|Home | About | Journals | Submit | Contact Us | Français|
The giant protein titin plays key roles in myofilament assembly and determines the passive mechanical properties of the sarcomere. The cardiac titin molecule has two mayor elastic elements, the N2B and the PEVK region. Both have been suggested to determine the elastic properties of the heart with loss of function data only available for the N2B region.
Investigate the contribution of titin’s PEVK region to biomechanics and growth of the heart.
We removed a portion of the PEVK segment (exons 219–225; 282aa) that corresponds to the PEVK element of N2B titin, the main cardiac titin isoform. Adult homozygous PEVK knockout (KO) mice developed diastolic dysfunction, as determined by pressure-volume loops, echocardiography, isolated heart experiments, and muscle mechanics. Immunoelectron microscopy revealed increased strain of the N2B element, a spring region retained in the PEVK-KO. Interestingly, the PEVK-KO mice had hypertrophied hearts with an induction of the hypertrophy and fetal gene response that includes upregulation of FHL proteins. This contrasts the cardiac atrophy phenotype with decreased FHL2 levels that result from the deletion of the N2B element.
Titin’s PEVK region contributes to the elastic properties of the cardiac ventricle. Our findings are consistent with a model in which strain of the N2B spring element and expression of FHL proteins trigger cardiac hypertrophy. These novel findings provide a molecular basis for the future differential therapy of isolated diastolic dysfunction versus more complex cardiomyopathies.
Titin is the largest protein in mammals and forms a continuous elastic filament along the myofibril (reviewed in 1). Due to its enormous size, titin is a prominent target for mutations that give rise to diseases such as familial dilated cardiomyopathy and muscular dystrophy 2, 3. Titin’s extensible region resides in the I-band of the sarcomere and consists of immunoglobulin (Ig)-like domains arranged in tandem, the heart specific N2B element, and the PEVK element 4. The PEVK element is thought to function as a largely unfolded polypeptide that extends at low force levels and that thereby provides an important source of elasticity at physiological sarcomere lengths 5–7. Unlike the one-exon heart specific N2B element, the titin gene contains 112 PEVK exons that are differentially expressed between muscle types 8. Of these PEVK exons 219–225 are expressed in the so-called N2B titin isoform, that constitutes the dominant cardiac isoform in the left ventricle of a wide range of species, including rodents and human 9.
Here we generated a mouse deficient in titin’s exons 219 – 225 that results in a deletion of the c-terminal PEVK region (282 aa), and determined its role in cardiac function using echocardiography, in vivo pressure-volume loops, isolated heart physiology, muscle mechanics, immuno-electron microscopy, and expression analysis. We investigated the hypertrophy phenotype and studied members of the four-and-a-half LIM family involved in atrophy/hypertrophy signaling - FHL1 and FHL2 10, 11. Our results reveal the strong effect of the PEVK element on diastolic function but also that the role of the PEVK extends beyond that of a mechanical spring including a novel role in hypertrophy signaling.
A targeting construct was assembled to replace exons 219–225 with a FRT flanked neomycin expression cassette, which was subsequently removed using the Flp deleter stain 12. Details on generation and genotyping of KO mice are provided as supporting information.
Animals were maintained on a mixed C57Bl/6 × 129S6 genetic background. Sex and age matched animals (4 months) were used for experiments. All experiments involving animals were carried out following institutional and NIH guidelines, “Using Animals in Intramural Research”.
For data acquisition with the Acuson Sequoia C512 Echocardiography imaging system and the 15 MHz linear transducer and functional calculations we followed American Society of Echocardiography guidelines. We used a micro-conductance pressure catheter (ARIA SPR-719; Millar-Instruments, Inc., Texas, USA) for continuous registration of LV pressure/volume (PV) loops in an open-chest model. Details on the in vivo cardiac analysis are provided in the online supplement.
We determined developed and passive pressure volume relationships using the isolated heart preparation and a single beat Frank-Starling (FS) protocol. The FS protocol was run first in normal tyrode solution, followed by beta-adrenergic stimulation (dobutamine, 0.2 µM) and beta-adrenergic blockade (propranolol, 0.1 µM) as described previously 13. Pressures were converted to LV wall stress (σ) using a thick-walled spherical model: σ=P/[(LVw/1.05V+1)2/3−1], where LVw is the weight of the LV wall. This conversion normalizes for differences in LV size and wall thickness, and obtained values reflect the intrinsic stress generated by the myocardium. We determined the equilibrium volume, Vequ, defined as the volume at which passive pressure is zero. Note that this is not the volume at diastasis which we cannot measure in the isolated heart setup. This causes a systematic shift of the curve (relative to the diastatic Pressure-volume relation), but maintains the relative measurement of stiffness 14. Thus, the diastolic properties can reliably be differentiated between genotypes.
Hearts from WT and KO littermates (1y) were sectioned across the ventricles (transverse) for histological analyses. Tissues were fixed with 10% buffered formalin and embedded in paraffin. After deparaffination in xylene, sections were rehydrated in a series of graded alcohols, rinsed in PBS, and stained with hematoxylin and eosin.
IEM to localize titin epitopes at different sarcomere lengths has been described previously 13. Primary anti-titin ABs used were rabbit polyclonal (UC, UN, I84, MIR). Epitope distances with respect to the middle of the Z-disc were measured from scanned negatives of electron micrographs using Scion Image software (v. 1.6).
Protein and RNA levels were quantified as described previously 16. For titin transcript analysis we used an established oligonucleotide array 17. Details on the expression analyses are available in the online supplement.
For statistical analysis, GraphPad prism software was used. All results are expressed as means ± SEM. An unpaired two-tailed t test was performed to assess differences between two groups. The significance level was p=0.05.
Using homologous recombination we removed the titin exons 219 – 225 (Figure 1A). Homozygous KO mice survive to adulthood and are fertile, without any obvious abnormalities. PCR, Southern and Western blotting, as well as transcript analysis confirmed the deletion of the PEVK region (Figure 1A and B; supplement Figure S1). The reading frame 3’ of exon 225 is maintained in the mutant titin as indicated by the presence of the C-terminal M-band region (Figure S1D). The deleted exons encode 282 amino acids (188 PEVK residues and an Ig domain), representing a ~ 30 kDa polypeptide or ~1% of the total titin molecule. Although we used the highest resolution gel system that exists 18 this small difference is below the detection limit and, thus, WT and KO titins co-migrate on these gels. Quantification of titin isoforms and the cleavage product T2 normalized to myosin heavy chain (MHC) indicates no significant change in total titin levels but reduced cleavage of KO-titin (Figure S2A–C). Thus, the cleavage site that produces T2 could be masked in KO titin or, more likely, the site is localized within the PEVK region and when the site is excised a more stable titin molecule is obtained. Consistent with our recent study 13, the N2BA to N2B titin isoform ratio in WT hearts is ~0.2. In response to the loss of the PEVK exons we detected a minor but significant decrease in the N2BA to N2B titin isoform ratio (Figure S2D). In summary, we successfully generated a KO model in which all PEVK exons of the N2B cardiac titin isoform have been deleted.
The PEVK KO shows a cardiac phenotype with hypertrophy resulting in a ~10% increase in heart to body weight ratio (Figure 1C). Lung to body weight ratio was unchanged (WT: 5.9±0.8 mg/g vs. KO: 5.5±0.3 mg/g; n=11, P=0.1). Histological analysis revealed an enlarged left ventricular cavity but normal morphology of the ventricular wall (Figure 1D). We also performed echocardiography and obtained calculated LV weights that were increased in the PEVK KO mice, further establishing the phenotype of cardiac hypertrophy (supplement Table S1). Both diastolic and systolic LV volumes were increased after deletion of the PEVK exons indicating chamber dilation (supplement Table S1 and Figure 2A). Diastolic function was evaluated by Doppler imaging of mitral inflow (supplement Table S2) and in vivo hemodynamic measurements (Figure 2). The reduced deceleration time and increased late filling velocity indicate increased LV stiffness and diastolic dysfunction (supplement Table S2), consistent with the increased end-diastolic pressures and the increased slope of the end-diastolic pressure volume relation (beta) as documented by conductance catheter analysis (Figure 2B). We isolated LV cardiac myocytes and found that unlike in the KO of the N2B element 13, the slack sarcomere length of cardiac myocytes was unaffected (Figure 3A). However, the cellular dimensions were increased in the KO-mice (Figure 3B–D), supporting the hypertrophy phenotype seen at the level of the LV with the increase in cell length consistent with the observed chamber dilation.
To further characterize cardiac function we performed isolated heart experiments. In KOs, the pressure-volume relationship revealed both increased diastolic pressures and an increased Vequ (here: volume at which pressure is zero), while contractile function was unaffected (Figure 4). To account for differences in LV wall thickness and chamber size, we converted pressures to wall stress and found that diastolic wall stress was increased in the KO (supplement Figure S3). This indicates that the diastolic pressure increase is not due to altered chamber geometry but is instead an intrinsic muscle property. We also measured systolic function under baseline conditions and in the presence of dobutamine or propranolol and found no significant change in the KO (Figures 4 B and C). Heterozygote animals displayed an intermediate phenotype (supplement Figure S9). Thus, results of the isolated heart studies support the echo data and establish increased diastolic LV wall stress as a major phenotype of the PEVK KO model.
To investigate the mechanism underlying altered diastolic function in PEVK KO animals, we determined the mechanical properties of skinned cardiac LV muscle. There was no significant difference in slack sarcomere length or maximal active tension between WT and KO (Figure 5A,). The total passive tension–SL relationship was significantly steeper at SL > 2.05 µm in KO as compared to WT myocardium (Figure 5B). While titin-based tension was significantly increased at all SLs > 2.0 µm (Figure 5C), there was no significant change in collagen-based tension (Figure 5D). Thus increased LV passive wall stress is due to increased titin-based tension. A preliminary analysis of skeletal muscle did not reveal an effect on weight or passive and active properties (supplement Figure S10), which is likely to reflect differences between cardiac and skeletal muscle titin-isoforms (with a larger PEVK segment in skeletal muscle, the relative contribution of the 7 deleted PEVK exons is considerably smaller).
To determine how loss of the PEVK element affects the cardiac sarcomeric I-band and the extensibility of the N2B element, an important remaining source of elasticity in the KO heart 19, we studied the ultrastructure of PEVK deficient sarcomeres (Figure 6). As the gross sarcomere structure is unchanged in the PEVK-KO (Figure 6A), we were able to quantify the I-band epitope distances by immuno-electron microscopy using antibodies that flank the N2B and PEVK elements, as indicated in Figure 6B. Because the excision of the PEVK region in the KO eliminates the elastic region between UC and I84, these epitopes are separated in the WT, but overlap in the KO (Figure 6C). In contrast the distance between epitopes UN and UC was increased in the KO mice, indicating an augmented contribution of the N2B element. Figures 6D show the length of the N2B element (Uc-Un epitope distance) and that of the PEVK (Uc-I84) as a function of SL. At all sarcomere lengths, the N2B element extends to a higher degree in the KO mice. For example at a SL of 2.3 µm, the N2B element extends ~30 nm more in the KO than in the WT, accounting thereby for nearly half of the extensibility that in the WT is provided by the PEVK (with the other half provided by the two tandem Ig segments). The increased extension of the remaining spring elements in the PEVK KO provides an explanation for the increased passive tension and ensuing diastolic dysfunction of PEVK deficient cardiomyocytes.
We previously described a KO model of the elastic N2B element that develops cardiac atrophy. In contrast, the PEVK KO displays cardiac hypertrophy and dilation. Because expression of FHL2, a protein linked to atrophy/hypertrophy signaling, is down-regulated in the N2B KO we tested the expression of FHL-proteins in the PEVK KO. While upon loss of the PEVK-region mRNA levels were increased by only ~10% for FHL1 and FHL2 (supplement Figure S4), protein levels were more than doubled (Figure 7A, B). Unlike FHL1, which is not expressed at high levels in the normal myocardium 20, cardiac FHL2 protein levels were sufficient to enable subcellular localization studies by immunofluorescence (supplement Figure S5). The comparison between wildtype, N2B-knockout, and PEVK-knockout animals confirmed the differential FHL2 expression and indicated its proper localization even with the strongly increased FHL2 levels in the PEVK-Knockout (supplement Figure S5). FHL2 was localized at the I-band of the sarcomere with no additional signal in the cytoplasm or nucleus.
The upregulation of both FHL1 and FHL2 protein in the PEVK-KO suggests an important role of FHL-proteins in titin–based cardiac atrophy/hypertrophy signaling further supported by the increased expression of the hypertrophy markers ANP and Mapkap2 in the KO (Figure S6). The heat-shock-protein αB-crystallin has also been associated with the hypertrophy response 21; furthermore it acts as a chaperone and might play a role in protecting the titin filament 22. We found αB-crystallin upregulated in the PEVK knockout (Figure 7 B,C). We also included markers of the fetal gene program in our expression analysis and found that skeletal muscle actin was upregulated, accompanied by an MHC isoform shift towards the βMHC isoform (Figure S7).
The mouse PEVK region is encoded by 97 exons (112 in human) that correspond to ~53 kbp genomic sequence 9. While this virtually rules out a complete PEVK KO using conventional gene targeting, alternative splicing in the mouse heart provides a titin N2B isoform that contains the 7 C-terminal PEVK exons (N2B-PEVK), which can be excised using standard gene targeting protocols. Thus, hearts deficient in these exons express the dominant titin N2B isoform, which is devoid of PEVK sequence and an N2BA isoform that constitutes <20% of total titin with a reduced number of PEVK exons. Together with the published KO of the elastic N2B element this new model allows us to compare in vivo how distinct elastic elements differentially affect cardiac function and growth.
Although the 7 C-terminal PEVK exons comprise only a modest fraction of the elastic region of cardiac titin, the KO shows a diastolic dysfunction phenotype and increased titin based passive tension. The large effect of excising a relatively small region can be explained by earlier work 23 that showed the PEVK as a major source of elasticity towards the upper limit of the physiological sarcomere length range (tandem Ig segments are relatively inextensible at these lengths and the N2B is the other major source). Eliminating this source of extensibility in the PEVK KO, results in increased extension of the N2B element (Figure 6D) explaining the increase in titin-based passive tension that we found. This increased passive tension of cardiac KO myocytes is a likely explanation for the large increase in diastolic LV wall pressure derived from the isolated heart experiments (Figure 4). The resulting diastolic dysfunction was documented by Doppler analysis (supplement Table S2) with a significant reduction in deceleration time (MV DT - early rapid filling phase), and a restrictive filling pattern as indicated by both reduced deceleration time and aortic ejection time. Because MV DT has been inversely correlated with LV stiffness in both animals and humans 24, 25, the reduction in MV DT of the PEVK KO supports our ex vivo data, which indicate a diastolic phenotype. It was recently shown that the deceleration time cannot solely be ascribed to chamber stiffening but is also affected by viscosity26 and future work is needed to establish an additional role of viscosity in the reduced MV DT in PEVK KO mice. In our in vivo functional analysis we found that the end-diastolic pressure and the slope of diastolic pressure volume relationship were both significantly increased in the KO heart (Figure 2). These findings are consistent with echo, isolated heart, and skinned muscle data that all indicate increased diastolic stiffness in the PEVK KO.
The N2B cardiac titin isoform is the dominant isoform in the ventricular myocardium of the mouse where it is co-expressed with a small amount of N2BA titin (a larger and more compliant isoform more abundantly found in the atria). The N2B titin isoform has a shorter contour length than N2BA titin and thus the fractional extension (end-to-end length divided by contour length), for a given sarcomere stretch will be higher in N2B than N2BA cardiac titin. Because titin’s force is a function of fractional extension, force will be much higher for N2B titin than N2BA titin. In the N2B KO a minor but significant increase in expression of N2BA titin is present that we interpreted as an attempt to compensate for the increased passive stiffness that results from deletion of the N2B element 13. In contrast we found in the PEVK KO a reduction in the N2BA/N2B ratio, which is expected to increase passive stiffness. A possible explanation for this reduction is that the mutant N2BA isoform is more vulnerable to degradation, relative to the mutant N2B isoform, because some of its PEVK exons (outside the 219–225 exons) are still present in the PEVK KO. We calculated the expected increase of both the PEVK excision and the isoform switch and found a predicted force increase in the KO of ~30% of which ~2% can be accounted for by the isoform switch (supplement Figure S8). Thus the major reason for the passive tension increase in the PEVK KO is the excision of the PEVK region and not a reduced N2BA/N2B expression ratio.
No significant changes were present in active tension of skinned muscle (Figure 5A) or in developed pressure of the isolated heart (Figure 4B–C). In contrast in vivo P-V loops and echo revealed a modest reduction in the EF in KO mice (Figure 2A and supplemental Table S1). Our interpretation is that the change in EF is likely to be a secondary effect, possibly triggered by anesthesia and that the PEVK KO is a primary diastolic dysfunction model with secondary changes in systolic function.
Although the loss of elastic elements in both the N2B and the PEVK KO results in increased stiffness, the net effect on slack sarcomere length is significant in the former, but not in the latter. In slack sarcomeres titin corresponds to a flexible chain at zero external force where the mean square end-to-end distance is a function of the contour length (CL) of the chain. A reduction in CL (knockout of N2B or PEVK) will reduce the end-to-end distance and hence slack sarcomere length. While this holds true for the N2B KO, the sarcomere length in the PEVK KO is unchanged, possibly due to the smaller deletion as compared to the N2B KO, reducing the effect to the point where it was undetectable in our experiments.
Hearts deficient in the elastic N2B or PEVK region do not show signs of fibrosis as determined by histology (Figure 1D) and the unchanged collagen based stiffness (Figure 5D). Both models share an increased titin based stiffness that results in diastolic dysfunction. However, their deletion differentially affects cardiac growth with atrophy in the N2B KO 13, and hypertrophy and chamber dilation in the PEVK KO. In part, this might be explained by the differential binding of proteins that relate to hypertrophy and stress signaling such as FHL1, FHL2, and αB-crystallin, which all bind the N2B element, but not the cardiac PEVK 11, 27, 28. All these proteins were upregulated in the PEVK knockout with differential implications for the phenotypic differences in the N2B and PEVK knockout. αB-crystallin is a small heat shock protein that acts as a chaperone to maintain the folded state of proteins. In addition to binding to the N2B element of titin, it binds the tandem Ig segment between the N2B element and the Z-disc in highly stretched sarcomeres and possibly protects titin from structural damage under conditions of increased vulnerability 28, 29. Upon elimination of PEVK exons, the strain of titin’s I-band region, at a given sarcomere length, is increased and the increased expression of αB-crystallin can be interpreted as a compensatory mechanism to maintain a functional I-band region.
The unexpected hypertrophy phenotype in the PEVK-KO could result both from the loss of a PEVK dependent signal or from a mechanosignal secondary to the increased strain of the remaining I-band titin. Recent work in both skeletal and cardiac muscle has indicated a critical role for FHL1 and FHL2 in hypertrophy signaling 10, 11. The differential regulation of FHL proteins in the PEVK- and N2B knockout could thus explain their disparate trophic phenotypes (Figure 1 and 13). Only recently has FHL1 been implied in cardiac pathology with protection of the FHL1 knockout from pressure induced hypertrophy 11. Gain of function studies have shown induction of hypertrophy by FHL1 in skeletal muscle 30 and we find a similar hypertrophy phenotype with increased cardiomyocyte diameter in the PEVK knockout heart. While future work is needed to dissect the mechanosignaling pathway, we propose that the N2B region plays a critical role in the trophic response to strain based on the data presented here and published work 11, 13, 27. Our localization and expression studies of FHL1 and 2 suggest a model where increased strain of the N2B region in the PEVK-KO induces structural changes in the N2B region that facilitate binding of FHL proteins and the assembly of a signaling complex to induce hypertrophy (Figure 8 and Figure S5). The N2B-KO lacks the I-band attachment sites for FHL2 and thus the basis for signalosome assembly, resulting in atrophy. We propose that the moderate changes in fetal and hypertrophy signaling (supplement Figures S6 and S7) are likely to reflect structural changes in the N2B region secondary to the PEVK deficiency and that these changes are translated via FHL proteins. Altered biomechanics due to the excision of titin’s PEVK region is thus the primary determinant of both the diastolic and the trophic PEVK KO phenotype. In summary, we have generated a novel PEVK KO and have shown that the PEVK region is an important source of elasticity within the physiological sarcomere length range of the heart. Its absence results in diastolic dysfunction and hypertrophy accompanied by upregulation of FHL1 and FHL2. Our findings indicate that the cardiac PEVK region is required for proper ventricular filling and the regulation of cardiac muscle mass. Insights provided by the PEVK KO should aid in the design of titin-based therapies for diastolic dysfunction and more complex cardiomyopathies.
We thank Beate Goldbrich, Gemaine Wright, Joe Popper, and Bryan Hudson for technical assistance.
Sources of Funding
This study was supported by the American Heart Association, the Deutsche Forschungsgemeinschaft, the Sofja-Kovalevskaya program of the Alexander von Humboldt Foundation (MG), and NIH (HL62881 to HG).
Subject codes: 145; 105; 130; 138; 11; 15