(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Inka2, a novel Pak4 inhibitor, regulates actin dynamics in neuronal development [1]

['Seiya Yamada', 'Laboratory For Molecular Neurobiology', 'Faculty Of Human Sciences', 'Waseda University', 'Tokorozawa', 'Saitama', 'Tomoya Mizukoshi', 'Akinori Tokunaga', 'Division Of Laboratory Animal Resources', 'Life Science Research Laboratory']

Date: 2022-12 

The actin filament is a fundamental part of the cytoskeleton defining cell morphology and regulating various physiological processes, including filopodia formation and dendritic spinogenesis of neurons. Serine/threonine-protein kinase Pak4, an essential effector, links Rho GTPases to control actin polymerization. Previously, we identified the Inka2 gene, a novel mammalian protein exhibiting sequence similarity to Inka1, which serves as a possible inhibitor for Pak4. Although Inka2 is dominantly expressed in the nervous system and involved in focal-adhesion dynamics, its molecular role remains unclear. Here, we found that Inka2-iBox directly binds to Pak4 catalytic domain to suppress actin polymerization. Inka2 promoted actin depolymerization and inhibited the formation of cellular protrusion caused by Pak4 activation. We further generated the conditional knockout mice of the Inka2 gene. The beta-galactosidase reporter indicated the preferential Inka2 expression in the dorsal forebrain neurons. Cortical pyramidal neurons of Inka2 -/- mice exhibited decreased density and aberrant morphology of dendritic spines with marked activation/phosphorylation of downstream molecules of Pak4 signal cascade, including LIMK and Cofilin. These results uncovered the unexpected function of endogenous Pak4 inhibitor in neurons. Unlike Inka1, Inka2 is a critical mediator for actin reorganization required for dendritic spine development.

Actin filaments are an essential part of the cytoskeleton defining cell morphology and regulating various cellular processes, such as cell migration and synapse formation in the brain. Actin polymerization is controlled by the kinase activity of the Pak4 signaling cascade, including LIMK and Cofilin. Previously, we identified the Inka2 gene, which is strongly expressed in the mammalian central nervous system and a similar sequence as Inka1. Inka1 was reported to serve as a Pak4 inhibitor in cancer cell lines; however, the physiological function of Inka2 is unclear. In this study, we found that (i) Inka2 overexpression inhibits the formation of cell-protrusion caused by Pak4 activation; (ii) Inka2 directly binds to the catalytic domain of Pak4 to inhibit intracellular actin polymerization; (iii) Inka2 is specifically expressed in neurons in the forebrain region, including the cerebral cortex and hippocampus that are known to be essential for brain plasticity, such as learning and memory; and (iv) cortical neurons of Inka2-deficient mice showed decreased synapse formation and abnormal spine morphology, probably due to the marked phosphorylation of LIMK and Cofilin. These results indicate that Inka2 is an endogenous Pak4 inhibitor in neurons required for normal synapse formation through the modulation of actin reorganization.

Funding: S. S. was supported by the Japan Society for the Promotion of Science grants-in-aid (26430042 and 19K06931) and the Waseda University Grants for Special Research Projects (2017K-301 and 2020C-372). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

In addition, the loss or malformation of spines is frequently associated with various psychiatric disorders, including schizophrenia, and autism spectrum disorder, including fragile X syndrome (FXS) caused by the mutation in the fragile X mental retardation 1 (FMR1) gene [ 18 ]. Despite the critical role of actin dynamics in spinogenesis, the mechanisms regulating F-actin polymerization and depolymerization in the dendritic spine remain lesser-known. Several studies have revealed that Pak/LIMK/Cofilin signaling is related to dendritic spine morphology in the pyramidal neurons [ 3 , 19 ]. Increased Pak signaling induces the formation of dendritic spines and the unfunctional dendritic protrusions [ 19 , 20 ]. FMR1 knockout (KO) mouse, a mouse model of FXS, displays the formation of immature and excessive dendritic spines on cortical neurons similar to the symptoms of FXS in humans [ 18 ]. The administration of a Pak inhibitor rescues the dendritic spine defect in FMR1 KO mice and ameliorates the autism-like behavioral phenotypes [ 21 ]. This finding indicates the relevance of the intrinsic system that negatively regulates Pak signaling to prevent abnormal spine formation, facilitating healthy brain development. However, such endogenous Pak inhibitor has not yet been identified in neurons.

In the mammalian CNS development, immature neurons generated from neural stem/progenitor cells (NSPCs) migrate into the cortical plate through sequential morphological changes and polarity formation [ 14 ]. In the postnatal period, pyramidal neurons arriving at the cortical plate begin to form dendritic spines [ 14 – 16 ]. Dendritic spines are actin-based dynamic small protrusions from dendritic shafts that continuously form, change shape, and eliminate throughout life [ 16 ]. Dendritic spines serve as a storage site for synaptic strength and help transmit electrical signals to the neuron cell body. The appropriate density and morphological type of dendritic spines are crucial for many physiological processes. For example, changes in spine morphology affect the efficacy of memory formation and learning [ 17 ].

In our previous study, we identified Inka2 (Fam212b) as an evolutionarily conserved gene in various vertebrates and preferentially expressed in the central nervous system (CNS) [ 10 ]. Inferring from the partial similarity of the predicted protein structure, Inka2 and Inka1 genes seem to constitute the Inka family, characterized by a short conserved motif called “Inka-box (iBox).” Nevertheless, the similarity in function between these two genes remains uncertain because they do not show any sequence similarity besides iBox [ 10 ]. Genetic ablation of Inka1 in mice results in embryonic exencephaly due to failure of neural tube closure [ 11 ]. Inka1 directly binds the catalytic domain of Pak4 in vitro via iBox, repressing the kinase activity of Pak4 [ 12 ]. Inka2 gene is transcriptionally controlled by a tumor suppressor and transcription factor p53 and acts as a potent inhibitor for cancer cell growth. In addition, the direct interaction of Pak4 with Inka2-iBox is attributable to the anti-tumor effect of Inka2 [ 13 ]; however, the physiological function of Inka2 in neurons remains unclear.

The actin filament (F-actin) is a fundamental part of the cytoskeleton regulating various physiological cellular processes involving the membrane dynamics, such as motility and cell morphology. For instance, the barbed ends of polymerizing F-actin push the plasma membrane to form cellular protrusions. The Rho family GTPases, including Rac and Cdc42, regulate this highly orchestrated bidirectional process of actin polymerization and depolymerization through several signaling cascades [ 1 ]. The p-21 activated kinase (Pak)/LIM kinases (LIMK)/Cofilin signaling cascade is the central axis regulating F-actin dynamics [ 2 , 3 ]. The Pak family, an evolutionarily conserved serine/threonine-protein kinase, phosphorylates and activates LIMK upon Rac and Cdc42 signaling. Then, the activated LIMK phosphorylates and inactivates the actin-binding protein Cofilin, which severs and depolymerizes the F-actin, resulting in the formation and stabilization of the actin cytoskeleton. Pak family consists of six isoforms subdivided into two groups based on the domain architecture: group I (Pak 1, 2, and 3) and group II (Pak 4, 5, and 6), both having distinct and overlapping functions. Pak4 is expressed in diverse tissues, and Pak4 knockout mice are embryonically lethal, suggesting the essential roles of Pak4 in normal development [ 4 , 5 ]. Previous studies have shown that constitutively active Pak4 mutant, Pak4 (S445N), phosphorylates LIMK and Cofilin to induce cell morphology changes by altered actin polymerization of C2C12 and NIH3T3 cells [ 2 , 6 , 7 ]. In contrast, Pak4 knockdown induces the cell rounding attributed to the actin depolymerization in tumor cell lines H1229 and HCT116 cells [ 8 , 9 ], suggesting that the control of the fine morphology of cells requires strict regulation of Pak4 activity.

Results

Disruption of Inka2 induces an aberrant dendritic spine formation Although homozygous mutant Inka2-/- mice showed no gross defects in development, the detailed histological analysis of the brain revealed the unusual morphological features of Inka2-/- neurons. Importantly, Nissl staining showed that the thick apical dendrites appeared lost or invisible in a considerable number of layer V pyramidal neurons of Inka2-/- neocortices (Fig 6A and 6B, arrows). To determine the morphological defects of pyramidal neurons, we performed Golgi silver impregnation staining using six-month-old Inka2-/- and littermate mice. In wild-type mice brains, each apical dendrite emerged from the apex of a pyramidal neuron and extended perpendicularly toward the upper layers as a thick process (Fig 6C). However, Inka2-/- pyramidal cells in layer V often had abnormal apical dendrites with thinner, irregularly distorted, and wavy morphology (Fig 6C, arrows). In addition to apical dendrites, such abnormal bending was frequently observed at many dendrite arbors formed by apical dendrites (Fig 6C and 6D, arrows) and basal dendrites (Fig 6D, arrows). Moreover, closer inspection manifested that the number of dendritic spines strikingly decreased in six-month-old Inka2-/- mice (number of spines/10 μm apical dendrite: wild-type, 16.2 ± 0.6; Inka2-/-, 11.8 ± 0.6; number of spines/10 μm basal dendrite: wild-type, 10.6 ± 0.6; Inka2-/-, 7.0 ± 0.5) (Fig 6E and 6F). In general, most spines formed on the dendritic arborization have a globular-shaped head (head) and a thin neck that interconnects the head to the shaft of the dendrite. However, the remaining dendritic spines in Inka2-/- neurons had smaller and irregularly distorted heads (Fig 6D). Decreased density of dendritic spines was observed in juvenile mice (one-month-old), implying the defect in the onset of spine formation (spinogenesis) rather than the degeneration of spines (number of spines per 10 μm of apical dendrite: wild-type, 13.7 ± 0.5; Inka2-/-, 9.3 ± 0.4; spines/10 μm of basal dendrite: wild-type,10.8 ± 0.4; Inka2-/-, 6.8 ± 0.4) (Fig 6D–6F). To exclude the possibility that the effects of the spine abnormalities in Inka2-/- neurons were secondary to neuronal degeneration, we measured the area of the lateral ventricles (LV), thickness of the corpus callosum (CC), cell density of cortical layer V, and number of apoptotic cells in juvenile and adult brains. The results showed no significant changes in the LV area, CC thickness, or number of cortical neurons in the Inka2-/- brains (S5A–S5F Fig). To confirm the impaired spinogenesis of Inka2-/- neurons, the primary cortical neurons (PCNs) were prepared from the embryonic cortices and cultured in vitro. EGFP expression vector was introduced into the dissociated neurons from Inka2-/- or Inka2+/+ embryos to visualize the fine morphology of each neuron. At 3 days in vitro (div), we could not observe any morphological difference between Inka2-/- and wild-type neurons, which exhibited the extension of SMI312+ axon and normal arborization of MAP2+ dendrites (S5G Fig). As early as 7 div, when active synaptogenesis began, post-synaptic density protein 95 (PSD95) and Homer1 (a post-synaptic scaffolding protein used as a synaptic marker for excitatory neurons) expression was observed as a granular signal in wild-type cortical neuron spines. Inka2-/- neurons had fewer PSD95+ or Homer1+ spines (Fig 6G). At 13 div, when most neurons had fully differentiated and made synaptic contacts with one another, wild-type neurons formed a lot of EGFP+ dendritic spines that were colocalized with PSD95 and/or Hoemr1 (Fig 6H). In contrast, Inka2-/- neurons formed fewer PSD95+ or Homer1+ spines (number of PSD95+ spines per 30 μm of dendritic shaft: wild-type, 16.8 ± 0.8; Inka2-/-, 11.8 ± 0.7; number of PSD95+ Homer1+ spines per 30 μm of dendritic shaft: wild-type, 14.7 ± 0.8; Inka2-/-, 10.1 ± 0.6) (Fig 6I and 6J). Altogether, these results indicated a critical function of Inka2 on dendritic spine formation.

Inka2 represses LIMK–Cofilin signaling pathway in neurons Our qPCR analysis showed that Inka2 mRNA was considerably expressed in cultured mouse embryonic fibroblasts (MEFs) (Fig 7A) as well as the PCNs and the cerebral cortex of wild-type mice (Fig 7B) and that its expression was completely eliminated in Inka2-/- MEFs and Inka2-/- cerebral cortex. Meanwhile, a high level of Inka1 expression was observed only in MEFs. This Inka1 expression level was unchanged between wild-type and Inka2-/- MEFs. Both the cerebral cortex and the cultured PCNs exhibited no expression of Inka1 (Fig 7B). These findings suggested redundancy between the two genes in MEFs, while Inka2 solely functions in the neural tissues. We hypothesized that altered downstream pathways of Pak signaling contributed to the spinogenesis defects observed in the Inka2-/- brain. LIMK–Cofilin signaling is a major downstream pathway of Pak; Pak4 mediates the phosphorylation and activation of LIMK (pLIMK), leading to the phosphorylation and inactivation of Cofilin (pCofilin) to promote stabilization and polymerization of the actin cytoskeleton [28]. We first examined the pLIMK and pCofilin levels in MEFs prepared from subcutaneous tissue of wild-type or Inka2-/- embryos. As a result, Inka2-/- MEFs showed no significant change in pLIMK or pCofilin levels compared with wild-type MEFs (Fig 7C–7E). Considering the possibility that Inka1 and Inka2 mutually suppress Pak4 signaling, we further performed the knockdown experiment of the Inka1 gene in wild-type or Inka2-/- MEFs using Inka1 shRNAs (S5H Fig). The double knockdown of Inka1 and Inka2 significantly raised the content of pCofilin, despite the unchanged level of pLIMK (Fig 7C). Considering the redundancy of the two genes in MEFs, these results suggested functional cooperation of Inka1 and Inka2 in Cofilin inactivation and actin dynamics in MEFs, mediated by the LIMK independent pathway. Next, we investigated the phosphorylation levels of LIMK and Cofilin in PCNs prepared from Inka2-/- embryonic cortices. Both pLIMK and pCofilin levels were drastically upregulated in Inka2-/- neurons at 7 div (Fig 7F–7H). Contrary to the MEF case, Inka1 knockdown in PCN showed no additive effect on phosphorylation of LIMK or Cofilin (Fig 7F–7H). This result might imply that Inka1 does not function in neurons, unlike fibroblasts. During the differentiation of cortical neurons, Inka2, but not Inka1, probably modulates the Pak4–LIMK–Cofilin pathway to regulate dendritic spine formation. In non-neuronal cells, such as fibroblasts, the Pak4 pathway might be cooperatively regulated by Inka1 and Inka2. To confirm the in vivo effect of Inka2-/- in LIMK–Cofilin signaling in synapses, we isolated the synaptoneurosome (SN) fraction, in which pre- and post-synaptic sites are enriched (Fig 8A). Concentrated PSD95 protein level indicated that SNs were accurately fractionated from the cerebral cortices of Inka2-/- or wild-type mice (Fig 8B). Both pLIMK and pCofilin levels were drastically elevated in Inka2-/- SN (Fig 8B–8D). Altogether, Inka2 acts as a neuron-specific inhibitor of the Pak4 signaling cascade. During the differentiation of cortical neurons, Inka2, but not Inka1, modulates the Pak4–LIMK–Cofilin pathway to regulate neuronal development, including dendritic spine formation (Fig 8E). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 8. Inka2 depletion in the brain enhances the LIMK-Cofilin signaling cascade in synaptoneurosome. (A) Schematic model of synaptoneurosome (SN). (B–D) Phosphorylation of LIMK and Cofilin were elevated in Inka2-/- SN. Total protein lysate of the cerebral cortex (Ctx lysate) and the SN fraction were prepared from wild-type or Inka2-/- and subjected to immunoblotting with anti-pCofilin, pLIMK, PSD95, and GAPDH antibodies. (C, D) Quantified comparison of pCofilin (C) and pLIMK (D) level normalized by GAPDH content. Five independent experiments were performed. *, P < 0.05; **, P < 0.01; Paired t-test. (E) Model for the Inka2 function in cortical neurons. Inka2, but not Inka1, inhibits Pak4 activity to suppress the LIMK/Cofilin pathway in neurons affecting dendritic spine formation and neuronal development via actin dynamics. https://doi.org/10.1371/journal.pgen.1010438.g008

[END]
---
[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010438

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/