A novel anti-apoptotic role for Cdc42/ACK-1 signaling in neurons

Neurodegenerative diseases such as amyotrophic lateral sclerosis, Alzheimer ’ s and Parkinson ’ s disease are caused by a progressive and aberrant destruction of neurons in the brain and spinal cord. These disorders lack effective long-term treatments that impact the underlying mechanisms of pathogenesis and as a result, existing options focus primarily on alleviating symptomology. Dysregulated programmed cell death (i.e., apoptosis) is a significant contributor to neurodegeneration, and is controlled by a number of different factors. Rho family GTPases are molecular switches with recognized importance in proper neuronal development and migration that have more recently emerged as central regulators of apoptosis and neuronal survival. Here, we investigated a role for the Rho GTPase family member, Cdc42, and its downstream effectors, in neuronal survival and apoptosis. We initially induced apoptosis in primary cultures of rat cerebellar granule neurons (CGNs) by removing both growth factor-containing serum and depolarizing potassium from the cell medium. We then utilized both chemical inhibitors and adenoviral shRNA targeted to Cdc42 to block the function of Cdc42 or its downstream effectors under either control or apoptotic conditions. Our in vitro studies demonstrate that functional inhibition of Cdc42 or its downstream effector, activated Cdc42-associated tyrosine kinase-1 (ACK-1), had no adverse effects on CGN survival under control conditions, but significantly sensitized neurons to cell death under apoptotic conditions. In conclusion, our results suggest a key pro-survival role for Cdc42/ACK-1 signaling in neurons, particularly in regulating neuronal susceptibility to pro-apoptotic stress such as that observed in neurodegenerative disorders.


Introduction
In neurodegenerative disease states, multiple forms of neuronal cell death, including apoptosis, are hyper-activated due to pathophysiological signals such as oxidative stress, mitochondrial dysfunction and excitotoxicity (Martin et al., 1998;Beal, 1998;Chen et al., 2012). Because neurons are post-mitotic, they cannot replenish themselves. The progressive loss of neuronal populations and their underlying circuits in diseases such as Alzheimer's, Parkinson's and amyotrophic lateral sclerosis (ALS) results in the deterioration of cognitive and motor function. Current treatments for these disorders are limited both in number and in efficacy, typically aiming to alleviate symptoms or delay disease onset while having little-to-no effect on the underlying disease pathogenesis. The detrimental and incurable nature of neurodegenerative diseases has prompted researchers to focus on more accurately delineating the factors underlying these disorders and identifying novel molecular targets for therapeutic development.
The Rho GTPase subfamily is a class of proteins belonging to the Ras superfamily of low molecular weight G-proteins. While they are best known for their role in regulating actin cytoskeletal dynamics, these molecular switches have also been implicated in many other functional areas, such as the regulation of various transcription factors (Hall, 2005), development and maintenance of cell polarity (Nobes and Hall, 1999) and cell cycle progression (Olson et al., 1995). Rho GTPases can be activated by many extracellular signals including growth factor receptors, cytokine receptors, G-protein coupled receptors (GPCRs) and integrins (Sinha and Yang, 2008). Once activated, they signal downstream to various effectors to promote biological responses. Rho GTPases are well known for their prominent roles in the nervous system including the regulation of dendritic spine morphogenesis, growth cone motility and axonal migration (reviewed by Linseman and Loucks, 2008).
More recently, a key role in neuronal survival has also been ascribed to two of the three most well-studied Rho family members. Rac1 and RhoA functionally antagonize each other in a survival context, where Rac1 primarily plays a pro-survival role while RhoA displays proapoptotic effects (Linseman and Loucks, 2008;Govek et al., 2005;Stankiewicz and Linseman, 2014). The pro-survival characteristics of Rac1 have been attributed in part, to signaling via its downstream effector p21-activated kinase (PAK), which promotes survival through activation of a mitogen-activated protein kinase MEK1/2-ERK1/2 pathway. This pathway induces degradation of the pro-apoptotic BH3only protein Bim and suppresses pro-apoptotic JAK/STAT signaling in cerebellar granule neurons (CGNs) (Loucks et al., 2006;Stankiewicz et al., 2012;Stankiewicz et al., 2015). In addition, Rac1 activates phosphatidylinositol-3-kinase (PI3K), which stimulates cell survival via Akt-mediated phosphorylation of the pro-apoptotic protein Bad (Datta et al., 1997). In contrast, RhoA propagates neuronal death by activating its major downstream pathway, Rho-associated protein kinase/phosphatase and tensin homolog (ROCK/PTEN), which opposes Rac activation of Akt (Lai et al., 2014). Thus, the balance of pro-survival Rac/PAK signaling and pro-apoptotic Rho/ROCK signaling plays a central role in maintaining cell survival in certain neuronal populations. To date, a role for Cdc42 in neuronal survival has not been clearly described.
Cdc42 transduces information from various extracellular signals including cytokines, growth factors, GPCRs, proteoglycans and integrins in order to exert a number of physiological effects (Symons and Settleman, 2000). The most well characterized function of Cdc42 is its actin polymerizing role which is critical to the formation of filopodia (Gupton and Gertler, 2007). Other functions of Cdc42 include G1/S phase cell cycle progression and mitotic development (Stengel and Zheng, 2011), regulation of cell polarity, cellular migration and chemotaxis, cell fate determination, transcriptional control and intracellular trafficking (Valdés-Mora et al., 2009). Most of what is known about Cdc42 in relation to mammalian physiology comes from studies using dominantnegative or constitutively-active mutants, though more recent genetic studies in mice have provided insight into other cell type-specific and tissue type-specific biological functions of Cdc42 (Melendez et al., 2011). Activated Cdc42-associated tyrosine kinase-1 (ACK-1) is a 120 kDa non-receptor tyrosine kinase that is a specific downstream effector of Cdc42. There are two isoforms (ACK-1 and ACK-2), both of which are expressed throughout the body, but are especially enriched in the brain (Hoffman and Cerione, 2000). Upon Cdc42 activation, ACK-1 stimulates a number of signaling cascades, such as the PI3K/Akt pathway and the androgen receptor to promote the survival and proliferation of cells . It is also involved in regulation of actin cytoskeletal arrangement and cell migration via signaling proteins such as p130Cas and Crk proteins, critical components of integrin-mediated cell motility (Modzelewska et al., 2006). Additionally, ACK-1 has been shown to phosphorylate and activate Wiskott-Aldrich syndrome protein (WASp) as another means of regulating actin dynamics (Yokoyama et al., 2005). In a neuronal context, we described how ACK-1 contributes to neurite outgrowth after muscarinic cholinergic receptor (mAChR) activation in a human neuroblastoma cell line (Linseman et al., 2001b). Finally, ACK-1 has demonstrated anti-apoptotic properties in the Drosophila eye, wherein it interacts with a Grb2 homolog and a transcriptional co-activator to promote proliferative and anti-apoptotic gene transcription (Schoenherr et al., 2012).
Due to their relatively simple isolation process and high level of culture homogeneity (~95 %), primary rat cerebellar granule neurons (CGNs) are widely used as a model system to study neuronal apoptosis (Linseman et al., 2001a). Relevant to the current study, Cdc42, Rac1 and RhoA all demonstrate high expression in various areas of the rat brain including the cerebellum (Olenik et al., 1997). Control growth medium for cultured CGNs contains 25 mM potassium chloride, as it has been shown that this neuronal population requires high levels of extracellular potassium to survive in vitro (D'Mello et al., 1993). This high potassium supplementation and consequent membrane depolarization, allows for a relatively high influx of Ca2+ into the cells, which is important in the activation of various pro-survival transcription factors (e.g., CREB and MEF2). Removal of depolarizing potassium and serum causes CGNs to undergo apoptosis through a principally intrinsic (i.e., mitochondrial) pathway (D'Mello et al., 1993;Linseman et al., 2002).

CGN culture and treatment
Rat cerebellar granule neurons (CGNs) were isolated and cultured from seven-day-old Sprague-Dawley rat pups of both sexes as described previously (Linseman et al., 2001a). CGNs were plated on 35-mm-diameter, six-well plastic dishes coated with poly-L-lysine, at a density of 4.0 × 10 6 cells/well in Basal Medium Eagle's containing 10 % FBS, 25 mM potassium chloride (KCl), 2 mM L-glutamine, and penicillinstreptomycin (100 U/mL/100 μg/mL). Cytosine arabinoside (10 μM) was added to culture medium 24 h after plating to limit the growth of N.C. Punessen et al. non-neuronal cells. The CGNs were incubated in 10 % CO 2 at 37 • C in culture medium for 6 to 7 days prior to experimentation. With this protocol, cultures were approximately 95 % pure for granule neurons. The animal component of this study with respect to the isolation of CGNs from rats was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Denver (IACUC approval #799313).
CGNs undergoing treatment in "healthy" control medium conditions remained in high potassium (25 mM KCl) culture medium containing 10 % FBS. The medium for CGNs undergoing treatment in "stressful" proapoptotic conditions was replaced with serum-free Basal Medium Eagle's containing low potassium (5 mM KCl). Cells were then treated with either a chemical inhibitor to Cdc42 or to one of its downstream effectors using in vitro concentrations derived from the literature (see Results). The final in vitro concentrations of each inhibitor used were as follows: 50 μM ZCL278, 5 μM Casin, 10 μM ML141, 10 μM AIM100, and 1 μM Dasatinib. Stock solutions of each inhibitor were prepared at 1000× concentration in DMSO. The final concentration of DMSO used in the cell culture medium was 0.1 %. For all experiments, untreated 25K + serum (control condition) and 5K-serum (apoptotic condition) treatment groups containing 0.1 % DMSO were used to compare cell death.

Hoechst staining and apoptotic quantification
CGNs in either 25K + serum (control condition) or 5K-serum (apoptotic condition) media were treated with chemical inhibitors for 24 h as described, prior to fixation and staining. Subsequently, the media was aspirated, CGNs were incubated for approximately 45 min at room temperature in 4 % PFM, then washed twice with phosphate-buffered saline (PBS; pH = 7.4), and finally stained with Hoechst 33258 (1 μg/ mL) for visualization of DNA. Prior to imaging, the stain was removed, and PBS was added to each well. All cells were imaged using a Zeiss Axiovert-200 M epifluorescence microscope. Five Hoechst images per well (in duplicate) were captured for each experiment to assay apoptosis. Cells were determined to be apoptotic based on nuclear condensation and/or fragmentation.

Immunocytochemistry and microscopic imaging
CGNs were co-treated in control or apoptotic media for 24 h as described, with or without one of the aforementioned chemical inhibitors. The cells were then washed once with PBS and fixed in 4 % PFM prior to a 1 h incubation in blocking buffer (5 % BSA in 0.2 % triton-X in PBS) at room temperature. This was followed by overnight incubation at 4 • C with either primary antibody against β-tubulin or primary antibody against active caspase 3, prepared in a 1:250 dilution and 1:500 dilution in 2 % BSA in 0.2 % triton-X in PBS, respectively. The following day, the cells were washed 5 times with PBS, and secondary antibodies were then applied for 1 h to each well. FITC-conjugated secondary antibody was used for β-tubulin stained cells, and Cy3-conjugated secondary antibody was used for active caspase stained cells (both prepared in a 1:250 dilution in 2 % BSA in 0.2 % triton-X in PBS with Hoechst stain at 1 μg/ mL). CGNs were then washed once more with PBS, and finally placed in p-phenylenediamine solution to prevent photo-bleaching prior to microscopic imaging. All cells were imaged using a Zeiss Axiovert-200 M epifluorescence microscope, with a minimum of 5 images captured per well (in duplicate for each treatment).

MTT cell viability assay
An MTT assay was used to determine cell viability after treatment with either ZCL278 or AIM100 in both control and apoptotic conditions. CGNs were treated as described previously for 12 h, with or without ZCL278 or AIM100. MTT reagent (12 mM) was then added directly to each well (100 μL/mL), and the cells were incubated at 37 • C and 10 % CO2 for 4 h. After incubation, 4 mL of dimethyl sulfoxide (DMSO) was added to each well to solubilize any purple formazan precipitate, and the CGNs were gently rocked for approximately 1 h at room temperature until all solid product had dissolved. Samples (300 μL) were then taken from each well after solubilization and re-plated (in duplicate) in a 96well clear-bottom plate for colorimetric detection at 570 nm using Gen5 Microplate Reader spectrophotometer software. Absorbance was expressed as a percentage of the untreated control.

Cell lysis and protein assays
Cell lysates were prepared and protein assays were performed prior to both Western blotting and the use of a GTPase-linked immunosorbent assay (G-LISA) kit, which is described below. CGNs were treated with chemical inhibitors in either control or apoptotic conditions as described previously for a period of 12 or 24 h. CGN whole cell lysates were prepared essentially as described previously (Loucks et al., 2006). Briefly, after aspiration of the media, cells were washed once with icecold PBS, and then incubated for 10-15 min in lysis buffer (Wahl buffer including leupeptin and aprotinin) prior to harvesting and a 2 minute centrifugation at 13,000 RPM. The supernatant was then taken and used to determine sample protein concentrations. Protein concentrations (μg/μL) were determined using a commercially available, colorimetric protein assay kit (BCA) and Gen5 Microplate Reader spectrophotometer software detecting at 562 nm.

GTPase-linked immunosorbent assay (G-LISA)
CGNs were treated in triplicate wells containing control or apoptotic media for 24 h with either ZCL278, Casin or ML141 (inhibitors to Cdc42) as described previously. Whole cell lysates were then prepared, and a protein assay was performed using the BCA colorimetric protein assay kit. All CGN lysate concentrations were equalized using ice-cold lysis buffer, and then 50 μL of either equalized lysate, buffer blank or Cdc42 positive control protein were added to replicate wells of a 96-well plate. The plate was then placed on an orbital microplate shaker at 400 RPM at 4 • C for exactly 15 min prior to washing and a 2 minute incubation in antigen presenting buffer at room temperature. Another set of washes was then performed, followed by the addition of anti-Cdc42 primary antibody (diluted 1:20 in antibody dilution buffer) and a 30 minute shaking period at room temperature. After washing again, secondary antibody conjugated to HRP (diluted 1:62.5 in antibody dilution buffer) was added to each well of the plate on the orbital microplate shaker at room temperature for another 30 min. After a final set of washes, an HRP detection reagent mixture was added to each well, and the plate was then incubated at 37 • C for 15 min before adding HRP stop buffer to each well. Sample absorbance was then immediately measured at 490 nm using Gen5 Microplate Reader spectrophotometer software.

Western blotting
CGNs were treated in triplicate wells of control or apoptotic media for 12 h with either ZCL278 or AIM100, whole cell lysates were prepared and a protein assay was conducted, as described previously. Lysate concentrations were then equalized to the same total protein concentration (30 μg) using MilliQ water. 5× Laemmli buffer (SDS, glycerol, Tris, beta-mercaptoethanol and bromophenol blue) was then added to each sample prior to boiling in water in order to break disulfide bonds. Samples were centrifuged for 2 min at 13,000 RPM and kept on ice until gel loading. A 7.5 % polyacrylamide gel was prepared, and each of 10 wells was filled with 1× running buffer (SDS, Tris, glycine and MilliQ water). The wells were then loaded with 125 μL of either 1× Laemmli buffer or equalized lysate, with one well also containing 20 μL of a molecular weight standard. The proteins were then quickly separated by electrophoresis for approximately 1 h at 35 mA, and then more slowly overnight at 7.5 mA. The following day, a PVDF membrane was activated with methanol and rinsed with 1× transfer buffer (SDS, Tris, glycine, MilliQ water and methanol) prior to preparing the transfer stack. The proteins were then transferred to the PVDF membrane for approximately 1.5 h before disassembling the stack and discarding the polyacrylamide gel. The membrane was then placed in blocking buffer (BSA, sodium azide and PBS-T) for 1 h at room temperature to prevent nonspecific binding, after which a primary antibody against ACK-1 (phospho Y284) was diluted 1:1000 in blocking buffer, and applied to the membrane to be incubated at 4 • C overnight. After removal of the primary antibody, the membrane was washed 3 times (15 min each) in phosphate-buffered saline with Tween 20 (PBS-T), and a secondary antibody conjugated to HRP (rabbit, polyclonal) was diluted 1:5000 in PBS-T. The membrane was shaken in secondary antibody for approximately 2 h, and another set of 3 PBS-T washes was performed. The membrane was then soaked in ECL reagent on a shaker at room temperature for approximately 10 min for luminol based detection. Hyperfilm was then exposed to the membrane for approximately 30 s, and the film was developed using a film processor.

Adenoviral shRNA infection and immunocytochemistry
Adenoviral constructs of Cdc42 short-hairpin RNA (shRNA) coexpressing green fluorescence protein (GFP), as well as a GFP coexpressing scrambled shRNA control adenoviral construct were purchased from Vector Biolabs (Malvern, PA). CGNs were infected for 72 h at a multiplicity of infection (MOI) of 1000 (400 × 10 7 viral particles/ well) with either scrambled or Cdc42-targeted adenoviral shRNA. The virus also co-expressed GFP, which was used to gauge the amount of cellular infection. After 48 h of infection in control plating medium, there was an apoptotic medium exchange in half of the plates treated. At 24 h later, the CGNs were fixed to the plate using 4 % PFM, washed twice with PBS, and incubated in blocking buffer (triton-X and 5 % BSA in PBS) for 1 h. Afterwards, primary antibody against Cdc42 (diluted 1:200 in blocking buffer) was applied to the plates for overnight incubation at 4 • C. The following day, the CGNs were washed 3 times with PBS, and a secondary antibody conjugated to Cy3 (diluted 1:250 in PBS) was applied to each well for approximately 2 h at 4 • C. After removal of the secondary antibody, the wells were washed 3 more times with PBS and placed in p-phenylenediamine solution to prevent photo-bleaching prior to microscopic imaging. All cells were imaged using a Zeiss Axiovert-200 M epifluorescence microscope, with a minimum of 15 images (including a total of approximately 200 GFP-positive cells) captured per well (in duplicate for each treatment).

Data analysis
Each experiment was performed either in duplicate or triplicate wells per treatment. Cell treatment and apoptotic quantification data were analyzed using a one-way analysis of variance (ANOVA) with a post hoc Tukey's test. MTT assay, Western blot densitometry, adenoviral shRNA infection and apoptotic quantification data were analyzed using an unpaired t-test. A p-value <0.05 was considered statistically significant. Data represent the means ± standard error of the mean (SEM) for the total number (n) of experiments performed. G-LISA data were analyzed as the mean ± range of 2 experiments.

Chemical inhibition of Cdc42 sensitizes CGNs to cell death under apoptotic stress
Three pharmacologically distinct Cdc42 specific inhibitors, ZCL278, Casin and ML141, were chosen based on dissimilarities in their molecular structures. Optimal in vitro concentrations were derived from literature detailing prior studies. ZCL278 directly targets the binding site for the Cdc42 guanine nucleotide exchange factor, intersectin, as well as GTP/GDP binding, to inhibit Cdc42. At a concentration of 50 μM in vitro, ZCL278 was previously shown to disrupt various Cdc42 directed activities (e.g., actin microspike formation) in both metastatic PC-3 prostate cancer cells and Swiss 3T3 fibroblasts (Friesland et al., 2013). Casin specifically inhibits PIP 2 -dependent actin assembly (Peterson et al., 2006), and was found to reduce the amount of active Cdc42 in aged hematopoietic stem cells at a concentration of 5 μM in vitro (Florian et al., 2012). ML141 is a Cdc42 specific inhibitor with low affinity for other Rho GTPase family members (Surviladze et al., 2010). It was shown to inhibit the migratory phenotype of ovarian carcinoma cell lines OVCA429 and SKOV3ip, attributed to Cdc42 activation, in a Boyden chamber assay at a concentration of 10 μM in vitro (Hong et al., 2013;Ip et al., 2011). Based on these prior studies, we used the following in vitro concentrations of each inhibitor to study Cdc42 function in CGN survival: 50 μM ZCL278, 5 μM Casin and 10 μM ML141.
CGNs were treated for 24 h with one of the three aforementioned Cdc42 inhibitors in either control (25K + serum) or apoptotic (5Kserum) media, and then microscopically inspected for nuclear condensation, nuclear fragmentation, and for disruption of the microtubule network to detect apoptotic changes. Regardless of which compound was used, inhibition of Cdc42 activity did not significantly increase apoptosis in control growth medium. However, all three Cdc42 inhibitors induced significant sensitization of CGNs to apoptosis caused by the removal of depolarizing potassium and serum (Figs. 1-3). Panel A in Figs. 1 and 2 shows the microtubule network in CGNs treated with ZCL278 and Casin, respectively. Panel A in Fig. 3 shows bright field images of CGNs treated with ML141. In all three cases, negligible differences were observed in the microtubule network (Figs. 1A and 2A) or the overall cell morphology (Fig. 3A) of cells incubated in control media in either the absence or presence of Cdc42 inhibitor. Some minor cellular damage was observed when CGNs were incubated in apoptotic medium. However, incubation in apoptotic medium in the presence of a Cdc42 inhibitor induced a marked increase in microtubule fragmentation (Figs. 1A and 2A) and overall disruption of cell morphology (Fig. 3A). Microtubule fragmentation was particularly evident in cells incubated with ZCL278 under apoptotic conditions (Fig. 1A).
Panel B in Figs. 1-3 shows the changes in nuclear morphology observed with each inhibitor. Under control conditions, CGNs displayed large healthy nuclei, regardless of the absence or presence of a Cdc42 inhibitor. As expected, we observed marked nuclear condensation and fragmentation under apoptotic conditions and these morphological effects were exacerbated by addition of each of the three Cdc42 inhibitors (Figs. 1B, 2B, 3B). CGN apoptosis was quantified by scoring cells with condensed and/or fragmented nuclei as apoptotic and these results are quantitatively presented in panel C of Figs. 1-3. Under control conditions, irrespective of the inhibitor used, apoptosis remained between 5 and 15 %. Incubation in apoptotic medium alone caused a significant increase in apoptosis to approximately 45-50 %. Addition of Cdc42 inhibitors consistently induced a statistically significant larger increase in apoptotic cell death than observed in the apoptotic media alone. CGN apoptosis under apoptotic conditions increased from 45 to 50 % in the absence of inhibitors to a mean value of ~80 % with the addition of ZCL278 (Fig. 1C), ~85 % with Casin (Fig. 2C), and ~60 % with ML141 (Fig. 3C). These results show that incubation with any one of three chemically distinct Cdc42 inhibitors has little effect on CGN survival under control conditions, but significantly sensitizes these cells to cell death under apoptotic conditions.

Chemical inhibition of the Cdc42 effector ACK-1 sensitizes CGNs to cell death under apoptotic stress
Two pharmacologically distinct ACK-1 specific inhibitors, AIM100 and Dasatinib, were chosen based on dissimilarities in their molecular structures. Optimal in vitro concentrations were derived from prior studies. X-ray crystallography and high-throughput screening studies characterized the ATP-mimicking ability, and thus ATP-site binding, of AIM100 through which it inhibits ACK-1 (DiMauro et al., 2007). AIM100 has been shown to induce apoptosis in a pancreatic cancer cell line (Panc-1) at an in vitro concentration of 10 μM (Mahajan et al., 2012). AIM100 was also found to inhibit cell growth in two human prostate cancer cell lines (LNCaP and LAPC4) via an increase in the quiescent G0/G1 cell phase . Multiple cancer cell types including prostate have been shown to depend on Cdc42/ACK-1 signaling for cell survival and proliferation . Dasatinib, also sold under the commercially available brand name Sprycel as an oral treatment for patients suffering from chronic myelogenous leukemia, was first noted as an inhibitor of kinases such as Src and Abl (Lombardo et al., 2004). A previous chemical proteomics study in lung cancer cells demonstrated that Dasatinib also inhibits ACK-1 tyrosine kinase activity by decreasing four of its major sites of autophosphorylation (Tyr284, Tyr518, Tyr857 and Tyr858) (Li et al., 2010). Dasatinib was previously shown to decrease growth of two human malignant melanoma cell lines (HT144 and Malme-3M) at an in vitro concentration of 1 μM (Eustace et al., 2008). Based on these previous studies, we chose to use the following in vitro concentrations of each inhibitor to study the role of ACK-1 in CGN survival: 10 μM AIM100 and 1 μM Dasatinib.
CGNs were treated for 24 h with one of the two aforementioned ACK-1 specific inhibitors in either control (25K + serum) or apoptotic (5Kserum) media, and then microscopically inspected for nuclear condensation, nuclear fragmentation, and for disruption of the microtubule network to detect apoptotic changes. Inhibition of ACK-1 with either AIM100 or Dasatinib did not significantly increase apoptosis in control growth medium. However, both ACK-1 inhibitors induced significant sensitization of CGNs to apoptosis caused by the removal of depolarizing potassium and serum . Panel A in Figs. 4 and 5 shows the microtubule network in CGNs treated with AIM100 and bright field apoptosis in CGNs treated with ZCL278 under either control or apoptotic conditions. Cells were counted and scored as apoptotic based on nuclear condensation and/ or fragmentation, and a percentage for this group of cells was calculated. Data were analyzed using one-way ANOVA and Tukey's post-hoc test. *** indicates p < 0.001 compared to the untreated 25K control. ### indicates p < 0.001 compared to the 5K apoptotic control. Data are represented as the mean ± SEM, n = 8 independent experiments performed in duplicate. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) images in CGNs treated with Dasatinib, respectively. Negligible differences were observed in either the microtubule network (Fig. 4A) or the overall cell morphology (Fig. 5A) of cells incubated in control media in either the absence or presence of ACK-1 inhibitor. Some minor cellular damage was observed when CGNs were incubated in apoptotic medium. However, incubation in apoptotic medium in the presence of either ACK-1 inhibitor induced a marked increase in microtubule fragmentation (Fig. 4A) and overall disruption of cell morphology (Fig. 5A). Microtubule fragmentation was particularly evident in cells incubated with AIM100 under apoptotic conditions (Fig. 4A).
Panel B in Figs. 4 and 5 shows the changes in nuclear morphology observed with each inhibitor. Under control conditions, CGNs displayed large healthy nuclei, regardless of the absence or presence of an ACK-1 inhibitor. As expected, we observed marked nuclear condensation and fragmentation under apoptotic conditions and these morphological effects were exacerbated by addition of either ACK-1 inhibitor (Figs. 4B and 5B). CGN apoptosis was quantified by scoring cells with condensed and/or fragmented nuclei as apoptotic and these results are quantitatively presented in panel C of Figs. 4 and 5. Under control conditions, irrespective of the inhibitor used, apoptosis remained between 5 and 15 %. Incubation in apoptotic medium alone caused a significant increase in apoptosis to approximately 45-55 %. Addition of either ACK-1 inhibitor consistently induced a statistically significant larger increase in apoptotic cell death than observed in the apoptotic media alone. CGN apoptosis under apoptotic conditions increased from 45 to 55 % in the absence of inhibitors to a mean value of ~90 % with the addition of AIM100 (Fig. 4C) and ~80 % with Dasatinib (Fig. 5C). These results show that incubation with either of two chemically distinct ACK-1 inhibitors has little effect on CGN survival under control conditions, but significantly sensitizes these cells to cell death under apoptotic conditions. with Casin under either control or apoptotic conditions. Cells were counted and scored as apoptotic based on nuclear condensation and/or fragmentation, and a percentage for this group of cells was calculated. Data were analyzed using one-way ANOVA and Tukey's post-hoc test. *** indicates p < 0.001 compared to the untreated 25K control. ## indicates p < 0.01 compared to the 5K apoptotic control. Data are represented as the mean ± SEM, n = 4 independent experiments performed in duplicate. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Confirmation of the efficacy of the Cdc42 inhibitors, ZCL278, Casin and ML141, and the ACK-1 inhibitor AIM100: analysis of downstream signaling molecules
Although microscopic imaging and analyses were performed to assess inhibitor effects on CGN survival, we wanted to confirm the efficacy of our chemical inhibitors using other experimental methods. A GTPase-linked immunosorbent assay (G-LISA) was performed to determine the efficacy of the inhibitors, ZCL278, Casin and ML141, in decreasing the amount of active (GTP-bound) Cdc42 in CGNs. The Cdc42 G-LISA utilizes a Cdc42-GTP-binding protein linked to the wells of a 96-well plate. Active, GTP-bound Cdc42 in cell lysates binds to the apoptotic conditions. Cells were counted and scored as apoptotic based on nuclear condensation and/or fragmentation, and a percentage for this group of cells was calculated. Data were analyzed using one-way ANOVA and Tukey's post-hoc test. *** indicates p < 0.001 compared to the untreated 25K control. ## indicates p < 0.01 compared to the 5K apoptotic control. Data are represented as the mean ± SEM, n = 11 independent experiments performed in duplicate.
wells while inactive GDP-bound Cdc42 is removed during washing steps. The bound active Cdc42 is detected with a Cdc42 specific antibody. There was no apparent difference in the amount of active (GTP-bound) Cdc42 detected in CGNs incubated under control (25K + serum) or apoptotic (5K-serum) conditions (Fig. 6A). However, incubation with any one of the three Cdc42 inhibitors caused a 50-75 % decrease in the mean amount of active (GTP-bound) Cdc42 detected in CGNs under apoptotic conditions, indicating that the inhibitors effectively blocked Cdc42 function (Fig. 6A).
Next, we evaluated the capacity of AIM100 to inhibit the activation of ACK-1. CGNs were incubated in the absence or presence of AIM100 and cell lysates were immunoblotted to detect ACK-1 phosphorylated on the autophosphorylation site (Tyr284) and total ACK-1. There was no apparent difference in the amount of phosphorylated ACK-1 detected in CGNs incubated under control (25K + serum) or apoptotic (5K-serum) conditions (Fig. 6B). However, incubation with AIM100 caused a marked reduction in the amount of phosphorylated ACK-1 detected in CGNs under apoptotic conditions, indicating that this inhibitor effectively blocked ACK-1 tyrosine kinase activity (Fig. 6B). Interestingly, incubation with the Cdc42 inhibitor, ZCL278, was not as effective as AIM100 at decreasing phosphorylated ACK-1. Perhaps this is because ZCL278 did not completely block Cdc42 function, as shown in the G-LISA experiment described above.
Finally, we focused our attention on analyzing signaling molecules downstream of Cdc42 and/or ACK-1. We initially considered the key pro-survival molecules, PAK, AKT and MAPK. However, PAK can be activated by both Rac and Cdc42, so it seemed unlikely that the Cdc42 effects we observed were mediated predominantly through this pathway. In a similar manner, we have shown that both AKT and MAPK are regulated in CGNs to at least some extent by Rac signaling (Loucks et al., 2006;Stankiewicz et al., 2015). Therefore, we did not think that these pathways could account for the distinct effects of Cdc42 on CGN survival. However, another protein known as WW domain containing oxidoreductase (Wwox) is a tumor suppressor that has been shown to be an ACK-1-interacting protein and is negatively regulated by ACK-1 signaling in prostate cancer cells (Mahajan et al., 2005). This brought up the interesting possibility that Cdc42/ACK-1 signaling might promote CGN survival by negatively regulating Wwox expression and thereby preventing its pro-apoptotic actions. Therefore, we predicted that inhibiting Cdc42 or ACK-1 in CGNs under apoptotic conditions may lead to an upregulation of Wwox expression which would in turn, cause a potentiation of apoptosis. In contrast to this prediction, although Wwox expression was unaffected by switching the CGNs from healthy control (25K + serum) to apoptotic (5K-serum) medium, the expression of this tumor suppressor was markedly reduced by the addition of either the Cdc42 inhibitor, ZCL278, or the ACK-1 inhibitor, AIM100 (Fig. 6C). These findings suggest that regulation of Wwox expression is an unlikely mechanism for the pro-survival effects of Cdc42/ACK-1 signaling in CGNs.

Assessment of cell death as a measure of mitochondrial metabolic function
Changes in chromatin, such as nuclear condensation and fragmentation, are morphological indicators of apoptotic cell death. Using this morphological criteria, we have established that chemical inhibitors of either Cdc42 or its downstream effector, ACK-1, significantly sensitize CGNs to cell death under apoptotic conditions. To confirm these findings using a biochemical measure of cell viability, we performed MTT assays. In viable cells, the tetrazolium dye MTT (yellow) is reduced to its insoluble formazan (purple) by NAD(P)H-dependent cellular oxidoreductase enzymes found primarily in mitochondria (Riss et al., 2016). With a significant reduction in the number of viable cells, the MTT reagent will remain yellow, indicating that cell death has occurred. CGNs   Fig. 4. Apoptotic sensitization of CGNs treated with the ACK-1 specific inhibitor AIM100. (A) Representative photomicrographs showing CGNs with or without AIM100 treatment. Leftmost panels show AIM100 treatment in control (25K + serum) conditions. Rightmost panels show AIM100 treatment in apoptotic (5K-serum) conditions. Green indicates β-tubulin. Blue indicates Hoechst staining of cell nuclei. (B) Representative photomicrographs of CGNs with or without AIM100 treatment, showing black and white Hoechst fluorescence images to visualize nuclear morphology. Leftmost panels show AIM100 treatment in control conditions. Rightmost panels show AIM100 treatment in apoptotic conditions. Images in (A) and (B) show 8 different fields. Scale bar = 30 μm. (C) Quantitative assessment of cellular apoptosis in CGNs treated with AIM100 under either control or apoptotic conditions. Cells were counted and scored as apoptotic based on nuclear condensation and/or fragmentation, and a percentage for this group of cells was calculated. Data were analyzed using one-way ANOVA and Tukey's post-hoc test. *** indicates p < 0.001 compared to the untreated 25K control. ### indicates p < 0.001 compared to the 5K apoptotic control. Data are represented as the mean ± SEM, n = 5 independent experiments performed in duplicate. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) were incubated with either ZCL278 or AIM100 in either control (25K + serum) or apoptotic (5K-serum) media, followed by incubation with MTT reagent. Afterwards, the absorbance of the purple formazan derivative at 570 nm was measured and converted to cell viability as a percentage of the control medium condition. A representative image of a typical MTT assay is shown in Fig. 7A. Quantitative analysis showed that although incubation with either ZCL278 or AIM100 tended to decrease viability slightly in control medium, neither effect was statistically significant (Fig. 7B,C). As expected, cell viability decreased markedly in apoptotic medium to a mean value of approximately 35-40 % of that observed under control conditions. Incubation with either ZCL278 or AIM100 significantly sensitized CGNs to cell death under apoptotic conditions, as indicated by statistically significant decreases in cell viability to ≤10 % of control values (Fig. 7B,C). These results confirm the data obtained using nuclear morphology as an index of CGN apoptosis and verify that chemical inhibitors of either Cdc42 or its downstream effector, ACK-1, indeed significantly sensitize CGNs to cell death under apoptotic conditions.

Knockdown of Cdc42 expression sensitizes CGNs to cell death under apoptotic stress
Chemical inhibitors are useful tools to reveal novel functions of proteins such as a role for Cdc42 in neuronal survival. However, the relative lack of specificity of chemical inhibitors and potential off target effects could lead to inaccurate conclusions. In order to definitively establish a role for Cdc42 in neuronal survival, we infected CGNs with an adenoviral vector that expresses shRNA targeted to Cdc42. This viral construct also co-expresses green fluorescent protein (GFP), which enabled us to visualize infected cells. CGNs were incubated with either an adenovirus expressing shRNA to Cdc42 or a scrambled control shRNA in control (25K + serum) culture medium for 48 h, before switching the medium to either control or apoptotic (5K-serum) conditions for a further 24 h. CGNs were fixed and stained at either 48 h or 72 h to measure the expression of Cdc42 and to quantify apoptosis, respectively. Following a total of 72 h of adenoviral infection, the percentage of infected (GFP-positive) CGNs was calculated for each construct by counting duplicate wells over 3 independent experiments. The percentage GFP-positive cells were 46 ± 8 % for the scrambled control or apoptotic conditions. Cells were counted and scored as apoptotic based on nuclear condensation and/or fragmentation, and a percentage for this group of cells was calculated. Data were analyzed using one-way ANOVA and Tukey's post-hoc test. *** indicates p < 0.001 compared to the untreated 25K control. ### indicates p < 0.001 compared to the 5K apoptotic control. Data are represented as the mean ± SEM, n = 6 independent experiments performed in duplicate.
The efficiency of shRNA-mediated knockdown of Cdc42 expression is shown in Fig. 8. The top two rows show two distinct fields of CGNs infected for 48 h with the scrambled control shRNA. Infected (GFPpositive) cells are indicated by the arrows and display very similar Cdc42 staining to other CGNs in the same field that were not infected (Fig. 8, top two rows). The bottom two rows show two distinct fields of CGNs infected for 48 h with the shRNA targeted to Cdc42. Infected (GFPpositive) cells are indicated by the arrows and are essentially devoid of Cdc42 staining. Note that other CGNs in the same field that were not infected showed normal Cdc42 staining (Fig. 8, bottom two rows). These data show that incubation for 48 h with the adenoviral vector expressing shRNA targeted to Cdc42 efficiently knocked down Cdc42 expression in CGNs.
As indicated above, approximately 46 % of the CGNs infected with the adenovirus expressing the scrambled control shRNA and 39 % of those infected with the adenovirus expressing the shRNA targeted to Cdc42 became infected (were GFP-positive) and were able to be quantified for apoptosis. Representative images of Cdc42 expression and nuclear morphology are shown at 24 h after switching the medium to either control or apoptotic conditions (Fig. 9A). The top two rows show two distinct fields of CGNs infected for 48 h with the scrambled control shRNA and then switched to apoptotic medium for an additional 24 h. Infected (GFP-positive) cells are indicated by the arrows and display very similar Cdc42 staining to other CGNs in the same field that were not infected. Moreover, these cells for the most part, had healthy nuclei (Fig. 9A, top two rows). The bottom two rows show two distinct fields of CGNs infected for 48 h with the shRNA targeted to Cdc42 and then switched to apoptotic medium for an additional 24 h. Infected (GFPpositive) cells are indicated by the arrows and are essentially devoid of Cdc42 staining. Moreover, these cells uniformly had condensed and/or fragmented nuclei indicative of apoptosis. Note that other CGNs in the same field that were not infected showed normal Cdc42 staining and many had healthy nuclei (Fig. 9A, bottom two rows). Finally, quantification of apoptosis demonstrated that CGNs infected with the scrambled control shRNA displayed an approximate 20-fold increase in apoptosis when switched to apoptotic medium compared to control medium. In contrast, CGNs infected with shRNA targeted to Cdc42 showed an approximate 80-fold increase in apoptosis when switched to apoptotic medium compared to control medium. Thus, CGNs in which Cdc42 expression was knocked down with shRNA, displayed a 4-fold increase in sensitivity to cell death under apoptotic conditions (Fig. 9B). These findings further validate our observations with Cdc42 inhibitors and support the conclusion that Cdc42 plays a key pro-survival role in CGNs.

Discussion
Considering their wide array of biological functions, such as regulation of the actin cytoskeleton, transcription factors, cell cycle progression, and cell motility, it is not surprising that new roles for the Rho GTPase family as key regulators of neuronal survival and apoptosis have emerged. Prior research has strongly established a pro-survival role for Rac and a pro-apoptotic role for Rho in neurons, although there are cell type-specific exceptions to these generalizations (e.g., apoptosis of sympathetic neurons following nerve growth factor deprivation is driven by Cdc42/Rac signaling; Bazenet et al., 1998;Xu et al., 2001). We have established Rac as an essential regulator of CGN survival (Linseman et al., 2001a;Le et al., 2005;Loucks et al., 2006;Stankiewicz et al., 2012;Stankiewicz et al., 2015). Moreover, the opposing roles of Rac and Rho on motor neuron survival have previously been demonstrated and the activity of these GTPases is dysregulated in neurodegenerative disorders like ALS (Jacquier et al., 2006;Stankiewicz et al., 2020). In contrast to these studies on Rac and Rho, the relative contribution of Cdc42 to neuronal survival has thus far been largely overlooked.
In the current study, we evaluated the involvement of Cdc42 in CGN survival. Under healthy control conditions (i.e., in the presence of growth factor-containing serum and depolarizing extracellular potassium), CGN survival was largely unaffected by inhibitors of Cdc42. These findings are in agreement with our previous work showing that adenoviral expression of dominant-negative Cdc42 has no effect on CGN survival under these same healthy growth conditions (Le et al., 2005). These results suggest that Cdc42 function is not essential to maintain CGN survival under healthy conditions and indicate that survival signals Fig. 6. Detection of active (GTP-bound) Cdc42, phosphorylated ACK-1, and Wwox expression after chemical inhibition. (A) G-LISA data of CGNs treated with Cdc42 specific inhibitors ZCL278, Casin and ML141 for 24 h in either control (25K + serum) or apoptotic (5K-serum) media. Data represent the mean ± the range of duplicate samples from a single experiment. (B) Representative Western blot for detection of both phosphorylated ACK-1 (phospho-Y284) and total ACK-1 (to serve as a loading control) after 12 h of treatment with either ZCL278 or AIM100. The blot shown is representative of similar results obtained in two independent experiments. (C) Representative Western blot for detection of Wwox and actin (to serve as a loading control) after 24 h of treatment with either ZCL278 or AIM100. The blot shown is representative of similar results obtained in two independent experiments. produced by serum and depolarizing potassium are sufficient to overcome any deleterious effects of Cdc42 inhibition. This is in marked contrast to the essential role of Rac in CGN survival as evidenced by the fact that Rac inhibition (either via Clostridial cytotoxins, chemical inhibitors, or adenoviral expression of dominant-negative Rac) is entirely sufficient to induce profound CGN apoptosis in the presence of serum and depolarizing potassium (Linseman et al., 2001a;Le et al., 2005;Stankiewicz et al., 2015).
Despite the fact that inhibitors of Cdc42 did not influence CGN survival under healthy control conditions, each of these inhibitors significantly sensitized CGNs to cell death under apoptotic conditions (i.e., in the absence of growth factor-containing serum and depolarizing extracellular potassium). This sensitizing effect was replicated by shRNAmediated knockdown of Cdc42 expression, validating the results obtained with chemical inhibition of Cdc42. These findings demonstrate that Cdc42 does indeed play a pro-survival role in CGNs and furthermore, loss of Cdc42 function compromises the capacity of CGNs to withstand an apoptotic challenge.
Cdc42 signals to a large number of downstream effector molecules including but not limited to, ACK-1, p21-activated kinase (PAK), IQGAP, and neural Wiskott-Aldrich syndrome protein (N-WASp). Some of these effectors, PAK and IQGAP, can be activated by both Cdc42 and Rac, whereas ACK-1 and N-WASp are unique effectors of Cdc42. PAK is known to play a pro-survival role in CGNs and so we did not examine it in this study (Johnson and D'Mello, 2005). Incubation of CGNs with two chemically distinct inhibitors of N-WASp (187-1 and Wiskostatin) had no significant effect on CGN survival or apoptosis under either control or apoptotic conditions, suggesting that N-WASp is not involved in Cdc42mediated neuronal survival (data not shown). However, two chemically distinct inhibitors of the tyrosine kinase ACK-1 (AIM100 and Dasatinib) did mimic the effects of Cdc42 inhibitors; they each had no effect on CGN survival under healthy control conditions but significantly sensitized CGNs to cell death under apoptotic conditions. These findings reveal a novel pro-survival role for a Cdc42/ACK-1 signaling pathway in CGNs. Intriguingly, Cdc42/ACK-1 signaling appears to play an antiapoptotic role in various cancer cells and overexpression of these proteins is associated with poor prognosis in certain cancers (Nur-E-Kamal et al., 2005;Arias-Romero and Chernoff, 2013;Zhu et al., 2021). In this context, we evaluated the tumor suppressor Wwox as one potential target of pro-survival Cdc42/ACK-1 signaling in CGNs, as this protein is negatively regulated by ACK-1 phosphorylation in prostate cancer cells (Mahajan et al., 2005). However, inhibitors of Cdc42 and ACK-1 induced a pronounced downregulation of Wwox in CGNs under apoptotic conditions, a result opposite of what would be expected if Wwox expression was regulated by ACK-1 in CGNs as it is in prostate cancer cells. Therefore, it seems unlikely that regulation of Wwox expression is the mechanism responsible for the pro-survival effects of Cdc42/ACK-1 signaling in CGNs. Identification of the precise downstream signals involved in this effect will require further investigation. Fig. 7. Assessment of cell viability as a measure of mitochondrial function in CGNs treated with ZCL278 and AIM100. (A) Representative MTT assay image of CGNs in either control (25K + serum) or apoptotic (5K-serum) media that have been treated with ZCL278 or AIM100. (B-C) Quantification of cell viability as a function of MTT absorbance data. The 25K control was set at 100 % cell viability and other treatments were calculated relative to that control. Data were analyzed using an unpaired t-test. "ns" indicates not significantly different compared to the 25K control, ** indicates p < 0.01 compared to the 5K control, n = 5 independent experiments performed in duplicate.

Conclusion
In summary, our results demonstrate that while Cdc42 function may not be an essential component to neuronal survival under healthy conditions, it does play a pro-survival role in neurons exposed to apoptotic stress. Moreover, the pro-survival effects of Cdc42 in CGNs appear to be mediated via its downstream effector ACK-1. Future studies will investigate the precise signaling cascade involved and how this pathway antagonizes pro-apoptotic signals, such as the expression of pro-death Bcl-2 family members (BH3-only proteins). In a broader sense, our findings suggest that as neuronal populations are progressively exposed to increasing levels of apoptotic stress during the course of neurodegenerative diseases; Cdc42/ACK-1 signaling could be altered. Future experiments are required to determine if this is indeed the case. Finally, it will be of interest to determine if Cdc42/ACK-1 signaling plays a prosurvival role in other types of neurons besides CGNs and whether this pathway is altered in other models of neurodegeneration.

Funding
Funding for this project was provided by internal Professional Research Opportunity Fund (PROF) and Knoebel Institute for Healthy Aging (KIHA) grants from the University of Denver to DAL. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. (caption on next column) Fig. 9. Effects of Cdc42 knockdown on CGN apoptosis in 25K control and 5K apoptotic conditions. (A) Immunocytochemistry for Cdc42 and assessment of CGN apoptosis after 72 h adenoviral knockdown and an apoptotic media exchange. Representative photomicrographs of CGNs that were incubated in control (25K + serum) media for an initial 48 h infection period with adenoviral scrambled (scrm) or Cdc42-targeted shRNA constructs co-expressing GFP, followed by a switch to apoptotic (5K-serum) media for an additional 24 h. From left to right, panels show virally infected cells (DAPI/GFP), Cdc42 content within cells (DAPI/Cdc42), black and white DAPI staining to highlight nuclear morphology, and a merged image. Blue fluorescence indicates nuclear staining. Green fluorescence indicates positive GFP expression and viral infection. Red fluorescence indicates Cdc42. White arrows designate virally infected cells. Top two rows: CGNs infected with scrambled (scrm) shRNA. Bottom two rows: CGNs infected with target Cdc42 shRNA. Note the lack of Cdc42 staining (red fluorescence) and prominent apoptotic nuclei in cells infected with shRNA targeted to Cdc42. Scale bar = 10 μm. (B) Quantification of apoptosis in CGNs after 72-hour adenoviral knockdown of Cdc42 and an apoptotic media exchange. Quantitative assessment of apoptosis (expressed as a fold change in the amount of apoptosis observed under 5K apoptotic vs 25K control conditions) in CGNs that were incubated in control (25K + serum) media for an initial 48 hour infection period with adenoviral scrambled (Ad-GFP-scrm) or Cdc42-targeted (Ad-GFP-target) shRNA constructs co-expressing GFP, followed by a switch to control (25K + serum) or apoptotic (5K-serum) media for an additional 24 h. The switch to apoptotic media caused an approximate 20-fold increase in apoptosis of CGNs infected with the scrambled control shRNA, whereas CGNs infected with the target Cdc42 shRNA showed an approximate 80-fold increase in apoptosis when switched to apoptotic media for 24 h. Data were analyzed using an unpaired t-test and are represented as the mean ± SEM, n = 3 independent experiments performed in triplicate. ** indicates p < 0.01 compared to scrambled control shRNA infection. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)