In experimental AKI models, blood-circulating Ly6C^high monocytes are recruited to the inflamed kidney as early as within 1 h (Li and Okusa, 2010; Zhang et al., 2012). The migration of Ly6C^high monocytes to the site of inflammation occurs through chemotactic mechanisms (e.g., via CCR2 and CX₃CR1). Therefore, deletion or blockage of chemotaxis receptors on monocytes is found to be protective against ischemia-induced AKI in mice (Furuichi et al., 2003; Li et al., 2008; Lu et al., 2008; Oh et al., 2008; Yang et al., 2019). Monocyte infiltration occurs in the first 48 h (Lu et al., 2008) and completely ceases before day 3 of AKI. Accordingly, studies showed that the number of Ly6C^high monocytes and M1-like Mϕ drastically declines before day 3 of IRI (Lee et al., 2011; Lever et al., 2019) and CX₃CR1-dependent monocyte migration is not detectable at day 3 of UUO (Peng et al., 2015). Interestingly, the peak of tubular injury [e.g., following IRI (Lee et al., 2011) and glycerol injection (Bussolati et al., 2005)] timely overlaps with the maximum presence of infiltrating monocytes, indicating a close spatial and temporal relationship between the tissue destruction and the accumulation of infiltrating monocytes.

Ly6C^high monocytes differentiate into Mϕ, which are primarily skewed toward an M1 phenotype. M1 Mϕ polarization is mediated by pro-inflammatory cytokines [e.g., IFN-γ, IL-6, IL-1β, IL-23, IL-17, C3, C5a, and C5b (Rabb et al., 2016)] and DAMPs [e.g., high mobility group protein B1 (HMGB1), adenosine triphosphate (ATP), uric acid, or hypomethylated DNA (Anders, 2010; McDonald et al., 2010)] released by dying cells or damaged ECM (Anders and Schaefer, 2014). Most recently, soluble epoxide hydrolase was identified as a proximal tubular factor driving M1 polarization of Mϕ in IgA nephropathy (Wang Q. et al., 2018). DAMPs activate various PRRs [e.g., TLR families (Kulkarni et al., 2014; Leaf et al., 2017), NLRP3 and purinergic receptors (Anders and Schaefer, 2014)] on Mϕ and parenchymal cells (Leaf et al., 2017). The importance of DAMPs in inducing innate immune responses is highlighted by findings that the inhibition of PRR signaling suppresses immune responses in AKI (Kim et al., 2013; Leaf et al., 2017). Similar mechanisms are known from acute injuries in other organs, corroborating that DAMPs are central to the immune activation during tissue injury (Egawa et al., 2013; Wang et al., 2014; Groves et al., 2018). Interestingly, TLR4 activation was also shown to induce the expression of IL-22 in Mϕ, which is protective against AKI accelerating kidney regeneration (Kulkarni et al., 2014). M1 polarization of infiltrating Mϕ is additionally supported by parenchymal factors [e.g., Krüppel-like factor 5 (KLF5) expressed by collecting ducts (Fujiu et al., 2011) and suppressor of cytokine signaling 3 (SOCS3) upregulated by proximal tubules in AKI (Susnik et al., 2014)]. Both KLF5 and SOCS3 promote M1 activation of Mϕ and inhibit the expansion of M2 Mϕ in AKI (Fujiu et al., 2011; Susnik et al., 2014). M1-activated Mϕ largely produce pro-inflammatory cytokines and mediators (e.g., IL-1α, IL-6, IL-12, IL-18, TNF-α, nitric oxide), in turn, exacerbating the kidney inflammation (Li and Okusa, 2010).

Inflammation following a transient insult is meant to prepare the tissue for healing. When the inflammation escalates (before day 3 of AKI), Mϕ seek to counteract overwhelming immune activation by skewing toward an immunosuppressive M2 Mϕ to restore tissue homeostasis (Figure 3) (Lee et al., 2011; Baek et al., 2015). However, this only depicts the global view of Mϕ dynamics and does not resolve how individual Mϕ subtypes change during AKI. Since Mϕ are highly plastic and rapidly adapt to the tissue microenvironment, it is difficult to trace the development of individual Mϕ subtypes during AKI. Nevertheless, we are steadily expanding our knowledge base through genetic fate mapping studies and parabiosis experiments. Earlier fate mapping studies revealed that Ly6C^high monocytes infiltrating the inflamed kidney give rise to Ly6C^low and Ly6C^int Mϕ, both phenotypically resembling tissue-resident Mϕ (Lin et al., 2009) (Figure 4). Several studies have shown that monocyte-derived Ly6C^int and Ly6C^low Mϕ populations display transcriptionally and functionally distinct M2 phenotypes, both implicated in immunosuppression and tissue regeneration. In the later stages of AKI, Ly6C^low Mϕ predominate over Ly6C^int Mϕ and are found to promote interstitial fibrosis (Lin et al., 2009; Clements et al., 2016; Lever et al., 2019; Yang et al., 2019) (Figure 4). More recent studies revealed that quiescent tissue-resident Mϕ remain in the tissue independently of monocyte-derived Ly6C^low Mϕ (Lin et al., 2009; Zhang et al., 2012; Lever et al., 2019) and are reprogramed in AKI toward a developmental state resembling perinatal Mϕ (Schulz et al., 2012; Mass et al., 2016), which are implicated in early kidney development (Lever et al., 2019). These cells display a unique transcriptional profile complying with neither canonical M1 nor M2 nor quiescent Mϕ phenotypes during the first 3 days after IRI. Interestingly, they activate the canonical wingless-type MMTV integration site family (Wnt) signaling by expressing Wnt ligand genes and downstream intracellular signaling mediators, implying that they mediate kidney healing after AKI (Lever et al., 2019). How reprogramed kidney-resident Mϕ further develop in the later stages of AKI and whether they are related to interstitial fibrosis following AKI deserve further investigation (Figure 4).

Beneficial effects of M2-activated Mϕ in AKI are supported by many findings: (1) M2 Mϕ clear intraluminal debris (e.g., by apoptosis inhibitor of Mϕ [AIM]-dependent mechanisms) (Arai et al., 2016); (2) secrete tissue-reparative factors, which limit cell cycle arrest (Lin et al., 2010) or apoptosis (Lin et al., 2010; Sola et al., 2011) or which support proliferation in tubular cells (Schmidt et al., 2013) [e.g., Wnt-7b (Lin et al., 2010), lipocalin-2 (Sola et al., 2011), breast regression protein 39 (BRP-39) (Schmidt et al., 2013)]; (3) secrete anti-inflammatory cytokines, which suppress effector T cells or activate regulatory T cells (e.g., IL-10, TGF-β) (Chen et al., 2019); and (4) reduce neutrophil infiltration by downregulating intracellular adhesion molecule-1 (ICAM-1) (Karasawa et al., 2015) and potentially also by sequestering the tissue damage through “cloaking” mechanisms as found in the peritoneal serosa (Uderhardt et al., 2019) etc. As M1 Mϕ are converted into M2 Mϕ in the resolution phase of AKI, a number of studies have focused on identifying stimuli driving Mϕ phenotypic switch during AKI. These stimuli include: (1) paracrine factors released from parenchymal and immune cells; (2) systemic factors, which are released into the blood circulation; and potentially (3) (tissue micro)environmental changes [e.g., apoptotic neutrophils (Filardy et al., 2010; Lee et al., 2011), oxygen (Raggi et al., 2017), and nutrient availability (Geeraerts et al., 2017)].

As mentioned above, proximal tubules are the main locale of the inflammation and potent producers of cytokines in AKI. Therefore, it would not be surprising if they substantially contributed to the phenotypic switch of Mϕ in AKI. In proximal tubule/Mϕ co-culture experiments (proximal tubules and Mϕ physically separated), quiescent proximal tubules are capable of polarizing both non- and M1-activated Mϕ toward an M2 Mϕ phenotype in a paracrine manner, similarly as known from mesenchymal stem cells (Lee et al., 2011; Huen et al., 2015). Unfortunately, the proximal tubular mechanisms of M2 polarization are still largely elusive, and it is also unclear whether the contribution of proximal tubules or proximal tubule-derived factors is indispensable for M2 Mϕ polarization in the resolution phase of AKI. Several studies suggest that proximal tubule-derived Mϕ survival factors (e.g., CSF-1, -2, IL-34) drive M2 Mϕ polarization, but these results remain controversial as discussed in the following paragraph. Proximal tubule-derived factors identified to polarize Mϕ toward an M2 phenotype are Wnt ligands and netrin-1, which are both upregulated during AKI (Reeves et al., 2008; Grenz et al., 2011; Ranganathan et al., 2013; Feng et al., 2018a, 2018b). It has been shown that the blockade of Wnt/β-catenin signaling diminishes M2 Mϕ polarization also reducing interstitial fibrosis in AKI (Feng et al., 2018a, 2018b). Netrin-1 deficiency was found to aggravate AKI, whereas the adoptive transfer of netrin-1-treated Mϕ was protective against AKI (Reeves et al., 2008; Grenz et al., 2011; Ranganathan et al., 2013). Proximal tubules also express both transforming growth factor β (TGF-β) and its receptors at high levels. While TGF-β with its pleiotropic effects acts on various cell types, it is known to polarize Mϕ toward an anti-inflammatory (Wang et al., 2005) and pro-fibrotic phenotype (Braga et al., 2015). However, it is unclear how beneficial the Mϕ-specific effects of TGF-β are on AKI as TGF-β can signal directly to proximal tubules and induce proximal tubular apoptosis (Nath et al., 2011; Gewin et al., 2012). In addition, TGF-β may promote the persistence of fibrotic M2 Mϕ and mediate interstitial fibrosis (Martinez et al., 2009; Lech and Anders, 2013; Chung et al., 2018).

CSF-1 and -2 are produced and up-regulated by proximal tubules during AKI. Many studies have pinpointed Mϕ survival factors, CSF-1 (Menke et al., 2009; Alikhan et al., 2011; Zhang et al., 2012; Wang et al., 2015) and -2 [also known as granulocyte-Mϕ CSF (GM-CSF)] (Huen et al., 2015), as factors driving the phenotypic switch toward an M2 phenotype, but this feature of Mϕ survival factors is highly controversial. CSF-1 and -2 are commonly used for generating in vitro Mϕ from bone marrow or blood monocytes, both being sufficient for Mϕ differentiation and maturation. Since CSF-1 and -2 mature and induce expression of distinct patterns of functional genes after a sufficient culture period, researchers have been incited to determine the polarization potential of CSF-1 and -2 and were led to propose that CSF-1 give rise to a more M2-like and CSF-2 to more M1-like expression patterns in Mϕ in vitro (Lacey et al., 2012). Nevertheless, it is important to understand that the translatability of in vitro data is limited as CSF-1 and -2 show in vitro M2 polarization potential only at high concentrations (Lutz et al., 2000; Hume and MacDonald, 2012; Huen et al., 2015) while being efficient at maintaining Mϕ already at low concentrations (Hamilton et al., 1988; Lutz et al., 2000; Meshkibaf et al., 2014). Whereas a number of studies have claimed that both CSF-1 and -2 drive M2 skewing of Mϕ in mice with AKI (Zhang et al., 2012; Huen et al., 2015; Wang et al., 2015), it was controversially found that IL-34, another ligand for CSF-1R, does not polarize Mϕ in murine AKI (Baek et al., 2015) and lupus model (Wada et al., 2019), indicating that CSF-1R signaling is dispensable in M2 Mϕ polarization. Supportive of this data, other studies have shown that: (1) increased CSF-1 expression in the resolution phase of AKI is not sufficient to prevent Mϕ from M1 polarization when Mϕ are exposed to an M1 stimulus or when they are deprived of an M2 stimulus during AKI (Fujiu et al., 2011; Susnik et al., 2014; Chiba et al., 2016); (2) quiescent and M2 Mϕ in the resolution phase of AKI differ in transcriptional profiles and functions (Lever et al., 2019; Yang et al., 2019); and (3) the sustained blockage of CSF-1R or the constitutive deletion of CSF-1 ameliorates AKI (Lenda et al., 2003; Ma et al., 2009; more discussion in Assessing Mϕ functions by depleting Mϕ section). Nevertheless, what is consistent throughout all studies (Zhang et al., 2012; Baek et al., 2015; Huen et al., 2015; Wang et al., 2015; Chiba et al., 2016) is that the deficiency in Mϕ survival factors reduces the number of Mϕ (including that of M2 Mϕ predominating in the resolution phase of AKI). It is interesting to note that the deletion of proximal tubule CSF-1 or the blockade of CSF-2 in AKI leads to a reduction in expression of M2 Mϕ-specific genes, which appears modest (<30%; except regarding Arg1 expression), which may have resulted from the altered ratio of infiltrating and kidney-resident Mϕ, as blood-circulating Mϕ are not affected by the deletion of proximal tubule CSF-1 or the blockade of CSF-2 (Huen et al., 2015; Wang et al., 2015). The observation that clodronate-induced Mϕ depletion increased initial AKI and reduced recovery in the absence of proximal tubule CSF-1 (Wang et al., 2015) provided additional evidence that CSF-1 is not required or sufficient for M2 Mϕ polarization (also commented in Perry and Okusa, 2015). Taken together, CSF-1, -2 and IL-34 are likely not sufficient to polarize Mϕ toward M2 phenotype, but promote the expansion of M2 Mϕ post-AKI (Chiba et al., 2016). It would be interesting for future work to explore whether Mϕ survival factors have only redundant functions and, if not, what the unique, non-overlapping, functions of these Mϕ survival factors are.

Th2 cytokines, IL-4, -10, and -13, are detected in the resolution phase of AKI, but they are not functionally expressed in proximal tubules (Andres-Hernando et al., 2017; Zhang et al., 2017). IL-4 and -13 are produced by Th2 T cells, basophils, mast cells, and granulocytes, whereas IL-10 is produced by regulatory T cells (Liu et al., 2011), B cells as well as Mϕ, induced by prostaglandins, glucocorticoids, apoptotic cells, and G protein-coupled receptor ligands (Zhang et al., 2017). IL-10, which is part of negative feedback response to inflammation and expressed along with pro-inflammatory cytokines, is released into the local tissue and blood circulation and contributes to the suppression of AKI (Deng et al., 2001; Wan et al., 2014; Greenberg et al., 2015; Zhang et al., 2015). Circulating pentraxin-2, also known as serum amyloid P, is found to facilitate the uptake of apoptotic cells and to bind to Fcγ receptors by opsonizing apoptotic cells. This process triggers IL-10 expression and M2 polarization in infiltrating Mϕ (Castano et al., 2009). IL-4 and -13 activate IL-4Rα and its downstream signaling molecule STAT6 and mediate tissue repair and IL-10 immunosuppression (Zhang et al., 2017). IL-4-stimulated Mϕ, not M1-stimulated Mϕ, promote tubular cell proliferation (Lee et al., 2011). Locally synthesized RA, most likely produced by peritubular Mϕ, represses M1 Mϕ and activates RA signaling in the injured tubular epithelium, which, in turn, promotes M2 polarization, thereby reducing Mϕ-dependent injury post-AKI (Chiba et al., 2016). As mentioned above, the proximal tubular mechanisms of M2 Mϕ polarization are only partially understood and deserve more investigation in the future.

AKI is reversible as long as the cause has been eliminated and the tissue has not been structurally damaged (Chawla et al., 2011; Basile et al., 2012). So far, it is largely unknown which mechanisms determine full recovery versus subsequent CKD after AKI. For a full recovery after AKI, two conditions regarding Mϕ need to be fulfilled: (1) re-transforming and/or removal of pro-fibrotic M2 Mϕ; and (2) the decline in Mϕ numbers to the basal level. Importantly, Mϕ numbers during AKI are supposed to be strictly controlled across all stages, and uncontrolled hyper-proliferation or inadequate removal of M2 Mϕ in the resolution phase of AKI may cause a non-resolving inflammation and chronic pathology as we observe in other disease areas (e.g., in muscle inflammation (Iwata et al., 2012; Baek et al., 2015; Baek et al., 2017). So far, virtually nothing is known about the fate of Mϕ after the tubular repair is complete, and this needs to be investigated in the future (Huen and Cantley, 2017). Since M2 Mϕ are considered to be protective in AKI, there is a growing interest to use M2 Mϕ and Mϕ-modulating agents as therapeutic tools to treat patients with AKI. However, it is to note that M2 Mϕ are considered to be instrumental in the development of pathological fibrosis and the progression of CKD (Duffield, 2010; Anders and Ryu, 2011). Especially, several studies have identified monocyte-derived Ly6C^low Mϕ, which predominate over Ly6C^int Mϕ in the later stages of AKI, as direct or indirect contributors to interstitial fibrosis (Lin et al., 2009; Anders and Ryu, 2011; Clements et al., 2016; Lever et al., 2019; Yang et al., 2019) and as a hallmark of CKD progression post-AKI. In line with this, CX₃CL1-CX₃CR1-mediated survival of Ly6C^low Mϕ correlates with interstitial fibrosis in obstructed kidneys (Peng et al., 2015). M2 Mϕ may take an important pro-fibrotic role (1) by promoting the formation of a provisional ECM (containing fibrin, fibrinogen, and fibronectin), which mediates the recruitment of fibrocytes, giving rise to myofibroblasts (= the effector cells in fibrosis, which, in turn, produce large amounts of ECM components); (2) by expressing matrix metalloproteases, some of which serve as essential drivers of fibrosis; and (3) by secreting large amounts of pro-fibrotic factors, which activate and differentiate resident fibroblasts and infiltrating fibrocytes into myofibroblasts [e.g., TGF-β1 and PDGF, vascular endothelial growth factor (VEGF), insulin-like growth factor 1 (IGF1), Galactin-3] (Vernon et al., 2010; Lech and Anders, 2013; Braga et al., 2015; Wynn and Vannella, 2016); and potentially (4) by directly transitioning into myofibroblasts to mediate interstitial fibrosis via a mechanism named “Mϕ-myofibroblast transition (MMT)” (Nikolic-Paterson et al., 2014; Meng et al., 2016; Wang et al., 2016; Wang Y.Y. et al., 2017; Liang et al., 2018; Tang et al., 2018).

To determine the role of global Mϕ, regardless of polarization state, in AKI, studies have been performed using depletion methods based on: (1) repeated administration of Mϕ-depleting agents (clodronate, small molecule CSF-1R inhibitors, neutralizing anti-CSF-1R antibodies, etc.); (2) genetic deletion of Mϕ survival factors (CSF-1, IL-34 or CSF-1R deficiency) (Lenda et al., 2003; Ma et al., 2009; Baek et al., 2015). In general, a (partial) global depletion of Mϕ was revealed to mitigate AKI in UUO and IRI experiments resulting in reduced tubular apoptosis (Lenda et al., 2003; Kitamoto et al., 2009; Ma et al., 2009; Baek et al., 2015) and interstitial fibrosis (Ma et al., 2009; Baek et al., 2015; Liu et al., 2018) (Table 1). In an experimental model of hypertension, the sustained depletion of global Mϕ was shown to attenuate hypertensive renal injury and fibrosis as well as to lower blood pressure (Huang et al., 2018). Much to our surprise, the reduced number of M2 Mϕ in Il34^–/– mice did not show a delay in the kidney recovery, but prevented kidney fibrosis, being clearly beneficial to the injured kidney (Baek et al., 2015). Remarkably, a UUO experiment showed that the global depletion of Mϕ reduces tubular apoptosis, but does not affect interstitial fibrosis (Ma et al., 2009), but this may be due to the specificity of UUO, where the renal insult is irreversible and the suppression of the injury driving the fibrotic response is more difficult than in other models (Nikolic-Paterson et al., 2014).

When AKI was induced by cell-specific depletion of proximal tubules in Ggt1-DTR mice, global depletion of Mϕ led to opposite results, aggravating AKI. In this specific AKI model, global depletion of Mϕ resulted in reduced survival of mice (Zhang et al., 2012), delayed functional and structural recovery from AKI and increase in interstitial fibrosis (Wang et al., 2015) (Table 1). However, it is important to mention that this AKI model does not involve a prominent Mϕ infiltration as seen in other IRI models (Figure 6 in Zhang et al., 2012), indicating that the initial injury is independent of M1 Mϕ and depletion of global Mϕ mainly targets the resident Mϕ-derived M2 pool.

Overall, these studies indicate that (partial) general depletion of Mϕ is rather beneficial than harmful to the injured kidney, especially in AKI settings where Mϕ infiltration and M1 Mϕ are prominent features (e.g., IRI and UUO models). The significance of Mϕ in tissue repair after AKI is unquestioned as Mϕ are known as extremely potent phagocytes supposed to accelerate the tissue recovery by clearing debris. However, studies showed that the kidney epithelium possess its own mechanisms to self-heal, e.g., by producing autocrine factors, which mediate tubular regeneration [CSF-1 (Menke et al., 2009), TGF-β1 (Gewin et al., 2012) etc.], and Mϕ may not be the only phagocytes in the injured tissue. So far, we do not know whether Mϕ are indispensable in tissue repair after AKI.

To examine the impact of M1 Mϕ on AKI and its outcomes, Mϕ were depleted by injecting clodronate into mice before the induction of AKI by either uni- or bilateral IRI (Day et al., 2005; Jo et al., 2006; Vinuesa et al., 2008; Lee et al., 2011; Ferenbach et al., 2012; Lu et al., 2012) or glycerol injection (Kim et al., 2014) (Table 2). All of these experiments demonstrated that reducing M1 Mϕ prevents immunopathology and improves kidney function in injured kidneys. In one of these studies, the depletion of M1 Mϕ paradoxically showed a reduced tubular regeneration at day 3 of bilateral IRI (Vinuesa et al., 2008); but, this may reflect that tubules were less damaged due to the depletion of M1 Mϕ. Independently, M1 Mϕ removal by immunotoxin (Fet et al., 2012) or neutralizing anti-CSF-1R antibody (Clements et al., 2016) prior to IRI uncovered similar findings including improved kidney function (Fet et al., 2012; Clements et al., 2016) and pathology and reduced oxidative stress (Fet et al., 2012) (Table 2). On the other hand, the reduction of M1 Mϕ by clodronate injection (Lu et al., 2008) and by DT injection in Cd11b-DTR mice (± clodronate) did not show any effect on cisplatin- and ischemia-induced AKI, respectively (Ferenbach et al., 2012; Lu et al., 2012) (Table 2). Interestingly, Mϕ depletion by clodronate injection improved AKI in the same IRI study (Ferenbach et al., 2012), indicating that DT-induced depletion of CD11b⁺ cells in Cd11b-DTR mice may have affected a larger variety of immune cells including immunosuppressive cell types. Overall, the impact of M1 Mϕ depletion on AKI and its outcomes can be considered as protective in AKI.

It may be easy to assume that the expansion of reparative M2 Mϕ in the resolution phase of inflammation would be beneficial to the injured tissue, but, in reality, M2 Mϕ can be both friends and foes in AKI (Braga et al., 2015). Indeed, studies focusing on evaluating the effect of M2 Mϕ depletion in AKI have led to controversial results. On the one hand, depletion of M2 Mϕ (e.g., by clodronate injection or DT-mediated conditional ablation of CD11b⁺ cells) was found to decrease kidney fibrosis (Lin et al., 2009; Kim et al., 2015; Yang et al., 2019), improve the kidney function and reduce the production of inflammatory and pro-fibrotic cytokines in some IRI and UUO experiments (Ko et al., 2008). In addition, M2 Mϕ depletion starting as early as at day 1 after UUO showed an improvement of the immunopathology limiting tissue injury (Yang et al., 2019) (Table 3). Notably, renal fibrosis was found to be reduced in all experiments where M2 Mϕ depletion improves the renal pathology after AKI, indicating that the most undesirable feature of M2 Mϕ in Mϕ-based therapeutic approaches for AKI and CKD is the capability to promote renal fibrosis. Some IRI and septic AKI experiments, on the other hand, led to completely opposite results suggesting that M2 Mϕ depletion is harmful to injured kidneys (Table 3). In these experiments, M2 Mϕ depletion worsened AKI (Li et al., 2018) or delayed the recovery from AKI (Jang et al., 2008; Lee et al., 2011); increased tubular damage and apoptosis (Jang et al., 2008; Menke et al., 2009; Clements et al., 2016) and oxidative stress (Jang et al., 2008); and impaired of kidney function (Menke et al., 2009; Lee et al., 2011; Karasawa et al., 2015; Li et al., 2018). In one study, M2 Mϕ depletion by DT-mediated ablation of CD11b⁺ cells even aggravated kidney fibrosis following IRI, which was contrary to the observations previously mentioned (Menke et al., 2009). In addition, DT-mediated depletion of CD169⁺ cells, which represent tissue-resident M2 Mϕ, markedly worsened the kidney injury and increased the lethality in mice after IRI, which, but, could be rescued by the adoptive transfer of Ly6C^– monocytes (Karasawa et al., 2015). Notwithstanding of all above, there was also a study showing that M2 Mϕ depletion (by conditional ablation in CD11b- or CD11c-DTR) does not have any impact on the development of fibrosis after unilateral IRI (Kim et al., 2015) (Table 3). In summary, the impact of M2 Mϕ depletion on AKI and its outcomes was not consistent throughout the experiments, illustrating that M2 Mϕ can be both beneficial and harmful to the injured kidney. These controversial results from AKI studies focusing on elucidating the role of M2 Mϕ corroborate the dual nature of M2 Mϕ. It is interesting to note that M2 Mϕ can to be rather disturbing than useful in the recovery process after AKI depending on the conditions given.