Biology and Signaling Alterations in AD

The Rho-associated coiled-coil containing kinases (ROCKs) are ubiquitously expressed serine/threonine kinases that are downstream effectors of the RhoA GTPase (Ishizaki et al., 1996b; Leung et al., 1996a; Leung et al., 1995a; Matsui et al., 1996a; Nakagawa et al., 1996a; Schackmann et al., 2011). ROCKs are key regulators of the actomyosin cytoskeleton dynamics, as well as neuronal morphology and synaptic plasticity (Schubert and Dotti, 2007). The two ROCK homologs (ROCK1 and ROCK2) share high homology in their kinase domains (92%) and overall amino acid sequence (64%) (Julian and Olson, 2014).

As mentioned previously, the ROCKs are ubiquitously expressed throughout development (Leung et al., 1996a; Nakagawa et al., 1996a), and are critical for these processes as demonstrated by embryonic lethality in ROCK homozygous knockout mice (Rikitake et al., 2005; Shimizu et al., 2005; Thumkeo et al., 2003). Here we will further define ROCK biology including structure and

localization, activation, function, and downstream signaling pathways.
The structure of the Rho-kinases consists of an amino-terminal kinase domain followed by an extensive coiled-coil region that contains the Rho-binding domain (RBD), and then finally a carboxy-terminal pleckstrin homology (PH) domain (Figure 1). (Riento and Ridley, 2003). The PH domain of the carboxy-terminal serves as an auto-inhibitory region that modulates ROCK activity by binding to the kinase domain (Amano et al., 1999). The PH domain is also thought to be critical for precise cellular localization of the ROCKs. The RBD is contained within the coiled-coiled region of the protein, which is also thought to regulate interactions between other ?-helical proteins. The kinase domain is located at the amino-terminal and can only be activated when auto-inhibition is prevented. In terms of ROCK localization and distribution, both isoforms are ubiquitously expressed. ROCK1 is more predominantly expressed in the bladder, blood, mammary, muscle, pancreas, peripheral nervous system, small intestine, spleen, testis, thymus, and tongue; ROCK2 is more predominantly expressed bone marrow, brain, colon, eye, heart, kidney, larynx, soft tissue, and stomach (Julian and Olson, 2014).

The ROCKs can be activated by one of two pathways: binding of GTP-bound RhoA or carboxy-terminal cleavage, which renders a constitutively active enzyme. Canonically, auto-inhibition is controlled by the binding of GTP-bound RhoA, which, when bound, relieves the interaction between the PH domain and the kinase domain allowing for ROCK activity (Amano et al., 1999; Ishizaki et al., 1996b; Leung et al., 1995a; Matsui et al., 1996a). ROCK activation can also be controlled by caspase-mediated cleavage of the carboxy-terminal (Ishizaki et al., 1997; Leung et al., 1996a). This typically occurs under apoptotic conditions, when caspases cleave the ROCKs to generate constitutively active enzymes (Coleman et al., 2001; Sebbagh et al., 2005; Sebbagh et al., 2001).

The ROCKs serve a variety of different functions related to actin stress fiber formation, apoptosis, development, and neurite outgrowth and retraction. One of the first functions of the ROCKs that was identified was the formation of stress fibers (Amano et al., 1997), which are cellular contractile structures composed of filamentous actin and myosin II (Murakami and Ishikawa, 1991; Sugimoto et al., 1991). Studies have shown that when constitutively active ROCKs are expressed the formation of stress fibers is induced, whereas kinase-dead or other ROCK mutants precluded the formation of stress fibers (Amano et al., 1997; Ishizaki et al., 1997; Leung et al., 1996a). The formation of these stress fibers is controlled through ROCKs regulation of myosin light chain (MLC) activity. ROCK2 phosphorylates and inactivates myosin light chain phosphatase, which results in increased activity of MLC (Kimura et al., 1996). Increased activity of MLC then results in enhanced cellular contractility via enhanced binding of myosin II to actin (Amano et al., 1998). This enhanced binding is what then promotes filamentous actin assembly into stress fibers (Chrzanowska-Wodnicka, 1996).

The ROCKs are also potent regulators of apoptotic processes (Coleman and Olson, 2002; Coleman et al., 2001). During apoptosis membrane blebbing, nuclear disintegration, and cellular fragmentation is driven by actin-myosin contraction. This membrane blebbing and actin cytoskeleton remodeling is driven by MLC activity that is controlled via caspase-3 mediated cleavage and activation of ROCK1 (Cotter et al., 1992; Mills et al., 1998; Mills et al., 1999; N et al., 2000; Zhang et al., 1999). Caspase-3 mediated cleavage of ROCK1 yields a constitutively active enzyme by relieving the auto-inhibitory action of the C-terminal domain, which then promotes membrane blebbing via phosphorylation and activation of MLC (Coleman et al., 2001; Sebbagh et al., 2001). ROCK2 plays a key role in specialized cases of cell death. In this case, natural killer cells induced Granzyme B mediated cleavage of ROCK2 that yields a constitutively active enzyme much like in the case of caspase-3 cleaved ROCK1. This induces a ROCK2-mediated, caspase-independent mechanism of apoptotic membrane blebbing (Sebbagh et al., 2005).

The ROCKs have redundant expression patterns in development, and are highly expressed in embryonic cardiac and neural tissue where they regulate such morphogenic events as cell migration and cell differentiation (Shi et al., 2011). Pan-ROCK inhibition via pharmacologic inhibitors has revealed the importance of ROCK signaling in development, but has not elucidated isoform-specific differences of ROCK1 and ROCK2 in development (Wei et al., 2001). Genetic depletion, however, has provided a mechanism to elucidate ROCK1 and ROCK2 differences in development. Mice lacking either ROCK1 or ROCK2 in development do not fully form eyelids and also present with an omphalocele phenotype, which is due to actin disorganization in the epithelial cells of the eyelids and umbilical region (Shimizu et al., 2005; Thumkeo et al., 2003; Thumkeo et al., 2005). It is important to note that while the ROCKs show distinct differences in development, they appear to have redundant function in most tissue types.

Central to this dissertation is the role of the ROCKs in neuronal morphogenesis and synaptic plasticity. Work in the late 1990s and early 2000s used neuronal cell culture systems to elucidate the role for the ROCKs in neurite retraction and axonogenesis (Bito et al., 2000; Duffy et al., 2009; Fujita and Yamashita, 2014; Hirose et al., 1998). ROCKs regulate neurite retraction via increased phosphorylation, and subsequent activation and contraction of MLC (Hirose et al., 1998; Katoh et al., 1998). ROCK inhibition not only produced new axonal outgrowth from neurons, but also stimulated axonal regeneration after detrimental spinal cord injury (Bito et al., 2000; Dergham et al., 2002). ROCK signaling is critical to the process of synaptic plasticity, with functions that regulate dendritic spine morphogenesis and synaptic vesicle endocytosis and exocytosis (Hodges et al., 2011; Newell-Litwa et al., 2015; Swanger et al., 2016; Taoufiq et al., 2013). Kinase activity of ROCKs influence dendritic spine morphology and synaptic plasticity through one of two pathways (Figure 2): (i) phosphorylation of LIMK?cofilin?actin remodeling, and/or (ii) MLC phosphatase?MLC?actin remodeling (Olson, 2008; Riento and Ridley, 2003). Either Rho GTPase-dependent or –independent pathways can activate ROCKs, but when ROCKs are active, neuronal structural plasticity is repressed (Bito et al., 2000; Woo and Gomez, 2006; Yuan et al., 2003; Zhang et al., 2003). Pharmacological inhibition with the pan-ROCK inhibitor Fasudil prevents age-related dendritic spine morphology defects (Swanger et al., 2016). ROCK1 and ROCK2 are localized to differing spine compartments under physiological conditions, with ROCK2 more localized to the spine head and ROCK1 to the base of the spine neck (Newell-Litwa et al., 2015). This allows for localized, activity-dependent regulation of dendritic spine dynamics and structural synaptic plasticity through RhoA-ROCK signaling pathways (Hodges et al., 2011; Murakoshi et al., 2011; Newell-Litwa et al., 2015).

The ROCKs have many downstream signaling targets that regulate the functions listed above. Those targets include: vimentin, GFAP, NF-L, CRMP2, ERM, adducin, and NHE1 (Riento and Ridley, 2003). These pathways regulate functions related to the actin assembly network of the cell. Due to high similarity of the kinase domain, it is believed that ROCK1 and ROCK2 serve redundant functions in many cell-signaling pathways (Julian and Olson, 2014). Imperative to this dissertation is how the ROCKs regulate the reorganization of the actin cytoskeleton via MLC and LIM kinase (LIMK). ROCK phosphorylates MLC at Serine-19 and Threonine-18, which then activates myosin ATPase to generate contractile forces in the cell by binding to F-actin (Amano et al., 1996; Vicente-Manzanares et al., 2009). ROCK also regulates this function by phosphorylating and inhibiting the activity of MLC phosphatase, resulting in increased phosphorylation of MLC (Iizuka et al., 1999; Sw‡ard et al., 2000). This allows for specific regulation of actin-myosin contraction dynamics in the cell by ROCK. ROCKs further regulate actin dynamics via the phosphorylation and activation of LIMK 1 and 2 at Threonine-508 and 505, respectively (Ohashi et al., 2000; Scott and Olson, 2007; Sumi et al., 2001a). Activated LIMK then phosphorylates and deactivates the actin severing protein cofilin, leading to the stalling of actin rearrangement in the cell (ADAMS et al., 1996; Arber et al., 1998). These downstream signaling pathways allow for the ROCKs to have exquisite control over actin dynamics in the cell.

Aberrations in ROCK signaling pathways have been demonstrated in AD (Henderson et al., 2016; Heredia et al., 2006; Petratos et al., 2008). Alterations in ROCK signaling are likely induced by A?, which likely activates the RhoA GTPase and its primary downstream effectors ROCK1 and ROCK2 (Henderson et al., 2016; Heredia et al., 2006; Petratos et al., 2008). As outlined above, ROCKs regulate actin–myosin-mediated cytoskeleton contractility (Ishizaki et al., 1996a; Leung et al., 1996b; Leung et al., 1995b; Matsui et al., 1996b; Nakagawa et al., 1996b), and the elevated activity of ROCKs in AD could have detrimental consequences on dendritic spine remodeling (Swanger et al., 2016). Furthermore, ROCK1 and ROCK2 protein levels are increased among MCI and AD patients, implying that ROCKs may contribute to synaptic loss in early disease stages (Henderson et al., 2016; Herskowitz et al., 2013). Pharmacologic studies with Fasudil and Y-27632, the most widely characterized pan-ROCK inhibitors, suggest beneficial effects of ROCK inhibitors in AD models (Rush et al., 2018; Sellers et al., 2018). However, these and other ROCK inhibitors are not isoform-specific and can inhibit other kinases of the AGC family, including protein kinase A (PKA) and protein kinase C (PKC) (DAVIES et al., 2000a). This dissertation will examine critical questions that remain regarding the role of ROCKs in AD and the contribution of ROCK1 or ROCK2 to the observed beneficial effects of pan-ROCK inhibitors. In particular, examining distinct isoform-specific mechanisms by which ROCKs may drive dendritic spine degeneration in MCI and AD.

Did you like this example?

252

11