**3. Results**

### *3.1. PRDXs Are among the Highest Expressed Antioxidant Genes in Neonatal Gonocytes and in Spermatogonia*

Gene array analysis of antioxidant genes in highly purified rat PND3 gonocytes and PND8 spermatogonia showed that, at both ages, Prdxs were among the most abundant antioxidant genes (Figure 1A, Table S1). Comparing the signal intensities of all genes in the arrays showed that thioredoxin 1 (*Txn1*), *Prdx1* and superoxide dismutase (*Sod*) 1 were among the 1% most abundant genes in these cells, comprising relative signal intensities around and above 2000. Next, *Prdx2*, *Prdx5*, *Prdx6*, *Sod2*, *Gstp1*, and *Gpx4* were among the most abundant genes, with relative signal intensities from 600 to 1200 (Table S1). Other highly expressed genes (signal intensities between 300 and 600) included *Gsto1*, *Prdx3*, *Txnl1*, *Mgst1*, *Txnrd1*, *Gstp2*, *Gpx1*, and *Nrf2*. As a comparison, these genes had much higher signal intensities than 65 percent of the genes in the cells, which presented intensities below or around 20, levels found for Sod3 and Catalase. The ranking of expression for PRDXs were *Prdx1* > *Prdx2* > *Prdx5* > *Prdx6* > *Prdx3*, in both germ cell types. Measurement of the relative expression of the PRDXs by qPCR in PND3 gonocytes indicated a similar ranking of expression levels, where *Prdx1* and *Prdx2* were the most abundant, followed by *Prdx6*, *Prdx5* and *Prdx3*, and finally *Prdx4* present at very low levels (Figure 1B). The similarities in expression profiles of antioxidant genes between gonocytes and spermatogonia suggested a conserved antioxidant machinery between the two phases of germ cell development.

**Figure 1.** Relative expression of antioxidant genes in postnatal day 3 (PND3) gonocytes and PND8 spermatogonia. ( **A**) gene array analysis of PND3 gonocytes and PND8 spermatogonia antioxidant genes. (**B**) qPCR analysis of PRDXs expression in PND3 gonocytes. Results represent the means ± SEM from 3 or 4 experiments, each using multiples animals.

This was different from the levels found in enriched PND60 adult rat germ cells, in which the relative gene expression of *Gstm5* (6235) was the highest, followed by *Gpx4* (5507), *Hagh* (1141), *Sod1* (1136), *mGst1* (446), *Prdx1* (407), *Prdx6* (345), *Gstm1* (341), *Txn1* (335), *Prdx2* (223), *Prdx5* (204), *Gsto1* (188), *Prdx3* (144), *Gstt2* (140), *Txnl1* (132), *Txn2* (81), *Sod3* (73), *Gsta2* (64), *Gstp1* (65), and *Sod2* (44), the remaining genes being at very low expression levels.

The immunological analysis of the protein expression of PRDX1, 2, and 6 in mixed suspensions of PND3 gonocytes, Sertoli and myoid cells confirmed that the three PRDX proteins were expressed in gonocytes, Sertoli, and myoid cells, but at lower levels in gonocytes than in most of the somatic cells (Figure 2A). PRDX1 showed a slightly stronger signal than PRDX2, both being higher than PRDX6 in gonocytes, in agreemen<sup>t</sup> with the transcript levels. PRDX6 protein was also visible by IHC analysis in the cytoplasm of gonocytes in PND3 testis sections, but it was lower than the levels observed in Sertoli cells and some interstitial cells (Figure 2B).

**Figure 2.** Protein expression of PRDX1, 2, and 6 in PND3 gonocytes. (**A**) Immunocytochemistry (ICC) analysis of PRDXs expression in cells collected right after the cell isolation procedure on cytospin slides. Low purity pooled fractions were used. White arrow: gonocytes. White arrowhead: somatic cells (Sertoli and peritubular myoid cells). (**B**) Immunohistochemistry (IHC) analysis of PND3 testis section. Representative pictures are shown. Black arrow: gonocytes. Black arrowhead: somatic cells. Scales are in μm.

### *3.2. PRDXs Are Required for PND3 Gonocyte Survival in Basal Conditions*

In view of the abundance of PRDXs in gonocytes, and knowing the importance of these peroxidases for spermatozoa viability and function, we examined whether inhibiting PRDXs would have an impact on the survival of PND3 gonocytes. For that, we performed dose-response and time-course studies in basal conditions where only endogenous ROS may be formed, using either the inhibitor of 2-Cys PRDXs, Conoidin A; the inhibitor of GSTP1 (that promotes PRDX6 glutathione-dependent re-activation), Ezatiostat; or the inhibitor of the phospholipase A2 activity of PRDX6, MJ33. Conoidin A induced a rapid dose-dependent decrease in gonocyte viability with concentrations of 5–10 μM, inducing 50% cell death after one hour of treatment (Figure 3A). There were no surviving cells with conoidin A concentrations of 1–10 μM after 18 h of treatments. By contrast, after 18 h of treatment with MJ33, there was no or minimally detrimental effects on cell viability with concentration from 1 to 20 μM (Figure 3B). However, at 50 μM, MJ33 did induce cell death after 18 h. Eztiostat dose responses were performed for two hours of treatment, because preliminary studies showed that longer time-periods induced the loss of a large amount of cells. Eztiostat inhibited cell survival in a dose-dependent manner (Figure 3C), with a concentration of 50 μM inducing nearly 50% of cell death. These data sugges<sup>t</sup> that the peroxidase activity of 2-cys PRDXs and PRDX6 play a critical role in gonocyte protection against excessive endogenous ROS, and that inhibiting these enzymes even for a short time is sufficient to induce significant gonocyte cell death. Thus, PRDXs are required to maintain ROS homeostasis in neonatal gonocytes. By contrast, inhibiting the PLA2 activity of PRDX6 with MJ33 did not exert rapid deleterious effects on cell viability.

**Figure 3.** Effects of PRDX inhibitors on PND3 gonocyte viability. Cell viability was determined by trypan blue exclusion method. (**A**) Dose response and time course of conoidin A effects. Control: green; Conoidin A at 0.1 μM (blue); 1 μM (orange); 5 μM (pink); 10 μM (red). (**B**) MJ33 dose response at 18 h treatments. (**C**) Ezatiostat dose response at two hours of treatment. Results represent the means ± SEM from at least three experiments. Statistical significance: \* *p* ≤ 0.05; \*\* *p* ≤ 0.01; \*\*\* *p* ≤ 0.001.

PRDXs are required to prevent oxidative stress-induced cell death in PND3 gonocytes. Oxidative stress can be generated by external factors, as a result of exposure of the animals and their reproductive system to oxidative agents, or through the production of ROS by other testicular cell types in the vicinity of the gonocytes. Thus, we tested the ability of PRDXs to protect PND3 gonocytes from exogenous ROS, using H2O2 as an oxidative agent. The treatment of gonocytes with H2O2 at 100 and 200 μM for two hours induced a dose-dependent decrease in cell viability, with a 40% and 50% reduction in viability, respectively (Figure 4).

These damaging effects were significantly worsened by treating the cells with PRDXs inhibitors, with MJ33 having the mildest and ezatiostat the worst effects (Figure 4). These results indicate that the oxidative stress generated by H2O2 exerts an adverse effect on gonocyte viability, and that 2 Cys PRDX and PRDX6 peroxidase activities are required to protect gonocytes from oxidative stress. This further implies that other antioxidant proteins expressed in gonocytes, even those present at high levels such as Txn1, Sod1, and Sod2, are not able to rescue gonocytes from exposure to exogenous oxidants. Moreover, the fact that ezatiostat and MJ33 exacerbated the adverse effect of H2O2 on gonocyte viability suggests that both the peroxidase and phospholipase A2 activity of PRDX6 are important for gonocyte survival.

**Figure 4.** Effects of H2O2 and PRDX inhibitors on PND3 gonocyte viability. H2O2 and PRDX inhibitors were added to the cells at the indicated concentrations, alone or together for 2 h. Cell viability was determined by trypan blue exclusion method. Results represent the means ± SEM from at least three experiments. Statistical significance: \*\*\* *p* ≤ 0.001 treatment vs control; & *p* ≤ 0.05, &&& *p* ≤ 0.001 treatment vs 100 μM H2O2; # *p* ≤ 0.05, ### *p* ≤ 0.001 treatment vs 200 μM H2O2.

### *3.3. H2O2 Induces Lipid Peroxidation, and PRDXs Prevent Endogenous ROS-Induced Lipid Peroxidation in PND3 Gonocytes*

Lipid peroxidation has been known for decades to be one of the sources of spermatozoa damage in infertile men [26]. Thus, we examined whether lipid peroxidation could also play a role in the deleterious e ffects of H2O2 and PRDX inhibitors on gonocyte viability, using a fluorescent fatty acid reporter system that allowed quantifying the proportion of cells where lipid peroxidation had occurred. While lipid peroxidation was detectable in 40% of the control gonocytes, a condition corresponding to 80% cell viability, treatment with H2O2 (a condition decreasing cell viability by 50%) doubled the numbers of gonocytes presenting lipid peroxidation to nearly 80% of positive cells in gonocytes (Figure 5). To examine the e ffects of PRDX inhibitors, we used concentrations of conoidin A and ezatiostat decreasing viability by ~50% (10 and 50 μM respectively), and a concentration of MJ33 that did not significantly alter cell viability compared to controls (20 μM). Treatments with conoidin A alone had the same e ffect as H2O2 on lipid peroxidation, doubling the percent of gonocytes presenting lipid peroxidation, while MJ33 and ezatiostat induced significant but lower increases in lipid peroxidation, reaching 70% of positive cells, a 1.55 increase over the levels in control cells (Figure 5).

**Figure 5.** Effects of H2O2 and PRDX inhibitors on lipid peroxide formation in PND3 gonocytes. H2O2 and PRDX inhibitors were added for two hours on gonocytes. Lipid peroxidation was measured using the Bodipy assay. Results are the means ± SEM from at least three experiments. Statistical significance: \*\* *p* ≤ 0.01; \*\*\* *p* ≤ 0.001.

Combining PRDX inhibitors with H2O2 did not further increase lipid peroxidation, which already affected most of the cells with individual treatments (data not shown). These results show that an oxidative stressor such as H2O2 induces similar levels of lipid peroxidation as endogenous ROS accumulating in the absence of PRDX activities in PND3 gonocytes. This suggests that lipid peroxides are produced physiologically in gonocytes, and that PRDXs are critical for preventing excessive levels that could a ffect gonocyte survival. The data also imply that PRDXs protect gonocytes from excessive endogenous ROS-induced lipid peroxidation. Moreover, the fact that 40% of control gonocytes showed lipid peroxidation without a ffecting their viability, and that lipid peroxidation was not proportional to viability in cells treated with PRDX inhibitors, sugges<sup>t</sup> that gonocytes can tolerate a certain level of lipid peroxidation. The finding that MJ33 treatment increased lipid peroxidation suggests that the PRDX6 iPLA2 activity also plays a role in repairing oxidized membranes [27].

### *3.4. PRDX Inhibition Blocks RA-Induced Di*ff*erentiation in PND3 Gonocytes*

Considering the physiological roles played by ROS in many tissues and cell types, including in spermatozoa, and our finding that inhibiting PRDXs increases endogenous ROS formation in PND3 gonocytes, we examined whether blocking PRDXs could a ffect RA-induced gonocytes di fferentiation. First, we determined cell viability in cells treated for two hours with RA alone or together with two di fferent concentrations of PRDX inhibitors, in order to select concentrations of inhibitors that would not overly decrease cell survival. RA at 1 μM did not have an e ffect on cell survival. Similarly, concentrations of 20 μM MJ33 or 10 μM ezatiostat did not a ffect cell viability (Figure 6). While 0.5 μM conoidin A decreased cell viability by 9% when used alone, and 17% when added with RA in comparison to control and RA treatments, these was still relatively minor e ffects.

**Figure 6.** Effects of trans-retinoic acid (RA) and PRDX inhibitors on PND3 gonocyte viability. RA (1 μM) and PRDX inhibitors at the indicated concentrations were added alone or together for two hours on gonocytes. Cell viability was determined by the trypan blue exclusion method. Results represent the means ± SEM from at least 3 experiments. Statistical significance: \* *p* ≤ 0.05, \*\*\* *p* ≤ 0.001, treatment against control; &&& *p* ≤ 0.001, treatment against RA alone.

Next, we measured the mRNA expression of Stra8, previously found to be a good marker for assessing neonatal gonocyte di fferentiation [15–17]. While two hours of treatment with RA significantly increased Stra8 expression by over three-fold, there was no e ffect of conoidin A, MJ33 or ezatiostat added alone to the cells (Figure 7). However, both conoidin A and ezatiostat significantly repressed RA-induced Stra8 increases, indicating inhibitory e ffects of PRDX inhibitors on gonocytes di fferentiation. However, MJ33 did not alter RA e ffect on Stra8 levels. These data sugges<sup>t</sup> that the oxidative stress resulting from PRDXs inhibition is detrimental to gonocytes di fferentiation, whereas PRDX6 iPLA activity does not appear to be required for gonocyte di fferentiation.

**Figure 7.** Effects of trans-retinoic acid (RA) and PRDX inhibitors on PND3 gonocyte differentiation. RA (1 μM) and PRDX inhibitors at the indicated concentrations were added alone or together for two hours on gonocytes. *Stra8* mRNA expression was measured by qPCR analysis. Results represent the means ± SEM from three experiments. Statistical significance: \*\*\* *p* ≤ 0.001, treatment against control; & *p* ≤ 0.05, && *p* ≤ 0.01, treatment against RA alone.
