Could Modification of Titin Contribute to an Answer for HFpEF?

Treatment of Heart Failure With Preserved Ejection Fraction: An Unmet Need

In 2016, treatment of heart failure with preserved ejection fraction (HFpEF) remains a critical unmet need. A recent guideline publication (1) stated: “No treatment has yet been shown, convincingly, to reduce morbidity or mortality in patients with HFpEF.” This fact stands in stark contrast to the disease burden borne by HFpEF patients (2): It is estimated that > 50% of all HF patients have HFpEF; 5 year mortality is > 60%; 6 month re-hospitalization rate for decompensated HFpEF > 50%; many patients have profound disability, often comparable to patients with reduced EF (HFrEF) who are candidates for transplantation. A recent opinion paper by Shah et al (3) discussed phenotypic diversity in the HFpEF population and proposed an approach to treatment based on specific aspects of pathophysiology present in individual patients. The accompanying paper by Methawasin et al (4), with its focus on titin, is an example of how such an approach might ultimately work.

Seven large randomized clinical trials of pharmacologic approaches have failed to demonstrate a reduction in morbidity and mortality (1,2). Why? Were the studies designed or conducted poorly? Did they include the wrong group(s) of patients? Did they test the wrong endpoints? It is certainly possible that each of these factors contributed to the neutral results. However, lack of demonstrated efficacy may be based on the possibility that the underlying mechanisms specific to the syndrome of HFpEF were not the target of the treatments tested and/or the treatment targets were too narrow. In large part, treatment for HFrEF targets left ventricular remodeling characterized by increased chamber volume and marked neurohumoral activation (1). As shown in Table 1, these are either not present or present to a lesser degree in HFpEF. Therefore, the development of a successful approach to HFpEF management may require targeting multiple underlying mechanisms responsible for its development, including abnormalities of myocardial function as well as a variety of non-myocardial alterations reported in these patients, e.g., insulin resistance, obesity, systemic inflammation/oxidative stress and chronic kidney disease.

Table 1.

Differences that distinguish HFpEF from HFrEF

	HFpEF	HFrEF
Left Ventricular (LV) Structure and Function
End Diastolic Volume (EDV)	↔	↑
End Systolic Volume (ESV)	↔	↑
Wall Thickness	↑	↔
LV Mass	↑	↑
Mass / EDV ratio	↑	↓
Remodeling	Concentric	Eccentric
Ejection Fraction (at rest)	↔	↓
Stroke Work (at rest)	↔	↓
End Systolic Elastance	↔	↓
Vascular Stiffness	↑	↓
End Diastolic Stiffness	↑	↓
Myocardial Ultrastructure
Cardiomyocyte Length	↔	↑
Cardiomyocyte Diameter	↑	↔
Remodeling	Concentric	Eccentric
Myocardial Fibrosis	Diffuse	Patchy, Focal
Effect on HF hospitalization and CV Death
B-blockers	No Effect	Improved
Angiotensin Converting Enzyme Inhibitor (ACE-I)	No Effect	Improved
Angiotensin Receptor Blocker (ARB)	No Effect	Improved
Digitalis	No Effect	Improved
Hydralazine / Nitrates	No Effect	Improved
Mineralocorticoid Receptor blocker (MRA)	No Effect	Improved
PDE-5 Inhibitor	No Effect	Not Studied

With respect to the myocardium, a number of molecular and cellular mechanisms have been defined in patients that individually and collectively contribute to the underlying pathophysiology of HFpEF. These include alterations in sodium-calcium handling (5) and acto-myosin dynamics (6), energy requiring processes which largely determine the extent and rate of relaxation, and collagen content and cross-linking and titin, which largely determine passive stiffness (7–9). Granzier, Zile, LeWinter, Paulus and collaborators (8,9) have reported changes in titin phosphorylation in human HFpEF that contribute to increased cardiomyocyte and myocardial passive stiffness (Figure 1). Moreover, reversal of these changes improves resting cardiomyocyte tension in vitro (9). Thus, titin represents a novel and attractive target for treatment of HFpEF. For readers not familiar with titin, understanding the ground-breaking work of Methawasin and co-workers (4) requires a brief introduction to this important protein.

Fig 1.

Mechanisms of increased passive myocardial stiffness in patients in with HFpEF (see text).

Structure and Function of Titin

Based on electron microscopic studies, additional myofilaments associated with the thick filament of the sarcomere in skeletal and flight muscle were recognized as early as the 1960s (11). During the mid-late 1980s these structures were characterized as the giant molecular spring protein titin. The “modern era” for cardiac titin began in the mid to late 1990s when details of the protein’s structure and mechanical function as a determinant of cardiomyocyte resting tension as well as understanding of its binding partners and role in stretch-based cell signaling began to emerge (12).

Titin is the largest known mammalian protein (11,12). Its N-terminus is embedded in the Z-disk and its C-terminus is bound to the thick filament in the M-band. The near Z-disk region is also bound to the thin filament. The I-band region consists of multiple segments, including tandem Ig, PEVK and N2A and N2B elements, which function as molecular springs of varying stiffness and together determine the composite stiffness of the entire protein. Alternative splicing of large fetal-type isoforms gives rise to the two isoforms found in adult mammalian myocardium, the larger and more compliant N2BA, containing both N2A and N2B segments, and the smaller N2B isoform, containing only the N2B segment. Adult human myocardium contains roughly equal amounts of N2BA and N2B titin, whereas rodent myocardium is largely composed of N2B titin. Splicing is mediated by the recently described splice factor, RNA binding motif-20 (RBM20) (13).

In addition to isoform variation, titin’s stiffness can be modified by other post-translational modifications (PTMs). Phosphorylation, the most widely studied PTM, allows for rapid changes in titin stiffness (11,12). Titin contains a large number of phosphorylatable sites. Cyclic-AMP dependent protein kinase (PKA) and cyclic-GMP dependent protein kinase (PKG) phosphorylate the same sites, resulting in decreased titin stiffness. Protein kinase C alpha (PKCα) phosphorylates different sites and increases titin stiffness. Extracellular-signal-regulated kinase-2 (ERK2) has its own sites and lowers titin stiffness, while CaMKII, a Ca²⁺/calmodulin dependent serine/threonine kinase, has a complex phosphorylation scheme whose net result is likely decreased titin stiffness.

Titin is the predominant source of cardiomyocyte (cellular) resting tension and stiffness (11,12). Together with extra-cellular matrix collagen it is the primary determinant of passive myocardial (tissue) stiffness. Titin has a number of other mechanical and non-mechanical effects. During contraction below slack sarcomere length (SL), titin is extended in the opposite direction from that which occurs during passive stretch, resulting in a cardiomyocyte restoring force and elastic recoil. This property may contribute to ventricular diastolic suction. SL-dependent activation mediated by increased myofilament Ca²⁺ sensitivity is magnified at higher levels of titin-dependent tension, likely related to titin binding to thick and/or thin filament. Titin interacts with various Z-disk proteins as well as with a number of other binding factors and may have an important role in biomechanical sensing and nuclear signaling (11,12). Finally, titin may provide scaffolding for normal sarcomere assembly and organization.

Titin in Human Heart Disease

Over the last 10–15 years an important role for titin in human heart disease has emerged (11,12,14,15). Genetic studies (15) reveal multiple titin mutations that cause dilated cardiomyopathy (DCM), and titin is the most commonly mutated gene associated with DCM. Changes in titin isoforms toward a higher N2BA:N2B ratio which reduce passive cardiomyocyte stiffness have been reported in patients with ischemic and non-ischemic DCM (11,12,14). Reduced PKA/PKG-mediated phosphorylation has also been reported, which would have the opposite effect. In these studies, passive myocardial stiffness of muscle strips from DCM patients was reduced, suggesting that the isoform effect is more important. In contrast, in HFpEF, significant changes in isoform ratio have not been demonstrated. However, as noted earlier phosphorylation changes appear to play a major role (8,9,12,14). Borbely and Paulus and co-workers were first to report a decrease in phosphorylation of titin’s PKA/PKG sites in LV endomyocardial biopsies from HFpEF patients (9), resulting in increased cardiomyocyte resting tension. They subsequently showed that this is associated with reduced PKG activity and have advanced the hypothesis that an inflammatory process centered in the endothelium is responsible (2,9). Recently, we showed that phosphorylation of a functionally important PKCα site is increased while, as expected, PKA/G site phosphorylation was decreased in HFpEF (8). Thus, these phosphorylation changes combine to increase cardiomyocyte resting tension.

Manipulation of Titin Isoform Expression in a Murine Model of HFpEF

Methawasin et al (4) are to be congratulated for a tour de force proof of principle investigation of the effects of manipulation of titin isoforms as a treatment for a transverse aortic constriction (TAC) – DOCA murine model of progressive LV hypertrophy leading to HFpEF. Conditional expression of a transgene with deletion of the RNA recognition motif for one of the RBM20 alleles (cRbm20^∆RRM) resulted in reduced splicing and a substantial increase in larger, more compliant titins. This novel approach allowed testing of the hypothesis that increasing the compliance of titin can reverse HFpEF after its establishment. The result was normalization of passive stiffness of isolated muscle strips as well as LV diastolic function and chamber stiffness assessed by echocardiography and pressure-volume analysis. There were no effects of “treatment” on collagen content or intracellular Ca²⁺ transients. The authors also showed that other splice targets of RBM20 did not contribute to their results. Thus, the beneficial effects were almost certainly entirely related to changes in titin isoforms. Interestingly, cRbm20^∆RRM mice also displayed increased relaxation rate. Improved relaxation may reflect an alteration in the interaction between titin and thin and/or thick filaments caused by reduced titin dependent tension. Importantly, treated animals displayed improved exercise tolerance, and there were no obvious adverse effects of the isoform switch. In summary, increasing titin compliance in this model resulted in marked improvement in multiple measures of cardiac diastolic function and performance.

There have been previous attempts to treat HFpEF patients with pharmaceutical agents that could alter titin phosphorylation. Unfortunately, trials of sildenafil (16), a phosphodiesterase V inhibitor, and isosorbide mononitrate (17), both of which would be expected to reverse the decrease in PKG-mediated titin phosphorylation observed in HFpEF, did not show benefit. This approach is not necessarily futile, as there may be other means to achieve this goal, for example, inhaled or oral nitrites. However, it is important to reiterate that Methawasin et al (4) targeted titin by engineering an isoform switch which, while markedly improving myocardial stiffness and normalizing cardiac function and exercise tolerance, did not reverse an abnormality found in HFpEF. Thus, this promising approach is not dependent on reversing a pre-existing titin abnormality and has the added benefit of speeding relaxation. However, it would be naïve to think that the TAC-DOCA model fully replicates human HFpEF. Because of the multiple underlying pathophysiologic mechanisms in HFpEF (2,3), it seems unlikely that targeting only one will either completely normalize diastolic function and/or fully resolve a complex, multi-level clinical syndrome. Application of these principles predicts that targeting each mechanism may improve but not fully normalize diastolic function and functional status and potentially decrease (but not eliminate) morbidity and mortality associated with HFpEF. This paradigm also does not take into account the likelihood that individual patients differ with respect to which underlying mechanisms are present and their respective magnitudes.

Conclusions

The work of Methawasin et al (8) provides evidence that a singular focus on increasing the compliance of titin may have a significant impact. This, in and of itself, is an important observation. Of course, successful adaptation of this approach to patients would require development of a practical means to inhibit RBM20. Moreover, the higher proportion of N2B titin in rodents means that the effect of switching to larger, more compliant isoforms would be more limited in patients. Finally, as discussed above, the undoubtedly greater complexity of the HFpEF syndrome in patients makes it less likely that such a focused approach will be completely successful on its own. Nonetheless, this is exactly the sort of novel and rigorous investigation which may point the way to real progress in the conundrum of HFpEF.