In the past decade, pharmaceutical companies have developed many life-saving drugs for patients with cancer. Unfortunately, some of these drugs are associated with nephrotoxicity, which remains an important and more frequent cause of acute kidney injury (AKI). Currently, AKI represents a serious global public health challenge, with significant adverse economic and medical burden, that affects approximately 13.3 million people annually (1, 2). Drug-induced AKI is a particularly important category, accounting for 19–26% of AKI cases (2). The use of nephrotoxic drugs is the main cause of severe AKI in critically ill patients. Moreover, drug-induced nephrotoxicity is one of the main factors leading to unsuccessful drug development, leading to loss of money and time in the pharmaceutical industry (3). Notably, approximately 19% of phase 3 clinical trial failures are caused by nephrotoxicity (3).

Recent studies have provided a deeper understanding of the mechanisms underlying drug-induced AKI. In particular, oxidative stress was shown to activate inflammatory responses, by promoting the release of pro-inflammatory cytokines and the accumulation of inflammatory cells in tissues, which supports the progression of AKI (4). Mesenchymal stem cell (MSC) therapy has been considered to hold great potential for the treatment of drug-induced AKI given its powerful immune regulatory effects (5). Previous studies on the immune pathogenesis of AKI and MSC therapy mostly focused on ischemia-reperfusion (IR), but drug-induced AKI is also important and needs more attention. The immune pathogenesis of IR and drug-induced AKI were reported to be different, which may also impact on the efficiency of MSC treatment on drug-induced AKI. Therefore, the differences between these two types of AKI could provide a good reference for scientists to explore MSC-based treatment strategies for drug-induced AKI in the future. In addition, it is very important to notice that MSCs can be derived from various sources and their immunomodulatory ability may differ based on their origin. When using MSCs for clinical treatment of patients, it should be considered the difficulty of obtaining these cells, as well as the number of cells and ethical limitations. Therefore, understanding the immunomodulatory ability, proliferative potential, and clinical application characteristics of MSCs from different sources can aid to choose the most appropriate MSC population for clinical use in AKI treatment.

In this review, we provide an in-depth, comprehensive summary of the immune pathogenesis of drug-induced AKI and subsequent molecular mechanisms, including the different types of immune cells, cytokines, and related signaling pathways. The immunomodulatory effects of MSCs for the treatment of drug-induced AKI were also explored based on the latest research progress. In particular, we compared the different immune pathogenesis of drug-induced and IR-induced AKI, as well as the different immune regulatory abilities of MSCs from different sources.

Several recent, large-scale epidemiological studies have shown that nephrotoxic drugs cause severe acute renal failure in many critically ill patients (2), for whom the use of potentially nephrotoxic drugs is, in most cases, unavoidable (6). Drug-induced AKI usually presents as one of two types: acute tubular necrosis, which is directly related to nephrotoxicity to the renal tubular epithelial cells, or immune-mediated acute interstitial nephritis, which develops from drugs that cause allergic reactions. Approximately 60% of patients with hospital-acquired AKI experience nephrotoxic induced by medications (7). Among the various cellular mechanisms of drug-induced AKI, the most important is the accumulation of inflammatory cells in the tissue, which in turn activates the inflammatory response and oxidative stress that triggers the release of pro-inflammatory factors. Nephrotoxic injuries often occur in the proximal renal tubules due to the administration of cisplatin, aminoglycosides (gentamycin, kanamycin, streptomycin, or tobramycin), amphotericin B, antiviral agents (adefovir, cidofovir, or tenofovir), radiocontrast, bisphosphonate, among other drugs (4).

Cisplatin is a chemotherapy drug, and is the most classic and commonly used model for basic research on drug-induced AKI, with a reported frequency of cisplatin-induced AKI ranging from 20% to 30% due to dose-limiting side effects in the treatment of various solid cancers, including testicular, ovarian, cervical, breast, lung, head and neck, bladder, and gastric cancer (8). Although there are other less common cisplatin-induced toxicities (such as ototoxicity, gastrotoxicity, and myelosuppression), nephrotoxicity has a major impact on the patient health and should be the first side effect to be monitored and evaluated in patients undergoing cancer chemotherapy. Notably, the accumulation of toxic intermediates in the kidney limits the use of cisplatin, which leads to incomplete chemotherapy. Therefore, the mechanism reviewed hereafter is mainly based on basic research conducted with the cisplatin-induced AKI model.

The typical histological features of drug-induced AKI are severe tubule damage comprising lumen dilation, significant cytoplasmic simplification, increased cytoplasmic eosinophilia, and disappearance of brush borders and of the tubule epithelium. First, several events, such as loss of polarity and cytoskeletal integrity, and cell necrosis and apoptosis, occur in the proximal tubular epithelial cells of the kidney over time (9, 10). Subsequently, necrosis induces inflammation, and damage-associated molecular patterns (DAMPs) and other pro-inflammatory molecules are released, resulting in the activation of the innate immune system via membrane surface receptors. Inflammatory cells, such as white blood cells, are recruited into the peritubular interstitium. Furthermore, inflammation accelerates the damage to the renal tubular tissues and causes necroptosis and the release of tumor necrosis factor alpha (TNF-α) and other inflammatory factors that continue to drive cell necrosis. This leads to tubular necrosis and renal insufficiency, forming an inflammation-necrosis amplification loop (11, 12). Another mechanism of drug-induced AKI is oxidative stress. Drug nephrotoxicity directly acts on the proximal renal tubules and causes cell damage, such as mitochondrial dysfunction, lysosomal hydrolase inhibition, phospholipid damage, and increased intracellular calcium concentrations, thereby leading to the formation of reactive oxygen species (ROS) (13). The pathogenic mechanisms of ROS have three main aspects: first, nephrotoxic drugs react with cellular antioxidants (such as glutathione) when they are in a highly reactive form (14, 15), thus, depleting or inactivating them, leading to the accumulation of endogenous ROS in cells. ROS activates the intracellular mitogen-activated protein kinases, p53, p21, and other pathways, leading to the death of renal tubular cells. Second, ROS directly or indirectly promotes fibrosis by promoting tissue inflammation. Fibrosis and inflammation will, in turn, promote positive feedback pathways, further increasing ROS production and stimulating the secretion of inflammatory factors. Third, nephrotoxic drugs affect the normal respiration of mitochondria, making them dysfunctional and increasing the production of ROS (16).

AKI is mainly triggered by IR injury, which causes high morbidity and mortality in both adults and children (17). IR-induced AKI results from acute hypoxia caused by reduced blood perfusion in the renal tissue, which is prone to occur in the renal tubule region. Reperfusion leads to the production of metabolites, such as nitric oxide and ROS, which can damage the cell membranes and lead to cell apoptosis. However, drug-induced AKI is more common in infants and older people with underlying cardiovascular diseases and renal dysfunction, such as intravascular volume depletion, diabetes, congestive heart failure, chronic kidney disease, and sepsis (18, 19). Drug-induced renal injury, which results from the direct damage to the renal tubular epithelial cells, occurs when the increasing concentration of nephrotoxic drugs in the renal tubule reaches a toxic level. Therefore, the degree of damage is related to the drug dose administrated. Noteworthy, there are several differences in the pathogenesis of IR-induced and drug-induced AKI; however, there are very limited systematic reviews comparing the differences in the pathogenesis between these two models. Understanding the differences in their immune pathogenesis may be helpful for the management of AKI. A summary of these differences is provided in Table 1 .

T cells play an important role in AKI. In IR-induced AKI, CD8⁺ cells have no obvious pathogenic effects. However, in the cisplatin-induced AKI model, CD8-deficient mice had significantly less damage than wild-type mice, indicating that CD8⁺ cells have an important pathogenic role in this model (20). No obvious differences in other subpopulations of T cells were found between the two models. During the inflammatory response to IR injury, the main role of renal dendritic cells (DCs) is to serve as antigen-presenting cells to T cells, release TNF-α, upregulate adhesion molecules, and promote leukocyte extravasation (21, 22). However, renal DCs play a protective role in drug-induced AKI. In one study, infiltration of DCs in damaged tissues resulted in secretion of cytokines such as interleukin (IL)-10 to prevent AKI progression (23). In the IR model, inhibition of neutrophil recruitment can significantly reduce kidney damage, which suggests that neutrophil infiltration is one of the pathogenic mechanisms of renal IR injury (24). In contrast, inhibition of neutrophil infiltration has no effect on cisplatin-induced kidney injury, indicating that neutrophil infiltration is not necessary for cisplatin-induced kidney damage (25). Importantly, macrophages have pro-inflammatory (M1) and anti-inflammatory (M2) cell phenotypes and their roles in the two AKI models are similar (26), with M1 cells being dominant in the damage stage, whereas M2 cells playing an important role in the repair stage. In addition, the relationship between the complement system and AKI has been studied more in the IR injury model than in the drug-induced AKI model. However, the mechanism of action of the complement system in both models is mainly mediated by the activation of the complement component 5a and its receptor (C5a/C5aR)–nuclear factor kappa B (NF-κB) pathway (27).

Cytokines and chemokines are differently expressed in drug-induced and IR-induced AKI, which may be due to the different cellular origins and actions of the molecules. IL-11 is upregulated in IR-induced (but not in cisplatin-induced) AKI (28). In addition, inhibition of IL-18 can significantly reduce IR-induced renal injury but not cisplatin-induced nephrotoxicity (29), which suggests that IL-18 may be an important factor in IR-induced AKI. Moreover, the expression of the C-C motif chemokine ligand (CCL) 5 and IL-1α is significantly increased in cisplatin-induced AKI, but only slightly increased in the IR injury model (28). Furthermore, TNF-α-deficient mice are resistant to cisplatin nephrotoxicity, thereby suggesting that TNF-α plays an important role in the pathogenesis of cisplatin-induced renal injury. Accordingly, TNF-α inhibitors can ameliorate renal dysfunction and reduce the structural damage caused by cisplatin, and inhibition of TNF-α more effectively reduce cisplatin-induced than IR-induced kidney damage (28). In addition, the nucleotide-binding oligomerization domain-like receptors (NLR) family pyrin domain containing 3 (NLRP3) pathway is critical in the pathogenesis of IR-induced AKI, but is less important in cisplatin-induced AKI (30).

MSCs are pluripotent stem cells with anti-inflammatory and immune tolerance potential (31), which were first identified in the bone marrow stroma where supported the development and differentiation of hematopoietic stem cells. In recent years, with research advances, MSCs have been increasingly used in clinical practice given their immunosuppressive functions and tissue repair ability (32). Hence, stem cell therapy is another medical revolution that may be considered after drug therapy and surgical treatment. MSCs can easily expand in vitro into high numbers in a short period of time. This is an important prerequisite for MSCs that are widely used in experimental research and clinical practice, including AKI treatment (33). Moreover, MSCs can be cultured from adipose tissue, cord blood, umbilical cord, placenta, and fetal lungs. However, the biological characteristics of the MSCs originating from these various tissues are different, especially concerning their immune regulation capacity (34). The immunological activity of MSCs from different tissues may differ because of the different original activation states of these cells in the source tissues (35–37). The differences in the immunomodulatory ability, proliferation potential, and clinical application characteristics of MSCs from different sources is summarized in Table 2 .

Current studies have shown that the immunological activity of MSCs originating from different tissues is “strong” or “weak” rather than “present” or “absent” (38). Yoo et al. compared the immune regulatory functions of adipose-derived MSCs (AD-MSCs), umbilical cord blood-derived MSCs, umbilical cord-derived MSCs (UC-MSCs), and bone marrow-derived MSCs (BM-MSCs) on T lymphocytes (34), and found that all four types of MSCs inhibited the proliferation of activated T cells and the secretion of interferon-γ and TNF-α. Moreover, Bochev et al. found that both BM-MSCs and AD-MSCs could inhibit the secretion of immunoglobulins by activating B lymphocytes (39). However, the inhibitory effects of AD-MSCs on immunoglobulin secretion were stronger than those of BM-MSCs. Some study also reported that AD-MSCs have higher indoleamine 2,3-dioxygenase (IDO) activity and secret more IL-6, IL-10, and transforming growth factor (TGF)-β than BM-MSCs (35, 40). Moreover, Barcia et al. found that compared with BM-MSCs, UC-MSCs have stronger immunomodulatory abilities, can more efficiently inhibit CD3- and CD28-induced lymphocyte proliferation, and can more efficiently induce the production of CD3⁺CD4⁺CD25⁺Foxp3⁺ regulatory T cells (Tregs) (41).

Nevertheless, some studies have shown that the in vitro expansion capabilities of MSCs from different tissues are not completely the same. Different microenvironments may be the reason for these differences. Li et al. compared the proliferative capacity of BM-MSCs and AD-MSCs, and found that AD-MSCs have stronger proliferation potential (42). However, as the age of the donor increases, the number and proliferation capacity of BM-MSCs and AD-MSCs significantly decreases. MSCs derived from perinatal tissues (such as umbilical cord, placental chorion, and amniotic membrane) have much stronger proliferation potential than BM-MSCs. Comparing MSCs isolated from umbilical cord tissue and bone marrow, more MSCs can be obtained from umbilical cord tissues, and with greater proliferation ability than BM-MSCs (43). After 30 generations, the proliferative capacity of UC-MSCs did not change significantly, whereas BM-MS showed weakened proliferation ability and prolonged doubling time after six passages.

In summary, perinatal MSCs (MSCs derived from the umbilical cord and placenta) generally have higher immunomodulatory capabilities, while BM-MSCs show higher potential to support the regeneration process, such as neuronal differentiation and development (43). These differences between perinatal and bone marrow-derived MSCs may impact on their clinical application. In addition, when MSCs are used in the clinical treatment of diseases, the feasibility of its acquisition, the number of cells, and ethical limitations must be considered. For example, BM-MSCs and AD-MSCs have several limitations. They are generally obtained from adults and the number of cells available is limited. Moreover, there are restrictions regarding the age of the donor. However, MSCs in the perinatal period have stronger expansion capacity and have no ethical restrictions. After expansion, the basic properties and morphology of MSCs have no obvious changes, so that a large number of MSCs can be obtained from the same sample to meet the needs of the clinical cell therapy (43). Therefore, compared with adult-derived stem cells that proliferate slowly, perinatal MSCs have unique advantages. From this perspective, MSCs derived from umbilical cord and placenta are an ideal choice.

The strong ability of MSCs to differentiate is one of the main mechanisms contributing for tissue damage repair. Although the initial focus on MSCs was on their regenerative ability to differentiate into various tissue cell types (33), more attention is currently addressed onto their ability to regulate immune responses (44). In vivo and in vitro studies have shown that MSCs can directly inhibit the proliferation, differentiation, and effector mechanisms of T cells by preventing the externalization and maturation of DCs. In addition to T cells and DCs, the target cells of MSCs include many other types of immune cells, including natural killer cells, B cells, neutrophils, and macrophages (45). MSCs regulate the function of these cells, including the secretion of cytokines and their subsequent cytotoxic effects, eventually exerting anti-inflammatory effects and/or inducing an immune tolerance state. To date, the immune regulation mechanisms of MSCs remain poorly understood, but it is generally believed that MSCs have an immunosuppressive function mainly through cell-to-cell contact with a variety of immune cells and secretion of soluble factors, such as prostaglandin E2 (PGE2) and IDO, thereby reducing tissue damage caused by inflammation (46). MSCs have low immunogenicity, that is, MSCs do not express major histocompatibility complex class II molecules or costimulatory molecules, and do not cause the activation and proliferation of allogeneic lymphocytes, which provides the possibility for the clinical application of allogeneic MSCs transplantation (47). These cells inhibit the proliferation of activated T lymphocytes and the secretion of interferon-γ, the proliferation of B lymphocytes and the secretion of immunoglobulins, the killing activity of natural killer cells, the differentiation of M1 macrophages, the secretion of pro-inflammatory factors, and the differentiation of monocytes into DCs (48).

There is a very close connection between the immune system and the kidney system (49). Pro-inflammatory DAMPs, pathogen-associated molecular patterns (PAMPs), toll-like receptors (TLRs), oxidative stress, complement system, resident DCs, neutrophils, T lymphocytes, macrophages, and secreted cytokines and chemokines, all participate in the immunopathological mechanism of drug-induced AKI and promote disease progression (20, 28, 50–53). Therefore, imbalance of the immune system has a direct or indirect impact on the progression of kidney disease. If the immunopathological process of AKI continues, this may lead to renal fibrosis and/or chronic kidney disease. The immunomodulatory mechanism of MSCs in AKI treatment is shown in Figure 1 .

To date, the safest route of administration of MSCs in clinical applications is through vascular injection; thus, the number of MSCs that can reach the kidney is very small (47). Several scientists have attempted to improve the efficacy of MSCs through various methods. Improving the immunomodulatory ability of MSCs is key to improving their therapeutic effect. Studies have been conducted to regulate the relevant immune responses by innovative pretreatment methods, thereby enhancing the kidney repairing effect (78, 79). Currently, most reviews focus on AKI due to ischemia-reperfusion. However, there are relatively few studies on drug-induced AKI, especially on the role of the complement system. Drugs-induced kidney injury is also one of the biggest causes of AKI (4). Future research in this area will help deepen our understanding of the immune regulation mechanism of MSCs therapy for AKI, and discover more potential therapeutic targets.

QH and XW had the original idea and wrote the first draft. HZ, JH, and GC reviewed the manuscript. HZ and XD provided critical input. All authors contributed to the article and approved the submitted version.

This research was funded by The National Natural Science Foundation of China (No. 61971441) and the National Key R&D Program of China (No. 2016YFC1305500, 2018YFA0108803).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.