Toll-Like Receptor 4–Induced IL-22 Accelerates Kidney Regeneration

Abstract

AKI involves early Toll-like receptor (TLR)–driven immunopathology, and resolution of inflammation is needed for rapid regeneration of injured tubule cells. Notably, activation of TLRs also has been implicated in epithelial repair. We hypothesized that TLR signaling drives tubule regeneration after acute injury through the induction of certain ILs. Systematic screening in vitro identified IL-22 as a candidate proregeneratory factor in primary tubular cell recovery, and IL-22 deficiency or IL-22 blockade impaired post-ischemic tubular recovery after AKI in mice. Interstitial mononuclear cells, such as dendritic cells and macrophages, were the predominant source of IL-22 secretion, whereas IL-22 receptor was expressed by tubular epithelial cells exclusively. Depleting IL-22–producing cells during the healing phase impaired epithelial recovery, which could be rescued entirely by reconstituting mice with IL-22. In vitro, necrotic tubular cells and oxidative stress induced IL-22 secretion selectively through TLR4. Although TLR4 blockade during the early injury phase prevented tubular necrosis and AKI, TLR4 blockade during the healing phase suppressed IL-22 production and impaired kidney regeneration. Taken together, these results suggest that necrotic cell–derived TLR4 agonists activate intrarenal mononuclear cells to secrete IL-22, which accelerates tubular regeneration and recovery in AKI.

AKI involves a sterile inflammatory response that contributes to the extent of tubular cell damage.¹ In turn, tubular cell necrosis is the predominant trigger for this associated inflammatory response, because dying cells release intracellular molecules, such as HMGB1, histones, uric acid, or ATP, that elicit immunostimulatory effects in the extracellular space, which are referred to as damage-associated molecular patterns (DAMPs).^2–5 Tubular damage also induces Tamm–Horsfall protein (THP)/uromodulin leakage from the tubular lumen into the interstitium, where it turns into an immunostimulatory danger signal.^6,7 DAMPs activate a set of pattern recognition receptors, such as Toll-like receptors (TLRs) or inflammasomes, on renal parenchymal cells as well as in interstitial dendritic cells.^2–9 The subsequent innate immune response involves the transcription of numerous proinflammatory cytokines and chemokines, which initiate the influx of various immune cell subsets into the kidney, that contribute to the early amplification of the inflammatory response and AKI by enhancing immune-mediated tubular cell death.^10,11 To limit overshooting immunopathology in sterile tissue injuries and allow tissue recovery,¹² a number of counterregulatory mechanisms exists that mostly limits immune activation of intrarenal dendritic cells.^13,14 For example, pentraxin-3 release from the microvasculature and dendritic cells limits leukocyte recruitment.^15,16 Most importantly, switching the phenotype of intrarenal mononuclear phagocytes away from classically activated (proinflammatory) to alternatively activated (anti-inflammatory/proregeneratory) cells is necessary for recovery on AKI.^17–19

Although surviving tubular epithelial cells (TECs) enter the cell cycle within few hours on injury,¹ a functional tubular recovery does not occur before the resolution of sterile inflammation has occurred and the tubulointerstitial microenvironments become dominated by proregeneratory factors.^18,19 They are provided in a paracrine manner by other surviving TECs, intratubular tubular progenitor cells, or bone marrow-derived stem cells.^1,20–23 Although dendritic cells and other immune cells play a dominant role in orchestrating the early injury phase of AKI, little is known about the contribution of intrarenal immune cells to the subsequent phase of kidney regeneration.²⁴

Three reports recently consistently showed that mononuclear phagocyte depletion in the healing phase of post-ischemic AKI impairs tubular regeneration,^18,19,25 but the phagocyte-derived mediators that drive tubular healing remain unknown. Zhang et al.¹⁹ showed that colony-stimulating factor (CSF)-1 activates M2 macrophage–dependent tubular repair, but CSF-1 is predominately produced by activated tubular cells.²⁶ M2 macrophages produce IL-10 and growth factors like TGF-β that have anti-inflammatory and profibrotic effects, but more specific myeloid cell-derived mediators that enhance epithelial healing should exist.

Generally, we follow the concept that renal mononuclear phagocytes orchestrate all phases of AKI—the onset and the resolution of inflammation as well as subsequent tissue regeneration or repair.²⁷ Here, we focused on the ILs, a paradigmatic family of leukocyte-derived mediators that regulates homeostasis and immunity in a paracrine manner. We hoped to identify yet unknown proregeneratory properties of ILs on TEC regeneration in AKI. Our unbiased in vitro screening approach revealed IL-22 as a candidate, which we subsequently validated as a renal dendritic cell–derived stimulator of tubular regeneration. To our surprise, we found IL-22 secretion to be selectively induced by TLR4 agonists released from necrotic tubular cells, which first documents a role of TLR signaling for not only renal immunopathology but also, kidney regeneration in vivo. Furthermore, our data imply that dying TECs involve interstitial dendritic cells to support their regeneration through a specific TLR4–IL-22 pathway.

Results

IL-22 Enhances Tubular Cell Re-Epithelialization by Jak-Stat3 and Erk1/2 In Vitro

First, we set up an experimental system that mimics primary tubular cell recovery on ischemia-reperfusion injury. We used kidney explanation from adult mice and the preparation of renal tubular cell suspensions for culture as a model of ischemia-reperfusion injury, in which, similar to severe AKI, less than 10% of the TECs survived.^19,28 We considered the regenerative outgrowth from the surviving TECs reforming a monolayer as a valuable in vitro system to screen for the impact of ILs on post-ischemic epithelial healing. Among all ILs tested, recombinant IL-22 had the strongest proregeneratory effect (Figure 1A). To better mimic epithelial monolayer injury and validate this candidate, we also tested IL-22 on mechanical scratching of primary TEC monolayers. Recombinant IL-22 dose dependently enhanced wound closure (Figure 1B), confirming that IL-22 enhances re-epithelialization on TEC injury. Which signaling pathways contribute to this process? Recombinant IL-22 induced phosphorylation of signal transducer and activator of transcription 3 (STAT3) and extracellular signal-regulated kinases 1/2 (ERK1/2) in primary TECs, a process that was blocked by an mitogen-activated protein kinase (MEK1/2) inhibitor (Figure 1C). Furthermore, Jak1 and MEK1/2 inhibition partially abrogated the proregeneratory effect of recombinant IL-22 on mechanical scratch–induced epithelial injury of primary TEC monolayers (Figure 1D). The inhibitors did not affect the wound healing process on their own. Thus, IL-22 enhances re-epithelialization on TEC injury through activation of the Jak/STAT3P and ERK1/2 pathways.

Figure 1.

IL-22 promotes epithelial healing in vitro. (A) Primary isolated tubular epithelial cells were studied for regeneration in the presence of various ILs. IL-22 significantly induced the regenerative outgrowth of isolated primary TECs. (B) Proregenerative activity of IL-22 was further confirmed using in vitro scratch assays on monolayers of primary TECs with and without recombinant IL-22. IL-22 (0.1, 1, and 10 ng/ml) significantly enhanced wound closure in a dose-dependent manner. (C) IL-22 (10 ng/ml) activates phosphorylation of STAT3 and ERK1/2 on primary TECs. (D) Pharmacological inhibition of STAT3 phosphorylation using Jak inhibitor I (10 nM) and ERK1/2 phosphorylation using PD98059 (10 µM) abolished the wound re-epithelialization capacity of IL-22 (1 ng/ml). Data are means±SEMs from three separate experiments. *P<0.05 versus control; **P<0.01 versus control.

Neutralizing Endogenous IL-22 Impairs Epithelial Recovery on AKI

To confirm the role of IL-22 on epithelial repair in vivo, we first characterized the expression of IL-22 in plasma (ELISA) and kidney (RT-PCR) 1, 5, and 10 days after unilateral renal pedicle clamping. Plasma IL-22 protein levels were significantly elevated at days 5 and 10 (Supplemental Figure 1A), whereas intrarenal IL-22 mRNA levels increased as early as day 1 (Figure 2A). IL-22 immunostaining revealed positivity exclusively in the interstitial compartment on day 1 as well as day 5 in post-ischemic but not contralateral control kidneys after unilateral renal pedicle clamping (Figure 2B). By contrast, IL-22R immunostaining displayed positivity exclusively in TECs (Figure 2C). We also found extrarenal expression of IL-22 in post-ischemic injury (e.g., in splenic dendritic cells) (Figure 2D).

Figure 2.

Renal ischemia clamping induces intrarenal IL-22 expression. (A) IL-22 expression was analyzed by RT-PCR. IL-22 is expressed significantly in ischemic kidney on all the observed time points on days 1, 5, and 10. (B) Interstitial expression of IL-22 was observed in ischemic kidneys on days 1 and 5, whereas contralateral kidney sections do not show IL-22 expression on these time points. (C) IL-22R expression was observed in immunohistochemical analysis on tubular epithelial cells, and (D) IL-22 was expressed at day 5 in spleen cells, which were expressed by CD11c⁺ cells identified by flow cytometry. Data are means±SEMs from three separate experiments. *P<0.05 versus control; **P<0.01 versus ischemic control (IR).

To investigate the functional significance of endogenous IL-22 for AKI recovery, we neutralized IL-22 in the healing phase of AKI by injecting anti–IL-22 antibody only from day 2 after renal pedicle clamping and euthanized the mice on day 5. This time point was selected, because in our model, the injury phase lasts for 48 hours; the recovery phase covers days 2–5 at the given ischemia time.²⁹ IL-22 blockade significantly impaired tubular recovery as defined by the index of injured tubules assessed on periodic acid–Schiff sections, the numbers of Lotus tetragonolobus lectin⁺ proximal tubules and THP⁺ distal tubules, and the number of proliferating THP⁺ proliferating cell nuclear antigen⁺ cells (Figure 3). IL-22 neutralization did not affect IL-22 expression as assessed by IL-22 staining (Figure 3A). This result was consistent with the profound increase in the renal mRNA expression levels of the tubular injury markers kidney injury molecule (Kim)-1, π−glutathione S-transferase (GST), α-GST, and fatty acid–binding protein (FABP) versus isotype controls (Figure 3B). To understand the role of IL-22 in regaining renal function after ischemic tubular injury, we carried out bilateral ischemic injury in Il-22–deficient mice. These mice displayed impaired tubular regeneration accompanied by higher plasma BUN levels on day 5 (Figure 3C). Of note, plasma BUN levels were identical at day 1, implying that IL-22 does not contribute to the injury phase of AKI. Together, the post-ischemic kidney induces IL-22 expression, which supports tubular recovery during the healing phase of AKI.

Figure 3.

IL-22 blockade attenuates tubular repair on post-ischemic kidney injury. (A) Tubular injury was quantified on a periodic acid–Schiff (PAS)-stained renal section at day 5 after unilateral renal artery clamping as described in Concise Methods. Isotype represents isotype IgG, and αIL-22 represents IL-22 IgG. L. tetragonolobus lectin staining identified proximal tubuli, and THP staining identified distal tubuli in post-ischemic kidneys. IL-22 expression was analyzed using staining of IL-22. Renal injury was quantified by analyzing PAS stain, and the quantitative assessment of tubuli with intact staining patterns is shown for each staining. Data are means±SEMs from six mice in each group. Original magnification, ×100. (B) Effect of IL-22 inhibition on renal injury and tubular repair was further determined by expression of various tubular injury markers (Kim-1, α-GST, π-GST, and L-FABP). (C) Il-22–deficient mice impaired tubular regeneration in bilateral ischemic injury, which was evident by elevated plasma BUN on day 5, but IL-22 does not contribute to the injury phase; plasma BUN levels were comparable on day 1. (D) Pharmacological inhibition of IL-22 reduced proliferating THP⁺ proliferating cell nuclear antigen⁺ (PCNA⁺) tubular cells. Data are means±SEMs from six mice in each group. *P<0.05 versus isotype IgG-treated mice.

Interstitial Mononuclear Phagocytes Are the Major Source of Renal IL-22 Expression

We used flow cytometry of renal cell suspensions prepared 5 days after kidney ischemia to better define the IL-22–producing cells. IL-22 was selectively produced by CD45⁺ leukocytes and not at all produced by CD45⁻ renal nonimmune cells (data not shown). Additional surface markers specified the IL-22–producing renal leukocyte subsets (Table 1). We characterized CD45⁺IL-22⁺ cells in various subsets of mononuclear phagocytes based on the surface expression of CD11b, CD11c, F4/80, and CD103. All CD11b⁺, CD11b⁺CD103⁺, CD103⁺, CD11c⁺, and F4/80⁺ mononuclear phagocytes expressed IL-22 in the post-ischemic kidney (Figure 4A, Supplemental Figure 3). Clodronate liposome did not reduce a minor population of natural killer (NK⁺) and CD3 T cells producing IL-22 (Supplemental Figure 2). Intravenous clodronate liposome injection on days 2 and 4 after renal pedicle clamping significantly depleted these IL-22–producing cells inside the post-ischemic kidney, except for CD11c⁺F4/80⁺ cells, but only during the recovery phase of AKI (Table 1) In contrast, clodronate injection did not affect the numbers of neutrophils, CD3⁺ T cells, and NK1.1⁺ cells inside the post-ischemic kidney (Supplemental Figure 2). Immunostaining confirmed the depletion of IL-22–producing leukocytes in the post-ischemic kidney at day 5 (Figure 4B). Clodronate-induced depletion of mononuclear phagocytes was recently reported to impair tubular repair in this model.^18,19,25 Our own experiments reproduced the phenotype of impaired AKI recovery reported by these studies, because clodronate depletion of renal mononuclear phagocytes reduced the proliferating cell nuclear antigen⁺ THP-1⁺ proliferating distal TECs and the amount of THP⁺ tubules (Figure 5A), showing insufficient tubular cell proliferation as a cause for poor AKI recovery, which was evidenced by a higher tubular necrosis index and elevated tubule injury marker (Kim-1, π‐GST, α‐GST, and FABP) mRNA expression at day 5 (Figure 5B). However, does this effect specifically relate to a lack of IL-22 release by the depleted mononuclear phagocytes? To address this question, we injected recombinant IL-22 from day 3 into clodronate liposome-treated mice. IL-22 reconstitution recovered tubular cell repair, which was indicated by normalization of all the aforementioned AKI parameters (Figure 5, A and B). We examined the effect of renal phagocytes depletion on renal function in post-ischemic injury. Renal phagocyte depletion impaired tubular recovery, which was indicated by elevated plasma BUN on day 5. Reconstitution of IL-22 normalized plasma BUN (Figure 5C). These findings show that renal mononuclear phagocytes contribute to AKI recovery during the healing phase by secreting IL-22.

Table 1.

Clodronate induced depletion of IL-22–producing cells in the post-ischemic kidney

Renal Mononuclear Phagocytes	Control	Clodronate
CD11b+IL-22+	0.13±0.06	0.03±0.01a
CD11b+CD103+IL-22+	0.08±0.04	0.02±0.01a
CD11b−CD103+IL-22+	0.03±0.02	0.01±0.004a
CD11c+IL-22+	0.10±0.04	0.01±0.01a
CD11c+F4/80+IL-22+	0.03±0.01	0.03±0.01
CD11c−F4/80+IL-22+	0.09±0.04	0.01±0.01a

Cell number is in cells per kidney (×10⁶). Data are means±SEMs from five mice in each group. ^aP<0.05 versus Lipo-Control.

Figure 4.

Clodronate liposomes deplete IL-22–producing myeloid cells from the post-ischemic kidney. (A) Ischemic renal injury induced intrarenal expression of IL-22, and renal mononuclear phagocytes (CD11b⁺, CD103⁺, CD11c⁺, and F4/80⁺) are the major source of IL-22 in ischemic renal injury. (B) Effect of clodronate-liposome depletion on IL-22 expression in ischemic kidney was analyzed using immunohistochemistry. Clodronate-liposome significantly reduced intrarenal expression of IL-22. Data are means±SEMs from six mice in each group. *P<0.05 versus PBS-liposome (Lipo-Control)–treated mice.

Figure 5.

Depletion of IL-22–producing cells attenuates tubular recovery. (A) Depletion of IL-22–producing phagocytes significantly increased renal injury, reduced THP⁺ and Lectin⁺ tubules, and reduced proliferating renal tubular cells identified as proliferating cell nuclear antigen⁺ (PCNA⁺) THP⁺ cells (green, PCNA⁺ cells; red, THP⁺ cells). Renal injury was quantified by analyzing PAS stain, and the quantitative assessment of tubuli with intact staining patterns is shown for each staining. Control represents IR control, Clodronate represents Clodronate-liposome, and Clodronate⁺IL-22 represents Clodronate-liposome⁺recombinant IL-22 on day 3. Original magnification, ×100. (B) Effect depletion of IL-22–producing cells on renal injury and tubular repair was further determined by mRNA expression of various tubular injury markers (Kim-1, α-GST, π-GST, and L-FABP). (C) Depletion of renal phagocytes using clodronate significantly impaired recovery of renal function in bilateral ischemic injury (25 minutes). Reconstitution of IL-22 on day 3 normalized plasma BUN levels. Data are means±SEMs from six mice in each group. ⁺P<0.05 versus control; *P<0.05 versus clodronate-treated mice.

Necrotic Tubular Cells Trigger IL-22 Release by Releasing Agonists for TLR4

What are the triggers for intrarenal mononuclear phagocytes to secrete IL-22 during AKI? AKI involves tubular necrosis that activates pattern recognition receptors on immune cells by releasing endogenous danger signals, which are referred to as DAMPs.⁹ Therefore, we first used supernatants of necrotic TECs to stimulate bone marrow–derived dendritic cells (BMDCs) and found them to induce IL-22 secretion even stronger than the known IL-22 inducers aryl hydrocarbon receptor agonist 6-formylindolo[3,2-b]carbazole (FICZ) and hydroxyl peroxide (Figure 6A). To study the involved pattern recognition receptors, we used synthetic receptor agonists (e.g., Pam3Cys [TLR2], LPS [TLR4], poly-IC RNA [TLR3], imiquimod [TLR7], and CpG-DNA [TLR9]). All of these TLR agonists induced IL-6 release from BMDCs, whereas only TLR4 activation induced IL-22 release (Figure 6B). To further confirm the role of TLR4, we blocked the necrotic supernatant stimulation with a neutralizing TLR4 antibody, which entirely suppressed IL-22 release in a dose-dependent manner (Figure 6C). Necrotic supernatant and TLR4 stimulation of primary renal dendritic cells also induced IL-22 expression (Figure 6D). In contrast, direct TLR4 stimulation of TEC monocultures did not affect their viability or growth (Supplemental Figure 4, B and C). These data document that necrotic TECs trigger IL-22 release in dendritic cells by TLR4 activation.

Figure 6.

Necrotic cell supernatant, oxidative stress, and AhR regulate IL-22 secretion. (A) Stimulation of BMDCs with necrotic supernatant in different concentrations induced IL-22 production. AhR agonist (FICZ) and H₂O₂ also produced IL-22 secretion by BMDCs. (B) BMDCs were stimulated for 16 hours with LPS (1 µg/ml), pI:C RNA (5 µg/ml), Pam3Cys (1 µg/ml), Imiquimod (1 µg/ml), and CpG (3 µg/ml). Cell culture supernatants were analyzed for IL-6 and IL-22. (C) Primary BMDCs cultured necrotic supernatant from TECs. IL-22 release was quantified in cell culture supernatant on 18 hours of stimulation in the presence of TLR4 blocking antibody at different concentrations. (D) Necrotic cell supernatant and TLR4 stimulation induced IL-22 expression in primary renal dendritic cells. Data are means±SEMs from three independent experiments. *P<0.05 versus NS control; *P<0.01 versus NS control. Med, medium; ND, not detected.

TLR4 Activation–Mediated IL-22 Release Drives Tubular Repair on Post-ischemic AKI

TLR4 activation triggers innate immunity, immunopathology, and the extent of AKI during the early injury phase of AKI.³⁰ To test a putative role of TLR4 in AKI recovery experimentally requires either a conditional knockout approach or a specific TLR4 antagonist. Because conditional TLR4^−/− mice are not available, we used a TLR4 blocking antibody. However, when we injected anti-TLR4 IgG only from 48 hours after renal pedicle clamping, it partially reduced intrarenal expression of IL-22, which was evidenced by decreased expression of IL-22 in the injured kidney (Figure 7A). In this way, TLR4 inhibition impaired tubular recovery on post-ischemic AKI, which was illustrated by an increased index of tubular necrosis, lower numbers of lectin⁺ or THP⁺ tubules, and the tubular injury marker π‐GST (Figure 7B). TLR4 blockade with CLI-095 initiated in the recovery phase was also associated with higher BUN levels on bilateral ischemic injury (P=0.06), which indicates that TLR4 blockade during recovery phase slows down the process of the tubular repair, whereas TLR4 blockade during the injury phase prevented AKI (Figure 7C). Together, releases from necrotic TECs activate TLR4 on intrarenal mononuclear phagocytes to induce IL-22, a mechanism that enhances tubular recovery during the healing phase of AKI.

Figure 7.

TLR4-mediated release of IL-22 contributes to tubular repair. TLR4 blocking antibody was injected on days 2, 3, and 4 after induction of ischemic and renal injuries, and tubular injury was analyzed by immunohistochemistry and RT-PCR. (A) Tubular injury was quantified on PAS-stained renal sections at day 5 after unilateral renal artery clamping as described in Concise Methods. Isotype represents isotype IgG, and αTLR4Ab represents TLR4 blocking IgG. L. tetragonolobus lectin staining identified proximal tubuli, and THP staining identified distal tubuli in post-ischemic kidneys. IL-22 expression was analyzed using staining of IL-22. Renal injury was quantified by analyzing PAS stain, and the quantitative assessment of tubuli with intact staining patterns is shown for each staining. Original magnification, ×100. (B) Effect of TLR4 blockade on renal injury and tubular repair was further determined by expression of various markers of tubular injury (Kim-1, α-GST, π-GST, and L-FABP). (C) We analyzed the effects of TLR4 inhibitor (CLI-095; 1 mg/kg) treatment on renal function in bilateral ischemic injury. Early blockade (pre) of TLR4 protected from renal damage, which was evident by plasma BUN on day 1, whereas late blockade (post) of TLR4 impaired recovery of renal function. Data are means±SEMs from six mice in each group. *P<0.05 versus isotype IgG-treated mice.

Discussion

We had hypothesized that some ILs may hold the potential to drive tubular regeneration during the recovery phase of AKI. Our unbiased in vitro screening approach revealed IL-22 as a candidate. IL-22 not only fostered the regenerative outgrowth from those few TECs that had survived the ischemia-reperfusion injury of primary TEC isolation but also, accelerated wound closure on scratching TEC monolayers in culture. We used post-ischemic AKI in mice to validate these findings and found that IL-22 accelerates tubular injury. We identified intrarenal mononuclear phagocytes (i.e., the various subsets of renal dendritic cells and macrophages) as the predominant source of IL-22 expression. Furthermore, we found TLR4 signaling as a specific stimulus for IL-22 induction. These data further specify how renal mononuclear phagocytes also orchestrate kidney regeneration beyond their well known function in renal immunopathology. These data surprisingly imply that TLR4 signaling in myeloid cells drives danger response programs and tissue inflammation as well as tissue regeneration.^31,32

IL-22 is a member of the IL-10 family of cytokines with regulatory roles in tissue repair.³³ IL-22 regulates the proliferation of keratinocytes in psoriasis,^34,35 lung epithelial cells in lung injury,³⁶ enterocytes on toxic mucosal injury,^37,38 hepatocytes in liver injury,^39,40 and thymic epithelial cells on total body irradiation.⁴¹ The data from epithelial organs, like the gut and the skin, are consistent with our findings, documenting the conceptual similarities between re-epithelialization of TEC monolayer injury inside renal tubules and re-epithelialization of epithelial surfaces in other body compartments. IL-22–mediated re-epithelialization involves binding to its receptor IL-22R on TECs followed by activation of STAT3 and ERK1/2 phosphorylation.⁴² Although T and NKT cells were first thought to represent the predominant IL-22–producing cells,^43,44 IL-22 secretion by dendritic cells has previously been recognized.^38,45 The dense network of interstitial dendritic cells in the healthy kidney seems to be the first source of intrarenal IL-22 secretion.⁴⁶

Three studies as well as our own results showed an unexpected contribution of renal mononuclear phagocytes to AKI recovery by depleting them with clodronate selectively during the recovery phase. This process depleted the IL-22–producing cells, although it is theoretically possible that clodronate indirectly reduced IL-22 production.^18,19,25 IL-22 reconstitution recovered this phenotype, which identifies IL-22 as a paracrine mediator of this effect. It remains unclear if IL-22 reconstitution also had a supraphysiological therapeutic effect. It is likely that IL-22 is just one of several phagocyte-derived factors that regulate the healing process. However, these data add on to the existing evidence that the various phenotypes of intrarenal dendritic cells and macrophages orchestrate all phases of tissue injury and recovery to allow tissues to regain and maintain homeostasis.²⁴

We also identified TLR4 signaling as a stimulus for IL-22 secretion in dendritic cells, similar to the endogenous DAMP concept of cell necrosis-induced activation of sterile inflammation.⁴⁷ This finding does not contradict the proinflammatory role of TLR4 in the injury phase of AKI. In fact, rapid repair of epithelial barriers is another important innate response mechanism of pathogen control and host defense (e.g., inside the intestinal epithelium or the skin), which implies that inflammation and rapid re-epithelialization are both needed and potentially triggered by the same signaling cascades during host defense. The potential of the TLR/MyD88 signaling pathway to accelerate epithelial healing was first discovered by Rakoff-Nahoum et al.,⁴⁸ who found that dextran-induced intestinal epithelial injury causes sepsis and death. Also, signals from the intestinal microbiota activate TLRs to drive rapid intestinal epithelial repair. In addition, activation of keratinocyte TLR2 by macrophage-activating lipopeptide-2 or topical application of the TLR3 agonist poly-I:C RNA accelerates dermal wound healing.^49,50 Similar mechanisms seem to apply to tubular recovery during AKI, where necrotic TECs release DAMPs with TLR4 agonistic activity that induce IL-22 secretion by renal dendritic cells. This mechanism would add on to the potential of endogenous TLR2 agonists to activate tubular progenitor cells to release microvesicles with proregeneratory proteins and microRNAs and drive their differentiation into TECs.^51,52 Finally, IL-22 increases the survival of such epithelial progenitor cells during injury.⁵³

Altogether, IL-22 is a previously unknown mediator of TEC regeneration during the AKI recovery phase. TEC necrosis provides a stimulus to TLR4-mediated induction of IL-22 secretion by intrarenal dendritic cells and other mononuclear phagocytes. The mitogenic effects of IL-22 on the surviving TECs are mediated by IL-22R and subsequent STAT3 and ERK1/2 phosphorylation. We conclude that TLR4 signaling drives epithelial regeneration on post-ischemic AKI through secretion of IL-22 from mononuclear phagocytes, which links the two danger response programs of renal inflammation and regeneration at the level of TLR4.

Concise Methods

Isolation of Tubular Epithelial Cells

We performed complete perfusion with sterile Dulbecco's phosphate-buffered saline (DPBS) on 6- to 7-week-old wild-type C57BL6 mice by puncture of the left ventricle. Subsequently, both kidneys were weighed, mashed, digested in collagenase (Collagenase A; Roche Diagnostics) for 30 minutes, and then sieved through a 70-µm sieve. After one step of centrifugation (1500 rpm for 5 minutes), the tubular fragments were resuspended in 2 ml DPBS, carefully layered on a 31% Percoll column, and centrifuged (3000 rpm for 10 minutes). The pellet (tubular fragments) was washed two times in DPBS and then cultured in hormone-conditioned media at 37°C and 5% CO₂ as described.²⁸

Assessment of Regeneration

We performed a large-scale screening experiment with over 20 cytokines (Immunotools). Right after isolation, an equal amount (according to the kidney weight) of tubular fragments was plated into 24-well plates. On the same 24-well plates, four squares (4×4 mm) with a gap between them were marked on the bottom of the plate before the cell culture. Photos of these squares were taken with a phase contrast microscope (ProgRes software; Leica) on day 5 after ischemia. By digital analysis (Photoshop; Adobe), the area surface covered by tubular cells was measured in percent to the total image size.

Scratch Assay

For scratch assays, artificial wounds were created through standardized scratching of the tubular cell monolayer with a pipette tip. IL-22 was added to the cells at different concentrations (0.1, 1, and 10 ng/ml). Two pictures per wound were taken on a phase contrast microscope at different time points (0, 16, 20, and 24 hours). Changes in wound size were analyzed digitally (Photoshop) and then translated into wound closure. To understand the functional role of various pathways, scratch assay was performed in the presence of JAK-STAT inhibitor (10 nM Jak inhibitor I; Santa Cruz Biotechnology) and MEK1/2 inhibitor (10 µM PD98059; Cell Signaling Technology), and cells were stimulated with 1 ng/ml IL-22.

Protein Isolation and Western Blotting

We extracted protein from cell lysate using RIPA buffer (Sigma-Aldrich) containing protease inhibitors (Roche Diagnostics) and phosphatase inhibitor (Sigma-Aldrich) and processed it for Western blotting as described.^29,54 Briefly, proteins were separated by SDS-PAGE and then transferred to a polyvinylidene difluoride membrane. Nonspecific binding to the membrane was blocked for 2 hours at room temperature with 5% BSA in Tris-buffered saline buffer. The membranes were incubated overnight at 4°C with primary rabbit antibodies against mouse p-stat3, p-ERK 42/44, and β-actin (Cell Signaling Technology). After washing, the membranes were incubated with peroxidase-conjugated secondary antibodies in Tris-buffered saline buffer. Secondary antibodies were peroxidase-conjugated anti-rabbit IgG (Cell Signaling Technology). The signals were visualized by an enhanced chemiluminescence system (Amersham).

BMDC Stimulation Experiments

BMDCs were generated by established protocols. Briefly, bone marrow cells were harvested from mouse femor and cultured with granulocyte macrophage CSF (2 ng/ml) for 8 days. Cell media were replaced on day 8, and adherent dendritic cells were stimulated. Renal dendritic cells were isolated from whole kidneys by magnetic bead separation (MACS) using CD11c MicroBeads (Miltenyi Biotech). The purity of CD11c⁺ renal cells was confirmed by FACS analysis using anti-CD11c antibody (BD Biosciences–Pharmingen). Necrotic cell supernatants were prepared from mouse tubular cells by repeated freezing and thawing. BMDCs were stimulated with necrotic supernatant with different concentration (50, 150, and 250 µl) as well as AhR agonist (FICZ; Enzo Life Sciences) and H₂O₂ (Merck). We purchased ultrapure LPS (from Escherichia coli strain K12), pI:C RNA, Pam3Cys, imiquimod, and CpG (InvivoGen). All cells were stimulated in serum-free RPMI 1640 medium (Invitrogen) at a density of 1×10⁶ cells/ml. Cells were stimulated for 16 hours with LPS (1 µg/ml), pI:C RNA (5 µg/ml), Pam3Cys (1 µg/ml), imiquimod (1 µg/ml), and CpG (3 µg/ml). Cell culture supernatants were analyzed for IL-6 (BD Biosciences–Pharmingen) and IL-22 (eBiosciences) cytokine secretion by ELISA according to the manufacturer’s instructions. In some experiments, BMDCs were preincubated with TLR4 antibody (MTS510; BioLegend) at different concentration 30 minutes before necrotic supernatant stimulation.

Animal Experiments

C57BL/6 wild-type mice were obtained from Charles River (Sulzfeld, Germany). Il-22–deficient mice were provided by S. Kobold. Mice were housed in groups of five in filter-top cages with unlimited access to food and water. Cages, food, and water were sterilized by autoclaving before use. Unilateral (45 minutes) and bilateral (25 minutes) renal ischemia-reperfusion injuries were induced under general anesthesia as previously described.²⁹ Mice were euthanized 1 and 5 days later, and pieces from ischemic and contralateral (sham) kidneys were snap-frozen in liquid nitrogen and fixed in 10% buffered formalin. In some experiments, mice received intraperitoneal injections with either 20 µg/injection of the anti–IL-22 antibody or isotype control (eBiosciences) on days 2, 3, and 4 postinjury. Clodronate-liposome and PBS-liposome (200 µl per mouse per injection), both of which were procured from Haarlem (The Netherlands), were injected on days 1, 2, and 4 postinduction of injury. Recombinant IL-22 (20 µg/mouse; Biolegend) was injected in clodronate-liposome–injected mice on day 3 post-ischemic injury. Anti-TLR4 antibody (15 µg per mouse per injection; Biolegend) and isotype control were injected as specified. TLR4 inhibitor (CLI-095; Invivogen) was injected at the dose of 1 mg/kg on day 0 (pre) as well as days 2, 3, and 4 post-ischemic injury. Plasma BUN was determined by taking blood samples on days 0, 1, and 5. IL-22 knockout mice were also examined to understand the role of IL-22 on renal function. All experiments were conducted according to German animal protection laws (equivalent to the National Institutes of Health Guide for the Care and Use of Laboratory Animals) and had been approved by the local government authorities.

Assessment of Kidney Inflammation and Injury

Kidneys were embedded in paraffin, and 2-μm sections were used for periodic acid–Schiff stains and immunostaining as described.⁵⁵ Post-ischemic tubular injury was scored by assessing the percentage of tubules in the corticomedullary junction that displayed tubular cell flattening, cell necrosis, loss of brush border, and luminal cast formation as 0, none; 1, ≤10%; 2, 11%–25%; 3, 26%–45%; 4, 46%–75%; 5, >76%. For histochemistry, we used biotinylated L. tetragonolobus lectin stain (Vector Labs), THP stain (Santa Cruz Biotechnology), proliferating cell nuclear antigen (Cell Signaling Technology), IL-22R (BIOSS), and IL-22 antibody (Santa Cruz Biotechnology). To quantify the IL-22 staining, the slides were scanned using an Olympus BX 61 microscope and recorded with CellP software. The scans underwent digital stain counting using Adobe Photoshop.

Real-Time Quantitative RT-PCR

Total RNA was isolated from kidneys using a Qiagen RNA extraction kit following the manufacturer’s instructions. After quantification, RNA quality was assessed using agarose gels. From isolated RNA, cDNA was prepared using reverse transcription (Superscript II; Invitrogen). Real-time quantitative RT-PCR was performed using SYBRGreen PCR master mix and analyzed with a Light Cycler 480 (Roche Diagnostics). All gene expression values were normalized using 18s RNA as a housekeeping gene, and negative controls were used for quality control as described.⁵⁶ All primers used for amplification were from Metabion and are listed in Table 2.

Table 2.

List of primers used for RT-PCR

	Right Primer (5′ → 3′)	Left Primer (5′ → 3′)
18S	AGGGCCTCACTAAACCATCC	GCAATTATTCCCCATGAACG
π-GST	ACACCGCCCTCGAACTGGGAA	CGCAGCACTGAATCCGCACC
α-GST	CTTCAAACTCCACCCCTGCTGC	CAATGGCCGGGAAGCCCGTG
IL-22	GCTCAGCTCCTGTCACATCA	TCGCCTTGATCTCTCCACTC
L-FABP	AGGCAATAGGTCTGCCCGAGGAC	CCAGTTCGCACTCCTCCCCCA
AhR	CTCCTTCTTGCAAATCCTGC	GGCCAAGAGCTTCTTTGATG
KIM-1	TGGTTGCCTTCCGTGTCTCT	TCAGCTCGGGAATGCACAA

Flow Cytometry

Anti-mouse CD11c, CD11b, CD103, F4/80, and CD45 antibodies (BD Biosciences–Pharmingen) were used to detect renal mononuclear phagocyte populations in the ischemic kidneys as described⁵⁷; 7/4 (Abd Serotec), Ly6G (BD), CD3 (BD), and NK1.1 (BD) antibodies were used to detect neutrophil, T cell, and NK cell in the ischemic kidneys. IL-22 expression was determined using intracellular staining with IL-22 antibodies (R&D Systems). Respective isotype antibodies were used to show specific staining of cell populations. Quantification of cell number was achieved using counting beads (Invitrogen).

Statistical Analyses

Data are presented as mean±SEM. Comparison of groups was performed using ANOVA or t test; post hoc Bonferroni correction was used for multiple comparisons. A P value <0.05 were considered significant.

Disclosures

None.