the control elements involved in molecular signaling pathways responsible for changes in contractility of muscle cells have been known for decades. Whether contractility is actin driven, in the case of cardiac and skeletal muscle with highly organized sarcomeric filaments, or myosin driven, for relatively disorganized smooth muscle, the intracellular second messenger Ca2+ provides the stimulus that allows for the interaction of actin and myosin, leading to cross-bridge cycling and contraction. In the case of vascular smooth muscle, Ca2+ binds calmodulin, which, when complexed, activates myosin light chain kinase (MLCK). MLCK phosphorylates myosin light chain (MLC), which is then capable of interacting with actin to produce contraction. More recently, studies by investigators such as van Breemen, Somlyo, and Kaibuchi (6, 11, 14, 16), with permeabilized preparations of blood vessels and vascular smooth muscle cells in which intracellular Ca2+ concentration is maintained constant, have shown that contraction in response to agonists can occur without an increase in Ca2+. Subsequent to this novel finding, a role for Rho, a member of the Ras superfamily of small G proteins, was elucidated in vascular contraction (3, 6, 12). In the absence of increases in intracellular Ca2+, activation of Rho by a subset of heterotrimeric G protein-linked receptors had been shown to lead to a cascade whereby Rho signaling through its effector Rho kinase leads to an increase in contraction. Substantial evidence indicates that Rho kinase phosphorylates the myosin-binding subunit of myosin phosphatase (8, 10, 17). This phosphorylation inactivates myosin phosphatase, resulting in an accumulation of phosphorylated MLC. Data obtained in vitro also suggest that Rho kinase might directly phosphorylate MLC (1). Therefore, there appear to be at least two pathways contributing to contraction: an initial phasic component of contraction initiated by Ca2+-dependent processes involving MLCK and a tonic component maintained by sustained cellular levels of phosphorylated MLC in response to Rho-dependent inhibition of myosin phosphatase (see Fig.1).
Whereas muscle cells are characterized by their contractile ability, contraction is also a crucial function of other cell types. For example, fibroblasts contract and migrate in the process of wound healing. Although experimental evidence is lacking regarding potential mechanisms involved in non-muscle cell contractility, the widely believed hypothesis is that MLC phosphorylation might play a role in contractile function of such non-muscle cells. Indeed, some reports suggest that fibroblasts do respond with an increase in phosphorylated myosin when stimulated with serum or agonists such as PDGF or lysophosphatidic acid (LPA) (2, 13, 21). For the most part, such experiments were performed on cells attached to culture dishes, rather than fibroblasts that are able to contract, thus preventing a direct comparison between force generation and MLC phosphorylation. Studies using fibroblasts grown on collagen gels highlight the importance of a direct comparison between biochemical signaling and force generation. Grinnell et al. (4) found numerous differences between fibroblasts that were floating vs. those on stressed collagen matrices. Whereas LPA produced contraction of either type of fibroblast, PDGF contracted only the floating fibroblasts. Of additional importance was a differential ability of inhibitors to prevent these contractions (4). For instance, KT 5926, an inhibitor of MLCK, blocked PDGF contractions, but it had no effect on LPA. Pertussis toxin, which ADP ribosylates and inactivates Gi/o proteins, only prevented LPA contraction of floating cells. Interestingly, C3, an inhibitor of Rho function, was able to block LPA contractions, suggesting the involvement of a Rho pathway that contributes to fibroblast contraction. A study by Yee et al. (21) further examined the role of the Rho/Rho kinase pathway in contraction and myosin phosphorylation of chicken embryo fibroblasts under standard cell culture conditions, i.e., without stressed collagen fibers. These authors found that serum and endothelin-1 produced cell contraction and myosin phosphorylation, but the Rho kinase inhibitor Y-27632 blocked myosin phosphorylation at a lower concentration than it blocked contraction, implicating alternative mechanisms in fibroblast contraction.
Additional studies point to a discrepancy between an ability to promote fibroblast contraction and stimulation of a traditional Ca2+/MLC phosphorylation pathway. Richard Paul's group [Obara et al. (19)] has shown that cytochalasin D, a microfilament disrupting agent, produces a decrease in contraction while having no effect on intracellular Ca2+ concentration. Conversely, the microtubule disrupter nocodazole has no effect on force but produces a slight reduction in intracellular Ca2+concentration, arguing against a role for increases in intracellular Ca2+ in fibroblast contraction. Likewise, the preponderance of evidence suggests that myosin phosphorylation is unrelated to contractility. Not only are inhibitors of MLCK ineffective at blocking fibroblast contraction, but numerous studies have also demonstrated that growth factor stimulation of contraction does not correlate with an increase in the phosphorylation state of MLC (5, 18, 20,21). For example, the cAMP-elevating agent forskolin, which inhibits MLC phosphorylation in response to LPA, has no effect on contraction (20). Consistent with this, fibroblasts overexpressing the catalytic domain of MLCK exhibit the same dose-response relationship for maximal force and force velocity following serum stimulation as do wild-type cells, implicating a lack of involvement of MLC phosphorylation in fibroblast contraction.
In the current article in focus (Ref. 16a, see p. C599 in this issue), Paul's group provides another provocative finding: an increase in contractility in the absence of myosin phosphorylation. While in light of past work it is not surprising that the Rho/Rho kinase pathway appears to be involved in fibroblast contraction, it is unprecedented that this pathway is found to mediate increases in contractility in the absence of myosin phosphorylation. The study is quite compelling in that the MLC phosphorylation experiments, while technically challenging, were performed on contracted fibroblast fibers, precluding the possibility that the signaling pathways utilized (or not) might be attributed to cell culture conditions. The findings point to a road less traveled, an underinvestigated line of research, regarding the non-myosin-driven effects of the Rho/Rho kinase pathway. Of equal importance is the new and unexpected observation that an effect inhibited by a Rho kinase inhibitor is not attributable to myosin phosphorylation.
If not myosin phosphorylation, then what signal(s) is(are) responsible for the increase in force generation of fibroblasts in response to serum or growth factors? This question remains to be answered, but a few candidates come to mind. For example, Rho kinase and myosin phosphatase have been shown to regulate the phosphorylation state of adducin, an actin-interacting protein involved in capping ends of F-actin (9). Kaibuchi's group demonstrated that expression of a mutant form of α-adducin, which cannot be phosphorylated, blocks cell migration. Another group has recently identified elongation factor-1α, a cofactor for polypeptide elongation, as a target for Rho kinase phosphorylation (7). They have shown that phosphorylation of elongation factor-1α by Rho kinase reduces its ability to interact with actin. Rho kinase has also been shown to phosphorylate another cytoskeletal regulator, LIM kinase, which phosphorylates cofilin, leading to actin cytoskeletal reorganization (15). Ezrin/radixin/moesin proteins are another group of cytoskeletal regulatory proteins that might be targets for phosphorylation by a Rho-dependent mechanism, although there is some question regarding whether Rho kinase or another kinase might mediate their phosphorylation (22).
Although the target of Rho kinase involved in fibroblast contraction has yet to be elucidated, the above-mentioned potential targets provide a good starting point for the investigation of this question. While direct assessment of the phosphorylation of such targets during fibroblast/collagen gel contraction is likely to be difficult, the use of knockout fibroblasts, which lack the targets of interest, would be quite useful in determining their necessity for this response. Of additional interest is whether phosphorylation of, and inactivation of, myosin phosphatase is an intermediary in this signaling pathway or, alternatively, whether direct phosphorylation by Rho kinase is responsible. By addressing and answering these important questions, candidate therapeutic targets, which specifically effect wound healing without untoward effects on vascular contraction, might be identified.
Address for reprint requests and other correspondence: Dept. of Pharmacology, Univ. of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0636 (E-mail:).
- Copyright © 2003 the American Physiological Society