TLR4 activation mediates kidney ischemia/reperfusion injury (original) (raw)
TLR4 expression is increased in the kidney following kidney ischemia. To determine whether kidney ischemia stimulates TLR4 upregulation, we measured mRNA expression of TLR4 in IRI kidney by real-time PCR. Normal kidney tissue expressed TLR4 at a basal level. TLR4 mRNA levels were significantly increased at days 1 and 3 (P < 0.05) after ischemia, with further upregulation through days 5 and 9 (P < 0.01; Figure 1A). Isolated TECs submitted to ischemia expressed significantly higher mRNA levels of TLR4 than the controls (P < 0.05; Figure 1B).
TLR4 mRNA expression is increased following ischemia/reperfusion. (A) Ischemia-induced upregulation of TLR4 mRNA expression in the kidney (n = 6–8) from day 1 to day 9 after IRI compared with sham-operated controls. n = 6–8. (B) TECs submitted to ischemia expressed significantly higher levels of mRNA for TLR4 than controls. P < 0.05. mRNA expression was measured by real-time PCR. The results have been normalized by expressing the number of transcript copies as a ratio to GAPDH. Data are mean ± SD. Ctrl, control; IR, ischemia/reperfusion. *P < 0.05; **P < 0.01.
Immunostaining demonstrated that TLR4 protein was expressed by TECs and intrakidney leukocytes at day 1 after IRI and thereafter predominantly by TECs (Figure 2, A and C). Sham-operated controls had very faint staining in occasional tubules only. TLR4+ leukocytes on day 1 after IRI may have represented resident kidney DCs/macrophages, infiltrating cells, or a combination of these (Figure 2B). The number of TLR4+ leukocytes diminished through days 3–5 (Figure 2D). Expression of TLR4 protein by the tubules was most evident in the injured proximal tubules of the medullary ray and the outer medullary strip (Figure 2, A and C). Expression of TLR4 protein in tubules was significantly increased from day 1 to day 9 after IRI compared with that in sham-operated controls (Figure 2E).
TLR protein expression is increased in the kidney following ischemia/reperfusion. (A) Immunostaining demonstrated that TLR4 protein was expressed by infiltrating cells and TECs at day 1 and then predominantly expressed by TECs on days 3, 5, and 9 after IRI. Original magnification, ×200. (B) TLR4 was expressed by intrarenal leukocyte cells at day 1 after IRI. Magnification, ×600. (C) TLR4 was expressed by tubular cells at day 5 after IRI. Original magnification, ×600. (D) The number of cells expressing TLR4 dramatically increased on day 1, declining rapidly thereafter. These cells could be resident renal DCs/macrophages or infiltrating leukocytes. (E) Expression levels of TLR4 protein in tubules were significantly increased from day 1 to day 9 after IRI compared with sham-operated controls. **P < 0.01; ***P < 0.001. Data are mean ± SD. n = 6–8 per group.
Endogenous ligands for TLR4 are expressed in the kidney during IRI. To determine whether ischemia induced upregulation of endogenous TLR ligands in our model, we measured mRNA expression for HMGB1, biglycan, all 3 isoforms of hyaluronan synthase (HAS) reflecting hyaluronan biosynthesis, and HSP70 in IRI kidney by real-time PCR. Levels of mRNA for HMGB1 and biglycan were significantly increased at day 1 after IRI, with further upregulation through days 3 and 5 compared with sham-operated controls (Figure 3). Sham-operated kidneys expressed negligible levels of HAS, whereas HAS1, -2, and -3 mRNA expression increased dramatically in IRI kidney from day 1 to day 5 (26- to 136-fold increase for HAS1; 23- to 68-fold increase for HAS2; 3- to 8-fold increase for HAS3) (Figure 3). In contrast, mRNA expression for the inducible form of HSP70 was not increased in IRI kidney on days 1 through 5 compared with sham-operated kidneys (Figure 3).
mRNA expression of endogenous ligands for TLR4 expressed in the IRI kidney by real-time PCR. mRNA levels for HMGB1, biglycan, and HAS1, -2, and -3 were significantly increased from day 1 to day 5 after IRI, but HSP70 mRNA levels were not increased. Data shown are mean ± SD. n = 7–10 per group. *P < 0.05, **P < 0.01, ***P < 0.001.
Consistent with the real-time PCR data, staining with biotinylated hyaluronan-binding protein (b-HABP) showed that IRI induced a progressive increase in interstitial hyaluronan expression in the medullary ray and the outer medullary strip from day 1 to day 9 (Figure 4A). There was minimal expression in sham-operated kidney. Morphometric analysis of hyaluronan expression showed that both the area and intensity of hyaluronan staining increased from day 1 to day 9 after IRI compared with sham-operated controls (Figure 4B). Immunofluorescent staining demonstrated expression of HMGB1 by TECs on days 1 through 5 after IRI with negligible expression in sham-operated controls (Figure 4C). Western blot showed that protein expression of HSP70 was not significantly increased between IRI and sham-operated kidneys (Figure 4D).
IRI induces a marked and progressive increase in interstitial hyaluronan expression in WT mice. Representative sections of the kidney are stained for hyaluronan (HA) using biotinylated hyaluronan-binding protein (b-HABP) (original magnification, ×200) from day 1 to day 9 after IRI (A). (B) Analysis of HA expression showed that the area of HA-positive staining was significantly increased in the renal interstitium from day 1 to day 9 after IRI compared with sham-operated controls. The comparison between IRI kidney from day 1 to day 9 and sham-operated control is indicated by asterisks. **P < 0.01; ***P < 0.001. (C) Immunofluorescent staining indicates upregulation of HMGB1 expression by TECs at day 3 after IRI compared with sham-operated controls. Original magnification, ×400. (D) Western blot showed that protein expression of HSP70 was not significantly increased in IRI versus sham-operated kidneys.
TLR4–/– mice are protected against kidney IRI. To determine whether the full development of kidney IRI was dependent on TLR4 signaling, we used TLR4–/– mice in the ischemia/reperfusion model. TLR4–/– mice were protected against the effects of ischemia, exhibiting significantly lower serum creatinine and less tubular damage than WT controls. As shown in Figure 5, IRI caused kidney dysfunction in WT mice with a peak serum creatinine of 109 ± 39.1 μmol/l at day 1 after ischemia/reperfusion, which gradually fell thereafter but remained elevated (27.2 ± 4.9 μmol/l) at day 9 after IRI as compared with sham-operated mice (14.8 ± 3.8 μmol/l). In contrast, there was a very modest rise in serum creatinine in TLR4–/– mice (29 ± 6.1 μmol/l) on day 1, and creatinine remained lower than in WT controls at all time points thereafter.
TLR4–/– (black bars) and MyD88–/–(white bars) mice were protected against renal IRI with significantly lower serum creatinine compared with WT controls (gray bars) from day 1 to day 9 after reperfusion. Sham-operated mice had normal serum creatinine (10–20 μmol/l). Data are mean ± SD. n = 6–8 per group. The comparison between TLR4–/– mice or MyD88–/– mice and WT controls is indicated by asterisks. *P < 0.05; **P < 0.01; #P < 0.001.
The functional data correlated with histological kidney tubular damage. As shown in Figure 6A, in WT mice there was severe tubular damage, as evidenced by widespread tubular necrosis, loss of the brush border, cast formation, and tubular dilatation at the corticomedullary junction, maximal at day 1 with gradual recovery by day 9, whereas TLR4–/– mice showed significantly less tubular damage as compared with WT controls from day 1 to day 5 after IRI (Figure 6B). Sham-operated mice incurred no tubular injury.
Tubular injury in TLR4–/– and MyD88–/– kidney was significantly less than that seen in kidney from WT mice. (A) Representative sections of outer medulla from sham-operated, WT, TLR4–/–, and MyD88 –/– mice 1 day after reperfusion (H&E stained). Original magnification, ×200. (B) Semiquantitative analysis of tubular damage in WT (gray bars), TLR4–/– (black bars), and MyD88–/– (white bars) mouse kidney from day 1 to day 9 after reperfusion. Data shown are mean ± SD. n = 6–8 per group. The comparison between TLR4–/– mice or MyD88–/– mice and WT controls is indicated by asterisks. **P < 0.01, #P < 0.001.
The adaptor protein MyD88 is important in TLR4-mediated IRI. The major signaling pathway for TLRs proceeds via an adaptor protein, MyD88. To determine whether the MyD88 signaling pathway is involved, we also used MyD88–/– mice in this ischemia/reperfusion model. MyD88–/– mice were also protected against kidney IRI with significantly lower serum creatinine versus WT controls from day 1 to day 9 as shown in Figure 5 and also manifested less tubular damage than WT controls as shown in Figure 6.
Interstitial infiltrates are reduced in TLR4–/– and MyD88–/– mice. We further analyzed the cellular infiltrates in IRI kidney. Prominent interstitial neutrophil infiltration was observed in WT kidney at day 1 after ischemia/reperfusion; this largely subsided by day 9 (Figure 7). Neutrophil infiltration was significantly less in TLR4–/– and MyD88–/– mice than in WT controls from day 1 to day 5 (Figure 7B).
Neutrophil accumulation within the interstitium of the kidney was significantly less in TLR4–/– and MyD88–/– mice versus WT controls from day 1 to day 5 after reperfusion. (A) Representative sections of kidney stained for neutrophils by immunohistochemistry. Original magnification, ×200. (B) Analysis of neutrophil infiltrate in WT (gray bars), TLR4–/– (black bars), and MyD88–/– (white bars) mouse kidney (numbers/10 HPFs). Data shown are mean ± SD. n = 6–8 per group. The comparison between TLR4–/– mice or MyD88–/– mice and WT controls is indicated by asterisks. *P < 0.05; #P < 0.001.
Interstitial macrophages progressively accumulated in WT animals from day 1 to day 5 after IRI, then declined slightly by day 9 (Figure 8). Macrophage infiltration was most pronounced in the outer medulla at day 1 and day 3 and extended nearly to the cortex at day 5 and day 9. Compared with WT controls, TLR4–/– and MyD88–/– mice had significantly less interstitial macrophages at all time points (Figure 8B).
Macrophage accumulation within the interstitium of the kidney was significantly less in TLR4–/– and MyD88–/– mice versus WT controls at all time points (P < 0.05). (A) Representative sections of the kidney stained for macrophages by immunohistochemistry. Original magnification ×200. (B) Analysis of macrophage infiltrate in WT (gray bars), TLR4–/– (black bars), and MyD88–/– (white bars) mouse kidney (numbers /10 HPFs). Data shown are mean ± SD. n = 6–8 per group. The comparison between TLR4–/– mice or MyD88–/– mice and WT controls is indicated by asterisks. **P < 0.01; #P < 0.001.
TLR4 mediates proinflammatory cytokine and chemokine expression in the kidney during IRI. To further determine the effects of TLR4 signaling in the IRI kidney model, we examined the expression of known TLR4 downstream cytokines and chemokines. IL-6 and _TNF-_α mRNAs were strongly upregulated in WT kidney on days 1–5, peaking on day 3 (P < 0.001 compared with sham-operated controls). Cytokine expression also increased in TLR4–/– and MyD88–/– mice but to a much lesser extent (P < 0.05–0.001; Figure 9). _IL-1_β mRNA levels in the kidney were significantly increased in WT mice on day 1 following ischemia, and this increase was abrogated in TLR4–/– and MyD88–/– kidneys. (Figure 9).
Proinflammatory cytokine and chemokine mRNA profile in the kidney measured by real-time PCR. mRNA expression of proinflammatory cytokines (IL-6, _TNF-_α, and _IL-1_β) and chemokines (MIP-2 and MCP-1) in the kidney was significantly reduced in TLR4–/– (black bars) and MyD88–/– (white bars) mice compared with WT controls (gray bars) from day 1 to day 5 after reperfusion. Results have been normalized by expressing the number of transcript copies as a ratio to GAPDH. Data shown are mean ± SD. n = 6–8 per group. The comparison between TLR4–/– mice or MyD88–/– mice and WT controls is indicated by asterisks. *P < 0.05; **P < 0.01, #P < 0.001.
Chemokine (macrophage inflammatory protein–2 [_MIP-2_] and monocyte chemoattractant protein–1 [_MCP-1_]) mRNA levels in the kidney increased by several hundred–fold in WT mice after IRI compared with sham-operated controls (P < 0.001). Upregulation of chemokine expression was greatly attenuated in TLR4–/– and MyD88–/– kidneys compared with WT controls at all time points (Figure 9). mRNA levels for the IFN-β inducible chemokine gene IFN-inducible protein 10 (IP10) also increased in the kidneys of WT mice after IRI on days 1–5 (P < 0.05–0.001). However, unlike the other chemokines, no significant differences in _IP1_0 mRNA expression could be found between TLR4–/– and MyD88–/– kidneys and those from WT mice (Figure 9).
Protein levels of cytokines (IL-6, IL-1β, and TNF-α) and chemokines (MIP-2 and MCP-1) in kidney homogenates from TLR4–/–, MyD88–/–, and WT mice on day 1 to day 5 after IRI, measured by ELISA, generally reflected mRNA expression (Figure 10). IL-6, IL-1β, and MCP-1 protein levels in IRI kidney were significantly increased in WT mice compared with sham-operated controls (P < 0.001). Limited increases in these protein levels occurred in TLR4–/– and MyD88–/– mice, and protein levels were always significantly reduced compared with those in WT mice. For MIP-2, protein levels were equivalent in all 3 groups on day 1 but remained elevated in the WT mice while declining rapidly in TLR4–/– and MyD88–/– mice. Low levels of TNF-α protein expression were detected in some IRI kidneys in WT mice but not in TLR4–/– and MyD88–/– kidney homogenates.
Cytokine and chemokine protein expression in the kidney measured by ELISA. Protein expression of cytokines (IL-6 and IL-1β) and chemokines (MIP-2 and MCP-1) in the kidney was significantly reduced in TLR4–/– (black bars) and MyD88–/– (white bars) mice compared with WT controls (gray bars) from day 1 to day 5 after reperfusion. *P < 0.05, **P < 0.01, #P < 0.001.
Cytokine and chemokine mRNA expression is also decreased in TLR4–/– and MyD88–/– primary cultured TECs undergoing transient ischemia in vitro. Kidney TECs in mice are known to express TLR1, -2, -3, -4, and -6 (15). Primary mouse kidney TECs in culture secrete CC chemokines and proinflammatory cytokines in response to stimulation via TLR2 or -4 (32). To examine mRNA expression of cytokines and chemokines by kidney TECs undergoing transient ischemia in vitro_,_ primary cultures of TECs were established. The purity of these cultures was defined by positive staining for the epithelial marker cytokeratin and always exceeded 95% (range 96%–100%; Figure 11). As shown in Figure 11B, WT TECs submitted to ischemia expressed significantly higher mRNA levels of cytokines (IL-6, _IL-1_β, and _TNF-_α) and chemokines (MIP-2 and MCP-1) as compared with nonischemic controls (P < 0.05). TLR4–/– and MyD88–/– TECs submitted to ischemia in vitro showed reduced cytokine and chemokine gene expression versus WT controls (P < 0.001) and indeed showed expression level equal to nonischemic TEC controls.
Proinflammatory cytokine and chemokine gene expression in primary cultured TECs submitted to 1 hour ischemia in vitro. (A) Immunofluorescence staining of primary cultured TECs from C57BL/6 mice with mAbs against cytokeratin (green) and nuclear staining with DAPI in blue. Original magnification, ×400. (B) mRNA expression of proinflammatory cytokines (IL-6, IL-1β, and TNF-α) and chemokines (MIP-2 and MCP-1) in TLR4–/– and MyD88–/– primary cultured TECs submitted to 1 hour ischemia in vitro was significantly reduced at 1 hour after medium replacement as compared with WT controls. The results have been normalized by expressing the number of transcript copies as a ratio to GAPDH. These data are representative of 3 experiments. The comparison between TLR4–/– or MyD88–/– TECs and WT controls is indicated by asterisks. #P < 0.001.
Apoptosis is decreased in TLR4–/– and MyD88–/– TECs undergoing transient ischemia in vitro. To determine whether TLR4–/– and MyD88–/– TECs were protected against ischemia-mediated cell death, TECs in primary culture were subjected to transient ischemia, then stained with annexin V to identify apoptotic cells (Figure 12A). Flow cytometric analysis showed that the proportion of apoptotic cells increased from 6.7% in the nonischemic controls to 18.7% after ischemia in the WT TECs (P < 0.001) (Figure 12B). In contrast, apoptotic cells did not increase significantly in TLR4–/– and MyD88–/– TECs after ischemia. TUNEL assay confirmed these findings, with an increase in apoptotic cells seen in the WT though not TLR4–/– or MyD88–/– TECs after ischemia (Figure 12C).
Apoptosis in ischemic TECs was significantly reduced in TLR4–/– and MyD88–/– mice. (A) Flow cytometric analysis of apoptosis in TECs from WT, TLR4–/–, and MyD88–/– mice (dotted line indicates unstained cells). Numbers represent the percentage of apoptotic cells among propidium iodide–negative, viable cells. TECs subjected to ischemia are represented in the lower panels compared with nonischemic controls in the upper panels. (B) In WT mice, the proportion of apoptotic cells increased from 6.7% to 18.7% after ischemia. In contrast, the proportion of apoptotic cells did not increase significantly in TLR4–/– and MyD88–/– TECs. Data are representative of 2 separate experiments in triplicate and are shown as mean ± SD. **P < 0.01; *** P < 0.001. (C) Apoptosis was further confirmed using a TUNEL assay. TECs subjected to ischemia are shown in the lower panels, with nonischemic controls in the upper panels. Original magnification, ×200.
TLR4-mediated kidney IRI requires functional TLR4 signaling on kidney parenchymal cells. As TLR4 was expressed on leukocytes and intrinsic kidney cells, we next determined the relative importance of TLR4 signaling through kidney parenchymal cells or BM-derived cells in the pathogenesis of kidney IRI by generating BM chimeric mice. We produced mice with TLR4 present on leukocytes but absent from parenchymal cells (TLR4–/– host received WT BM: TLR4–/–/WTBM), and mice with TLR4 present on parenchymal cells but not leukocytes (WT host received TLR4–/– BM: WT/_TLR4–/–_BM). Two additional groups of mice were produced with TLR4 present on all cells (WT/WTBM) or with a complete absence of TLR4 (_TLR4–/–/TLR4–/–_BM). These sham chimeras were produced using methods identical to those used for the chimeric mice. Eight weeks after BM transplantation, chimeric mice were subjected to kidney ischemia. At that time, PCR analysis of genomic DNA extracted from leukocytes showed that in WT/_TLR4–/–_BM chimeras, 94% ± 5.1% of genomic DNA was derived from the TLR4–/– strain, and in TLR4–/–/WT chimeras, 91% ± 6.8% of genomic DNA was derived from WT animals. WT/WTBM mice showed significant kidney dysfunction and injury at day 1 after ischemia/reperfusion, while TLR4–/–/_TLR4–/–_BM chimeric mice were protected from kidney IRI as measured by day 1 serum creatinine and tubular damage (Figure 13). Creatinine levels and tubular injury scores recapitulated those observed in WT and TLR4–/– mice (Figures 5 and 6), excluding an effect of the BM transplant procedure per se on the response to kidney ischemia. Moreover, TLR4–/–/WTBM chimeras were protected from kidney IRI to the same degree as TLR4–/–/_TLR4–/–_BM mice, while WT/_TLR4–/–_BM mice enjoyed only partial protection as assessed by day 1 serum creatinine and tubular damage (Figure 13). These results suggest that functional TLR4 on kidney parenchymal cells makes the more significant contribution to kidney damage although leukocytes are clearly also important in IRI.
Functional TLR4 on intrinsic kidney cells makes the more significant contribution to kidney damage. WT/WTBM mice showed significant kidney dysfunction and injury at day 1 after ischemia/reperfusion, while TLR4–/–/_TLR4–/–_BM chimeric mice were protected from kidney IRI as measured by day 1 serum creatinine (A) and tubular damage (B). Creatinine levels and tubular injury scores replicated those observed in WT and TLR4–/– mice (Figures 5 and 6), excluding an effect of the BM transplant procedure per se on the response to renal ischemia. Moreover, TLR4–/–/WTBM chimeras were protected from kidney IRI to the same degree as TLR4–/–/_TLR4–/–_BM mice, while WT/_TLR4–/–_BM mice enjoyed only partial protection as assessed by day 1 serum creatinine and tubular damage (A and B). Data shown are mean ± SD. n = 7–10 per group. *P < 0.05, **P < 0.01, *** P < 0.001. Full replacement of hematopoietic cells in the chimeric mice was confirmed by genotyping of genomic DNA from whole blood using PCR. PCR products shown on representative gels (C). The top panel represents PCR products for the WT allele DNA, and the bottom panel represents PCR products for the mutated allele DNA (lanes 1–2: WT/WTBM; lanes 3–4: WT/_TLR4–/–_BM; lanes 5–6: TLR4–/–/WTBM; lanes 7–8: TLR4–/–/ _TLR4–/–_BM; lane 9: TLR4 heterozygous blood as positive controls; and lane 10: negative controls).