 5, 9, 7,, 7, 33, 34, 46, 53). ε has been shown to localize to the cross-striated structures when activated (7, 7, 5) and is known to phosphorylate troponin I (TnI) and troponin T (TnT), as well as myosin light chain and myosin binding protein C (MyP-C), which modulates calcium sensitivity of the myofilaments in the heart (34). Similarly, is known to translocate to myofibrillar structures (67) and can modulate the function of TnT, TnI, and MyP-C (34). Finally, ε and have been shown to be necessary for ischemic preconditioning (3, 7), and has been H /3 Copyright 3 the merican Physiological Society

2 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION H383 FGFR CSQ HMW FGF LMW FGF FGF HMW LMW FGFR protein expression normalized to calsequestrin Fig.. Protein expression levels of fibroblast growth factor- (FGF) isoforms, isoforms and FGF receptor (FGFR) in wild-type (), high molecular weight knockout (), KO, and x KO hearts. : representative immunoblot of FGF and isoform protein expression. oth and HMW FGF expression is ablated in x KO hearts, with no alteration in the expression of ε or or low molecular weight (LMW) FGF. : representative immunoblot of FGFR protein expression normalized to calsequestrin (CSQ). There is no significant difference in FGFR expression in all groups similar to that previously published by our laboratory (4); n 6 hearts per genotype. shown to be activated in the heart associated with sevofluorane- and sildenafil-induced cardioprotection (6, 3). However, these kinase cascades are complex and stimulus dependent (, 9), which is why it is essential to delineate the isoforms of, as well as their downstream targets, that mediate LMW FGF s protective actions. ased on our previous findings that demonstrated that the individual FGF isoforms have opposing action on postisch- Treatment (drug or vehicle) Treatment (drug or vehicle) CE CE CE CE CE Heart snap frozen for for activation and myofibril protein level and activity aseline Ischemia Reperfusion Treatment (drug or vehicle) Treatment (drug or vehicle) aseline Ischemia Reperfusion aseline Treatment (drug or vehicle) Ischemia Heart snap frozen for activation and myofibril protein level and activity Treatment (drug or vehicle) aseline Fig.. Isolated working heart protocols for analysis of cardiac function (arrowhead) and lactate dehydrogenase (LDH) release (line; ) and for examination of the activity and phosphorylation of myofibrillar proteins () at baseline and during the course of I/R injury. CE, coronary effluent collection for LDH levels. JP-Heart Circ Physiol doi:.5/ajpheart.63.

3 H384 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION emic cardiac function but not on infarct size (4, 4), the current study focuses on the cardiac function effects modulated by LMW FGF during acute I/R injury. Here, our laboratory seeks to elucidate, for the first time, a mechanism by which endogenously expressed LMW FGF protects the heart, identifying the isoforms of that are directly responsible and determining their role in modulating myofibrillar protein function. Our study reveals that and -ε are selectively activated and translocate to the myofibril at early ischemia or reperfusion, respectively, and that ablation of and inhibition of ε abrogate the protective effects of LMW FGF against postischemic cardiac dysfunction. These data also demonstrate that expression of only LMW FGF results in differential phosphorylation of troponin I and T and alteration of actomyosin MgTPase activity during I/R, which is dependent on and suggest that the improvement in postischemic contractile function mediated by LMW FGF is due to direct targeting of the and ε to the myofibril. METHODS Mouse Models Generation and characterization of HMW FGF knockout () mice, which only express LMW FGF, on a SV 9/lack Swiss mixed background, have been previously described (3, 4). To test the role of, mice were bred to mice lacking expression of ( KO) on a mixed background (SV 9/lack Swiss/ FVN), which has been previously described (9). Mice that did not express either the HMW isoforms of FGF or (x- KO) were found to have no changes in the expression of other isoforms or LMW FGF compared with hearts (Fig. ). Similarly, there were no differences in fibroblast growth factor receptor (FGFR) expression, the predominant FGFR in mouse heart, in these genotypes (Fig. ) and as previously demonstrated by our laboratory (4). ll mice were randomly assigned to studies and compared with their wild-type littermates (n 3 6 per study/ genotype). ll animal protocols were submitted to and approved by the University of Cincinnati Institutional nimal Care and Use Committee. Mice were housed in pathogen-free housing according to standard use protocols, animal welfare regulations, and the National Institutes of Health Guide for the Care and Use of Laboratory nimals (ethesda, MD). Exclusion Criteria In the I/R studies, aortic leak (pressure 6 mmhg on retrograde perfusion), pulmonary vein leak (aortic flow of. ml/min and a perfusate gas PO 38 mmhg, with low atrial pressure 4 mmhg), or a visible leak in the heart itself (i.e., hole in ventricle or atrium) was the basis for exclusion, and a total of 7 mice were excluded. ortic leak was represented as an aortic pressure 6 mmhg on Langendorff, retrograde perfusion mode. Pulmonary vein leak was demonstrated as an aortic flow. ml/min, low ( 4 mmhg) atrial pressure, and a perfusate gas PO 38 mmhg or a visible leak (i.e., hole in ventricle or atrium) in the heart. I/R Injury Ten- to twelve-week-old, sex-matched mice were anesthetized with pentobarbital sodium (8 mg/kg ip), and anesthesia was confirmed by a lack of response to painful stimulus (toe pinch). Hearts were rapidly excised, the aorta and pulmonary vein were rapidly cannulated to induce a working-heart system with a cardiac output of 5 ml/min and a constant afterload of 5 mmhg, and establishment of baseline C D Phosphorylated/Total alpha (arbitrary units) Phosphorylated/Total epsilon (arbitrary units) p Total Phosphorylated/Total delta (arbitrary units) p Total p Total aseline 5 I 6 I + 5 R 6 I + R HMW HMW HMW KO KO KO Wildtype 5 I 6 I + 5 R 6 I + R Wildtype p<.5 vs. Wildtype p<.5 vs. 5 I 6 I + 5 R 6 I + R 5 I 6 I + 5 R 6 I + R HMW KO Fig. 3. Phosphorylation () of isoforms during ischemia and reperfusion as a measure of kinase activation for (), ε (C), and (D) in wild-type hearts (}, solid line) and (, solid line). Phosphorylation is normalized to total amounts of each kinase and given as arbitrary values (n 5 8 per group). ll samples were derived at the same time and processed in parallel. I, ischemia; R, reperfusion. P.5 vs. wild-type cohorts. JP-Heart Circ Physiol doi:.5/ajpheart.63.

4 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION H385 cardiac and hemodynamic parameters of cardiac output, aortic pressure, aortic and coronary flow, left ventricular systolic and diastolic pressure, and contractility ( dp/dt, positive derivative of the left ventricular pressure with respect to time), and relaxation ( dp/dt, negative derivative of the left ventricular pressure with respect to time) occurred. Mouse hearts were subjected to global low-flow I/R injury as previously described (6). Ischemia was induced by reducing total cardiac output from 5 to ml/min for h resulting in a 9% reduction in coronary flow, after which the heart was reperfused at 5 ml/min for h. Postischemic recovery of contractility and relaxation was depicted as a percentage of dp/dt at -h reperfusion normalized to baseline dp/dt. n ex vivo model has been chosen for these experiments to evaluate cardiac muscle function in the absence of neural or humoral influences. Following I/R, hearts were subsequently frozen and analyzed for phosphorylation of downstream targets in LMW FGF- mediated cardioprotection. Lactate dehydrogenase (LDH) levels released into coronary effluent were determined at designated time points of baseline, ischemia, and reperfusion (Fig. ). Protease inhibitor tablets (Roche) were added to the collected coronary effluent. The amount of LDH release was determined using the Promega CytoTox 96 Cytotox- +dp/dt (mmhg/s) Wildtype KO x KO % Contractile Recovery (+dp/dt) alpha present Wildtype p<.5 vs. p<.5 vs. absent aseline Ischemia (minutes) e inh e inh Vehicle Wt % Contractile Recovery (+dp/dt) Untreated Reperfusion (minutes) Vehicle nm Wildtype p<.5 vs. p<.5 vs. inhibitor nm Fig. 4. Recovery of contractile function as measured by dp/dt in wild-type hearts (Œ and inset: black bar) and hearts expressing only LMW FGF (, o and for inset: white bar) in the absence (, ) and presence of expression () and in the absence and presence ( and Œ for wild-type and, respectively) of a selective inhibitor of ε (). Insets: percent (%) recovery of contractility ( dp/dt) is normalized to baseline dp/dt (n 5 per group). P.5 vs. wild-type cohorts. P.5 vs. hearts with normal expression of or treated with TT vehicle. +dp/dt (mmhg/s) Vehicle 5 aseline Ischemia (minutes) Reperfusion (minutes) JP-Heart Circ Physiol doi:.5/ajpheart.63.

5 H386 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION icity colorimetric assay per the manufacturers instructions and normalized to heart weight and coronary flow. Pharmacological gents To examine the role of ε, TT-conjugated translocation inhibitor εv and the TT peptide (vehicle) were dissolved in water and administered to the heart via circulation of the Kreb s perfusate buffer 5 min before ischemia to 5 min after the onset of ischemia, as well as 5 min before reperfusion to 5 min after the onset of reperfusion (Fig. ). concentration response curve ( nm- M) was performed to identify the maximal efficacy of inhibitor without adverse effects. concentration of nm was found to effectively block the translocation of ε, while not affecting the activation of isoforms,, and (data not shown). Cardiac Preparation and Immunoblotting for Detection of FGF Snap-frozen nonischemic hearts were powdered and homogenized in homogenization buffer ( mm Tris, mm EDT, M NaCl, % NP4, and PMSF), and FGF was extracted as previously described (6). Western blot analysis of the extracted FGF was performed with a rabbit polyclonal antibody to FGF (:,; Santa Cruz iotechnology). Levels of FGF protein isoforms were visualized by chemiluminescence (mersham ioscience). Densitometry of protein bands was performed using a Fluorchem 88 gel imager (lpha Innotech). Whole Heart Preparation to Detect FGFR ctivation Snap-frozen nonischemic hearts (wild type) were powdered and homogenized in homogenization buffer (5 mm HEPES, ph 7.5, % glycerol, 5 mm NaCl, % Triton X-, 5 mm EDT, mm sodium orthovandate, 5 mm -glycerolphosphate, 5 mm sodium fluoride solid,.5 M okadaic acid, diisopropylfluorophosphate, M calpain inhibitor, Pefabloc stock and, and Sigma phosphatase inhibitor) as previously described (3, 4, 4). The homogenate was centrifuged at 3, g for 5 min, and the supernatant was collected. Protein concentration was determined via io-rad Lowry protein assay (io-rad, Hercules, C). Western Immunoblotting for FGFR Expression One hundred micrograms of whole cell homogenate were loaded onto an 8% SDS-PGE gel and then transferred onto. m nitrocellulose membrane. Transfer efficiency and loading equality were examined by staining the membrane with.% Ponceau S in 5% acetic acid. The membrane was blocked in 5% dry milk and then incubated with primary antibody against total FGFR (:5; Cell Signaling) followed by incubation with horseradish peroxidase-conjugated secondary mouse (:,; io-rad) antibody. FGFR expression was visualized by ECL, and densitometry of protein bands was quantitated using a Fluorchem 88 gel imager and normalized to calsequestrin [primary antibody dilution :,5 (Thermo Scientific), and horseradish peroxidase-conjugated secondary goat anti-rabbit antibody dilution :5, (Santa Cruz iotechnology)]. Phosphorylation and Translocation to the Myofibril ctivation of isoforms by phosphorylation or translocation to the myofibril was measured as previously described (68). Hearts were snap frozen at key time points during I/R (see Fig. ). activation in the whole heart samples or myofilaments was analyzed with immunoblotting, as described below. The time points were chosen to reflect times when,, and ε have been shown by our laboratory to be activated during I/R (6). phosphorylation via immunoblotting (whole heart). Hearts were homogenized in a Tris buffer containing a nonionic detergent (Tween-) with protease and phosphatase inhibitors and analyzed with SDS-PGE electrophoresis using % separating gels. Nitrocellulose membranes were probed for antibodies for phosphorylated ε,, and (:5; Santa Cruz) and then stripped and reprobed with antibodies against the total form of the kinase (:5; Santa Cruz). ctivation is indicated as a ratio of phosphorylated kinase to total kinase. translocation via immunoblotting (myofilament). Myofilament proteins (75 g) were resolved by SDS-PGE electrophoresis using % separating gels. Proteins were transferred to nitrocellulose membranes and probed with antibodies for,, and ε (D iosciences, Mississauga, ON, Canada; and Upstate, Mississauga, ON, Canada). Protein loading was assessed by probing with an anti-actin antibody (Millipore, Etobicoke, ON, Canada). Secondary antibodies were conjugated to horseradish peroxidase. Primary antibodies were used at :, dilution, except for anti-actin, which was :5,. Secondary antibodies were used at :5, dilution. bands were detected using Western Lightning (PerkinElmer Life and nalytical Sciences, Woodbridge, ON, Canada), and analysis of band density was done using ImageJ software. ll bands were normalized to their respective controls and are expressed as a percent change in Table. Preischemic and postischemic functional parameters for and hearts in the presence or absence of expression HR, beats/min MP, mmhg LVP, mmhg LVEDP, mmhg dp/dt, mmhg/s dp/dt, mmhg/s TPP, ms/mmhg RT /, ms/mmhg Preischemia , , , , KO ,89 9 3, x KO ,4 54 3, min Ischemia KO x KO min Ischemia -min reperfusion ,59 678, ,8 4, KO ,53 485, x KO ,437 3, ll values are means SE; n 6 per genotype and treatment. HR, heart rate; MP, mean aortic pressure; LVP, peak systolic left ventricular pressure; LVEDP, left ventricular end-diastolic pressure; dp/dt, rate of contractility; dp/dt, rate of relaxation; TPP, time-to-peak left ventricular pressure normalized to contractility; RT /, half relaxation time normalized to relaxation; LP, left atrial pressure;, wild type; HMW, high molecular weight; KO, knockout. P.5 vs. preischemic cohort. P.5 vs. cohort. P.5 vs. cohort. LP, mmhg JP-Heart Circ Physiol doi:.5/ajpheart.63.

6 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION Table. Lactate dehydrogenase release for and hearts in the presence or absence of expression H387 aseline Ischemia Early Reperfusion Late Reperfusion KO x KO ll values depicted as means SE of n independent experiments; n 6 per genotype or treatment. LDH, lactate dehydrogenase release (in U l g min ). Early reperfusion: 4 min of reperfusion. Late reperfusion: 4 min of reperfusion. P.5 vs. baseline. band density. s the immunoblots were conducted using myofilament isolations, sarcomeric -actin was used as a loading control. Myofilament Isolation Cardiac myofilaments were prepared as previously described (33). Hearts collected at specific time points of I/R (see Fig. ) were homogenized in ice-cold standard buffer [6 mm KCl, mm MgCl, 3 mm imidazole (ph 7.),. mm PMSF,. mm leupeptin, and. mm pepstatin ]. fter centrifugation (, g for 5 min at 4 C), the pellets were resuspended in ice-cold skinning buffer [% Triton X-, mm EGT, 8. mm MgCl, 4.4 mm KCl, 6 mm imidazole (ph 7.), 5.5 mm TP, mm creatine phosphate, U/ml creatine phosphokinase,. mm PMSF,. mm leupeptin, and. mm pepstatin ] and gently mixed for 45 min at 4 C. The suspension was centrifuged (, g for 5 min at 4 C), and the pellets were washed three times in cold standard buffer. Protein concentration was measured using a radford assay. ctomyosin MgTPase nalysis Myofilaments isolated from snap-frozen hearts (see Fig. ) were subjected to an actomyosin MgTPase activity as previously described (68). Purified myofilaments (5 g) were incubated in activating buffers containing various amounts of free Ca at 3 C for 5 min. ctivating buffers were made by mixing maximally activating and relaxing TPase buffers. Maximally activating TPase buffer contained 3.5 mm KCl, 5 mm MgCl, 3. mm TP, mm EGT, mm imidazole,. mm CaCl,. mm PMSF,. mm leupeptin, and. mm pepstatin (ph 7.). The relaxing TPase buffer contained 6 mm KCl, 5. mm MgCl, 3. mm TP, mm EGT, mm imidazole, 4.9 M CaCl,. mm PMSF,. mm leupeptin, and. mm pepstatin (ph 7.). Free Ca was calculated using the program of Patton et al. (48). Reactions were quenched with equal volumes of % trichloroacetic acid. The samples were centrifuged (4, g for 3 min), and the supernatant was removed for phosphate analysis. The amount of inorganic phosphate released was determined colorimetrically by adding an equal amount of developing solution (.5% FeSO 4-.5% ammonium molybdate in.5 M H SO 4) to the supernatant and measuring absorbance at 63 nm. Myofibrillar Protein Phosphorylation nalysis Myofilament proteins (4 g) were separated on.5% SDS- PGE gels. Gels were fixed in 5% methanol-% acetic acid at room temperature overnight, and phosphorylated proteins were stained with Pro-Q Diamond (Molecular Probes, Eugene, OR), according to the manufacturer s instructions. Imaging was done using a Typhoon gel scanner (GE Healthcare, aie d Urfé, QC, Canada), and analysis was done using ImageJ software (National Institutes of Health). Protein loading was assessed by Coomassie staining of gels after imaging. Statistical nalysis Values are presented as means SE of n independent experiments. The number of individuals for each study was determined by power analysis referencing previous studies that made use of each method. One-way NOV was performed for the postischemic functional recovery of hearts before and after I/R injury, followed by a post hoc Student s t-test. One- or two-way NOV was used to analyze the significance of the results among three or more independent groups within groups of untreated samples (one-way) or the presence or absence of treatments for evaluation of time course data (two-way). Probabilities of.5 or less are considered statistically significant. Table 3. Preischemic and postischemic functional parameters for vehicle-treated and inhibitor-treated and hearts HR, beats/min MP, mmhg LVP, mmhg LVEDP, mmhg dp/dt, mmhg/s dp/dt, mmhg/s TPP, ms/mmhg RT /, ms/mmhg LP, mmhg Preischemia Vehicle (TT)-treated , , Vehicle (TT)-treated ,358 3, ε inhibitor-treated , , ε inhibitor-treated ,36 3, min Ischemia Vehicle (TT)-treated Vehicle (TT)-treated ε inhibitor-treated ε inhibitor-treated min Ischemia -min reperfusion Vehicle (TT)-treated ,46 98, Vehicle (TT)-treated ,88 83, ε inhibitor-treated ,46 36, ε inhibitor-treated ,7 333, ll values are means SE; n 8 per genotype and treatment. P.5 vs. preischemic cohort. P.5 vs. cohort. P.5 vs. cohort. JP-Heart Circ Physiol doi:.5/ajpheart.63.

7 H388 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION Table 4. LDH release for vehicle-treated and inhibitor-treated and hearts aseline Ischemia Early Reperfusion Late Reperfusion Vehicle (TT)-treated Vehicle (TT)-treated ε inhibitor-treated ε inhibitor-treated ll values depicted as means SE of n independent experiments; n 8 per genotype or treatment. Early reperfusion: 4 min of reperfusion. Late reperfusion: 4 min of reperfusion. P.5 vs. baseline. RESULTS Expression of only LMW FGF results in differential activation of isoforms during I/R injury. Previous studies from our laboratory have established that is required for cardioprotection induced by overexpression of all isoforms of FGF (6), but it is unclear if this is also the case for the improvement in postischemic function seen when only LMW FGF is expressed. To evaluate changes in activation at key time points during I/R, the phosphorylation of isoforms was examined in and wild-type hearts (Fig. ). Hearts expressing only the LMW isoform of FGF and wildtype hearts showed similar activation of,, and ε at baseline (Fig. 3). t the onset of ischemia, showed higher activation ( %) in hearts only expressing LMW FGF () compared with wild type; this degree of activation correlates to that seen in other forms of cardioprotection sufficient to significantly reduce I/R injury, both in terms of activation by phosphorylation (6) and translocation (49). Similarly, hearts showed higher activation of ε compared with wild type at early reperfusion (P.5). Improved postischemic functional recovery of LMW expressing hearts is ablated in the absence of or presence of ε inhibition. It was next determined whether these changes in the phosphorylation state of at early ischemia, and ε at early reperfusion (Fig. 3), play a role in LMW FGF-mediated protection from cardiac dysfunction. s previously seen (4), hearts showed a significantly higher postischemic recovery of cardiac function than wild type (Fig. 4, and ). reeding these mice with genetically modified mice that lack expression ( KO) resulted in a complete abrogation of this protection (P.5; Fig. 4). No difference in postischemic functional recovery was seen in hearts with ablation, but with HMW FGF expression intact ( KO) compared with wild-type cohorts, suggesting that has little influence on function following I/R injury in our models, without activation by LMW FGF. Similarly, while a TT peptide alone had no effect on either wild-type or hearts, a TT-conjugated ε inhibitor administered during I/R attenuated the recovery of dp/dt, but not dp/dt, of heart (P.5; Fig. 4). Other preischemic and postischemic functional parameters for normal expression and ablated or untreated and ε inhibitor-treated wild-type and hearts are shown in Tables,, 3, and 4, respectively. There were no differences in preischemic functional parameters among all genotypes or treatment groups. Following I/R injury, there was significant systolic and diastolic dysfunction as measured by left ventricular pressure, dp/dt and, dp/dt in mice expressing LMW FGF but not (xa; P.5; Table ). Similarly, inhibition of ε in mice expressing only LMW FGF resulted in postischemic systolic and diastolic dysfunction as represented by dp/dt and dp/dt (P.5; Table 3). Myocardial cell death was represented as LDH release measured from coronary effluent at designated time points of baseline, ischemia, and reperfusion (Fig. ). There was a significant increase in LDH release at early reperfusion compared with baseline values in all groups (P.5; Tables and 4), suggesting this is an irreversible I/R model as previously described by our laboratory (6). Yet, there was no difference in LDH release at early reperfusion between the groups. These findings confirm our previously published studies, suggesting that the primary protective effects of this FGF isoform lie in the prevention of contractile dysfunction, whereas both the LMW and HMW are involved in infarct reduction (4, 4). Cross talk between and ε in hearts does not occur during I/R injury. s both ε and are neces- Phospho/total (arbitrary units) Phospho/total (arbitrary units) p<.5 vs TT vehicle x KO + inhibitor present KO TT vehicle inhibited p p Fig. 5. Cross talk between and ε. Phosphorylation of ε in the absence of at early reperfusion () or phosphorylation of in the presence of ε inhibition (). Phosphorylated is normalized to total amounts of each kinase and given as arbitrary values (n 6 8 per group). ll samples were derived at the same time and processed in parallel. P.5 vs. hearts with present. JP-Heart Circ Physiol doi:.5/ajpheart.63.

8 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION H389 sary for the protection of the heart from I/R-mediated functional injury in mice, cross talk between the two was evaluated by measuring the activation of ε in the absence of and the activation of in the presence of ε inhibition. There was no decrease in ε activation in the absence of, and in fact there was a slight increase, indicating that does not activate ε further downstream (Fig. 5). Similarly, there was no change in the phosphorylation of in the presence of ε inhibition (Fig. 5), suggesting that the activation of is not dependent on the epsilon isoform. Translocation of isoforms to the myofibril is altered in LMW expressing hearts during I/R. s LMW FGF expression increases the function of the heart after I/R injury, the relevance of myofibrillar proteins in LMW FGF-mediated signaling during cardiac ischemia was investigated. Translocation of isoforms to the myofibrils was measured at key time points during I/R. In sham hearts, significant differences in the levels of, ε, and at the myofibrils were seen in the compared with wild type (Fig. 6), although no difference in the phosphorylation of these isoforms was seen at this time point (Fig. 3), suggesting that the localization to the myofibril but not phosphorylation of these s is altered. fter the onset of ischemia, myofilament levels increased and levels of decreased in hearts compared with wild-type cohorts. fter reperfusion, levels of ε at the myofibril increased. The timing of the translocation of and ε mirrors the phosphorylation of these isoforms (Fig. 3), suggesting that upon phosphorylation, and ε translocate to the myofilaments during I/R and may target proteins that are relevant to cardiac function. Phosphorylation of myofilament proteins is increased during I/R in LMW expressing hearts. Given that both and ε translocate to the myofibrils during I/R injury (Fig. 6), the phosphorylation state of the myofibrillar proteins troponin I and T was analyzed (Fig. 7), which can affect calcium binding and cross-bridge cycling of the myofibrils (59). This is total protein phosphorylation, which may be the result of changes in the activity of several kinases/phosphatases. Interestingly, the phosphorylation of both troponins increased significantly in hearts expressing LMW FGF after the onset of ischemia through early reperfusion (Fig. 7, and C). The initial increase in phosphorylation was seen after 5 min of ischemia, which corresponds to the same time of activation and translocation of, suggesting that this kinase may be responsible for these changes. To test this, the phosphorylation of troponin I and T at early ischemia was measured in the absence of and found to be significantly reduced in the hearts at the Myofilament associated (arbritrary units) 'I 5 I 6'I +5'R 6 I + 5 R 6 I + R p<.5 vs. p<.5 vs. 5 I 6 I + 5 R 6 I + R -actin Myofilament associated (arbritrary units) 'I 5 I 6'I +5'R 6 I + 5 R 6 I + R 5 I 6 I + 5 R 6 I + R -actin C Myofilament associated (arbritrary units) 'I 5 I 6'I +5'R 6 I + 5 R 6 I + R 5 I 6 I + 5 R 6 I + R Fig. 6. Myofilament-associated isoforms during I/R. (), ε (), and (C) in wild-type (black bar) and (white bar) hearts (n 3 6 per group). ll samples were derived at the same time and processed in parallel. P.5 vs. wild-type cohorts. P.5 vs. sham. -actin JP-Heart Circ Physiol doi:.5/ajpheart.63.

9 H39 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION TnT 5 I 6 I + 5 R 6 I + R 5 I TnI -actin Phosphoroylated troponin (arbritrary units) Troponin I p<.5 vs. 5'I 5 I 6'I +5'R 6 I + 5 R 6'I 'R 6 I + R.5.5 Troponin I 5 I DKO C Phosphoroylated troponin (arbritrary units) Troponin T 5'I 5 I 6'I +5'R 6 I + 5 R 6'I + 'R 6 I + R D Troponin T Fig. 7. Phosphorylation of troponin I (TnI; and ) and T (TnT; C and D) during ischemia and reperfusion in wild-type (black bar) and (gray bar) hearts expressing or not expressing (DKO, white bar; n 3 6 per group). ll samples were derived at the same time and processed in parallel. P.5 vs. wild-type cohorts. P.5 vs. sham. 5 I DKO onset of ischemia, back to wild-type levels (Fig. 7, and D). Phosphorylation of the troponins was also increased at early reperfusion, the point that corresponds to the time of ε activation (Fig. 3). Interestingly, ε inhibition with εv did not lower the phosphorylation state of the troponins (data not shown). Therefore, these data indicate that this isoform was not involved in the posttranslational modifications of troponin I or T at early reperfusion, when ε is activated and translocated to the myofilament. This does not rule out a role, however, for ε in posttranslational modification of other myofibrillar proteins. ctomyosin MgTPase activity is altered in hearts during I/R. To determine if LMW FGF-mediated activation of isoforms during I/R injury affects the cross-bridge cycling activity of the myofibrils, actomyosin MgTPase activity was evaluated (Fig. 8). The maximum activity of actomyosin MgTPase was elevated in hearts during reperfusion (Fig. 8), which was not abrogated by the absence of or inhibition of ε (Fig. 8 and Fig. 9, respectively). lternately, the EC 5 was raised in hearts compared with wild type at baseline, early ischemia, and early reperfusion, suggesting that calcium activates the myofibril with lowered potency; these changes in myofilament calcium sensitivity in hearts were ablated in the absence of at baseline and early ischemia, indicating that this decreased potency is dependent on (Fig. 8). However, inhibition of ε did not have an effect (Fig. 9). Overall, this suggests that in hearts only expressing LMW FGF, the presence of results in a decrease in myofilament calcium sensitivity. DISCUSSION Our laboratory has previously demonstrated opposing roles for HMW and LMW FGF in the functional recovery of the heart after I/R (4, 4). This study is the first to show that LMW FGF requires the activity of two isoforms of, ε and, which are differentially activated during I/R through distinct parallel pathways. dditionally, the expression of only LMW FGF alters the activity of the actomyosin MgTPase in ischemic hearts, corresponding to differences in the phosphorylation of troponin I and T. These results describe a novel mechanism of FGF-mediated protection from I/R injury through activation of protective isoforms, which then phosphorylate myofibril proteins to modulate postischemic cardiac function (Fig. ). s we have previously described using gain-of-function and loss-of-function models, expression of only LMW FGF improves postischemic cardiac function compared with wild-type and Fgf KO hearts. These protective effects are dependent on the activity of the FGFR (4), which can activate s via phospholipase C (45, 47) or other scaffolding proteins such as FRS/Grb (35), which also have the potential to interact with s (4). Here, evidence sug- JP-Heart Circ Physiol doi:.5/ajpheart.63.

10 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION H39 ctomyosin TPase activity (nm Pi/min/mg protein) 5 I 6 I + 5 R ctomyosin TPase activity (nm Pi/min/mg protein) C ctomyosin TPase activity (nm Pi/min/mg protein) D ctomyosin TPase activity (nm Pi/min/mg protein)c Free Calcium (µm) Maximum Efficacy 5'I 5'R 5 I 6 I + 5 R HKO DKO Free Calcium (µm) p<.5 vs. p<.5 vs. Free Calcium (µm) E Hill coefficient (arbitrary units) Hill Coefficient HKO DKO F µmol Calcium EC 5 HKO DKO!"#$ %&' %&( 5 I 6 I + 5 R 5 I 6 I + 5 R Fig. 8. ctomyosin MgTPase activity in wild-type ( ) and ( ) hearts (,, and C). ctivity at maximum calcium levels (D), Hill coefficient (E), and EC 5 (F) were measured in wild type (black bar) (gray bar) hearts expressing or not expressing (DKO, white bar; n 3 6 per group); ll samples were derived at the same time and processed in parallel. P.5 vs. wild-type cohorts. P.5 vs. hearts with present. gests that expression of only LMW FGF results in the differential activation of and ε to protect the heart from postischemic dysfunction. While our model identified as necessary for FGF- induced cardioprotection, early studies examining the role of in protection against I/R injury, particularly ischemic preconditioning, produced conflicting findings (for review, see Ref. 56). This phenomenon has, in part, been attributed to the different and sometimes opposing roles of multiple isoforms of that have been uncovered in recent years (for review, see Refs., 57). For example, while inhibition of all isoforms of with staurosporine fails to prevent ischemic preconditioning in porcine (65), activation of simply the ε-isoform of protects the porcine heart from arrhythmia and infarct development (8). It was later determined that the activation of various isoforms of may result in diverse and opposing effects during I/R, with acting as an injurious element and ε behaving as a protective molecule in infarct development during ischemic preconditioning (9). For this reason, a thorough and complete characterization of the signaling mechanisms of FGF must include a description of the specific isoforms that are activated. Therefore, investigation of the targets of and ε was driven by the observation that, while hearts that only express LMW FGF are protected from I/R-induced dysfunction, they are not similarly protected from infarct development and tissue death (4, 4) (and see Tables and 4). The data presented indicate that is activated to a higher degree during ischemia, while ε is activated after reperfusion in hearts only expressing LMW FGF compared with wild-type and Fgf KO [data not shown and previously published (7)] hearts. It should be noted that this differs from a mouse model overexpressing all isoforms of FGF, where the activity of is reduced during I/R (6), suggesting that this increase in activity is a unique result of only LMW FGF. Conversely, in both models, ε activity is increased at early reperfusion, suggesting that this isozyme may contribute to protection from both postischemic cardiac dysfunction and infarct development (seen when all isoforms of FGF are overexpressed). While some isoforms can activate one another (58) and cross talk between and ε has been shown to occur in ischemic preconditioning models of JP-Heart Circ Physiol doi:.5/ajpheart.63.

11 H39 MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION C ctomyosin TPase activity (nm Pi/min/mg protein) D Hill coefficient (arbitrary units) ctomyosin TPase activity (nm Pi/min/mg protein) Maximum Efficacy Free Calcium (µm) 6 5'R I + 5 R 6 I 5'R+ 5 R HKO e inh HKO e inh ctomyosin TPase activity (nm Pi/min/mg protein) 6 I + 5 R Free Calcium (µm) e inh Hill Coefficient E EC HKO e inh 5 HKO e inh.8 HKO e inh.6 µmol Calcium.4. p<.5 vs. HKO 6 I 5'R + 5 R Fig. 9. ctomyosin MgTPase activity in wild-type ( ) and ( ) hearts ( and ). ctivity at maximum calcium levels (C), Hill coefficient (D), and EC 5 (E) were measured in hearts in the presence of vehicle treatment (wild type, black bar;, light gray bar) or a ε inhibitor (wild type, dark gray bar;, white bar; n 3 6 per group). ll samples were derived at the same time and processed in parallel. P.5 vs. wild-type cohorts. cardioprotection (3), this does not appear to be the case here, as and ε are activated independently of one another (Fig. 5). Paradoxically, while it is demonstrated in this study that expression of results in higher contractility after I/R injury in hearts, it has been established that expression is associated with lowered contractility in nonischemic hearts (9). Furthermore, phosphorylation of myofibril targets by is maladaptive for heart failure and remodeling (5). This phenomenon illustrates the importance of the conditions under which isoforms are activated. was activated differentially by LMW FGF only after the onset of ischemia, an event that precipitates a unique set of intracellular conditions, characterized by acute loss of TP, and transient but substantial increases in intracellular acidity, and sodium and calcium concentrations, which leads to a distinct manner of injury associated with calcium overload (5). It is well characterized that isoforms can play varying roles under different circumstances in the heart; for example, is necessary for cardioprotection from I/R triggered by opioid agonists () and adenosine-mediated late preconditioning (36), while this same isoform is detrimental in conditions of ischemic preconditioning (9) and ethanol-induced protection from ischemia (3). Similarly, it is during the unique events of I/R injury in the presence of LMW FGF that the activation of protects the postischemic heart and results in improved cardiac function. We speculate that the reduced levels of myofilament seen in the sham mice represent a compensatory change in which myofilament inhibition by is reduced, allowing for a more efficient contractile state. Paradoxically, the increase in myofilament-associated following an acute ischemic challenge may also represent a compensatory change. Molnar and colleagues (46) found that maintains cardiac myofilament force development under conditions that occur during I/R. y contrast, Shintani- Ishida and Yoshida (55) report that post-i/r activation of at the sarcoplasmic reticulum exacerbates I/R damage, arguing for subcellular-specific effects of. We propose that the reduction in myofilament-associated allows the myofilament complex to function with less hindrance by JP-Heart Circ Physiol doi:.5/ajpheart.63.

12 FGFR Myofibril MYOFIRILLR FUNCTION ND /ε IN LMW FGF CRDIOPROTECTION LMW FGF Transient decrease in calcium sensitivity during ischemia Improved post-ischemic function Fig.. Schematic of the hypothesized role of and ε at the myofibril in LMW FGF-mediated protection from postischemic dysfunction. LMW FGF activates, which phosphorylates troponin I and T at the myofibril and transiently reduces calcium sensitivity (EC 5), to protect the heart during ischemia. but that transient activation of myofilament-associated during I/R provides a cardioprotective advantage, possibly through an energy sparing mechanism (5). This hypothetical model incorporates the importance of both timing and subcellular localization of activation, as has been advanced elsewhere (4, 69). The time-dependent changes in isoform translocation to the myofilaments are interesting and suggestive of a dynamic system during I/R. lthough this represents the first report showing time-dependent changes in myofilament-associated isoforms, earlier work by Yoshida and investigators (7) reported similarly dynamic alterations in membrane-associated isoforms. The significance of these changes is difficult to assess, as it would require inhibiting not only the specific isoforms, but at certain times and within the myofilament compartment, to determine the significance of the specific changes. t the moment, this represents an untenable technical challenge and is beyond the scope of the current paper. The impact of s on cardiac functional recovery after I/R is manifested by changes in myofibrillar proteins. s shown here for the first time, alterations in both actomyosin MgTPase activity, and troponin I and T phosphorylation in hearts only expressing LMW FGF are dependent on the presence of. Troponin I and T have several phosphorylation sites that regulate the calcium affinity of the thin filaments and contractile force (59). Phosphorylation of the known sites on troponin T produce a reduction in contractile force and calcium sensitivity (6, 6), while the -targeted residues in troponin I result in a more complex response. Phosphorylation of S3/4, a PK and target, increases lusitropy and inotropy, while the targets T44 and S43/45 reduce calcium binding cooperativity and contractile force, respectively (34, 38, 46, 6). Unfortunately, the specific residues that contribute to increased JP-Heart Circ Physiol doi:.5/ajpheart H393 phosphorylation of TnI remain unknown until the availability of reliable and specific phospho-antibodies. The molecular basis of myocardial dysfunction has been under investigation for a significant period of time, but a singular, definitive mechanism has yet to be identified. mong the most well-supported theories are the oxyradical and calcium hypotheses (8, 4). lthough each model has component elements that are unique, both hold cardiac myofilaments as central players in the development of the contractile dysfunction that underlies stunning (4). MacGowan and group (44) support a key role for myofilament protein phosphorylation in the response to acute myocardial ischemia by demonstrating that alterations in troponin I phosphorylation alter the susceptibility of hearts to ischemic injury. However, a recent study by vner and colleagues () shows that myofilament protein oxidation with ischemia mediates a similar reduction in myofilament calcium sensitivity. In chronic models of heart failure, protein oxidation has been linked to functional impairment (), suggesting that a prolonged reduction in myofilament calcium sensitivity is not beneficial. Together, these studies suggest that acute changes in the phosphorylation of myofilaments to temporarily decrease myofilament calcium sensitivity confer a protective effect, but that long-term damage to cardiac myofilaments by oxidation is associated with impaired function. Phosphorylation of regulatory proteins, including troponin I and T, myosin light chain, and myosin-binding protein C, have been implicated in myofilament calcium sensitivity, which subsequently affects force development (5, 34, 44, 6, 6, 66). In the studies presented here, increased troponin phosphorylation during I/R mirrors an increase in the EC 5 of actomyosin MgTPase, suggesting that LMW FGF activation of results in a lowered potency of calcium-mediated activation of the TPase. This is particularly important during ischemia, where calcium overload is a primary effector of the cardiac dysfunction seen during late reperfusion; by requiring a higher calcium concentration to achieve the same level of TPase activity, myofilament activity and TP consumption are reduced, and the damage caused by calcium overload is slowed. Thus a temporary increase in EC 5 during ischemia results in a mitigation of calcium-induced injury and stunning. This -dependent transient depression of actomyosin MgTPase calcium sensitivity has previously been demonstrated to result in cardioprotection (5). The effects of on actomyosin MgTPase activity are complex, dependent on the site of phosphorylation and the conditions of activation. urkart et al. () showed that mutating S43 and S45 in cardiac troponin I to E43 and E45 to mimic phosphorylation decreased maximum actomyosin MgTPase activity. y contrast, a similar mutation at T44 had no effect on maximum myofilament TP consumption, demonstrating that the -dependent depression of maximum actomyosin MgTPase activity is specific to T44 phosphorylation. More recently, Engel and investigators (9) found that myofilament regulation via -preferred sites is altered under conditions that occur during prolonged ischemia. Thus, while physiological and biochemical investigations of dependent regulation of myofilament function present a sound basis from which to build an understanding of myocardial signaling, understanding the role of this molecular messenger