Potential Cytoprotective and Regulatory Effects of Ergothioneine on Gene Expression of Proteins Involved in Erythroid Adaptation Mechanisms and Redox Pathways in K562 Cells

This study aimed to establish the importance of ergothioneine (ERT) in the erythroid adaptation mechanisms by appraising the expression levels of redox-related genes associated with the PI3K/AKT/FoxO3 and Nrf2-ARE pathways using K562 cells induced to erythroid differentiation and H2O2-oxidative stress. Cell viability and gene expression were evaluated. Two concentrations of ERT were assessed, 1 nM (C1) and 100 µM (C2), with and without stress induction (100 µM H2O2). Assessments were made in three periods of the cellular differentiation process (D0, D2, and D4). The C1 treatment promoted the induction of FOXO3 (D0 and 2), PSMB5, and 6 expressions (D4); C1 + H2O2 treatment showed the highest levels of NRF2 transcripts, KEAP1 (D0), YWHAQ (D2 and 4), PSMB5 (D2) and PSMB6 (D4); and C2 + H2O2 (D2) an increase in FOXO3 and MST1 expression, with a decrease of YWHAQ and NRF2 was observed. in C2 + H2O2 (D2) an increase in FOXO3 and MST1, with a decrease in YWHAQ and NRF2 was observed All ERT treatments increased gamma-globin expression. Statistical multivariate analyzes highlighted that the Nrf2-ARE pathway presented a greater contribution in the production of PRDX1, SOD1, CAT, and PSBM5 mRNAs, whereas the PI3K/AKT/FoxO3 pathway was associated with the PRDX2 and TRX transcripts. In conclusion, ERT presented a cytoprotective action through Nrf2 and FoxO3, with the latter seeming to contribute to erythroid proliferation/differentiation.


Introduction
One of the most critical features in human physiology is redox homeostasis, playing a pivotal role in physiological and pathophysiological situations [1]. This biological redox equilibrium influences several cell signaling pathways, a common pathophysiological mechanism that underlies a wide range of chronic and degenerative diseases such as immunological, neurodegenerative, cardiovascular, oncological, hemolytic disorders, and age-related diseases [2][3][4][5][6]. The accumulation of oxidative damage presents an intrinsic relationship with the severity of these diseases [7,8], which leads to the assumption that antioxidant therapies represent a promising path for alternative treatment. However, the simple administration of exogenous antioxidants did not have the expected benefits [9][10][11].

Identification of Erythroid Differentiation
The erythroid differentiation of K562 cells was assessed by a benzidine cytochemical test [36]. A total of 50 µL of 30% hydrogen peroxide mixed with 10 µL of benzidine solution was added to 100 µL of cell suspension. The mixture was incubated in the dark at room temperature for 25 min, and the percentage of benzidine-positive cells was determined by microscopic examination. Gamma (γ)-globin mRNA was detected by the RT-PCR method, and the mRNA levels were also used as an indicator of erythroid differentiation.

Viability
Cell viability was assessed using the Trypan blue (Sigma-Aldrich, St. Louis, MO, USA) exclusion assay in the final concentration of 0.1%. The ratio between the number of living (L) and dead (D) cells provided the percentage (L/D × 100) of cell viability [37].

Experimental Design
The study had three pseudo-replicated samples followed during five days of cell differentiation, with collections of biological samples in three periods of the differentiation process: before the start (D0), at the beginning of differentiation (D2), and the end of the process (D4), with the prediction transcripts and protein level analyses, as well enzyme activities of the biomarkers proposed. On day 0 of the experiment, 3.1 × 10 6 cells/mL (100% viable and 70-80% confluent) were placed in polystyrene bottles in a final volume of 20 mL of culture medium, induced to erythroid differentiation through the addition of hemin (50 µM) and hydroxyurea (100 µM) and incubated at 37 • C/5% CO 2 . In addition, each sample was individually divided into the following groups: erythroid cells without inducing oxidative stress or ERT treatment (Reference); cells under stress induction with hydrogen peroxide (100 µM H 2 O 2 ); cells treated with a low concentration (C1) of ERT 1 nM and a high concentration (C2) of ERT 100 µM; and finally, two sets of cells treated with the same concentrations of ERT associated with stress induction (C1 + H 2 O 2 and C2 + H 2 O 2 , respectively).

Administration of Ergothioneine (Protective Agent) and Hydrogen Peroxide (Stressing Agent)
The concentration adopted of 100 µM H 2 O 2 had been previously tested and validated by de Paula (2020) for this experimental model [35]. For the ERT treatments, we adopted 1 nM as the low dose since the pM-nM scales are usually observed in regulatory molecules, such as EPO and others, as well as based on some of our previous studies with a different antioxidant, which was melatonin [34,38]. For the high dose (100 µM), we used a therapeutic one, according to the levels found in the plasma (~400 nM) [39] and erythrocytes (~50 µM) [40] of healthy individuals and its well-tolerance and low toxicity [27,29]. The administrations of the antioxidant and peroxide were carried out daily and in a staggered manner in the associated groups (C1 + H 2 O 2 and C2 + H 2 O 2 ); that is, one hour after the Genes 2022, 13, 2368 4 of 23 administration of the respective concentrations of ERT, H 2 O 2 was added. The samples were collected 1 h after the administration of the stressor, on days 0 (D0), 2 (D2), and 4 (D4). The experimental procedure is summarized in Figure 1. with a different antioxidant, which was melatonin [34,38]. For the high dose (100 µM), we used a therapeutic one, according to the levels found in the plasma (~400 nM) [39] and erythrocytes (~50 µM) [40] of healthy individuals and its well-tolerance and low toxicity [27,29]. The administrations of the antioxidant and peroxide were carried out daily and in a staggered manner in the associated groups (C1 + H2O2 and C2 + H2O2); that is, one hour after the administration of the respective concentrations of ERT, H2O2 was added. The samples were collected 1 h after the administration of the stressor, on days 0 (D0), 2 (D2), and 4 (D4). The experimental procedure is summarized in Figure 1.

Real-Time PCR
RNA samples were subjected to DNAseI treatment (Invitrogen, Rockville, MD, USA) and reverse transcribed with a High-Capacity cDNA Reverse Transcription kit (Thermo Scientific, Waltham, MA, USA) using oligo dTV and random primers blend. The sequence of primers is shown in Table S1 in the supplementary material. The concentration of each primer was determined, and the amplification efficiency was calculated according to the equation E (1/slope) to confirm the accuracy and reproducibility of the reactions. Amplification specificity was by the dissociation protocol. qPCRs were performed in a StepOne Plus Real-time PCR System (Thermo Scientific, Waltham, MA, USA), using SYBR Green GoTaq Master Mix (Promega, Madison, WI, USA). The fold change in mRNA level was calculated using 2 −ΔΔCt [41], and all the values were normalized to the expression of the beta-actin (ACTB) gene.

Statistical Analyses
Univariate analyzes were performed using the Statistica 9.0 software (Statsoft Inc., Tulsa, OK, USA), while the graphs were made using the GraphPad Prisma software version 5.01 for Windows (GraphPad Software, La Jolla, CA, USA). Data normality was ver-

Real-Time PCR
RNA samples were subjected to DNAseI treatment (Invitrogen, Rockville, MD, USA) and reverse transcribed with a High-Capacity cDNA Reverse Transcription kit (Thermo Scientific, Waltham, MA, USA) using oligo dTV and random primers blend. The sequence of primers is shown in Table S1 in the supplementary material. The concentration of each primer was determined, and the amplification efficiency was calculated according to the equation E (1/slope) to confirm the accuracy and reproducibility of the reactions. Amplification specificity was by the dissociation protocol. qPCRs were performed in a StepOne Plus Real-time PCR System (Thermo Scientific, Waltham, MA, USA), using SYBR Green GoTaq Master Mix (Promega, Madison, WI, USA). The fold change in mRNA level was calculated using 2 −∆∆Ct [41], and all the values were normalized to the expression of the beta-actin (ACTB) gene.

Statistical Analyses
Univariate analyzes were performed using the Statistica 9.0 software (Statsoft Inc., Tulsa, OK, USA), while the graphs were made using the GraphPad Prisma software version 5.01 for Windows (GraphPad Software, La Jolla, CA, USA). Data normality was verified using Normal Probability Plots of Residuals. Thus, some data were transformed by log10 when necessary. For comparison between groups, General Linear Models (GLM) were adopted in the ANOVA two-ways design, allowing to verify the effects of treatments, incubation periods, as well as any interactions between these predictors on the dependent variables. Each experiment was analyzed relative to its own reference group [42]. As a multivariate alternative for analyzing the degree of association, the General Regression Model (GRM) analysis with multiple regression design was adopted, implementing more about the relationship between several independent variables and a dependent one. This analysis also provided the partial correlation, that is, the individual contribution of a certain independent variable from the set on the dependent one, after the control for all the other variables in the equation [43]. The results were expressed as mean ± SEM of their biological values, and all statistical analyses were considered significant at p < 0.05.

Cell Viability
Contrary to expectations, no significant deleterious effect was observed in K562 cells subjected to oxidative stress induction with H 2 O 2 , making it impossible to assess possible protective effects of ERT treatments on cell viability in groups C1 + 100 µM H 2 O 2 and C2 + 100 µM H 2 O 2 ( Figure S1-Supplementary material). On the other hand, it is worth noting that treatments with ERT (C1 and C2) did not present toxicity with compromised viability of K562 cells in any of the differentiation periods. Thus, viability ranged from 77.7% (D4-100 µM H 2 O 2 ) to 100% (D0-all treatments) in all experimental situations tested.

Erythroid Differentiation
The comparisons of the relative gene expression of mRNA γ-globin ( Figure 2) showed an increased expression during the erythroid differentiation process (days 2 and 4), with higher transcript levels (up to~6.5-fold increase) at the end of the process (D4). Additionally, it is interesting to notice that almost all ERT treatments, regardless of hydrogen peroxide administration, showed higher mRNA levels than their respective controls during the erythroid differentiation process. Such observations suggest that ERT may influence this gene expression. were adopted in the ANOVA two-ways design, allowing to verify the effects of treatments, incubation periods, as well as any interactions between these predictors on the dependent variables. Each experiment was analyzed relative to its own reference group [42]. As a multivariate alternative for analyzing the degree of association, the General Regression Model (GRM) analysis with multiple regression design was adopted, implementing more about the relationship between several independent variables and a dependent one. This analysis also provided the partial correlation, that is, the individual contribution of a certain independent variable from the set on the dependent one, after the control for all the other variables in the equation [43]. The results were expressed as mean ± SEM of their biological values, and all statistical analyses were considered significant at p < 0.05.

Cell Viability
Contrary to expectations, no significant deleterious effect was observed in K562 cells subjected to oxidative stress induction with H2O2, making it impossible to assess possible protective effects of ERT treatments on cell viability in groups C1 + 100 µM H2O2 and C2 + 100 µM H2O2 ( Figure S1-Supplementary material). On the other hand, it is worth noting that treatments with ERT (C1 and C2) did not present toxicity with compromised viability of K562 cells in any of the differentiation periods. Thus, viability ranged from 77.7% (D4-100 µM H2O2) to 100% (D0-all treatments) in all experimental situations tested.

Erythroid Differentiation
The comparisons of the relative gene expression of mRNA γ-globin ( Figure 2) showed an increased expression during the erythroid differentiation process (days 2 and 4), with higher transcript levels (up to ~6.5-fold increase) at the end of the process (D4). Additionally, it is interesting to notice that almost all ERT treatments, regardless of hydrogen peroxide administration, showed higher mRNA levels than their respective controls during the erythroid differentiation process. Such observations suggest that ERT may influence this gene expression.

Keap1/Nrf2/ARE Signaling Pathway
The comparisons of the relative gene expression of the Keap1/Nrf2/ARE signaling pathway showed a 6.5-fold increase of Nrf2 levels in C1 + H 2 O 2 ( Figure 3A). Furthermore, a peak of mRNA expression of Keap1 (~2-fold increase) can be observed in C1 + H 2 O 2 on D0 ( Figure 3B). These results reveal an intriguing effect of the treatment (C1 + H 2 O 2 ) within the differentiation periods; the effect of the lower concentration of ERT (C1) compared to the highest (C2), both associated with the addition of H 2 O 2 within each evaluated period. In addition, in the same treatment (C1 + H 2 O 2 ), there is a significant reduction (~1.5-fold) in the transcript levels of Keap1 at the beginning of the differentiation process (D2) in relation to the same treatment in D0. A similar pattern (although not statistically significant) can be observed on day 4. Such observations reinforce the cytoprotective effect of ERT when cells are in the presence of H 2 O 2 , suggesting a conditioned action of ERT.
Effect of the higher concentration of ERT (C2) associated with 100 µM H2O2 compared to treatment C2; ‡ Effect of ERT C1 + 100 µM H2O2 treatment compared to all treatments; ** Effect of treatment within each period, compared to treatment Reference; and *** Effect of treatment within each period, compared to treatment 100 µM H2O2. The representative values of the three pseudoreplicates were expressed as mean with standard error (±SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.

Keap1/Nrf2/ARE Signaling Pathway
The comparisons of the relative gene expression of the Keap1/Nrf2/ARE signaling pathway showed a 6.5-fold increase of Nrf2 levels in C1 + H2O2 (Figure 3a). Furthermore, a peak of mRNA expression of Keap1 (~2-fold increase) can be observed in C1 + H2O2 on D0 (Figure 3b). These results reveal an intriguing effect of the treatment (C1 + H2O2) within the differentiation periods; the effect of the lower concentration of ERT (C1) compared to the highest (C2), both associated with the addition of H2O2 within each evaluated period. In addition, in the same treatment (C1 + H2O2), there is a significant reduction (~1.5-fold) in the transcript levels of Keap1 at the beginning of the differentiation process (D2) in relation to the same treatment in D0. A similar pattern (although not statistically significant) can be observed on day 4. Such observations reinforce the cytoprotective effect of ERT when cells are in the presence of H2O2, suggesting a conditioned action of ERT. The symbols indicate the following statistical differences: * Effect of treatment C1 + 100 µM H2O2 within each period; # Effect of the differentiation period (D2) compared to the same treatment (C1 + 100 µM H2O2) in the other differentiation periods; & Effect of lower concentration of ERT (C1) compared to higher (C2), both associated with 100 µM H2O2; ƒ Increased levels of transcripts in the C1 + 100 µM H2O2 group compared to the others within the D0 period (except for the 100 µM H2O2 group); and £ Effect of the differentiation period (D2) compared to the same treatment (C1 + 100 µM H2O2) only on D0. The representative values of the three pseudoreplicates were expressed as mean with standard error (± SEM). General Linear Models (GLM) with twoway ANOVA format, complemented by the Bonferroni test.

FOXO3 and Its Subcellular Location Regulators
The treatment with the lowest concentration of ERT (C1) promoted the induction of FOXO3 expression (~3.5-fold increase) when compared to the other treatments in the same differentiation period (D0) and, interestingly, only in this period. The highest concentration tested (C2) promoted a~2.5-fold reduction in the levels of transcripts at the beginning of the differentiation process (D2), remaining low until the end of the experiment ( Figure 4A). Although not statistically significant, a pattern of increased expression of this transcription factor can be observed in the ERT C2 + 100 µM H 2 O 2 when compared to ERT C1 + 100 µM H 2 O 2 (D0 and 2).
FOXO3 expression (~3.5-fold increase) when compared to the other treatments in the same differentiation period (D0) and, interestingly, only in this period. The highest concentration tested (C2) promoted a ~2.5-fold reduction in the levels of transcripts at the beginning of the differentiation process (D2), remaining low until the end of the experiment ( Figure 4a). Although not statistically significant, a pattern of increased expression of this transcription factor can be observed in the ERT C2 + 100 µM H2O2 when compared to ERT C1 + 100 µM H2O2 (D0 and 2). The symbols indicate the following statistical differences: $ Effect of the lowest concentration of ERT (C1) within each period; ¥ Effect of lower concentration of ERT (C1) compared to higher (C2), both associated with 100 µM H2O2; § Effect of the higher concentration of ERT (C2) associated with 100 µM H2O2 compared to treatment C2; @ Effect of the differentiation period within each treatment, compared to its counterpart in D0; * Effect of treatment C1 + 100 µM H2O2 within each period, compared to the treatments 100 µM H2O2 and ERT C1; ** Effect of treatment within each period, compared to the Reference; and *** Effect of treatment within each period, compared to treatment 100 µM H2O2. The representative values of the three pseudoreplicates were expressed as mean with standard error (± SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.
In relation to the YWHAQ gene (Figure 4b), C1 + 100 µM H2O2 seems to induce the expression during the cell differentiation process, while the opposite behavior is observed for C2 + 100 µM H2O2. Additionally, a ~2.5-fold increase of gene expression was observed in the Peroxide and ERT C1 treatments (D0) when compared to the reference, demonstrating an effect of these treatments that is no longer observed in the other evaluated periods. Finally, when observing the expression of the MST1 kinase gene (Figure The symbols indicate the following statistical differences: $ Effect of the lowest concentration of ERT (C1) within each period; ¥ Effect of lower concentration of ERT (C1) compared to higher (C2), both associated with 100 µM H 2 O 2 ; § Effect of the higher concentration of ERT (C2) associated with 100 µM H 2 O 2 compared to treatment C2; @ Effect of the differentiation period within each treatment, compared to its counterpart in D0; * Effect of treatment C1 + 100 µM H 2 O 2 within each period, compared to the treatments 100 µM H 2 O 2 and ERT C1; ** Effect of treatment within each period, compared to the Reference; and *** Effect of treatment within each period, compared to treatment 100 µM H 2 O 2 . The representative values of the three pseudoreplicates were expressed as mean with standard error (± SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.
In relation to the YWHAQ gene ( Figure 4B), C1 + 100 µM H 2 O 2 seems to induce the expression during the cell differentiation process, while the opposite behavior is observed for C2 + 100 µM H 2 O 2 . Additionally, a~2.5-fold increase of gene expression was observed in the Peroxide and ERT C1 treatments (D0) when compared to the reference, demonstrating an effect of these treatments that is no longer observed in the other evaluated periods. Finally, when observing the expression of the MST1 kinase gene ( Figure 4C), a decrease in the expression of this kinase was observed in the peroxide and ERT C1 treatments (D2). These expression profiles allude to MST1-FoxO3 pathway activation (increase in FOXO3 and MST1, concomitantly with a decrease in 14-3-3), which suggests a possible cytoprotective effect of ERT for the ERT C2 + H 2 O 2 treatment. 4c), a decrease in the expression of this kinase was observed in the peroxide and ERT C1 treatments (D2). These expression profiles allude to MST1-FoxO3 pathway activation (increase in FOXO3 and MST1, concomitantly with a decrease in 14-3-3), which suggests a possible cytoprotective effect of ERT for the ERT C2 + H2O2 treatment.

Gene Expression of Antioxidants
The comparisons of the relative gene expression of SOD1 ( Figure 5) showed an increase (up to ~3.5-fold) expression in the Peroxide, ERT C1, ERT C2, and ERT C2 + 100 µM H2O2 treatments (D2), while in the other treatments (Reference and ERT C1 + 100 µM H2O2) gene expression remained stable regardless of the differentiation period. Additionally, a combined effect of C1 + 100 µM H2O2 can be perceived when compared to the treatments Peroxide and ERT C2 + 100 µM H2O2. The symbols indicate the following statistical differences: * Effect of treatment C1 + 100 µM H2O2 within each period, compared to treatment with peroxide (100 µM H2O2); ** Effect of treatment C1 compared to the reference; & Effect of lower concentration of ERT (C1) compared to higher (C2), both associated with the induction of oxidative stress (100 µM H2O2); and @ Effect of the differentiation period within each treatment compared to its counterpart in D0 and D4. The representative values of the three pseudoreplicates were expressed as mean with standard error (±SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.

Catalase (CAT)
The comparisons of the relative gene expression of mRNA CAT ( Figure 6) showed on day 2, a ~2.5-fold increase of the Peroxide treatment (compared to the reference) and a ~3.5-fold increase in expression in the ERT C1 + 100 µM H2O2 treatment, in comparison with the treatments Reference (pattern also observed in D0) and ERT C2 + 100 µM H2O2, demonstrating an effect of ERT dependent on the addition of peroxide (100 µM H2O2). The symbols indicate the following statistical differences: * Effect of treatment C1 + 100 µM H 2 O 2 within each period, compared to treatment with peroxide (100 µM H 2 O 2 ); ** Effect of treatment C1 compared to the reference; & Effect of lower concentration of ERT (C1) compared to higher (C2), both associated with the induction of oxidative stress (100 µM H 2 O 2 ); and @ Effect of the differentiation period within each treatment, compared to its counterpart in D0 and D4. The representative values of the three pseudoreplicates were expressed as mean with standard error (±SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.

Catalase (CAT)
The comparisons of the relative gene expression of mRNA CAT ( Figure 6) showed, on day 2, a~2.5-fold increase of the Peroxide treatment (compared to the reference) and ã 3.5-fold increase in expression in the ERT C1 + 100 µM H 2 O 2 treatment, in comparison with the treatments Reference (pattern also observed in D0) and ERT C2 + 100 µM H 2 O 2 , demonstrating an effect of ERT dependent on the addition of peroxide (100 µM H 2 O 2 ).

Glutathione Peroxidase 1 (GPX1)
Treatments ERT C1 and C2 showed a~1.5-fold increase in the expression of GPX1 at the beginning of erythroid differentiation (D2). Furthermore, changes of this transcript in a dose-dependent manner were observed, with C1 + H 2 O 2 drastically inhibiting the expression (~8-fold decreased) while C2 + H 2 O 2 exhibited a substantial increase (4-fold). However, the opposite is observed, more subtly, on day 4. Although not with statistical significance, D0 presents the same pattern (Figure 7).

Thioredoxin (TRX)
Peroxide treatment showed a~3-and a 2-fold decrease in expression of TRX (days 2 and 4, respectively), while ERT C1 treatment exhibited an increased expression (~3.5-fold on days 0 and 2). Although not significant, a pattern of reduction prevention (day 2) can be observed when analyzing the relative expression of the treatments C1 + 100 µM H 2 O 2 and C2 + 100 µM H 2 O 2 ( Figure 8). Treatments ERT C1 and C2 showed a ~1.5-fold increase in the expression of GPX1 at the beginning of erythroid differentiation (D2). Furthermore, changes of this transcript in a dose-dependent manner were observed, with C1 + H2O2 drastically inhibiting the expression (~8-fold decreased) while C2 + H2O2 exhibited a substantial increase (4fold). However, the opposite is observed, more subtly, on day 4. Although not with statistical significance, D0 presents the same pattern (Figure 7). Treatments ERT C1 and C2 showed a ~1.5-fold increase in the expression of G at the beginning of erythroid differentiation (D2). Furthermore, changes of this script in a dose-dependent manner were observed, with C1 + H2O2 drastically inhib the expression (~8-fold decreased) while C2 + H2O2 exhibited a substantial increas fold). However, the opposite is observed, more subtly, on day 4. Although not with tistical significance, D0 presents the same pattern (Figure 7).

Peroxiredoxins (PRDXs)
The comparisons of the relative gene expression of the transcript PRDX1 ( Figure 9A) did not show statistically significant variations between treatments or periods of differentiation. Despite this, an antagonistic expression pattern can be observed according to the concentration of ERT administered in association with peroxide, in which C1 induces, and C2 inhibits PRDX1 expression, especially on days 2 and 4. PRDX2 ( Figure 9B) showed a constant expression in all treatments during the observed periods, except for ERT C1 + 100 µM H 2 O 2 treatment (day 0), in which there is a~5-fold increase in gene expression. Although not significant, the same pattern can be observed, on day 2, for the ERT C2 + 100 µM H 2 O 2 treatment. On day 2 of PRDX6 gene expression ( Figure 9C), an increased level of transcripts (up to~3.5-fold) in the treatments Peroxide, ERT C1, ERT C2, and ERT C2 + 100 µM H 2 O 2 was observed. Furthermore, in the same period, a decreased expression in ERT C1 can be observed, with the physiological concentration being restored.
Peroxide treatment showed a ~3-and a 2-fold decrease in expression of TRX (days 2 and 4, respectively), while ERT C1 treatment exhibited an increased expression (~3.5fold on days 0 and 2). Although not significant, a pattern of reduction prevention (day 2) can be observed when analyzing the relative expression of the treatments C1 + 100 µM H2O2 and C2 + 100 µM H2O2 (Figure 8).

Proteasome
The comparisons of the relative gene expression of PSMB5 ( Figure 10A) showed a constant expression in all treatments during D0. Although, during the beginning of the erythroid differentiation process (D2), a~2.5-and 4.5-fold increased expression can be observed in the Peroxide and ERT C1 + 100 µM H 2 O 2 treatments, respectively. However, at the end of the differentiation process (D4), the expression returns to a constant state in all treatments associated with H 2 O 2 , even though a seemed elevated expression can be observed in the Peroxide treatment. All treatments with ERT associated with the oxidation induction, regardless of the period of differentiation, present expression levels equivalent to the physiological concentration, except for the ERT C1 + 100 µM H 2 O 2 treatment in D2. Regarding PSMB6 expression ( Figure 10B), on D0 and D4, a decreased pattern of expression can be seen in the treatments ERT C1 + 100 µM H 2 O 2 and C2 + 100 µM H 2 O 2 when compared with the Peroxide, ERT C1, and C2, respectively. Lastly, on the D4, higher transcripts levels of both PSMB5 and 6 (~3-and 8-fold increase, respectively) in the treatment ERT C1 can be observed.
pression. Although not significant, the same pattern can be observed, on day 2, for the ERT C2 + 100 µM H2O2 treatment. On day 2 of PRDX6 gene expression (Figure 9c), an increased level of transcripts (up to ~3.5-fold) in the treatments Peroxide, ERT C1, ERT C2, and ERT C2 + 100 µM H2O2 was observed. Furthermore, in the same period, a decreased expression in ERT C1 can be observed, with the physiological concentration being restored.

Overview of the Expression Pattern of the Genes Involved in the Redox Adaptation Mechanisms in K562 Cells
We created a Heat map to improve the overview visualization of our data set. Thus, Figure 11 summarizes the transcription level changes observed in our work, with colors ranging from green (lowest) to red (highest), indicating the level of gene expression in each treatment and period evaluated. equivalent to the physiological concentration, except for the ERT C1 + 100 µM H2O2 treatment in D2. Regarding PSMB6 expression (Figure 10b), on D0 and D4, a decreased pattern of expression can be seen in the treatments ERT C1 + 100 µM H2O2 and C2 + 100 µM H2O2 when compared with the Peroxide, ERT C1, and C2, respectively. Lastly, on the D4, higher transcripts levels of both PSMB5 and 6 (~3-and 8-fold increase, respectively) in the treatment ERT C1 can be observed. The symbols indicate the following statistical differences: @ Effect of the differentiation period within each treatment, compared to its counterpart in D0; @' Effect of the differentiation period within each treatment, compared to its counterpart in D0 and D2; ** Effect of treatment within each period, compared to treatment Reference; *** Effect of treatment within each period, compared to treatment 100 µM H2O2; ‡ Effect of ERT C1 + 100 µM H2O2 treatment compared to all treatments; ¶ Increased levels of transcripts in the ERT C1 group compared to all treatments; ƒ Effect of ERT C1 group compared to the others (except for the 100 µM H2O2 group in D2-non-

Overview of the Expression Pattern of the Genes Involved in the Redox Adaptation Mechanisms in K562 Cells
We created a Heat map to improve the overview visualization of our data set. Thus, Figure 11 summarizes the transcription level changes observed in our work, with colors The representative values of the three pseudoreplicates were expressed as mean with standard error (±SEM). General Linear Models (GLM) with two-way ANOVA format, complemented by the Bonferroni test.

Associations between Expressions of Redox Pathways Genes and Antioxidants in Erythroid Cells
The GRM analysis provides the degree of association between a set of interrelated independent variables and a dependent variable through multiple associations (R). In other words, the coefficient R indicates the contribution of the group indicated as independent in the increase or decrease in the levels of the dependent variable. Among the results of GRM, the importance of the redox pathways analyzed in the production of antioxidant enzymes and proteasomal genes (except for PRDX6 and PSMB6, whose correlations were non-significant). Consequently, the maintenance of cell integrity and redox homeostasis in K562 erythroid cells stands out. Additionally, it is important to highlight that the expression of γ-globin was not correlated with any of the analyzed genes. The results obtained in this analysis can be seen in Table 1.

Associations between Expressions of Redox Pathways Genes and Antioxidants in Erythroid Cells
The GRM analysis provides the degree of association between a set of interrelated independent variables and a dependent variable through multiple associations (R). In other words, the coefficient R indicates the contribution of the group indicated as independent in the increase or decrease in the levels of the dependent variable. Among the results of GRM, the importance of the redox pathways analyzed in the production of antioxidant enzymes and proteasomal genes (except for PRDX6 and PSMB6, whose correlations were non-significant). Consequently, the maintenance of cell integrity and redox homeostasis in K562 erythroid cells stands out. Additionally, it is important to highlight that the expression of γ-globin was not correlated with any of the analyzed genes. The results obtained in this analysis can be seen in Table 1. Furthermore, the GRM analysis also provides the partial correlation (r), i.e., the individual contribution of each independent variable on the increasing or decreasing of the dependent variable, after the control for all other independent variables in the equation (Table 2). Regarding the expression of PRDX1, SOD1, CAT, and PSMB5, the Keap1/Nrf2/ARE pathway showed a greater contribution or involvement in the production of mRNA of these genes. On the other hand, the FoxO3-MST1 pathway seems to be associated with the production of PRDX2 and TRX enzymes in K562 erythroid cells. Table 2. Individual associations between members of the redox signaling pathways and the antioxi-Furthermore, the low mortality rate observed in the treatment with the stressor agent can be explained by the endogenous production of antioxidant agents during the differentiation process induced by hemin [56]. This production is associated with the set of morphological and physiological changes triggered by the administration of hemin and hydroxyurea that increases the complexity of this cell type [57][58][59][60].
A known regulator of crucial activities for the maintenance of erythroid cells is FoxO3. This transcription factor is considered an essential factor for the maintenance of cellular homeostasis during the hematopoiesis process. FoxO3 signaling networks, their regulators, and coactivators in hematopoietic stem cells and erythroid progenitors act in order to integrate and transmit multiple signals that cooperate in regulating the erythroid cell gene expression program [61,62]. Among the activities carried out by this transcription factor are proliferation, differentiation, γ-globin expression, and transcriptional regulation in terminal erythroblastic maturation [62][63][64][65][66]. Although, in this study, no correlation was seen between γ-globin and the expression of FoxO3 (or Nrf2-a known HbF stimulator), in disagreement with the literature [66,67]. However, an increase (up to~6.5-fold) in the γ-globin mRNA levels can be associated with ERT treatments, thus suggesting its association with HbF expression in K562 cells. Along these lines, it is proposed that ERT treatment would be especially beneficial for hemoglobinopathies, such as sickle cell disease (SCD), a group of inherited diseases whose therapeutic options currently available are extremely limited. The increased expression of HbF, a known modulator of SCD, is associated with decreased clinical manifestations [68].
In this context, by a mechanism not yet elucidated, the study performed by Nakamura et al. (2007) suggests a role of ERT in the differentiation, maturation, and growth processes of erythroid cells since the absence of this thiol culminated in the reduction of these processes in K562 cells [69]. The present study corroborates this hypothesis, suggesting, for the first time in the literature, a regulatory axis for erythroid differentiation in K562 cells, in which the low concentration of ERT acts indirectly, on erythroid proliferation, through the induction of FOXO3 expression associated with the induction of one of its target genes, TRX, evidenced by a~3.5-fold increase of their mRNA levels, on D0 and 2 of ERT C1 treatment.
In addition to modulating different cellular processes, including cell proliferation [70], Trx is one of the primary cellular redox buffers whose relative expression is provenly associated with FOXO3, as demonstrated by the multiple association analysis. To this author's knowledge, such association was only reported in bovine aorta endothelial cells [71]. From these results, there is significant evidence for the establishment of the ERT-FoxO3-Trx axis as a potential therapeutic target by modulating the expression of FOXO3, a positive regulator not only of erythroid differentiation but also essential for the maintenance of redox homeostasis, through controlling Trx bioavailability, and consequentially to PRDX reduction (except for PRDX6) [72,73], which reinforces the therapeutic benefits of ERT.
The influence of ERT on the modulation of Nrf2 has already been reported by Hseu et al. (2015) [31]. We also observed an indirect action, demonstrating an interesting cytoprotective role against the oxidative environment generated by adding peroxide. K562 cell treatment, ERT C1 + 100µM H 2 O 2 , showed increased levels of NRF2, PRDX1, SOD1, and CAT transcripts on all evaluated days and PRDX2 on day 0. The association of this transcription factor with the expression pattern of these antioxidants was confirmed through the GRM analysis. On the other hand, in K562 cells treated with ERT C2 + 100 µM H 2 O 2 , activation of the redox protection pathway MST1-FoxO3 can be seen. The FOXO3 expression seems to be related to the induction of PRDX2, PRDX6, and GPX1, on day 2, suggesting an indirect action of ERT in the combat of H 2 O 2 by stimulating the expression of the FOXO3 and MST1 associated with a decreased expression of YWHAQ (14-3-3 chaperone).
Considering these results, we proposed a regulatory mechanism in which ERT influences the expression of antioxidant genes and, consequently, the antioxidant response of erythroid cells incubated with 100 µM H 2 O 2 . It is worthy of mention that some limitations of the present study need to be acknowledged, such as the evaluation of post-translational modifications of Nrf2 and FoxO3, as well the protein levels of the biomarkers. Nevertheless, this work is an outstanding hypothesis-generating study, assuming similar patterns to the transcripts expressed at the protein level. In addition, the proposed regulation model suggests a dynamic relationship exists between the redox signaling pathways of the investigated transcription factors triggered as part of an adaptive response to the addition of H 2 O 2 .
This propositioned regulatory mechanism involves multiple steps, starting with the cell signaling generated by the added peroxide itself. H 2 O 2 can lead to the reversible oxidation of redox-sensitive Cys residues in the active sites of the phosphatase and tensin homologue (PTEN), a known AKT negative regulator, resulting in the negative modulation of FoxO3 ( Figure S2-supplementary material) [74][75][76][77]. ETT transports the ERT with very high efficiency (50-200 µL/min/mg of protein; K m value of about 20 µM) into cells [32]. Additionally, its uptake increases linearly in K562 cells. These facts allow the assumption that both ERT concentrations administered are internalized minutes after their administration [69]. It is assumed that in the low concentration (ERT C1 + 100 µM H 2 O 2 treatment), the ERT cannot significantly reduce the H 2 O 2 intracellular signaling. Thus, maintaining FoxO3 in the nucleus. In the ERT C1 + 100 µM H 2 O 2 treatment, a cytoplasmatic FoxO3 concentration is proposed (first regulatory level). The modulation of the FoxO3 allows the proposition that a second transcription factor is activated, the Nrf2, whose expression was increased in this treatment. Guan et al. were the first to suggest a dynamic regulation between these transcription factors by describing a negative regulation of Keap1 caused by FoxO3 depletion [78], a situation that could be occurring here, with our results showing an increase in NRF2 expression (D2 of the ERT C1 + 100 µM H 2 O 2 treatment) and, consequently, suggesting an activation of the Nrf2-ARE pathway.
According to Dare et al. (2021), ERT can be intimately related to the activation of the Nrf2-ARE pathway. The authors showed, in silico, that ERT could directly bind to Nrf2 by substituting the IVV-ligand with ERT in the Nrf2-IVV crystal complex, leading to Nrf2-Keap1 complex dissociation, nuclear accumulation of the freed Nrf2, thus the activation of antioxidant genes [79]. Moreover, indirectly, ERT could activate the Nrf2 pathway through the oxidized intermediates of the ERT (i.e., sulfonated [ESO 3 [80]. These intermediates could destabilize the Keap1-Cul3-Rbx1 complex (e.g., through oxidation of Keap1 cysteines), stimulating the nuclear accumulation of Nrf2 ( Figure S3-supplementary material). Regardless of the action of the proposed mechanisms, the greater expression of NRF2 suggests a probable nuclear accumulation of this TF, where it will act as a positive regulator of antioxidants genes in the ERT C1 + 100 µM H 2 O 2 treatment (second level of the proposed regulatory mechanism). Thus, an ERT's dose-dependent response is observed via the activated redox pathway selected to combat oxidative stress.
Due to its role as an oxidative stress sensor and in cell signaling, the PRDX1-whose expression increased in this treatment (data corroborated by GRM), leads to the proposition of the last step for the mechanism. As described by Hopkins et al., PRDX1 can interact with FoxO3 in an oxidative stress-dependent manner influencing its subcellular location [81]. Furthermore, PRDX1 can also propagate peroxide-mediated signals in a spatially confined location, generating H 2 O 2 gradients around sites where signaling proteins are concentrated [10,82]. These characteristics allow H 2 O 2 to accumulate at substantial levels under certain circumstances, thus facilitating H 2 O 2 -dependent signaling [82,83]. Moreover, a slower rate of reactivation of hyperoxidized PRDXs (via sulfiredoxin-dependent reduction) is proposed to be a built-in mechanism, which allows sufficient time for H 2 O 2 accumulation and signal propagation [84]. Thus, in the treatment ERT C1+ 100 µM H 2 O 2 , it is assumed that PRDX1 acts reinforce FoxO3 s cytoplasmic concentration. This retention could be associated with recycling or degradation (levels of PSMB5 in this treatment support the latter).
As previously stated, in the C2 + H 2 O 2 treatment, the ERT concentration should significantly reduce the H 2 O 2 -signaling, leading to the activation of the PI3K/AKT pathway and FoxO3 nuclear retention. Accordingly, an increase in FOXO3 transcripts is observed, especially on D2. Additionally, a decreased expression of NRF2 is observed, possibly due to the dynamic relationship between FoxO3-Nrf2 previously mentioned. Therefore, by indirectly decreasing H 2 O 2 -signaling, ERT modulates the FoxO3 signaling pathway and its target genes. Additionally, PRDX1 low transcription levels reinforce this proposed theory [20,85]. The regulatory mechanisms are summarized in Figure 12.  Grune, Reinheckel, and Davies (1996) were the first to demonstrate that mild oxidative stress can increase the intracellular degradation of both short-lived and long-lived proteins in K562 cells. They proposed that the oxidative modification of substrate proteins is the major cause of increased cellular protein degradation following oxidative stress [86]. The observed proteasomal genes expression (PSMB5 and 6) allowed for the proposal of an association between ERT and proteasome in a period and dose-dependent manner. To the author's knowledge, this association is the first to ever be made in human cells. The only mention between ERT and proteasome so far was made by Takeda et al. (2010) in a study involving the fission yeast, Schizosaccharomyces pombe, in which was observed accumulation of ERT in the proteasome mutants MTS3-1 and PTS1-727 [87].
The obtained data suggested that the high concentration of ERT is responsible for the maintenance of the physiological concentration or, in some cases, led to a reduction of the transcription rate of the analyzed proteasomal subunits (especially for PSMB6 during oxidative stress). This decline could be associated with the direct depletion of H 2 O 2 by ERT, which could lead to a lower concentration of oxidized proteins and, consequently, lessen the need for new proteasome synthesis. An exception to this pattern is seen in ERT C1 + 100 µM H 2 O 2 treatment (D2), whose increased expression (of~4.5-fold for PSMB5 and 2-fold for PSMB6, respectively) could be associated with the Keap1/Nrf2/ARE pathway activation, since PSMB5 and 6 are well-known targets genes of this transcript factor [17,18] or to the proposed degradation of FoxO3. Lastly, an interesting case is observed in the D4, where high transcripts levels of both PSMB5 and 6 in ERT C1 treatment can be observed (~3-and 8-fold increase, respectively). The elevated transcript levels of these proteasome subunits suggest a molecular moment of increased protein degradation. Due to the lack of oxidative damage in this treatment (no addition of H 2 O 2 ), this increase could be associated with the primary proteolytic route for short-lived, misfolded, and damaged proteins [88].

Conclusions
The administration of ERT in erythroid cells K562 showed period and dose-dependent effects against oxidative stress induced by H 2 O 2 , with direct actions on the detoxification of the stressor and indirect on the levels of the evaluated transcripts. These observations enabled the proposition of a regulatory mechanism among FoxO3-Keap1-Nrf2 with a fundamental role in redox homeostasis in the biological model studied, in which both Nrf2-ARE and MST1-FoxO3 are involved in redox homeostasis. In contrast, the ERT-FoxO3-Trx pathway is involved in erythroid proliferation and differentiation. Lastly, ERT is proposed to be also associated with γ-globin expression. It is worth noting that future studies of protein expression and subcellular immunostaining of the analyzed pathways are necessary to prove the proposed crosstalk hypothesis between FoxO3 and Nrf2 for erythroid cells. Finally, it is suggested that ERT should be considered a key element in future investigations of therapeutic alternatives for treating hematological diseases.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author, D.S., upon reasonable request.