JNK Bond Regulation

• human
• gap junction
• disease
• c-Jun N-terminal kinase
• transgenic

Large numbers of gap junctions interconnect working myocytes in the ventricle of adult humans.^1–5⇓⇓⇓⇓ Until recently, it was assumed that the reason for the abundance of these clusters of intercellular channels was straightforward. This phenomenon was thought to be principally a matter of ensuring that action potentials were conducted efficiently from myocyte to myocyte.^6–8⇓⇓ However, it has long been known that the mammalian heart is atypical with respect to that of other vertebrates in the high frequency and large size of gap junctions found between its constituent muscle cells.^1–5,9,10⇓⇓⇓⇓⇓⇓ In nonmammalian chordates ranging from the affixed sea-squirt¹⁰ to the nimble sparrow,⁹ spread of activation is expedited and maintained in the main pumping chamber of the heart with only a fraction of the channels that are apparently required in humans.

The molecular composition of gap junction channels in the adult mammalian ventricle is also a phylogenetic oddity. By far, the most abundantly expressed subunit protein of the numerous gap junctions found in this tissue is connexin43 (Cx43),^1–8,10,11⇓⇓⇓⇓⇓⇓⇓⇓⇓ a member of the connexin family of proteins that is probably all but absent from the mature cardiac muscle of other species on the vertebrate family tree.¹⁰ Thus, of the many species and classes of animal within our phylum, we mammals seem literally to be out on a limb in terms of the characteristics of the electrotonic couplings linking our ventricular muscle cells.

In addition to being expressed at high levels in the mammalian ventricle, significant resources are allocated to regulation of the trafficking, assembly, sarcolemmal distribution, and turnover of Cx43. For example, over postnatal growth, Cx43 gap junctions undergo remodeling into polarized and precisely organized geometries at intercalated disks^12,13⇓—specialized domains of electromechanical coupling between myocytes. The progressive change in coupling geometry occurs in a repeatable and highly choreographed sequence, suggesting close regulation of this developmental process. In a further illustration, the half-life of Cx43 is short, indicative of significant investments in mechanisms ensuring rapid turnover of this protein.¹⁴ Perhaps the most striking instance of just how closely Cx43 is tended in the myocardium of mammals is provided by its exquisite sensitivity to pathological disturbance.^{2,4,6–8,15–28}⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Localized remodeling of Cx43 gap junction distribution and overall reductions in Cx43 level are common features of ischemic, hypertrophic, and other cardiomyopathic diseases of the heart in humans.^{2,4–8,15–21}⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Similarly, there are now numerous reports of acute and long-term disruptions to Cx43 expression, pattern, and phosphorylation in a variety of mammalian models of cardiac disease.^{6–8,14–18,22–25}⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Dysregulation of ventricular Cx43 is also frequently observed in studies of mice manipulated by knockout and transgenic techniques in which a connexin gene is not the target.^26–28⇓⇓ In many cases, alterations to Cx43 gap junctions provoked by physiological stress, pathological insult, or deleterious genetic manipulation are accompanied by manifest disturbance in the conduction and stability of electrical activation in the heart, including sudden cardiac death from ventricular arrhythmia.

The c-Jun N-terminal kinase (JNK) signal transduction cascade is one of a number of intracellular signaling networks activated by pathological stress in mammalian ventricular myocytes. In this issue of Circulation Research, Petrich and coworkers²⁹ are the first to give evidence that JNK may be a key mediator of the sensitive response of cardiac Cx43 to acute pathological disturbance and chronic disease processes. In a comprehensive piece of work, these authors demonstrate that JNK activation accompanies the downregulation of gap junctional intercellular communication and Cx43 levels in murine ventricular myocytes. To underscore their point, the authors use a range of techniques, from chemical inducers of JNK phosphorylation in cultured cells to ventricular-specific expression of a constitutively active JNK in vitro and in vivo. How the relationship between JNK and Cx43 is orchestrated remains unclear. Based on the close temporal correlation between activation of the kinase and the accompanying rapid decrease in the levels of the gap junction protein, the authors conclude that JNK is most likely involved in regulation of Cx43 at the posttranscriptional level. Whether Cx43 is a direct downstream target of JNK or whether there are multiple steps intervening between activation of JNK and reduction in Cx43 remains to be discerned. It is nonetheless noteworthy that increases in phosphorylation of Cx43 over a similar time course have previously been reported to accompany the loss of cell membrane localized gap junctions in different cell types, including rodent ventricular myocytes.^30,31⇓

In terms of explaining a pathogenic response, the contribution by these authors makes a lot of sense. Their conclusion that JNK is “… an important mediator of stress-induced Cx43 downregulation and impaired intercellular communication in the failing heart” (page 640)²⁹ appears to have merit. However, to return to the broader issues concerning Cx43 raised earlier, one has to ask just what (if any) are the advantages of myocardial Cx43? Why do mammalian myocytes expend such effort to coordinate and maintain abundant Cx43 gap junctions? The hearts of other species function unerringly with far fewer intercellular channels. Why indeed, do mammals “bother” with this apparently problematic gap junction protein? As Petrich et al²⁹ teach us, there is likely a mechanism primed to see briskly to the disappearance Cx43 at the first sign of trouble. In turn, this downregulation of Cx43 and attendant remodeling of Cx43 gap junctions may eventually cause fatal electrophysiological dysfunction of the heart.

Complete answers to the questions posed above do not readily suggest themselves and perhaps it is an accident of nature that mammalian hearts express Cx43. However, there is further information that may begin to clarify these issues. Homozygous knockout of the Cx43 gene in mice is lethal at birth owing to obstruction of pulmonary outflow from the right ventricle.³² Interestingly, this fatal cardiac defect appears to result mainly from loss of Cx43 function in nonmyocardial tissues. Namely, in the coordinate and directed migration of motile cohorts, such as epicardially derived cells and neural crest immigrating, into the developing heart.^6,33,34⇓⇓ Thus, expression of Cx43 by the myocytes themselves does not seem to be essential for cardiac development (at least in the prenatal mouse), although knockout of Cx43 does measurably influence the rate and reliability of activation spread in ectopically paced and isolated hearts.³⁵

Studies of mice in which Cx43 has been conditionally knocked-out of cardiac lineages provide further unexpected evidence that we may have underestimated the subtleties of connexin function.³⁶ Mice with cardiac-specific knockout of Cx43 are born live without the heart defect seen in the germline knockout animal—reinforcing the concept that Cx43 in cardiac muscle cells does not play a major role in cardiac morphogenesis. Moreover, a good number of the conditionally targeted animals (homozygous for deletion of Cx43) remain postnatally viable for weeks after birth. The myocardial cellular structure and contractile function of these surviving Cx43 mutants appears to remain largely intact, although the rate of activation spread in the ventricles is somewhat slower than that of wild-type littermates. Nonetheless, this modest reduction in propagation velocity of action potential does not correspond to what would have been anticipated from a catastrophic loss of 95% or more of the gap junctional structures coupling ventricular myocytes. In 20/20 hindsight, the precedence set by lower vertebrate animals should have perhaps alerted us to the likely outcome of cardiac-specific targeting of Cx43 in mouse. Importantly, however, these mutant mice do succumb increasingly with postnatal age to sudden cardiac death from ventricular arrhythmias. Thus, although abundant gap junctions composed of Cx43 may not be necessary for intercellular conduction of cardiac action potential from beat-to-beat in mammals, these intercellular junctions may be a necessary “safety factor” for the long-term stability of ventricular activation. The nagging issues nonetheless remain. Why do nonmammals not require the putative “safety factor” conveyed by high expression of Cx43? Why do mammals quickly ditch and then not reestablish at the same level this safety factor after the very circumstances (ie, after pathological disturbance) in which it would seem to serve its most useful purpose? Does this paradox suggest additional roles for Cx43 that are independent of intercellular coupling as suggested by Lo and colleagues?³⁴ No doubt further understanding of how intracellular pathways such as the JNK signaling cascade are activated in cardiac development, health, and disease may provide answers to some of these questions.