点击显示 收起
【摘要】 Acute renal failure (ARF) induces tubular hyperresponsiveness to TLR4 ligands, culminating in exaggerated renal cytokine/chemokine production. However, the fate of TLR4 protein during acute tubular injury remains unknown. The study sought new insights into this issue. Male CD-1 mice were subjected to 1 ) unilateral ischemia-reperfusion (I/R), 2 ) cisplatin (CP) nephrotoxicity, or 3 ) glycerol-induced myohemoglobinuric ARF. Renal cortical TLR4 protein (Western blotting, immunohistochemistry) and TLR4 mRNA levels (RT-PCR) were determined thereafter (90 min-4 days). Urinary TLR4 excretion post-I/R or CP injection was also assessed. To gain proximal tubule-specific results, TLR4 protein and mRNA were quantified in posthypoxic or oxidant (Fe)-challenged isolated mouse tubules. Finally, TLR4 mRNA was determined in antimycin A-injured cultured proximal tubular (HK-2) cells. Acute in vivo renal injury reduced proximal tubule TLR4 content. These changes corresponded with the appearance of TLR4 fragment(s) in urine and a persistent increase in renal cortical TLR4 mRNA. Isolated proximal tubules responded to injury with rapid TLR4 reductions, dramatic extracellular TLR4 release, and increases in TLR4 mRNA. Glycine blocked these processes, implying membrane pore formation was involved. HK-2 cell injury increased TLR4 mRNA, but not protein levels, suggesting intact transcriptional, but not translational, pathways. Diverse forms of acute tubular injury rapidly reduce proximal tubular TLR4 content. Plasma membrane TLR4 release through glycine-suppressible pores, possibly coupled with a translation block, appears to be involved. Rapid postinjury urinary TLR4 excretion suggests its potential utility as a "biomarker" of impending ARF.
【关键词】 acute renal failure iron oxidant stress cisplatin biomarkers
IT IS WELL ACCEPTED that renal inflammation significantly contributes to the pathogenesis of nephrotoxic and postischemic acute renal failure (ARF) (e.g., 3, 7, 12, 14, 15, 22). The innate immune response/Toll-like receptor (TLR) pathway may be intimately involved in this process. TLRs are a family of 11 plasma membrane glycoproteins that bind a variety of pathogen (e.g., endotoxins TLR4; lipoteichoic acids TLR2)- and nonpathogen (e.g., damaged RNA/DNA; heat shock proteins)-associated molecules ( 1, 2, 10, 11, 13, 17, 18, 21 ). Once ligation occurs, downstream adaptor proteins (MyD88; Trif; Traf, NF- B) are recruited, culminating in increased cytokine and chemokine production ( 4, 8, 9 ). This process is hyperactive in the acutely damaged kidney ( 25 - 27 ). For example, when mice with diverse forms of ARF [ischemia-reperfusion (I/R), cisplatin (CP), myoglobinuria, urinary tract obstruction] are challenged with specific TLR ligands (endotoxin; lipoteichoic acid), greatly exaggerated intrarenal cytokine (TNF-; IL-10), chemokine (MCP-1), and nitric oxide production result ( 25 - 27 ). This process may increase circulating cytokine and chemokine levels, potentially impacting extrarenal organs, and hence, ARF-associated multiorgan failure.
Despite evidence that the ARF kidney hyperresponds to TLR ligands ( 25 - 27 ), the direct effect of tubular cell injury on TLR protein expression remains a subject of debate. For example, it has been suggested that TLR4, the most thoroughly studied member of the TLR family, is upregulated in the kidney 3-5 days postischemia ( 10, 21 ) and after 28 days of cyclosporine treatment ( 13 ). However, recent observations from our laboratory indicate that increased TLR4 expression is not necessarily a generalized renal injury response ( 27 ). For example, TLR4 abundance, as gauged by Western blotting, was suppressed by 50% in both oxidant-challenged cultured proximal tubular (HK-2) cells and in the glycerol model of myoglobinuric ARF ( 27 ). Furthermore, early (18 h) CP nephrotoxicity did not change renal cortical TLR4 levels (Western blots) despite the presence of the TLR4 ligand hyperresponsive state ( 27 ).
Given these uncertainties as to the fate of TLR4 following acute renal injury, the present study was undertaken to gain further insights into this issue. The results that will be presented indicate that TLR4 is rapidly shed from proximal tubules (PT) as part of the acute injury response, resulting in relative TLR4 depletion. Indeed, this shedding process raises the possibility that urinary TLR4 appearance could have potential utility as a "biomarker" of ischemic and toxic forms of ARF.
METHODS
Animal Models of Renal Injury
I/R. A previously reported model of unilateral renal I/R injury was employed ( 25 ). In brief, male CD-1 mice (25-35 g; Charles River Laboratories, Wilmington, MA) were anesthetized with pentobarbital sodium ( 30-40 mg/kg ip), the left renal pedicle was identified through a midline abdominal incision, and then occluded with a smooth vascular clamp ( x 17.5 min, unless stated otherwise). Body temperature was maintained at 37°C with heating lamps. After clamp removal and confirmation of reperfusion, the abdominal incision was sutured and the mice were allowed to recover from anesthesia. Either 18 h or 4 days later, the mice were reanesthetized, and the postischemic left kidney and contralateral right kidney were resected. The kidney pairs were iced and subjected to one of the following procedures: 1 ) renal cortical total RNA extraction for subsequent TLR4 mRNA analysis by RT-PCR (Ref. 20 ); n = 5 determinations at each time point, 2 ) renal cortical protein extraction for subsequent TLR4 Western blot analysis (Ref. 27 ); n = 5 at each time point, or 3 ) immersion in 10% buffered formalin for TLR4 immunohistochemical analysis.
To ascertain whether the contralateral (nonpostischemic) kidney in the above experiment had altered TLR4 expression, four mice were subjected to sham surgery and their kidneys were removed 18 h or 4 days later. The kidneys were used for TLR4 protein and mRNA assessments (left and right kidneys, respectively), and compared with values observed in kidneys resected from four normal mice. Additionally, kidneys from two sham surgery mice and two control mice were subjected to TLR4 immunohistochemical analysis.
Glycerol-Induced Myohemoglobinuric ARF
It has previously been demonstrated that by 18 h post-IM glycerol injection (50%; 9 ml/kg), severe ARF develops [blood urea nitrogen (BUN) 129 ± 11 mg/dl] and that this is associated with a 50% decrease in renal cortical TLR4 levels, as assessed by Western blot ( 27 ). This was despite a corresponding 2 x increase in renal cortical TLR4 mRNA ( 27 ). To further assess TLR4 expression following induction of glycerol ARF, four mice were subjected to 9 ml/kg glycerol injection ( 27 ). Eighteen hours later, the mice were anesthetized with pentobarbital sodium, a plasma sample was obtained from the inferior vena cava for BUN determination, and then one kidney per animal was removed for TLR4 immunohistochemical analysis. Four normal kidneys provided control histology samples.
CP-Induced ARF
18-h Assessments. We previously reported that by 18 h post-CP administration, TLR4 mRNA levels are increased but that there are no associated changes in renal cortical TLR4 protein levels (Western blotting) ( 27 ). To further define TLR4 expression, four mice were injected with CP (30 mg/kg ip) ( 26, 27 ); 18 h later, they were anesthetized, a blood sample was obtained for BUN analysis, and one kidney per mouse was processed for TLR4 expression by immunohistochemistry. The results were contrasted to those observed in four sham-treated controls.
72-h Assessments. To further define the impact of CP on TLR4 expression, assessments at a delayed time point were used. Four mice were injected with CP, as above. Three days later, they were anesthetized, a blood sample was removed for BUN analysis, and one kidney per mouse was removed for TLR4 immunohistochemistry. The other kidney was iced, the cortex dissected, and divided into two parts. The two parts were used for either TLR4 protein or mRNA analysis. Renal samples obtained from four normal mice were used as controls.
Urinary TLR4 Assessments
Acute I/R. The following experiment was undertaken to ascertain whether TLR4 might be acutely shed from tubules, followed by urinary TLR4 excretion. To this end, seven mice were deeply anesthetized with pentobarbital sodium and the kidneys and urinary bladder were exposed through a midline abdominal incision. The urinary bladder was emptied by direct compression to obtain a "baseline" urine sample. Bilateral renal ischemia was then induced x 10 min via application of vascular clamps. Following clamp removal, each mouse received 10 µg of furosemide IM to stimulate a diuresis. The abdominal incisions were closed with surgical clamps and the mice underwent a 4-h period of vascular reperfusion while anesthesia and body temperature at 37°C were maintained. At the completion of the reperfusion period, a second urine sample was collected followed by animal death by aortic transection. Five additional mice served as surgical controls, being subjected to the above procedures, but without renal vascular occlusion. Urine creatinine concentrations were determined and the concentrations were equalized by water addition. The samples were then probed for TLR4 by Western blotting ( 24 ). To exclude potential nonspecific secondary antibody binding, the urine samples were also probed with the secondary antibody (horseradish peroxidase-labeled goat anti-rabbit IgG from donkey; Amersham Biosciences; 1:12,000) without primary antibody addition (anti-mouse TLR4; ImGenix, San Diego, CA; cat. no. 578A). The postischemic urine results were compared both with baseline samples and with results obtained from sham surgical controls. [Notes: only a 10-min period of renal ischemia was used in these experiments to prevent severe ARF, thereby precluding early urine collection and hence, urine TLR4 assessment; bilateral clamping was used, since unilateral clamping would favor collection of urine from the noninjured kidney.]
CP nephrotoxicity. Four mice were injected with CP (30 mg/kg ip). Urine samples were then collected 18 and 48 h later by administering furosemide at each time point, as noted above. The results were contrasted to those observed in five urine samples obtained from controls.
Renal Cortical TLR4 Immunohistochemical Analyses
Kidneys were fixed in 10% neutral buffered formalin and paraffin processed. Four-micrometer sections were cut and deparaffinized. The slides were rehydrated in Dako Wash Buffer (Carpinteria, CA) and all staining steps were performed at room temperature using the Dako Autostainer. Endogenous peroxide activity was blocked using 3% H 2 O 2 for 8 min followed by the Avidin/Biotin Blocking Kit (Vector Laboratories, Burlington, CA) x 10 min for each solution. The sections were then incubated with Tris-buffered saline containing 1% BSA, 15% donkey serum, and 5% mouse serum (Jackson ImmunoResearch, West Grove, PA) for 10 min. TLR-4 was detected using a goat polyclonal antibody that was raised against a synthetic COOH-terminus peptide and that was then purified against that peptide by immunoaffinity chromatography (SC-12511, Santa Cruz Biotechnologies, Santa Cruz, CA; 10 µg protein per ml; stock solution of antibody, 200 µg/ml). After a 30-min incubation, the slides were washed with buffer, and then they were incubated with biotinylated donkey anti-goat (Jackson ImmunoResearch) x 30 min and Vectastain Elite R.T.U ABC (Vector Laboratories). The staining was visualized with 3,3'-diaminobenzidine (DAB; Dako) x 7 min, and the sections were counterstained with hematoxylin (Dako) x 2 min. Concentration-matched isotype control slides were run for each tissue sample (Jackson ImmunoResearch Laboratories) using the same IgG concentration that was present in the anti-TLR4 antibody (10 µg/ml).
Isolated PT Experiments
The following experiments were undertaken to further test whether acute tubular injury causes acute TLR4 shedding. Toward this end, TLR4 expression in isolated PT, and in their incubation medium, was assessed in the following acute cell injury experiments.
Hypoxia-reoxygenation/TLR4 Western blot assessments. PTs were harvested from normal CD mice by a process of cortical dissection, collagenase digestion, and differential centrifugation as previously described ( 24 ). Each preparation ( n = 6) was suspended in incubation buffer ( 24 ), equally divided into four aliquots, and placed in 10-ml Erlenmeyer flasks. After rewarming from isolation temperature (4°C) to experimentation temperature (37°C), each preparation were divided into two treatment groups (in duplicates): flasks 1 and 2 ): control incubation (95% O 2 -5% CO 2 ) x 30 min; flasks 3 and 4 ): 15-min hypoxia (95% N 2 -5% CO 2 ) followed by 15 min of reoxygenation (95% O 2 -5% CO 2 ). At the end of the incubation, lethal cell injury was determined by percent lactate dehydrogenase (LDH) release. The tubules were then centrifuged, and the pellets underwent protein extraction in the presence of protease inhibitors ( 25 ). The supernatants were concentrated threefold using syringes attached to 10,000-kDa filters. The tubule supernatants and pellet extracts were then subjected to TLR4 Western blot analysis ( 24 ). Half of the samples were simultaneously probed for TLR4 and -tubulin, the latter serving as an internal reference control ( 24 ). Detection and quantitation were performed by ECL followed by densitometry ( 24 ).
Impact of glycine-mediated cytoprotection on TLR4 release. Glycine addition to isolated tubules mitigates hypoxia-induced lethal cell injury without altering the extent of ATP depletion ( 5, 6, 19, 20 ). The following experiment assessed whether glycine-mediated protection alters hypoxia-induced TLR4 release. Four sets of isolated tubules were each divided into four experimental groups: 1 ) control incubation x 30 min; 2 ) incubation with 2 mM glycine; 3 ) 15-min hypoxia/15-min reoxygenation; and 4 ) hypoxia/reoxygenation in the presence of glycine. Percent LDH release and pellet/supernatant TLR4 levels were assessed, as above.
Fe-induced oxidative stress/TLR4 assessments. To further gauge the impact of cell injury on tubule TLR4 expression, four sets of isolated tubules were prepared and subjected to one of the following: 1 ) control x 30 min; 2 ) control incubation x 45 min; 3 ) incubation x 30 min with 25 µM ferrous ammonium sulfate complexed to the siderophore hydroxyquinoline (FeHQ) ( 16, 23 ); or 4 ) incubation with FeHQ x 45 min ( n = 4 per group). At the end of the incubations, percent LDH release and tubule/medium TLR4 levels were assessed.
Acute TLR4 mRNA Responses to Injury
Isolated tubules. Four sets of isolated tubules were prepared and subjected to the above-noted hypoxia/reoxygenation ± glycine addition experiments with the exception that a 7.5-min period of hypoxia, followed by 35 min of reoxygenation, was used (to induce less severe injury plus more time for mRNA changes postinjury). After completion of this protocol, the tubule pellets underwent RNA extraction using a commercially available kit (Invitrogen; cat. no. 12183-018) to which was added an RNase free DNase (Promega; cat. no. M6101; 1 U/100 µg RNA). The RNA samples were then subjected to TLR4 mRNA RT-PCR analysis ( 27 ). [Note: the absence of genomic DNA was confirmed by the absence of product on ethidium bromide-stained gels when the RNA samples were subjected to PCR without prior RNA reverse transcription.]
In vivo renal cortex. The following assessed whether in vivo renal ischemia also induces acute TLR4 mRNA increases. Four mice were subjected to 17.5 min of unilateral renal ischemia and 90 min of reperfusion. The postischemic and contralateral kidneys were then removed and the cortices were analyzed for TLR4 mRNA, as noted above.
TLR4 mRNA expression in HK-2 cells. To further gauge TLR4 mRNA responses to injury, eight near-confluent T75 flasks of HK-2 cells, maintained in serum-free medium, were subjected to either control incubation ( n = 4) or exposure x 18 h to 7.5 µM antimycin A ( n = 4) ( 27 ). This protocol has previously been reported by this laboratory to induce minimal lethal cell injury (due to ongoing glycolytic support of ATP) and no change in TLR4 protein levels ( 27 ). After the antimycin A challenge was completed, total RNA was extracted and analyzed for TLR4 mRNA by RT-PCR.
Calculations and Statistics
All values are presented as means ± 1 SE. Statistical comparisons were performed by paired (left vs. right kidney results; aliquots from isolated tubules) or unpaired Student?s t -tests. If more than one comparison was made, the Bonferroni correction was applied. All TLR4 mRNA results were expressed as ratios of TLR4/GAPDH products ( 27 ).
RESULTS
TLR4 mRNA Assessments Following In Vivo Renal Injury
I/R. As shown in Fig. 1, left, I/R significantly increased renal cortical TLR4 mRNA, compared with values seen in either contralateral (CL) control kidneys or sham (S)-operated controls. The degree of elevation was greater at day 4 vs. the 18-h time point (2 x vs. 4 x increases, respectively; P < 0.01). As expected, the unilateral I/R protocol did not induce significant renal insufficiency (BUN: 31 ± 3 mg/dl vs. controls: 28 ± 2 mg/dl).
Fig. 1. Toll-like receptor (TLR)4 mRNA assessments at 18 h and 4 days postunilateral ischemia-reperfusion (I/R) and at 3 days following cisplatin (CP) injection. The unilateral postischemic kidney results were contrasted with those from: 1 ) contralateral (CL) control nonischemic kidneys and 2 ) sham (S)-operated controls. At both the 18-h and 4-day time points, significant increases in TLR4 mRNA (factored by GAPDH product) were observed. The values were approximately twice as high at day 4 vs. the 18-h time point. Right : 4-fold increase in TLR4 mRNA was also seen at 3 days post-CP injection. NS, not significant; n = 5 per group.
CP toxicity. At 3 days post-CP injection, a fourfold increase in renal cortical TLR4 mRNA was observed ( Fig. 1, right ). This 3-day CP protocol also induced severe renal failure (BUN: 112 ± 21 mg/dl vs. controls: 28 ± 2 mg/dl; P < 0.01).
Western Blot Analyses
Figure 2 depicts the results of renal cortical Western blotting. TLR4 levels appeared to be slightly decreased by the I/R protocol, whether the assessments were made at the 18-h or 4-day time points. However, these results were not statistically different. Equal protein loading was confirmed by equivalent -tubulin (internal standard) ( 27 ) detection and by equivalent total protein staining with India ink ( 27 ). In contrast to I/R, by 72-h post-CP injection, an approximate 50% decrease in TLR4 levels was observed ( Fig. 2, right ). [Representative Western blot images will be presented below.]
Fig. 2. Renal cortical TLR4 Western blot analyses: 18-h and 4-day postischemic renal injury and 3-day post-CP injection. TLR4 was relatively well preserved in 18-h and 4-day postischemic renal cortical samples, compared with control (C) kidney samples (from sham-operated mice or CL controls; n = 5 per group). Conversely, an 50% reduction in TLR4 was observed at 3 days post-CP injection ( n = 4 per group). (Representative blots are presented in Figs. 5, A and B, and 6 A.)
TLR4 Immunohistochemical Analyses
Normal renal TLR4 expression. As shown in Fig. 3, A and B, prominent TLR4 staining was observed in PT, both within the cortex and outer medullary stripe. The brush border appeared to have the greatest TLR4 content ( Fig. 3 C ). Conversely, there was minimal, or no, staining of glomeruli, with the exception of the visceral and parietal aspects of Bowman?s capsule ( Fig. 3 C ). Cortical distal tubules stained less intensely than their PT counterparts. With the exception of outer medullary stripe PT, far less prominent TLR4 expression was observed within the medulla, compared with cortex ( Fig. 3 B ). When normal kidney histological sections were probed with control (non TLR4 immune) goat IgG, followed by secondary anti-goat IgG antibody, no staining was observed ( Fig. 3 D ), indicating that the anti-TLR4 staining was not due to nonspecific secondary antibody adherence.
Fig. 3. TLR4 expression in normal kidneys. A : low power ( x 10 objective) of normal renal cortex demonstrating intense proximal tubular staining for TLR4. B : junction of outer medullary stripe and inner medulla of normal kidney ( x 10 objective) demonstrating less intense TLR4 staining of the latter. C : normal cortex ( x 40 objective), illustrating proximal tubular TLR4 staining, with particular involvement of the brush border. The glomerulus is largely negative, except for linear staining of the visceral aspect of Bowman?s capsule. D : negative control of renal cortex outer medulla (nonimmune IgG followed by secondary antibody treatment).
CP toxicity. The kidneys examined 18 h post-CP administration showed no evidence of tubular necrosis (consistent with a lack of severe renal injury/azotemia at this time point), and no renal cortical TLR4 losses were observed ( Fig. 4 A ). However, mild reductions in PT TLR4 expression within the outer medullary stripe were seen ( Fig. 4 A ), consistent with the fact that this nephron segment is most prominently impacted during CP nephrotoxicity. This reduction existed in the absence of overt tubular necrosis. By 3 days post-CP administration, PT had sustained widespread reductions in TLR content, both within the cortex and the outer medullary stripe ( Fig. 4 B ). Additionally, cortical and outer medullary stripe proximal tubular necrosis was also observed. TLR4-positive material was seen in tubular lumina, consistent with TLR4 sloughing. Residual nonnecrotic tubular cells also showed reduced TLR4 staining.
Fig. 4. Renal TLR4 expression in the setting of acute renal failure (ARF). A : 18-h post-CP treatment ( x 10 objective). Relatively normal renal cortical TLR4 appearance is depicted, with a slight decrease in its expression in the outer medullary stripe. B : 72-h CP treatment ( x 20 objective) demonstrating TLR4 loss from selected tubules and with its appearance in tubular lumina. C : 18-h postrenal ischemia ( x 10) demonstrating focal TLR4 loss from proximal tubules in outer medullary stripe with relative preservation within renal cortex. D : 18-h postischemia ( x 40) demonstrating differing degrees of reduction in proximal tubular TLR4 content. E : 18-h postglycerol-induced ARF ( x 10) demonstrating widespread reductions in TLR4 staining. F : high-power ( x 40) view of renal cortex at 18-h postglycerol treatment. Loss of TLR4 from proximal tubules and sloughing into lumina are illustrated. Within proximal tubular cross sections, there are variable degrees of cellular TLR4 loss.
I/R. By 18 h postischemia, reductions in outer medullary stripe PT TLR4 staining were apparent. Conversely, most cortical PT maintained seemingly normal TLR4 content ( Fig. 4 C ). Sloughed tubular debris stained positively for TLR4. Focal segments of PT that remained intact within both the outer medullary stripe and cortical medullary rays also manifested reduced TLR4 content ( Fig. 4 D ). These changes were seen at both 18 h and 4 days postishcemia.
Glycerol-induced ARF. The 18-h postglycerol ARF tissue samples also demonstrated dramatic PT TLR4 reductions. Unlike the 18-h postischemic results, the TLR4 reductions were apparent in both the cortex and the outer medullary stripe ( Fig. 4 E ). Less severely injured (nonnecrotic) tubular cells also demonstrated reduced TLR4 staining (e.g., Fig. 4 F ). As with the other injury models, TLR4-positive material was prominently expressed in lumina, in the form of casts ( Fig. 4 F ).
Urine TLR4 Western Blots
Probing of the normal urine samples, obtained either with or without furosemide administration, revealed no immunostaining for TLR4. In contrast, each of the 4-h postischemic urine samples showed a clear immunoreactive TLR4 band with an approximate 30-kDa mass, consistent with a TLR4 cleavage product ( Fig. 5 ). When the urine samples were probed with the secondary antibody in the absence of the primary TLR4 antibody, completely negative blots were obtained. Urine samples obtained 18 h and 48 h post-CP injection also revealed TLR4 staining ( Fig. 6 ). The 18-h samples revealed approximate 60- and 30-kDa bands (presumably cleavage products of the native 90-kDa TLR4 moiety). By 48 h post-CP, only the 30-kDa band was seen (comparable to the postischemic urine samples). [Representative renal cortical TLR4 Western blots from the I/R and CP experiments are presented in Figs. 5 and 6.]
Fig. 5. Western blot of renal cortex and of urine samples collected after renal ischemia-reperfusion (I/R). A : representative renal cortical Western blots after in vivo ischemia/18 h of reperfusion (I/R) and in control kidneys. No significant change in TLR4 was apparent, as assessed by 2 closely related 90-kDa bands (that reflect different degrees of TLR4 glycosylation). B : representative renal cortical Western blots after in vivo ischemia/4-day reperfusion (I/R) and in control kidneys. No significant changes were apparent. C : urine samples obtained pre- and postrenal ischemia + 4 h of reperfusion (I/R). The control urine samples (C) had no discernable TLR4-immunoreactive material. Conversely, each of the postischemic urine samples contained a clear immunoreactive band, seen at 30 kDa (reflecting a presumptive TLR4 cleavage product).
Fig. 6. Western blot of renal cortex and of urine samples collected after CP administration. A : renal cortical TLR4 Western blots 3 days after CP administration. Marked reductions in TLR4 levels were observed. B : urine samples that were collected 18 h post-CP injection demonstrated 2 TLR4-immunoreactive bands ( 60 and 30 kDa; presumptive cleavage products, the sum of which equates with native 90-kDa protein). At 48 h, the 60-kDa band was no longer evident, but there was persistence of the 30-kDa band (the latter of which recapitulated the postischemic urine findings that are demonstrated in Fig. 5 C ). Conversely, control urine (C) samples revealed no detectable TLR4/fragments.
Isolated Tubule/Western Blot Experiments
Hypoxia-reoxygenation. Hypoxia-reoxygenation (H/R) caused marked cell injury, as assessed by percent LDH release ( 48% vs. 11% in control tubules; P < 0.005; Fig. 7 ). When hypoxia was imposed with glycine present, a marked reduction of lethal cell injury resulted ( Fig. 7 ).
Fig. 7. Percent LDH release and tubule TLR4 loss following hypoxia/reoxygenation (H/R) and Fe-mediated injury, as assessed in isolated proximal tubule segments. Left : % LDH release:Fe ( n = 4) and H/R ( n = 6 sets of tubules) each caused lethal cell injury ( 34 and 48% LDH release, respectively, vs. 11% for control, C, oxygenated tubules). Hypoxic injury was significantly attenuated by glycine (G; n = 4 per group) treatment (which did independently alter tubular cell viability in oxygenated control tubules; data not shown). Right : tubule pellet TLR4 content. A correlate of hypoxic and Fe-induced injury was a significant reduction in tubule TLR4 content (assessed in tubule pellets). G-mediated protection was associated with partial preservation of tubule TLR4 content.
Control tubule Western blots demonstrated TLR4 in all pellets, but not in the suspending medium ( Fig. 8 ). Tubules responded to H/R injury with a marked decrease in their TLR4 content (assessed in tubule pellets; Fig. 8 ) and a reciprocal increase in TLR4 content within the suspending media ( Fig. 8 ). The latter was observed in the absence of -tubulin. Glycine-mediated cytoprotection caused a significant preservation of tubule TLR4 content and prevention of TLR4 release ( Fig. 7, right, and Fig. 8 ).
Fig. 8. TLR4 levels in pelleted proximal tubules and in their suspending buffer after either control incubations, H/R, or Fe incubations. A : following control (C) incubations, no TLR4 was apparent in tubule buffer. TLR4 was readily apparent in the tubule pellets, appearing as 2 closely related bands of 90 kDa by Western blot. H/R injury caused striking tubule TLR4 loss, with reciprocal TLR4 appearance in the incubation buffer. B : Fe incubation caused modest TLR4 loss from tubules, with TLR4 appearing in the suspension buffer. Fe-induced tubule TLR4 reductions were less dramatic than those resulting from H/R injury, consistent with less extensive tubular cell death with the Fe vs. hypoxic challenge (34 vs. 48% LDH release, respectively; P < 0.01). C : buffer TLR4 following control, oxygenated, incubation (C), control incubation with 2 mM glycine (Gly), H/R injury, and H/R + glycine. TLR4 was not detectable in buffer after completing control incubations either with or without glycine. Marked TLR4 release into the buffer followed hypoxic injury. Hypoxic TLR4 release was almost completely blocked by glycine treatment.
Fe-mediated oxidative stress. Incubating the isolated tubules with Fe caused 34% LDH release, vs. 11% for control aliquots ( Fig. 7; results of the 30- and 45-min Fe incubations, combined). Correlates of this injury were TLR4 release into the tubule buffer (in the absence of -tubulin) and a corresponding reciprocal decrease in tubule pellet TLR4 content ( Fig. 7, right, and Fig. 8 B ).
Acute TLR4 mRNA Responses to Injury
Isolated tubules. As shown in Fig. 9, H/R injury caused about a threefold increase in TLR4 mRNA, factored by simultaneous GAPDH product. Glycine addition to control tubules did not independently affect TLR4 mRNA levels. However, a correlate of glycine-mediated protection was prevention of the H/R-induced increases in TLR4 mRNA.
Fig. 9. TLR4 mRNA in isolated tubules after completing control incubation ± glycine and H/R ± glycine. H/R caused a marked increase in TLR4/GAPDH ratios. Although glycine did not independently alter TLR4 mRNA expression, it almost completely abrogated the H/R-induced increases ( n = 4 per group).
Acute in vivo I/R. Ischemia + 90 min of reperfusion induced a significant increase in renal cortical TLR4 mRNA levels, compared with contralateral cortical controls (1.26 ± 0.12 vs. 0.7 ± 0.04, respectively; P < 0.05).
HK-2 cells. TLR4 mRNA rose greater than twofold in response to antimycin A treatment (1.60 ± 0.09 vs. control values of 0.75 ± 0.09; P < 0.005). As previously noted, this same antimycin A protocol failed to alter HK-2 cell TLR4 protein levels, as assessed by Western blotting ( 27 ). Thus, an antimycin-mediated increase in mRNA without a concomitant increase in TLR4 protein implies a translational block.
DISCUSSION
The present study places renal cortical TLR4 in a new perspective: rather than being strictly viewed as a component of the innate immune response, it now appears that it is a highly "injury-sensitive" protein, being lost from tubular cells in response to stress. Perhaps the most compelling, and direct, evidence for this new concept comes from the above-described isolated PT experiments. First, when PTs were subjected to H/R, TLR4 content fell by 75%. Second, TLR4 release was not specific for hypoxic tubular injury, given that acute reductions also occurred during Fe-driven oxidant stress. Third, documentation of large amounts of TLR4 in the extracellular space (incubation media) after both hypoxic and Fe-mediated injury indicates that rapid plasma membrane release was at least partially responsible for the tubule TLR4 depletion state. Fourth, TLR4 shedding is seemingly dependent on the development of lethal cell damage, given that glycine addition mitigated both hypoxic LDH release and tubule TLR4 loss. It is noteworthy that glycine-mediated cytoprotection is thought to occur via its low-affinity interaction(s) with multimeric protein channels, preventing pore formation with subsequent macromolecular efflux ( 5, 6, 19, 20 ). Thus, it would appear that injury-induced tubule TLR4 release does not simply represent nonspecific plasma membrane protein sloughing; rather, a gated process through injury-induced plasma membrane pores appears to be involved. That TLR4 loss from injured tubules occurred in the absence of -tubulin release further suggests that this process is not simply a nonspecific consequence of tubular cell death.
The above in vitro observations of proximal tubular TLR4 loss during lethal cell injury are consistent with our previously reported in vivo Western blot results ( 27 ). When mice were subjected to severe tubular necrosis and filtration failure via glycerol injection, a corresponding 50% reduction in renal cortical TLR4 content (Western blot) occurred ( 27 ). Conversely, when mice were studied at 18 h post-CP injection, a time that precedes the onset of tubular necrosis and filtration failure, no renal cortical TLR4 Western blot reductions were observed ( 27 ). To further explore whether severe proximal tubular injury evokes in vivo TLR4 loss, we have now performed Western blots on renal cortical extracts obtained 18 h and 4 days post-I/R injury and at 72 h post-CP injection. We hypothesized the following: 1 ) since a relatively modest ischemic insult (17.5 min) induces lethal injury in outer medullary stripe, but not most cortical PT, renal cortex should retain relatively normal TLR4 content; and 2 ) at 72 h post-CP injection, widespread proximal tubular necrosis exists within renal cortex, and thus, reduced cortical TLR4 content should be observed (i.e., analogous to the glycerol model). Indeed, these two hypotheses were supported by the presently gathered data: the postischemic kidneys retained near normal renal cortical TLR4 levels at both 18 h and 4 days postreflow; conversely, a 50% reduction in TLR4 levels were observed 72 h post-CP injection. Thus, all of the isolated tubule and renal cortical Western blot results support the following conclusion: that severe tubular injury culminates in a renal cortical TLR4 depletion state.
Renal cortical TLR4 Western blot results, such as gathered above, can arise from changes in any cortical tissue compartment (e.g, glomeruli or distal nephron segments). Therefore, to gain more direct insights into in vivo PT TLR4 expression in response to injury, full-length kidney sections obtained from control kidneys and kidneys from the above I/R, glycerol, and CP experiments were subjected to TLR4 immunohistochemical analysis. Interestingly, the control kidneys demonstrated an apparent anatomic "intrarenal inner medullary staining being observed. Within both renal cortex and outer medulla, PT expressed the greatest TLR4 content. However, with I/R, glycerol-induced ARF, and with fully developed (72 h) CP nephrotoxicity, marked reductions in proximal tubular TLR4 expression were observed. The TLR4 decrements were not simply confined to necrotic tubules, since intact cells within severely damaged tubule segments expressed markedly reduced TLR4 content (e.g., Fig. 5, D and F ). Conversely, with mild injury (e.g., at 18 h post-CP injection), tubular TLR4 content was relatively well preserved. Thus, these results are highly consistent with the conclusions derived from the above renal cortical Western blot and isolated tubule experiments: i.e., that renal cortical TLR4 reductions correspond with the onset of severe tubular injury/tubular cell death. It is noteworthy that the antibodies used to probe renal histological sections (immunohistochemistry) and renal cortical extracts (Western blots) were raised against different portions of the TLR4 molecule (COOH terminus, and amino acids 39-56, near the NH 2 -terminus end, respectively). That completely complementary data were obtained with these divergent approaches/TLR4 probes strengthens the conclusions that have been made.
A number of observations gleaned from the above experiments raise the possibility that urinary TLR4 sloughing might have potential utility as an early biomarker of impending ARF. First, the process of tubular TLR4 sloughing is a rapid one, as indicated by TLR4 appearance into isolated tubule suspension media with the onset of severe hypoxic and oxidant damage. Second, TLR4 loss appears to be a marker of lethal, rather than sublethal, injury, as indicated by the fact that glycine abrogated its release. Third, the immunohistochemical results indicate that following tubular injury, large amounts of TLR4 reside within tubular lumina. This presumably allows for urinary excretion, and hence, detection. To gain initial support for this latter possibility, control urine samples, urine samples obtained 4 h after ischemic renal injury, and urine obtained during the evolution of CP nephrotoxicity (18/48 h) were probed for TLR4 by Western blotting. In the absence of renal injury, no TLR4-immunoreactive material was detected in urine samples. Conversely, all postinjury urine samples contained 30 ± 60-kDa TLR4 fragments (presumably reflecting intrarenal protein cleavage). Thus, when these findings are interpreted along with the above observations, it appears that urinary TLR4 excretion might have utility as a marker of lethal tubular injury and impending ARF.
Although tubular cell sloughing could account for much, if not all, of the injury-induced renal cortical TLR4 reductions, depressed TLR4 gene transcription/mRNA translation might also contribute. To explore these possibilities, the acute impact of H/R on TLR4 mRNA expression was assessed in isolated tubule experiments. Surprisingly, a dramatic increase was observed. Another remarkable result was that glycine-mediated cytoprotection prevented these TLR4 mRNA increases. This implies that acute cell injury can rapidly induce TLR4 gene transcription and that the stimulus for this arises from some component of the glycine-suppressible injury pathway. To gain in vivo support for this in vitro observation, TLR4 mRNA was also measured in renal cortex after ischemia + 90 min of reperfusion. Again, acute TLR4 mRNA elevations were observed. To ascertain whether the latter might simply reflect a transient, or ischemia-specific response, renal cortical TLR4 mRNA was also measured at more delayed time points (18 h and 4 days) and 3 days post-CP injection. In all instances, enhanced TLR4 mRNA expression was observed. Given the rapidity of the mRNA increase (e.g., 15 min posthypoxic injury), increased transcription, and not simply postinjury mRNA stabilization, was likely involved. In a previous study, we noted that 18 h of antimycin-induced sublethal injury failed to raise TLR4 protein content in HK-2 cells ( 27 ). In the present study, we now demonstrate that this same antimycin injury protocol doubles HK-2 cell TLR4 mRNA. Thus, when viewed together, the TLR4 mRNA increase, without a corresponding change in that TLR4 protein levels, suggests the presence of a translational block. If true, then a failure of new TLR4 synthesis, in concert with ongoing TLR4 loss/shedding, could produce the postinjury TLR4 depletion state. It should be noted that other laboratories have also found postinjury TLR4 mRNA elevations ( 10, 21 ), and possibly, corresponding increases in TLR4 protein levels. What accounts for these differences between laboratories (e.g., differences in employed animal strains, assessment times, injury models) remains to be defined.
In conclusion, the present study advances a new concept: that TLR4 is an injury-sensitive component of the innate immune response complex, and as such, it can be rapidly depleted from PT during hypoxic/ischemic injury and oxidative stress. The evidence gathered suggests that TLR4 shedding via a glycine-suppressible pathway, possibly coupled with decreased TLR4 mRNA translation, is responsible. The development of tubule TLR4 depletion appears to denote the transition from a sublethal to a lethal injury stage. This fact, plus the rapid appearance of TLR4 fragments in pathologic urine, suggests that its excretion might have utility as a potential biomarker of lethal tubular injury and impending ARF. Finally, the results of the present experiments underscore that renal tubular hyperresponsiveness to TLR4 ligands (e.g., endotoxin) cannot simply be explained by increased TLR4 protein expression. Rather, a more downstream or alternative signaling pathway(s) is likely to be involved. Indeed, it could be that TLR4 loss during/after cell injury in a sense represents a beneficial response, blunting what is an already hyperreactive TLR4-ligand pathway in the setting of ARF ( 25 - 27 ).
GRANTS
This work was supported by National Institutes of Health Research Grants R37-DK-38432-20 and R01-DK-68520-03.
【参考文献】
Akira S, Takeda K. Toll-like receptor signaling. Nat Rev Immunol 4: 499-511, 2004.
Beutler B. Innate immunity: an overview. Mol Immunol 40: 845-859, 2004.
Bonventre JV, Zuk A. Ischemic acute renal failure: an inflammatory disease? Kidney Int 66: 480-485, 2004.
Chi H, Barry SP, Roth RJ, Wu JJ, Jones EA, Bennett AM, Flavell RA. Dynamic regulation of pro-, and anti-inflammatory cytokines by MAPK phosphatase 1 (MKP-1) in innate immune responses. Proc Natl Acad Sci USA 103: 2274-2279, 2006.
Dong Z, Patel Y, Saikumar P, Weinberg JM, Venkatachalam MA. Development of porous defects in plasma membrane of adenosine triphosphate depleted Madin Darby canine kidney cells, and its inhibition by glycine. Lab Invest 78: 657-668, 1998.
Dong Z, Venkatachalam MA, Weinberg JM, Saikumar P, Patel Y. Protection of ATP-depleted cells by impermeant strychnine derivatives. Am J Pathol 158: 1021-1028, 2001.
Friedewald JJ, Rabb H. Inflammatory cells in ischemic acute renal failure. Kidney Int 66: 486-490, 2004.
Hacker H, Redecke V, Blagoev B, Kratchmarova I, Hsu LC, Wang GG, Kamps MP, Raz E, Wagner H, Hacker G, Mann M, Karin M. Specificity in Toll-like receptor signaling through distinct effector functions of TRAF3 and TRAF6. Nature 439: 204-207, 2006.
Kawai T, Akira S. Toll-like receptor downstream signaling. Arthritis Research & Therapy 7: 12-19, 2005.
Kim SB, Lim SW, Kim JS, Sun BY, Ahm KO, Han SW, Kim J, Yang CW. Ischemia-reperfusion injury activates innate immunity in rat kidneys. Transplantation 79: 1370-1377, 2005.
Leemans JC, Stokman G, Claessen N, Rouschop KM, Teske GJ, Kirschning CJ, Akira S, van der Poll T, Weening JJ, Florquin S. Renal-associated TLR2 mediates ischemia/reperfusion injury in the kidney. J Clin Invest 115: 2894-2903, 2005.
Li S, Gokden N, Okusa MD, Bhatt R, Portilla D. Anti-inflammatory effect of fibrate protects from cisplatin-induced ARF. Am J Physiol Renal Physiol 289: F469-F480, 2005.
Lim SW, Li C, Ahn KO, Kim J, Moon IS, Ahn C, Lee JR, Yang CW. Cyclosporine-induced injury induces Toll-like receptor and maturation of dendritic cells. Transplantation 80: 691-699, 2005.
Ramesh G, Reeves WB. TNF- mediates chemokine and cytokine expression and renal injury in cisplatin nephrotoxicity. J Clin Invest 110: 835-842, 2002.
Ramesh G, Reeves WB. Inflammatory cytokines in acute renal failure. Kidney Int 91: S56-S61, 2004.
Sogabe K, Roeser NF, Venktatachalam MA, Weinberg JM. Differential cytoprotection by glycine against oxidant damage to proximal tubule cells. Kidney Int 50: 845-865, 1996.
Takeda K, Akira S. Toll receptors and pathogen resistance. Cell Microbiol 5: 143-153, 2003.
Takeda K, Kaisho T, Akira S. Toll-like receptors. Annu Rev Immunol 21: 335-376, 2003.
Venkatachalam MA, Weinberg JM, Patel Y, Saikumar P, Dong Z. Cytoprotection of kidney epithelial cells by compounds that target amino acid gated chloride channels. Kidney Int 49: 449-460, 1996.
Weinberg JM, Davis JA, Rajan T. Cytoprotective effects of glycine and glutathione against hypoxic injury to renal tubules. J Clin Invest 80: 1446-1454, 1987.
Wolfs TH, Buurman WA, Van Schadewijk A, de Vries B, Daemen MA, Hiemstra PS, Van?t Veer C. In vivo expression of Toll-like receptor 2, and 4 by renal epithelial cells: IFN- and TNF- mediated upregulation during inflammation. J Immunol 168: 1286-1293, 2002.
Ysebaert DK, De Greef KE, De Beuf A, Van Rompay AR, Vercauteren S, Persy VP, De Broe ME. T cells as mediators in renal ischemia/reperfusion injury. Kidney Int 66: 491-495, 2004.
Zager RA, Burkhart KM, Johnson ACM, Sacks BM. Plasma membrane cholesterol: a critical determinant of cellular energetics and tubular resistance to attack. Kidney Int 56: 193-205, 2000.
Zager RA, Johnson ACM, Hanson SY. Proximal tubular cytochrome c efflux: determinants, and potential marker, of mitochondrial injury. Kidney Int 65: 2123-2134, 2004.
Zager RA, Johnson ACM, Hanson SY, Lund S. Ischemic proximal tubular injury primes mice to endotoxin induced TNF- generation and release. Am J Physiol Renal Physiol 289: F289-F297, 2005.
Zager RA, Johnson ACM, Hanson SY, Lund S. Acute nephrotoxic and obstructive injury primes the kidney to endotoxin-driven cytokine/chemokine production. Kidney Int 69: 1181-1188, 2006.
Zager RA, Johnson ACM, Lund S, Hanson SY. Acute renal failure: determinants and characteristics of the injury induced hyperinflammatory response. Am J Physiol Renal Physiol 291: F546-F556, 2006.
作者单位：1 Department of Medicine, University of Washington, and the Fred Hutchinson Cancer Research Center ( 2 Clinical Division; 3 Research Pathology), Seattle, Washington