**4. Discussion**

Cisplatin (CDDP) is highly deleterious at proximal tubule level given its high density of mitochondria because of its reabsorption function [13]. Its nephrotoxicity is attributable to high CCDP accumulation in the kidneys and adverse impacts on the renal transport system. CDDP entry into the tubular cells by passive diffusion or by a number of cellular transporters including human copper transport protein 1 (Ctr1) and the organic cation transport 2 (OCT2), which are highly expressed on renal tubular cells [39]. Although this transport is expressed in other tissue, diverse strategies have demonstrated its importance in nephrotoxicity development [2]. Inside cells, CDDP metabolism results in its activation to a more potent toxin and glutathione-cisplatin-conjugate derivate and compromises glutathione level and synthesis [40]. CDDP can bind to proteins of cytosol, mitochondria, and ER [2,40,41]. Furthermore, CDDP binding to nuclear and especially mitochondrial DNA, trigger mitochondrial dysfunction [2]. Although hydration and diuretics are used as a strategy to prevent CDDP-induced nephrotoxicity, the optimal strategy to prevent this pathology is still being sought [1]. For this reason, new strategies are being developed that include using phytochemical compounds such αM [5].

αM has many biological functions [6,8]. There are three studies that demonstrated nephroprotection versus CDDP, without interfering with urinary secretion or antiproliferating activity [3,7,9]. Alpha-mangostin is not specific to tubular cells; however, it preserves the renal function. Pharmacokinetics studies have shown that it can be quantified in the kidney and liver [6] as well as its nephroprotective effects [3,7,9,11]. αM protection in CDDP nephrotoxicity related to its antioxidant properties has been described [3,7,9]. To study αM mechanism we used the LLC-PK1 cell line because it is very stable, undergoing little to no transformation or neoplastic change after numerous passages [42] and it shows tubular cells characteristics [43,44]. We corroborated in LLC-PK1 cells that αM co-incubation protects against CDDP-induced (30 μM) viability decrease (Figure 1C). The protection was found at 4 μM and 5 μM, however, no differences were found between 4 and 5 μM (Figure 1), this time using a much purer αM extract than the previous report [9,25]. It has been previously reported by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) viability assay 7.5 μM αM reduced viability cell [10], we reproduced this experiments with 1, 2.5, 5, 7.5, and 10 μM and found that αM can reduce near 50% viability at 7.5 μM (data not show). For this reason, we decide to use the minimum αM concentration in the present study. Our group has previously demonstrated that CDDP induces oxidative stress in vitro, and αM was able to prevent the increase in ROS and the decrease in glutathione (GSH) [9]. In our study, αM was unable to prevent MDA increase (Figure 2A). We are tempted to speculate that the use of different markers (MDA versus ROS and GSH) may explain the differences in both studies.

In this study, we demonstrated that CDDP induced mitochondrial bioenergetics alterations, characterized by the reduction in the respiratory parameters: routine, leak, E, and P (Figure 3), as well as CI-linked respiration (Figure 4). The absence of change in RC (Figure 3E), suggests that the CDDP-induced respiratory state alterations were related to a decrease in whole ETS activity rather than a reduction in a specific complex. Which was congruent with the absence of a reduction in the percentage of basal respiration attributable to CI (Figure 4B) and would be responsible for the decrease in OXHPOS capacity (Figure 3C). It was also demonstrated for the first time that αM prevents the bioenergetics alterations induced by CDDP as we can see by the preservation of all evaluated mitochondrial parameters in the co-treatment group (Figures 3 and 4). Mitochondrial dysfunction has been widely recognized as an important factor in renal disease, especially in tubular damage [13]. Furthermore, in the CDDP model, the oxidative stress and the mitochondrial dysfunction favor the activation of apoptosis, which contributes to renal damage [15,17,45]. So, the protection by αM would be, at least in part, attributable to mitochondrial preservation (Figures 3 and 4), which explains the protection of tubular epithelial cells.

Furthermore, the decrease in the levels of VDAC and respiratory complex subunits (Figure 5A–F), implies that the observed mitochondrial bioenergetics alterations are attributable to mitochondrial mass reduction and may also be attributable to mitochondrial network fragmentation. This is congruent with a previous report that demonstrates a shift of mitochondrial dynamics towards fission in CDDP-induced nephropathy [14]. Although the co-treatment with αM had no effect in the evaluation of mitochondrial protein subunit levels (Figure 5A–F), it prevented the alteration in the morphology of the mitochondrial network (Figure 5G), which is congruent with the preservation of mitochondrial dynamics, demonstrated by other antioxidants, such as curcumin in CDDP-induced damage nephropathy [14].

Our results also suggest that the observed reduction in mitochondrial mass is attributable to decreased mitochondrial biogenesis. We found a decrease in TFAM levels in CDDP-treated cells (Figure 6C,D), but PGC1α levels did not change (Figure 6A), suggesting a reduction in mitochondrial biogenesis. We suggest this, first, because although PGC-1α expression did not change, its activity also has to be considered. PGC-1α is regulated via phosphorylation by protein kinase B (AKT) and adenosine monophosphate-activated protein kinase (AMPK), as well as deacetylation by sirtuins [46]. In fact, it has been reported by Ortega-Dominguez [14] that CDDP decreased sirtuin-3 the levels, which triggers the increase in lysine acetylated levels. Therefore, this evidence supports the idea that PGC-1α is inactive in the cisplatin group. Second, as with PGC-1α, NRF1 expression does not necessarily correlate with its activity. NRF1 activity can be regulated by phosphorylation and/or by interactions with PGC-1α, PGC-1β, and PRC [14,46]. In this sense, TFAM levels determination is an indirect proof of its activity. About it, CDDP exacerbated NRF1 levels, whereas TFAM decreased. So, it can be suggested that CDDP-induced NRF1 overexpression is not functional, unlike the co-incubation with αM. Taken together, we can suggest cisplatin suppress the mitochondrial biogenesis pathway. Interestingly, we found increased levels of NRF1 in CDDP-treated cells (Figure 6B). This induction can be explained by the fact that the promoter of the NRF-1 gene possesses antioxidant response elements. Therefore, the nuclear factor erythroid 2-like 2 (Nrf2) can induce its expression [47]. Although we did not evaluate Nrf2 induction in this work, our group has previously reported that CDDP induces the increase in heme oxygenase-1 (HO-1) levels in LLC-PK1 [26], leading to Nrf2 activation, so the observed NRF1 increase can be the consequence of an adaptive response of Nrf2 activation. On the other hand, it was documented that αM can induce Nrf2 activation in retinal cells and liver tissue [48,49], which can also explain the observed increase in NRF1 levels in co-treatment and αM groups (Figure 6B,D), but more experiments are necessary to demonstrate Nrf2 induction.

In view of the fact that mitochondrial mass is regulated by biogenesis and mitophagy [13], we evaluated the mitophagy markers Pink1/Parkin. Co-incubation with αM avoided Parkin (Figure 7B) and PINK1 (Figure 7A) increase induced by CDDP. Recently, mitophagy was described as a response mechanism in CDDP-induced nephrotoxicity, and the PINK/Parkin pathway was suggested as the main route to remove damaged mitochondria [22–24]. The mitophagy machinery can be shared with the macroautophagy, but can also be independently-activated [19]. We show that pharmacological autophagy inhibition with CQ or with wortmannin did not have any effect in the protection by αM and did not increase CDDP toxicity (Figure 8). However, the increase in p62 levels (Figure 9C,D) and the no change of the LC3 ratio in CDDP-treatment at 24 h suggest that increases of PINK/Parkin can use the p62 protein as an adaptor, as suggested in a previous report [37]. However, it is also known that the accumulation of misfolded proteins leads to aberrant p62 expression [38]. Altogether, our results support the idea of a partial preservation of mitochondrial mass and function by αM-treatment.

We found 4 μM of αM did not change the respiratory parameters (Figures 3 and 4) nor the expression of mitochondrial proteins (Figure 5), except CII-SDHB (Figure 5B) and CIV-MTCO1 (Figure 5C). Unfortunately, nowadays, we do not have enough information to explain the origin of these changes. However, it is important to note, that although these subunits decreased in αM treatment, the functional respirometry studies did not show changes in any of the evaluated parameters with respect to the control group (Figures 3 and 4). This implies the observed CII-SDHB and CV-ATP5A decrease did not change the mitochondrial bioenergetics in the αM treatment with respect to the control. And we did not observe changes in the mitochondrial network by MTG (Figure 6A). Additionally, αM alone did not induce mitochondrial biogenesis (Figure 7) or mitophagy activation (Figure 8), implying that at least at this concentration, αM has no negative effects on mitochondrial function. In support of this, studies in heart tissue (200 mg/kg) and hepatic cells (30 μM) describe that αM can avoid mitochondrial dysfunction (preventing the decrease in activity and levels of antioxidant proteins and in RC, as well as ROS increase) induced by isoproterenol or free fatty acids increase [50,51]. By contrast, reports also describe that high αM concentrations (above 25 μM) can induce mitochondrial alterations in mitochondria isolated from kidneys [52], indicating that αM effects are tissue- and concentration-dependent. Although there are no clues regarding αM induced proliferation in normal cells, we evaluated GRP94, an ER chaperon, with 4 μM αM and found can it induce proliferation at 24 h of incubation (data not show). GRP94 is a glycoprotein related to protein folding, stores of calcium and targeting proteins to ER-associated degradation (ERAD), the client protein of GRP94 is related to specialized functions in immunity, grown signaling, and cell adhesion [53]. Probably, GRP94 induction

in αM treatment can be related to grown signaling and cell adhesion, and this can explain this result, but more experiments are necessary.

As we show in the integrative scheme (Figure 11) our results suggest that 30 μM CDDP-treatment induces alterations in LCC-PK1 cells principally by reducing mitochondrial function and mass, which is partially attributable to mitochondrial biogenesis reduction and mitophagy activation. αM can prevent mitochondrial bioenergetics alterations and the induction of mitophagy. These mechanisms would be involved in the protective effects observed in LLC-PK tubular cells.

**Figure 11.** Integrative scheme. CDDP induces (1) the increase MDA level, (2) decrease respiratory states (except respiratory control, RC), the mitochondrial network fragmentation, the mitochondrial fusion protein 2 (MNF2) decrease and induce mitophagy (PINK1/Parkin and p62 increase). Furthermore, CDDP induces (3) TFAM reduction which reduces mitochondrial protein (VDAC and OXPHOS) and probably mitochondrial transcription, and also induces an NRF1 increase. The αM prevented all alterations, except NRF1 induction. Created with BioRender.com.
