ROS produced by NOX promote the neurite growth in a PI3K/Akt independent manner

Reactive oxygen species (ROS) function as signaling molecules in several physiologic and pathologic processes. In central nervous system, ROS are critical for differentiation, migration, polarization, and neurite growth. These actions are mediated by reversible oxidation of target proteins. On the other hand, PI3K/Akt signaling pathway is susceptible to be modulated by ROS and it has been implicated in neurite growth. In this study, we evaluated the participation of ROS in the neurite growth of cultured rat cerebellar granule neurons (CGN), as well as the possible regulation of the PI3K/Akt pathway by ROS during neurite outgrowth. For this purpose, CGN were treated with cellular or mitochondrial antioxidants, or an NOX inhibitor and neurite growth was evaluated. Moreover, to assess the participation Akt in this process, the p‐Akt levels were measured in CGN treated with antioxidants or a NOX inhibitor. The effect of antioxidants on the neurite growth in the presence of a PI3K inhibitor was also measured. We found that cellular antioxidants and the NOX inhibitor decreased the neurite growth, but not the mitochondrial antioxidant. Interestingly, the antioxidants increased the p‐Akt levels; however, the effect of antioxidants on neurite growth was no dependent on the Akt activity since the inhibitor of PI3K did not modify the antioxidant action on neurite growth. Our results show that the PI3K/Akt pathway participates in neurite growth and that ROS produced by NOX could function as signals in this process; however, this action is not mediated by a redox regulation of Akt activity.

ROS are produced by mitochondria and the NADPH oxidases (NOX).NOX is a family of enzymes that produce superoxide anion and H 2 O 2 from oxygen.These enzymes are expressed in the central nervous system early during the brain development (Coyoy et al., 2013).
Additionally, several studies have shown a critical role of ROS and NOX in the proliferation of neural cells (Le Belle et al., 2011), the differentiation of neural progenitors (Tsatmali et al., 2006), the migration of neurons (Ali et al., 2019), the programmed neuronal death (Zaragoza-Campillo & Morán, 2017), the neural polarization and the neurite growth (Wilson et al., 2015).
Regarding neurite growth, it has been observed that MICAL, an enzyme that catalyzes redox reactions to control cytoskeleton dynamics (Van Battum et al., 2014) induces the generation of H 2 O 2 in response to specific signals involved in axonal growth such as semaphorin 3A that allows the collapse of the growth cone (Schmidt et al., 2008).In neuronal cultures of Aplysia, the growth cones express NOX2 and show high concentrations of ROS and the treatment with NOX inhibitors or lipoxygenase inhibitors induce a marked decrease of the axonal growth (Munnamalai et al., 2014;Munnamalai & Suter, 2009).Similarly, the treatment of different mammalian neuronal types with antioxidants or NOX inhibitors decreases the neurite growth (Chandrasekaran et al., 2015;Wang et al., 2011;Wilson et al., 2015) and NOX2 knockout neurons show shorter neurites than wild type neurons (Olguín-Albuerne & Morán, 2015;Weaver et al., 2018).
The mechanisms underlying these ROS actions are not fully understood, but some evidence suggest that ROS could regulate directly or indirectly on some proteins of signaling pathways involved in neuronal differentiation (Le Belle et al., 2011).In PC12, an increase in ROS levels induces the activation of the pathway RAS-MAPK and the activity of Erk1/2 (Goldsmit et al., 2001;Gopalakrishna et al., 2008) and PI3K (Kamata et al., 2005), which are essential for neuronal differentiation.These actions can be mediated by reversible modifications of specific cysteine residues of some proteins, leading to S-glutathionylation or inter-and intra-molecular disulfide bridges formation (Fra et al., 2017;Thannickal & Fanburg, 2000).In immature neurons, hydrogen peroxide produced by NOX indirectly activate the PI3K/Akt/mTOR pathway through of inhibition and the formation of a disulfide bridge in PTEN, an antagonistic enzyme of this pathway (Leslie, 2006).
In addition to the neuronal differentiation, the PI3K/Akt pathway also participates in the neuronal polarization and neurite growth (Jin et al., 2012;Read & Gorman, 2009).For example, in several neuronal types and cell lines the pharmacological or genetic inhibition of this pathway significantly reduces the neurite growth (Jiang et al., 2005;Markus et al., 2002;Wang et al., 2017;Yan et al., 2006).As mentioned previously, this pathway can be modulated by ROS.Therefore, in this study we evaluated the participation of ROS in the neurite growth and the possible participation of the PI3K/Akt pathway in this action.Using antioxidants and NOX inhibitors, we showed that ROS produced by NOX are required for the neurite growth of cerebellar granule neurons (CGN).We also found that the inactivation of PI3K did not affect the antioxidants actions.These results suggest that ROS and the PI3K/Akt pathway activity participate in the neurite growth, but these actions are not related.It is known that several pathologies related to neurodevelopmental disorders are linked to an abnormal neuritic growth during critical times of brain development (Liaci et al., 2021;Prem et al., 2020); thus, the understanding of the mechanisms involved in this process could be useful in finding some therapies to treat some of the neuropathologies.

| Plasmids
GFP plasmid was purchased from Lonza, p47phox plasmid was purchased from Sino Biological, and DNp22phox plasmid were kindly

Significance
Abnormal neurite growth is associated with pathologies of the neurodevelopment.The study of the mechanisms that participate in these processes could be useful to understand and propose therapies to treat these pathologies.
Here, we show that although both Akt activity and ROS produced by a NOX participate in the neurite growth of CGN this effect is not mediated by the redox regulation of Akt activity.This study contributes to the understanding of the role of ROS as physiological signals in the development of CGN.donated by Christian González-Billault (Universidad de Chile, Santiago de Chile).DNp22phox plasmid was changed to pIRES-hrGFP-1a vector (Agilent Technologies) for identification of transfected neurons.

| Cell culture
All animals used in the present study were treated according to the accepted standards of animal care and with the procedures Eight-day-old Wistar rats of regardless the sex were used for cerebellar granule neuron cultures.The whole litter were housed in enrichment cages and kept under controlled temperature (22-24°C) and humidity (40%-70%), with a regulated 12-h light-dark cycle, with water and food ad libitum.Cerebellar granule neuron cultures were prepared as previously described (Moran & Patel, 1989).Briefly, cell suspensions dissociated from cerebella of 8-day-old Wistar rats from both sexes, were plated at a density of 265 × 10 3 cells/cm 2 in plastic dishes coated previously with poly-l-lysine (5 μg/mL).The culture medium contained basal Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM glutamine, 25 mM KCl, 50 μg/mL streptomycin, and 50 U/mL penicillin.The cultures were kept at 37°C in an atmosphere of CO 2 (5%) and saturated air with water vapor (95%).
Cytosine arabinoside (10 μM) was added 24 h after seeding to prevent the proliferation of non-neuronal cells.

| Cell transfection
For the experiments of gain-or loss-of-function of the NOX2 subunits p47 phox and p22 phox , respectively, CGNs were transfected with either a p47 phox -GFP plasmid or the dominant negative of p22 phox (DNpr22-GFP) plasmid, respectively.GFP transfection was used as a control.CGN were electroporated before plating by using a Lonza nucleofector 2B.Briefly, 5 × 10 6 cells were centrifuged at 82× g for 5 min.The pellet was resuspended in 100 μL transfection buffer (120 mM KCl, 10 mM KH 2 PO 4 , 2 mM EGTA, 25 mM HEPES, 5 mM MgCl 2 , .5 mM CaCl 2 ) and mixed with the plasmid.Cells were transferred to cuvette, electroporated using the C-13 program and resuspended in basal Eagle's medium.CGN were transfected and plated mixed with untransfected cells in the following proportion: at 2 days after kept the culture (in vitro day, DIV), 6 × 10 4 transfected and 1.5 × 10 6 untransfected cells; at 3 DIV, 3 × 10 4 transfected and 1.5 × 10 6 untransfected cells.The DNp22 transfected cells have a low expression of GFP, for this reason, these cells were plated in the proportion of 2 × 10 5 transfected for 1.5 × 10 6 untransfected cells.

| Neurite outgrowth evaluation
CGN transfected were treated at different times as follows: The day of culture, CGN were treated with 50 μM Euk-134, 10 μM Ebselen, catalase (20 μg/mL), 10 μM MitoTEMPO or 5 μM Vas2870 and neurite growth was evaluated at 2 or 3 DIV.The first day after seeding, CGN were treated with .1 μM H 2 O 2 , 50 μM Euk-134, 10 μM LY294002 or pretreated with 10 μM LY294002 for 30 min and treated with 50 μM Euk-134 and neurite growth was evaluated at 2 DIV.To second day after seeding, CGN were treated with 50 μM Euk-134 or 5 μM Vas2870 and neurite growth were evaluated at 3 DIV.Furthermore, CGN were pretreated with 10 μM LY294002 for 30 min and treated with 50 μM Euk-134 the second day after seeding and neurite growth was evaluated at 3 DIV.Some CGN were treated with water as a control for Euk-134 and catalase, and DMSO as controls for Ebselen, MitoTEMPO, Vas2870 and LY294002 treatments.The neurite growth was evaluated at the indicated times (Figure 1a).For experiments of gain or loss of function the CGN were transfected as explained previously and the neurite growth was evaluated at 3 DIV (Figure 1b).Cells were fixed with 4% paraformaldehyde for 15 min and washed three times in PBS.CGN were photographed in a fluorescence microscope Nikon eclipse Ti-U, using an excitation filter wavelength/dichromatic mirror cut-on wavelength/barrier filter wavelength of 450-490/500/515.For experiments of gain or loss function, the cells were photographed in a two-photon microscope Zeiss LSM710.The neurites were measured manually with the plugin Neuron J of Image J (Meijering et al., 2004).The measurement of each neurite corresponds to two replicates of at least three independent experiments.
Viable cells have intracellular esterases that convert the permeant calcein-AM to calcein, a green fluorescent product that is impermeant in living cells, whereas IP crosses the plasmatic membrane of damaged cells.Red fluorescence signal increases when IP bind to DNA.CGN were treated with 50 μM Euk-134, 10 μM Ebselen, Catalase (20 μg/ mL), 10 μM MitoTEMPO or 5 μM Vas2870 the day of culture and cell viability was evaluated at 2 or 3 DIV.In some experiments, CGN were treated with .1 μM H 2 O 2 or 10 μM LY294002 the first day of the culture and cell viability was evaluated at 2 DIV (Figure 1a).
CGN were incubated with calcein-AM (.5 μM) and PI (5 μM) for 15 min at 37°C.Cells were photographed in a fluorescence microscope Nikon eclipse Ti-U using filters with the following characteristics: excitation filter wavelength/dichromatic mirror cut-on wavelength/barrier filter wavelengths of 450-490/500/515 and 510-560/565/590 nm for calcein and propidium iodide, respectively.Quantification of calcein-AM and propidium Iodide-positive cells was performed using the image-processing software ImageJ.The results are expressed as the percentage of viable cells (calcein-positive cells) from the total number of cells (calcein and propidium iodide-positive cells).At least three independent experiments were analyzed with two replicates for each experimental condition and three random fields for each replicate.

| Immunoblot
CGN were treated immediately after plated with 50 μM Euk-134 or 5 μM Vas2870 for 24, 48 and 72 h.Some cultures were treated at 2 DIV with 50 μM Euk-134, catalase (20 μg/mL) or 5 μM Vas2870 for 5, 10, 20, 40, 60 or 90 min or with .1, 1, 10, 100 or 200 μM H 2 O 2 for 1 h (Figure 1c).CGN cultures were washed with ice-cold PBS, F I G U R E 1 Experimental design.CGN were cultured from cerebella obtained of 8-day-old rats and were (a) treated with antioxidants or an Akt inhibitor or (b) transfected with p47 phox or DNp22 phox for the evaluation of neurite outgrowth and cell viability.Treatments for evaluated the participation of ROS in the neurite growth are shown in black.Treatments for evaluated the participation of PI3K/Akt signaling pathway in the ROS effects on neurite growth are shown in blue.(c) Some cultures were treated with antioxidants and the levels of proteins and ROS were measured.homogenized in lysis buffer (50 mM Trizma, pH 7.4, 100 mM NaCl and .5% NP40) added with protease and phosphatase inhibitor, followed by centrifugation at 12,000× g for 5 min.The protein concentration of cellular homogenates was determined by the Lowry method using bovine serum albumin (BSA) as the standard.25 μg of protein was separated into 10% SDS-PAGE gels and electrophoretically transferred onto PVDF membranes for 1 h for 15 min at 100 V.
Membranes were blocked with Tris-buffered saline/Tween 20 buffer (100 mM Tris-HCl, 150 mM NaCl and .1% Tween, pH 7.4, TTBS) containing 5% non-fat milk at room temperature for 1 h and incubated overnight at 4°C with antibodies against p-Akt S473 (Table 1) or Akt pan (Table 1) in TTBS containing 1% non-fat milk.Primary antibodies were detected with horseradish peroxidase conjugated (HRP) antirabbit antibody (Table 1) in TTBS containing 1% non-fat milk for 1 h at room temperature.Bands were visualized using chemiluminescence according to the manufacturer's recommendations and exposed to Kodak BioMax-Light Film.
Blots were stripped with stripping buffer (.1 M glycine, 1% SDS, 1% Tween, pH 2.2) at room temperature for 40 min and reproved with anti-GAPDH antibody (1:100,000).Blots were analyzed with Image J software and the results were expressed as fold change respect to 0 min.

| Immunocytochemistry
CGN were plated onto poly-l-lysine-coated glass slides, treated at 1 DIV with 50 μM Euk-134 or 5 μM Vas2870 for 90 min and fixed with 4% paraformaldehyde for 15 min.Cells were permeabilized with .5% Triton X-100 in PBS for 30 min.Subsequently, cells were blocked with blocking solution (PBS containing 1% normal goat serum) for 1 h and then were incubated for 3 days at 4°C in ASE incubation solution (10 mM glycine, .05%Tween 20, .1% Triton X-100 and .1% H 2 O 2 in PBS) with an antibody against p-Akt S473 (Table 1).The primary antibody against p-Akt S473 was detected with an Alexa Fluor 594 anti-rabbit IgG (H+L) secondary antibody (1:500) incubated in an incubation solution (.05% Tween 20, .1% Triton X-100 in PBS) for 4 h at room temperature.Coverslips were mounted using Vectashield mounting media.Images were photographed in the microscope Zeiss LSM710.

| Determination of ROS levels
Cells were incubated with 50 μM Amplex Red and .1 U/mL of HRP and H 2 O 2 levels were detected every minute for 30 min by using a microplate reader (BioTek Co., USA) at an excitation and emission wavelengths of 530/560 and 590 nm, respectively.CGN cultures were treated with catalase, DMSO or Vas2870 after plating and ROS levels were measured after 48 h (Figure 1c).Slopes were calculated and the results were expressed as fold change respect to control.
Three independent experiments with three replicates for each condition were carried out.
TA B L E 1 Antibody information.2), but not the number of branches per cell (Figure 2e).

Antibody
The treatment with Ebselen also decreased the longest neurite length by 17.97% (Figure 2b and Table 2) as compared to DMSO without any effect on the minor neurites length, the number of primary neurites and the number of branches per cell (Figure 2c-e and  1); however, the effect of Ebselen is no longer observed (Figure S1 and Table 2).
The previous results suggest the ROS participation in the regulation of neurite growth.The effect of ROS on neuritic growth could be through the initiation of neurite growth or the neurite elongation.To explore this possibility, we evaluated the ROS participation on neurite growth of CGN after the initiation of neurite growth (2 DIV) and we evaluated the neurite growth 24 h later.Under these conditions, we found that treatments with Euk-134 decreased the number of These results suggest that ROS participate in the neurite elongation and probably in the initiation of the neurite growth, which is supported by the finding that the antioxidant treatments decrease the number of cells with neurites from 1 DIV (data not shown).
To further explore the ROS participation in neurite growth, we evaluated the direct effect of ROS levels on neurite growth.
To this, CGN were treated with .3b), whereas the minor neurites length, the number of primary neurites and the number of branches per cell did not change (Figure 3c

| ROS generated by NOX, but not mitochondria promote neurite growth
Among the identified ROS sources in the cell, the mitochondria and NADPH-oxidases (NOX) seem to play a critical role.4b and Table 3).On the other hand, treatment with MitoTEMPO did not exert any effects on the neurite growth (Figure 4b-e and Table 3).When we treated the CGN with these antioxidants for 72 h we found that the treatment with Vas2870 reduced the number of primary neurites, whereas MitoTEMPO treatment did    3).
In previous studies, NOX was found in the plasma membrane of growth cones (Munnamalai et al., 2014;Olguín-Albuerne & Morán, 2015).Therefore, it is possible that ROS produced by NOX are released to extracellular space and then reentry to the cell or neighbor cells to exert their effects.To explore the participation of extracellular ROS in neurite growth, we treated CGN with catalase, a membrane-impermeant antioxidant enzyme, after seeding and then we evaluated the neurite growth 48 and 72 h later.We found that after 48 h this treatment decreased the number of primary neurites and the number of branches, as well as the longest and minor Table 3) as compared to control.When we treated the CGN for 72 h we found that the longest neurite length was reduced by 22.47% (Figure S4 and Table 3).These results suggest that the ROS present in the extracellular space participate in the neurite growth.
To evaluate the NOX participation in the neurite elongation we treated CGN with Vas2870 at 2 DIV and then we evaluated the neurite growth 24 h later.We found that CGN treated with we evaluated the H 2 O 2 levels by using the Amplex red assay.Under these conditions, we found that ROS levels were decreased by catalase treatment, but not by Vas2870 treatment (F (3,12) = 68.26,p < .0001and p < .0001(Bonferroni test), Figure S5e).These results suggest that: (1) NOX could be a source of the ROS involved in neurite growth; (2) extracellular ROS participate in the neurite growth; and (3) these effects are independent on cell viability.
NOX is composed by a catalytic subunit, a transmembrane subunit and three cytosolic subunits.To further corroborate the participation of NOX in the neurite growth, we carried out experiments of loss and gain function of NOX.To that, CGN were transfected with DNp22 phox , a mutant of p22 phox that inhibits the NOX1, 2 and 3 activities (Kawahara et al., 2005), or p47 phox , a cytosolic subunit of NOX that it is reported to increase H 2 O 2 levels (Wilson et al., 2016).Cells were transfected after plating and then we evaluated the neurite growth at 3 DIV.Under these conditions, the levels of p47 phox increased from 6 h after Together, these results suggests that NOX is a source of ROS involved in neurite growth.
TA B L E 3 Effect of a NOX inhibitor or an extracellular or mitochondrial antioxidant on neurite growth.

| Antioxidants increase p-Akt levels
Some studies have shown that ROS act through a modulation of the activity of PI3K/Akt pathway (Hervera et al., 2018;Le Belle et al., 2011) that is an important signaling pathway for neurite growth.For this reason, we evaluated whether the Akt activity was modified by Euk-134, catalase or Vas2870.These experiments showed that the treatments with all these antioxidants increased the levels of p-Akt (Ser473).Euk-134 showed an increase at 40 min Western blot showed that 10, 100 and 200 μM H 2 O 2 reduced the p-Akt(Ser473) levels by 45%, 69% and 86%; however, only the results with 200 μM H 2 O 2 was statistically significant (F (5,12) = 5.299, p = .0084and p = .0496(Bonferroni test); Figure S6c,d).We did not find any effect on total Akt levels (F (5,12) = .4705,p = .7913; Figure S6c,d).On the other hand, we evaluated the localization and p-Akt levels by immunocytochemistry in CGN treated with Euk-134 or Vas2870 for 90 min.We found that in control and DMSO conditions p-Akt is mainly localized in the neuronal soma, whereas the treatments with Euk-134 or Vas2870 increased the p-Akt levels in the soma, but also in the neurites (Figure S6e).

| The neurite growth induced by ROS does not depend on Akt activation
Akt has been related to have positive effects on neurite growth (Kumar et al., 2005;Okada et al., 2010), but some reports also suggest that this protein has a negative effect on neurite growth (Bang et al., 2001;Miyamoto et al., 2013).To explore the possibility that the effects of antioxidants on neurite growth could be mediated by an increase of p-Akt (Ser473) levels, cultures were treated at 2 DIV with the PI3K inhibitor LY294002 for 30 min and then cells were incubated with Euk-134 and 24 h after neurite growth was measured.As we previously showed, we found that Euk-134 reduced the longest neurite  4) respect to control.CGN treated with LY294002 did not exert any effect on neurite growth respect to DMSO (Figure 7b-e and Table 4).Additionally, we did not observe any difference between cells treated with Euk-134 plus LY294002 and those incubated with Euk-134 alone, whereas this treatment decreases the number of primary neurons respect to LY294002 alone (p = .0013(Dunn's test), Figure 7b-e and Table 4).
To further evaluate the participation of PI3K/Akt pathway on the neurite growth during the time, we explored the effect of LY294002  8a-e).
We also treated cultures with the PI3K inhibitor LY294002 at 1 DIV for 30 min and then with Euk-134 and neurite growth was measured after 24 h.We found that CGN treated with Euk-  5).These results were different from those found when CGN were treated with LY294002 plus Euk-134 at 2 DIV.Under these conditions we found a largest reduction in the number of branches per cell in CGN treated with LY294002 plus Euk-134 than in those incubated with Euk-134 alone (p = .0061(Dunn's test), Figure 8j and Table 5).
We also observed a tendency to reduce the minor neurites and the  previous studies showed that ROS levels increase in CGN at 2 and 3 DIV (Olguín-Albuerne & Morán, 2015).In the present study we found that the treatment of CGN with antioxidants at these days, significantly decreased the neurite growth, particularly the longest neurite length, in line with studies in other neuronal types (Chandrasekaran et al., 2015;Kamata et al., 2005;Munnamalai & 2009;Wang et al., 2011;Weaver et al., 2018).These effects on the length of neurites are not dependent on cell viability as the largest effects of antioxidants in neurite growth are at 2 DIV and the effects on cell viability are more evident at 3 DIV.
The observed action of antioxidants on the reduction of neurite growth can be mediated by regulation of the initiation of neurite growth or on the neurite elongation.Our results suggest that ROS participate in both processes.We found that the treatment for 48 h (from 0 to 2 DIV) with Euk-134 and catalase decreased both the mayor neurite length and the number of primary neurites (Figures 1   and 3) and previous studies of our group showed that antioxidant treatments decreased the number of cells with neurites from 1 DIV, as it was observed in other cell types (Gopalakrishna et al., 2008;Wang et al., 2011).These data support the idea of an involvement of ROS in the initiation process.Treatments with antioxidant for 24 h (from 2 to 3 DIV) decreased the neurite growth when almost all the neurons have already developed neurites (Powell et al., 1997).that glutathione (GSH) content changes during the cerebellar cortex development (Nanda et al., 1996;Rice & Russo-Menna, 1997).
Furthermore, in a previous study (Olguín-Albuerne & Morán, 2015), we found that the GSH content in cultured cerebellar granule neurons (CGN) increased during the first 2 DIV.Most of the observed increase (90%) occurred between 0 and 1 DIV.Interestingly, as with GSH, ROS levels also steadily increased over time from 0 to 1 DIV (160%) and from 1 to 2 DIV (110%).After this time ROS levels did not change, but start decreasing after 3 DIV.These levels correlated with the expression of NOX, which is known to produce O ⋅− 2 that dismute to H 2 O 2 , as demonstrated by its detection with the H 2 O 2 sensor HyPer (Wilson et al., 2015).
ROS can be generated as a by-product of cellular metabolism, principally by the mitochondria or by some enzymes, particularly the NOX family (Terzi & Suter, 2020;Thannickal & Fanburg, 2000).
Here, we found that NOX, but no mitochondria, are the source of ROS implicated in neurite growth.This notion was supported by the observation that a NOX inhibitor and a negative dominant the catalytic subunit of NOX (DNp22 phox ) (Kawahara et al., 2005) decreased the neurite growth.It has been observed that NOX2 is localized in the neurites, associated to growth cones (Munnamalai et al., 2014;Olguín-Albuerne & Morán, 2015).Furthermore, neurites of NOX2 KO CGN are smaller than neurite of wild-type CGN (Olguín-Albuerne & Morán, 2015).In hippocampal cells the expression of DNp22 phox and treatment with NOX inhibitors decrease the neurite growth (Wilson et al., 2015), as found in Aplysia neurons (Munnamalai & Suter, 2009).On the other hand, although mitochondrial ROS are important for many physiologic processes such as the neuronal differentiation (Rharass et al., 2014), we found that mitochondrial ROS are not required for neuritic growth, as shown in DRG neurons where the neuritic growth induced by sciatic nerve lesion was dependent of ROS produced by NOX, but not mitochondrial ROS (Hervera et al., 2018).
An interesting observation was that the non-permeant antioxidant catalase added exogenously to the culture significantly decreased the neurite growth.Catalase is an enzyme that catalyzes the decomposition of H 2 O 2 that cannot enter the cell.Therefore, ROS generated by NOX are released to the extracellular space, where they are dismutated to H 2 O 2 where they could have an autocrine/ paracrine action (Dhawan et al., 2021) Interestingly, catalase decreased the longest and minor neurites length, the number of primary neurites and the number of branches per cell, whereas the NOX inhibitor decreased only the longest neurite length, suggesting other ROS sources could also participate as ROS sources in the neurite growth.It is worth mentioning that the ROS (and their possible sources) involved in growth are particularly localized in specific areas of neurites and, therefore, the measurement of ROS levels by techniques such as Amplex Red (extracellular) or DHE (marking the nuclei in the soma) could not be appropriated to correlate with neuritic growth.Finally, we cannot rule out that null effect of Vas2870 could be by interference with the Amplex red assay, as has been found that dietary antioxidants interfere with this assay (Serrano et al., 2009).
To further explore the role of ROS produced by NOX in neurite growth we overexpressed the NOX subunit p47 phox .In hippocampal neurons the overexpression of this subunit increases the levels of ROS measure by Hyper and the neurite growth (Wilson et al., 2016).In the present study we did not find any effect on neurite growth.One explanation to discrepancy could be a difference on ROS sensibility of different neuronal types.Agree within line with this point, in DRG neurons the increase of neurite growth was only observed in neurons TA B L E 5 Effect of Euk-134 with or without LY294002 pretreatment on neurite growth treated at 1 DIV.transfected with a constitutively active p47 phox mutant, but not with a wild type p47 phox (Hervera et al., 2018).Additionally, we shown that the effects on neurite growth are concentration-dependent of ROS It is possible that overexpression of p47phox increase the ROS levels above of optimal level and the effect on neurite growth was lost like those found in CGN treated with H 2 O 2 1 μM.Finally, another possibility could be the cell density of the culture.Experiments with hippocampal neurons were done at a low density, while our cultures were prepared at a higher density.In that regard, the neurite growth is dependent of cellular density (Radio et al., 2010), and it is possible that in our conditions of culture the neurite growth is maximal and enhance the neurite growth is not possible.
It is well established that most physiological and pathological actions of ROS are mediated by reversible redox modifications of some amino acid residues, particularly cysteines, of various proteins involved in cellular activities, including signaling pathways (Fra et al., 2017;Gough & Cotter, 2011).In that regard, some studies have demonstrated that PI3K/Akt pathway might be modulated by ROS (Ahmad et al., 2014;Hervera et al., 2018;Le Belle et al., 2011).
Interestingly, several reports have found that PI3K/Akt pathway participates in the neurite growth of many neuronal types (Read & Gorman, 2009), for this reason we evaluated if this signaling pathway participate in the neurite growth mediated by ROS.It is known that the Akt activity is regulated by the tyrosine phosphatase PTEN and ROS can increase the Akt activity by directly inhibiting PTEN (Hervera et al., 2018;Kim et al., 2018;Le Belle et al., 2011;Lee et al., 2002).
To evaluate the role of Akt in neurite growth modulated by ROS, we inhibited the Akt activity with LY294002, a PI3K inhibitor (Vlahos et al., 1994) at 1 or 2 DIV and we measured the neurite outgrowth after 24 h.As mentioned previously, the antioxidant treatments increased the p-Akt and decreased the neurite growth, suggesting a negative effect of Akt in neurite growth as reported as reported in other models (Bang et al., 2001;Miyamoto et al., 2013).On the other hand, contrary to what was found in DGR neurons (Hervera et al., 2018) and PC12 cells (Park et al., 2020), in the present study the inhibition of Akt activity did not alter the inhibitory effect of antioxidants on the neurite growth in cells treated at 1 or 2 DIV and enhanced the effect of antioxidants at 1 DIV on the number of branches per cell.We therefore found opposite results between the effect of antioxidants on Akt activation (measured by an increase in p-Akt) and the effect of Akt activation on neurite growth.Since antioxidants inhibit growth and peroxide promotes it, and peroxide inactivates Akt, we expected that the inactivation of Akt would promote growth; however, the experiments with LY294002 inducing an inactivation of Akt showed a reduced neuritic growth.The results on the cotreatment with LY294002 and EUK-134 in neuritic growth confirm this idea.
All these results suggest that the neurite growth mediated by ROS is dependent of other targets different of Akt.In that regard, one of these targets for ROS during neuritic growth could be the cytoskeleton, mainly actin.For example, the administration of NOX inhibitors to Aplysia neurons decreased the actin retrograde flow and F-actin, which resulted in a smaller neurite growth (Munnamalai & Suter, 2009).Furthermore, in hippocampal neurons the use of a negative dominant for NOX decreased the lamellar area of the growth cone and the number, length and lifetime of actin filaments (Wilson et al., 2015).Thus, a possibility is that the observed actions of ROS can be mediated by a direct redox of actin (Farah et al., 2011;Fiaschi et al., 2006;Grintsevich et al., 2017) or some other proteins that regulate the actin dynamics, such as Rac1 (Hobbs et al., 2015), Cdc42 (Xu et al., 2021) and/ or RhoA (Aghajanian et al., 2009;Heo et al., 2006).To explore these possibilities could be important experiments with phalloidin to quantify the F-actin content after the treatment with antioxidants or experiments for evaluated if the actin dynamic changes after the treatment with antioxidants.
In summary, we showed that ROS generated by a NOX, but not mitochondrial ROS, are needed for the neurite growth of CGN.
Although we found that the PI3K/Akt pathway can be regulated by oxidative conditions and that Akt activity can regulate neurite growth, our results suggest that the Akt action in neuritic growth does not seem to be mediated by ROS.

D ECL
approved by the local Committee of Research and Ethics of the Instituto de Fisiología Celular, Universidad Nacional Autónoma de México (protocol number: JMA175-21).The protocol used followed the Guidelines for the Care and Use of Mammals in Neuroscience as well as guidelines released by the Mexican Institutes of Health Research and the National Institutes of Health guide for the care and use of laboratory animals.All efforts were made to minimize animal suffering and to reduce the number of animals used.
-e).Conversely, 10 μM H 2 O 2 treatments decreased the longest neurite length by 15.77% and reduce the number of branches per cell (Figure3b,e), whereas 1 μM H 2 O 2 treatments had no effects on the neurite growth (Figure3b-e).These results suggest that ROS promote the neurite growth, but this effect depends on the ROS concentration.To discard a possible cell viability influence during the neurite growth, we evaluated the effect of H 2 O 2 or antioxidant treatments on cell viability.Treatments with Ebselen had no effect on cell viability at any time evaluated as compared to DMSO, whereas the treatments with Euk-134 decreased the cell viability by 39.48% (F (3,12) = 14.32; p = .0003,FigureS2b,d) at 3 DIV, but not at 2DIV (F (3,12) = 3.373; p = .0546,FigureS2a,c).On the other hand, .1 and 1 μM H 2 O 2 treatments had no effect on cell viability, whereas 10 μM H 2 O 2 treatments slightly decrease by 4.3% cell viability (F (3,12) = 5.976; p = .0099,FigureS3a,b).Together, our results suggest that the effect of ROS on the neurite growth are independent of cell viability.
not affect neurite growth (Longest neurite length Kruskal-Wallis statistic = 20.35,p = .0004;minor neurites length Kruskal-Wallis statistic = 5.6, p = .23.11; number of primary neurites Kruskal-Wallis F I G U R E 2 Antioxidants treatment decreases neurite growth.CGN were treated with 50 μM Euk-134 or 10 μM Ebselen after seeding and the neurite growth was measured at 2 DIV (48 h treatment) or treatment was made 2 days after seeding and the neurite growth was evaluated at 3 DIV (24 h).(a) Representative fluorescence microphotographs of CGN at 2 DIV.(b-e) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and the number of branches per cell at 2 DIV.(f) Representative fluorescence microphotographs of CGN at 3 DIV.(g-j) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and the number of branches per cell at 3 DIV.Solid and dashed lines represent the median and quartiles, respectively.DIV 2 n = 180; DIV 3 n = 100.Treatments were compared to control or DMSO using Kruskal-Wallis test, followed by Dunn's test or Mann-Whitney test.*p < .05,**p < .01 or ***p < .001.Scale bar, 100 μm.TA B L E 2 Effect of antioxidants on neurite growth.

Note:
Neurite growth of CGN treated with antioxidants for 2 or 3 DIV.Values show the mean ± SEM (upper value) of the longer and minor neurites length (μM), the number of primary neurites or the number of branches per cell of n = 180 (DIV 2) or n = 100 (DIV 3) of 3 independent experiments.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test.Bottom values show p-value of Dunn's test.Euk-134 and DMSO values are compared to control; Ebselen is compared to DMSO values.F I G U R E 3 H 2 O 2 shows a concentration-dependent effect on neurite growth.CGN were treated with .1, 1 or 10 μM H 2 O 2 for one day in vitro.(a) Representative fluorescence microphotographs of CGN 2 DIV.(b-e) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and number of branches per cell at 2 DIV.Solid and dashed lines represent the median and quartiles, respectively.n = 250.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test or Mann-Whitney test.*p < .05 or **p < .01.Scale bar, 100 μm.statistic = 13.45,p = .0093;number of branches Kruskal-Wallis statistic = 11.67,p = .02,Figure S4 and Table

F
I G U R E 4 A cell-impermeant antioxidant and a NOX inhibitor, but not a mitochondrial antioxidant decreases neurite growth.CGN were treated with 20 μg/mL catalase, 5 μM Vas2870 or 10 μM MitoTEMPO after the culture and the neurite growth evaluated at 2 DIV, or cells were treated 2 days after and the neurite growth measured at 3 DIV.(a) Representative fluorescence microphotographs of CGN 2 DIV.(b-e) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and number of branches per cell at 2 DIV.(f) Representative fluorescence microphotographs of CGN at 3 DIV.(g-j) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and the number of branches per cell at 3 DIV.Solid and dashed lines represent the median and quartiles, respectively.DIV 2 n = 150; DIV 3 n = 100.Treatments were compared to control or DMSO using Kruskal-Wallis test, followed by Dunn's test.*p < .05,**p < .01 or ***p < .001.Scale bar, 100 μm.neurite length by 35.55% and 45.46%, respectively (Figure 4b-e and Vas2870 decreased the number of branches per cell and the longest neurite length by 15.16% (Longest neurite length Kruskal-Wallis statistic = 14.20, p = .0008and p = .0028(Dunn's test); minor neurites length Kruskal-Wallis statistic = 4.79, p = .0912;number of primary neurites Kruskal-Wallis statistic = 2.27; p = .3214;number of branches Kruskal-Wallis statistic = 7.73, p = .0209and p = .0281(Dunn's test), Figure4f-j).The observed effect of catalase and Vas2870 were independent on cell viability (DIV 2 F (4,15) = 1.579, p = .2311;DIV 3 F (4,15) = 3.073, p = .0492;FigureS5a-d).Additionally, the transfection and then it remained stable until 3 DIV (data not shown), whereas the ROS levels increased at 3 DIV (data not shown).On the other hand, DNp22 phox transfection decreased ROS levels only at 1 DIV (data not shown).CGN transfected with DNp22 phox showed a reduction in the number of branches per cell and a decrease in the longest neurite length by 23.07% as compared to control (Longest neurite length Mann-Whitney statistic = 845, p = .0142;minor neurites length Mann-Whitney statistic = 922, p = .0579;number of primary neurites Mann-Whitney statistic = 1088, p = .4101;number of branches Mann-Whitney statistic = 872, p = .0236,Figure 5a,b,e).In contrast, no effects were observed in CGN transfected with p47 phox (longest neurite length Mann-Whitney statistic = 1295, p = .3426;minor neurites length Mann-Whitney statistic = 1358, p = .5730;number of primary neurites Mann-Whitney statistic = 1213, p = .1192;number of branches Mann-Whitney statistic = 1321, p = .4285,Figure 5b-e).

(
FigureS6a,b).The p-Akt levels showed a tendency to increase after the treatments with all the antioxidants at all times measured, showing with a peak after 24 h.Akt levels remained without change in all treatments (F (8,18) = 1.232, p = .3366;FigureS6a,b).Additionally, we evaluated the effect of H 2 O 2 on p-Akt(Ser473) and Akt levels.

F
Antioxidants treatment increases p-Akt levels.(a) Representative Western blots of p-Akt (Ser473), Akt (pan) and GAPDH of CGN treated with 50 μM Euk-134; (c) 20 μg/mL catalase or (e) 5 μM Vas2870 at 2 DIV during 0, 5, 10, 20, 40, 60 or 90 min.(b) Densitometric analysis of p-Akt/Akt (left) and Akt/GAPDH (right) of CGN treated with Euk-134, (d) catalase or (f) Vas2870.Data were normalized respect to 0 min.Data are expressed as mean ± SEM of 3 (Euk-134 and catalase) or 4 (Vas2870) individual experiments.Treatments were compared to 0 min using ANOVA, followed by Bonferroni test.*p < .05. number of branches per cell respect to LY294002 alone (p = .0916and p = .0504(Dunn's test), respectively, Figure 8g,j).Since the effect of LY294002 plus Euk-134 is not different or enhance the effects of Euk-134 alone in neurite growth, these results suggest that the reduction of neurite growth by antioxidants is not dependent of the PI3K/Akt activity.4 | DISCUSS ION ROS function as signaling molecules in several processes of development of CNS, including neurite growth (Biswas et al., 2022; Olguín-Albuerne & Morán, 2018; Terzi & Suter, 2020).These effects are mediated by reversible modification of specific proteins.One of these proteins could be Akt acting through the signaling pathway PI3K/Akt.In this study, we evaluated the participation of ROS in the neurite growth if this effect is through of the modulation of Akt activity in a model of cultured cerebellar granule neurons.We found that ROS production by NOX enzymes is necessary for proper neurite growth of CGN.This effect was independent of Akt activity since the administration of an inhibitor of PI3K does not modify the antioxidant effects on neurite growth.It has been observed that the neurite growth of cultures occurs mainly during the first 3 DIV (Radio et al., 2010).Interestingly, F I G U R E 7 PI3K inhibitor does not change the effects of antioxidants on neurite growth.CGN were treated with 50 μM Euk-134, 10 μM LY or LY plus Euk-134 at 2 DIV.For Euk-134 plus LY group, LY were added 30 min before Euk-134.(a) Representative fluorescence microphotographs of CGN at 3 DIV.(b-e) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and number of branches.Solid and dashed lines represent the median and quartiles, respectively.n = 100.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test.*p < .05,**p < .01 or ***p < .001.LY, LY294002.Scale bar, 100 μm.TA B L E 4 Effect of Euk-134 with or without LY294002 pretreatment on neurite growth treated at 2 DIV.

F
The effect of antioxidants on neurite growth does not depends on Akt activity.CGN were treated with 10 μM LY at 1 DIV.(a) Representative fluorescence microphotographs of at 2 CGN DIV.(b-e) Violin plot of quantification of longest neurite length, longest length, number of primary neurites and number of branches.CGN were treated with 50 μM Euk-134, 10 μM LY or LY plus Euk-134 at 1 DIV.For Euk-134 plus LY group, LY were added 30 min before Euk-134.(f) Representative fluorescence microphotographs of CGN at 2 DIV.(g-j) Violin plot of quantification of longest neurite length, minor neurites length, number of primary neurites and number of branches.Solid and dashed lines represent the median and quartiles, respectively.n = 200.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test.*p < .05,**p < .01 or ***p < .001.LY, LY294002.Scale bar, 100 μm.
Furthermore, treatment with .1 μM H 2 O 2 at 1 DIV increased the longest neurite length suggesting that ROS also participate in the neurite elongation independently of the effect on initiation of neurite growth.Although all the antioxidants used reduced the neurite growth, the effects of Euk-134 is largest and more sustained respect to Ebselen.The differences could be explained by the different mechanisms of action of the antioxidants, as Euk-134 is a SOD/catalase mimetic and Ebselen is a glutathione peroxidase mimetic.It is possible that glutathione is needed only during the first 2 DIV of the development of CGN.In line with this point, it has been reported A R ATI O N O F TR A N S PA R EN C Y The authors, reviewers and editors affirm that in accordance to the policies set by the Journal of Neuroscience Research, this manuscript presents an accurate and transparent account of the study being reported and that all critical details describing the methods and results are present.AUTH O R CO NTR I B UTI O N S J. M.-Z.conducted the experiments, participated in the design of the study, and contributed to the analysis and interpretation of the data, as well as to the writing of the manuscript.J.M. conceived, designed and coordinated the study, raised funds for the study, and contributed to the writing of the manuscript.All authors have read and approved the submitted version of the manuscript.ACK N OWLED G M ENTS We thank Christian González Billault (Universidad de Chile, Santiago, Chile) for the generous donation of the DNp22phox plasmid.We are also grateful to Guadalupe Dominguez Macouzet, Gabriela Itzel Medina Ruiz, Alfredo Cárdenas Rivera, Yazmín Ramiro Cortés, and Abraham Rosas Arellano for the excellent technical assistance.We acknowledge the online Biorender tool used for making the graphical abstract and Figure 1.This study is part of the doctoral thesis of Janeth Mora-Zenil in the Doctorado en Ciencias Bioquímicas from the Universidad Nacional Autónoma de México.FU N D I N G I N FO R M ATI O N This work was supported by the Consejo Nacional de Humanidades, Ciencias y Tecnologías (CONAHCYT) [Grant number 285184] and the Dirección General de Asuntos del Personal Académico, UNAM (DGAPA-PAPIIT, UNAM) [Grants IN212019 and IN216422].Janeth Mora-Zenil received a fellowship from CONAHCYT.

Table 2
). Comparable results were obtained when cells were treated for 72 h with antioxidants (longest neurite length Kruskal-Wallis statistic = 139, p < .0001;minorneuriteslengthKruskal-Wallisstatistic = 73.51,p<.0001;number of primary neurites Kruskal-Wallis statistic = 74.66,p<.0001;number of branches Kruskal-Wallis statistic = 49.52,p=.1523,FigureS1).Euk-134 treatment showed a reduction of the number of branches and primary neurites and of the longest neurite length by 59.32% and minor neurites length by 61.77% (FigureS1and Table Data show the mean ± SEM (upper value) of the longer and minor neurites length (μM), the number of primary neurites or the number of branches per cell of n = 125 of 3 independent experiments.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test.Bottom values show p-value of Dunn's test.Euk-134 and DMSO are compared to control, LY is compared to DMSO and Euk-134+LY is compared to Euk-134.LY: LY294002.
Data show the mean ± SEM (upper value) of the longer and minor neurites length (μM), the number of primary neurites or the number of branches per cell of n = 200 of 3 independent experiments.Treatments were compared using Kruskal-Wallis test, followed by Dunn's test.Bottom values show p-value of Dunn's test.Euk-134 and DMSO are compared to control, LY is compared to DMSO and Euk-134+LY is compared to Euk-134.LY: LY294002.