Toll-like receptor-4 coordinates the innate immune response of the kidney to renal ischemia/reperfusion injury

Abstract

Toll-like receptors (TLRs) can detect endogenous danger molecules released upon tissue injury resulting in the induction of a proinflammatory response. One of the TLR family members, TLR4, is constitutively expressed at RNA level on renal epithelium and this expression is enhanced upon renal ischemia/reperfusion (I/R) injury. The functional relevance of this organ-specific upregulation remains however unknown. We therefore investigated the specific role of TLR4 and the relative contribution of its two downstream signaling cascades, the MyD88-dependent and TRIF-dependent cascades in renal damage by using TLR4-/-, MyD88-/- and TRIF-mutant mice that were subjected to renal ischemia/reperfusion injury. Our results show that TLR4 initiates an exaggerated proinflammatory response upon I/R injury, as reflected by lower levels of chemokines and infiltrating granulocytes, less renal damage and a more preserved renal function in TLR4-/- mice as compared to wild type mice. In vitro studies demonstrate that renal tubular epithelial cells can coordinate an immune response to ischemic injury in a TLR4-dependent manner. In vivo we found that epithelial- and leukocyte-associated functional TLR4 contribute in a similar proportion to renal dysfunction and injury as assessed by bone marrow chimeric mice. Surprisingly, no significant differences were found in renal function and inflammation in MyD88-/- and TRIF-mutant mice compared with their wild types, suggesting that selective targeting of TLR4 directly may be more effective for the development of therapeutic tools to prevent I/R injury than targeting the intracellular pathways used by TLR4. In conclusion, we identified TLR4 as a cellular sentinel for acute renal damage that subsequently controls the induction of an innate immune response.

Figure 1. Renal function parameters of wild type and TLR4−/− mice.

Renal function of TLR4−/− mice (black bars) was improved compared with that of the wild type mice (white bars) one day after renal I/R as reflected by lower serum urea (A) and creatinine (B) levels. Data are mean±SEM of 8 mice per group (sham-operated animals: n = 3/group). * p<0.05.

Figure 2. Scoring renal tubular damage of wild type and TLR4−/− mice.

Score for histopathology after renal I/R injury (A) using PAS-D-stained renal tissue sections (B; representative for t = 1). Tubular damage was significantly lower in the outer medulla of kidneys of TLR4−/− mice (black bars) than in kidneys of wild type mice (white bars) one day after I/R injury. Data are mean±SEM of 8 mice per group (sham-operated animals: n = 3/group). * p<0.05.

Figure 3. Renal influx of granulocytes in wild type and TLR4−/− mice.

Influx of granulocytes in kidneys from wild type (white bars) and TLR4−/− (black bars) kidneys 1, 5 and 10 days after renal I/R injury or sham operation. One and ten days after I/R injury the number of granulocytes was significantly lower in kidneys of TLR4−/− mice than in kidneys of wild type mice as counted in 10 randomly selected high-power fields (HPFs) on outer medulla (magnification ×400). The amount of granulocytes from 8 mice per group were counted on renal tissue sections stained for Ly-6G and presented as mean±SEM. * p<0.05.

Figure 4. Levels of proinflammatory cytokines and chemokines in wild type and TLR4−/− mice.

Expression of proinflammatory chemokines KC and MCP-1 in kidney homogenates of wild type (white bars) and TLR4−/− (black bars) mice one day after I/R injury or sham-operation. KC levels were significantly lower in homogenate of TLR4−/− kidneys compared with wild type kidneys, whereas there were no differences observed in MCP-1 levels. Data are mean±SEM of 8 mice per group measured in duplicate (sham-operated animals: n = 3/group). * p<0.05.

Figure 5. In vitro production of proinflammatory cytokines and chemokines by wild type and TLR4−/− TECs.

Production of proinflammatory cytokines and chemokines by primary murine TECs from TLR4−/− (black) and wild type (white) mice, subjected to simulated ischemia. TLR4−/− TECs produce significantly lower amounts of KC as compared with TECs from wild type mice when subjected to simulated ischemia, whereas levels of MCP-1 were similar. Data are mean±SEM of 4–5 mice per group, measured in duplicate. * p<0.05. Ctr, control; Isch, Ischemia.

Figure 6. Equal contribution of epithelium-associated and leukocyte-associated TLR4 on renal function and injury one day after I/R injury.

Renal function and injury of wild type mice reconstituted with TLR4−/− bone marrow (WT+KO BM, white bars, n = 7) was comparable to TLR4−/− mice reconstituted with wild type bone marrow (KO+WT BM, black bars, n = 9) one day after renal I/R injury, as reflected by equal levels of ureum (A left) and creatinine (A right) and tubular injury (B). Data are mean±SEM. * p<0.05.

Figure 7. Apoptosis and proliferation of TECs of wild type and TLR4−/− mice.

Apoptotic (A) and proliferating (B) tubular epithelial cells in kidneys from wild type (white bars) and TLR4−/− (black bars) mice at one, five and ten days after I/R injury or sham-operation. The amount of apoptotic tubular cells was significantly higher in the kidneys of TLR4−/− mice than in wild type mice five days after I/R injury, whereas ten days after I/R injury the amount of proliferating tubular cells was significantly lower in the kidneys of TLR4−/− than in kidneys of wild type mice as counted in 10 randomly selected non-overlapping high-power fields on outer medulla (magnification ×400). Positive tubular epithelial cells from 8 mice per group were counted on renal tissue sections stained for active caspase-3 (apoptosis) and BrdU (proliferation). Data are presented as mean±SEM. * p<0.05.

Figure 8. Renal function, injury and inflammatory influx in wild type, MyD88−/− and TRIF-mutant mice.

Renal function of MyD88−/− and TRIF-mutant mice (black bars) did not differ compared with that of their wild type mice (white bars) one day after renal I/R as reflected by serum urea (A) and creatinine (B) levels and tubular injury (C). In addition, no differences were observed in the number of infiltrating granulocytes in kidneys of MyD88−/− and TRIF-mutant mice compared with their wild type mice (D). Data are mean±SEM of 6 (TRIF) or 8 (MyD88) mice per group (sham-operated animals: n = 2/group (TRIF) or n = 3/group (MyD88)). * p<0.05.