**4. Discussion**

Hypoxic preconditioning significantly upregulated HIF-1 α/VEGF expression in the MSCs, however, it downregulated several growth factors in the hypoxic conditioned medium implicating an enhanced HIF-1 α/VEGF signaling. As far as we know, there is no in vivo data to confirm that HMSCs transplantation can upregulate HIF-1 α and VEGF in the rat GN model. However, HMSCs transplantation increased expression of pro-survival and pro-angiogenic factors including HIF-1 and VEGF in the rat model of myocardial infarction [23] or ischemic/reperfusion kidney [24]. Our present data displayed that intrarenal arterial administration of either MSCs or HMSCs ameliorates the severity of glomerulosclerosis and levels of proteinuria in the anti-Thy-1.1 induced rat GN model. HMSCs showed a better therapeutic e ffect than MSCs on the amelioration of glomerulosclerosis, inflammatory cell infiltration, ER stress, apoptosis, and autophagy. In addition, hypoxic preconditioning enabled MSCs to activate nuclear Nrf2 expression and rescued ROS scavengers in kidneys after Thy-1.1 lesion. Our results indicate that hypoxic preconditioning further enhances the therapeutic e ffects of MSCs through multiple mechanisms including increasing intra-nuclear Nrf2 expression in the target organ.

Several studies have demonstrated that stem cells derived from various origins ameliorate kidney injury in GN animal models [2,25–29]. MSCs have the ability to cause anti-inflammation, anti-fibrosis, and inhibition of cell death and this is the basis for cell therapies [2,25,26,29,30]. In this study, both MSCs and HMSCs showed anti-inflammatory e ffects by decreasing macrophage/monocyte infiltration in glomeruli of treated kidneys and by inhibiting NF-κB translocation into nucleus. A significant further decrease in numbers was found in the HMSC treated group, indicating that hypoxic preconditioning is an e ffective strategy to promote the anti-inflammatory e ffect. The results of Masson stains confirm the therapeutic e ffect of MSCs on anti-fibrosis, and this e ffect was further enhanced by hypoxic preconditioning. We also demonstrated that intra-glomerular cell apoptosis and autophagy were decreased by MSC infusion, and a further reduction was noted in the HMSC treat group. The therapeutic anti-inflammation, anti-fibrosis, anti-apoptosis, and anti-autophagy mechanisms of MSCs are compatible to previous studies [31,32]. Hypoxic preconditioning showed a consistent enhancement of these therapeutic mechanisms seen in MSCs.

Another important finding in this study is that hypoxic preconditioning enabled MSCs to increase nuclear Nrf2 and decrease NF-κB expression. Oxidative stress generated by the immune reaction is believed to be one of the crucial mechanisms that cause injuries to glomeruli in the GN. The Keap1–Nrf2 pathway signaling and the anti-oxidant responsive elements play a central role in protection against oxidative stresses. In this study, anti-Thy1.1 lesion was found to suppress the antioxidant activity by decreasing ROS scavenger expression and elevating nuclear NF-κB, which contributed to inflammatory cytokine cascades. The nuclear Nrf2 expression remained unchanged after anti-Thy1.1 lesion, indicating that the master intrinsic anti-oxidative regulator, Nrf2/Keap1 pathway, was not triggered. Nrf2 expression levels increased with MSCs transplantation to GN rats. However, it failed to reach a significant di fference. At the same time, ROS enzymatic scavengers and other ARE proteins were not rescued. With HMSC transplantation, nuclear Nrf2 expression increased and ROS scavengers and ARE proteins were rescued in diseased kidneys. From these results it seems that conventional MSC transplantation may not trigger enough Nrf2 pathway signaling activity to enhance ARE protein expressions. Therefore, hypoxic preconditioning enabled MSCs to activate the Nrf2 pathway signaling and to rescue the ROS scavengers in kidneys which were suppressed by the anti-Thy1.1 infusion. Our findings provide evidence supporting the viewpoint of Ezquer et al.'s study [33], in which MSCs were believed to possess the main enzymatic mechanisms to detoxify the reactive oxygen species and to prevent oxidative damage in rat nephritis based on some in vitro cellular studies [34,35].

The major challenges that underlie the application of stem cell therapy to GN patients are safety concerns and e fficacy issues. Enhancing the anti-oxidative e ffect of stem cells is one promising strategy to promote their e fficacy for inflammatory or oxidative stress related disease such as GN. Nrf2 is a crucial regulator of the antioxidant defense system and governs the expression of genes associated

with redox homeostasis. The beneficial e ffects of targeting the Nrf2 pathway for nephritis have been demonstrated in animal studies through tranduction of the OR1 gene to enhance antioxidation [36] or via Keap-1 gene knockout to activate the Nrf2 system [37]. Bardoxolone, an Nrf2 activator, has been shown to improve renal functions in type 2 diabetic patients with chronic kidney disease in a human phase 2 trial [38], though it had cardiovascular safety issues [39]. The current study shows that HMSC transplantation is an e ffective measure to enhance Nrf2 pathway signaling and therapeutic e ffects in damaged kidneys. Hypoxic preconditioning, i.e., stem cell cultured in an ischemic condition which mimics the bone marrow niche environment, is a common way to preserve stemness, enhance homing, and increase e fficacy of stem cell therapy. In cellular studies, hypoxic preconditioning has enhanced stemness and expanded cell numbers [40]. Hypoxic mimetic preconditioning enhances MSC migration and prolongs kidney retention through promoting CXCR4 expression [24]. The results of this study link hypoxic preconditioning to anti-oxidative injury. We demonstrated that HMSCs promote Nrf2 signaling and resultant ARE protein elevation. The cytokines involved in the Nrf2 pathway signaling activation promoted by HMSCs need to be further investigated.

MSC homing to injured tissue is the first step and crucial because the therapeutic application of stem cell therapy is predicated on the transplanted cells migrating and participating in tissue repair. Enhancing the homing capabilities and the self-defense potential of stem cells can promote their therapeutic e fficacy. As shown in Figure 1C, the MSCs indeed transplanted to the damaged glomeruli implicating its e fficient homing to the damaged site. Hypoxic preconditioning enhanced stem cells with several possible defense mechanisms including antioxidant, ant-apoptosis, anti-ER stress, and anti-inflammatory potential and the homing HMSC may protect itself and the adjacent cells against oxidative stress through the possible autocrine and/or paracrine e ffect to release growth factors and other protective mediators. This hypothesis requires further experiments to confirm. In future, we will perform the in vitro and in vivo experiments to determine the released protective molecules from the homing HMSCs and to explore the exact mechanism for attenuating glomerular injury in the GN model. In the present study, intrarenal arterial administrations of MSCs or HMSCs lead to positive CD44 staining existence in the glomeruli. The advantage of intrarenal arterial administration can demonstrate the direct delivery and location of stem cells to the kidney and prevent the risk of stem cells trapping in the lung or other non-target tissue/organ by systemically intravenous administration. The average stem cell numbers tra fficked in glomeruli increased with hypoxic preconditioning and higher infused cell numbers. In previous studies, hypoxic precondition has been associated with increased CXCR4, CX3CR1 expression in a cellular study [17]. CXCR4 and CX3CR1 respond to SDF-1 α, activate the Akt signal pathway [41] and elevate matrix metalloproteinases [42] contributing to transmigration. Our findings confirm that hypoxic preconditioning is an e ffective strategy to enhance the homing e ffect of MSCs in the rat GN model.

The mesangial cell proliferation peaks at about 1 week, and the nephritis spontaneously repairs after 2–3 weeks in the anti-Thy1.1 nephritic model. To prevent the possible spontaneous repair e ffect, we demonstrated the clinically available HMSCs transplantation on reducing glomerular matrix accumulation, attenuating proteinuria, and ameliorating glomerular sclerosis in rats with anti-Thy1 disease at the early stage within 1 week (sacrificed on day 5). These results sugges<sup>t</sup> that renal arterial administration of HMSCs in vivo may have promise as an anti-inflammatory, anti-proliferative, and anti-fibrotic strategy in the treatment of acute phases or relapses of mesangial proliferative glomerulonephritis. There are some limitations to the current study. First, we only investigated the Nrf2 pathway and ARE expressions in the regulation of MSCs antioxidant status. Functions of other oxidative stress-related pathways, such as PI3K/Akt and FoxO/TXNIP need further elucidation. Secondly, mechanisms that influence the enhanced repairing e fficacy of MSCs after transplantation were not fully elucidated. Further studies focusing on the cytokines involved in the anti-oxidant enhancement are needed. Third, we used an acute GN model in this study. Whether these results can be applied to chronic GN, mandates further investigations.
