Local Anesthetics Reduce Mortality and Protect against Renal and Hepatic Dysfunction in Murine Septic Peritonitis

We used male C57BL/6 mice (approximately 20–25g) purchased from Harlan Laboratories (Indianapolis, IN). All animal protocols were approved by the Institutional Animal Care and Use Committee of Columbia University (New York, NY).

Mice were anesthetized with intraperitoneal pentobarbital (50 mg/kg or to effect). A mini-osmotic pump (Model #2001; Durect Corporation, Cupertino, CA) was implanted into the subcutaneous space on the back of each animal. All pumps continuously delivered 1 μl/h of 10% lidocaine (∼5.0 mg·kg⁻¹·h⁻¹), 5% lidocaine (∼2.5 mg·kg⁻¹·h⁻¹), 2% bupivacaine (∼1.0 mg·kg⁻¹·h⁻¹), 1% bupivacaine (∼0.5 mg·kg⁻¹·h⁻¹), or saline vehicle for 7-day survival studies. Subsequent functional and molecular studies were then performed with 10% lidocaine and 1% bupivacaine at 24 h after CLP. These doses of local anesthetics were chosen to mimic clinically achieved plasma concentrations in patients receiving epidural local anesthetics for postoperative pain control. To confirm that clinically relevant local anesthetic plasma concentrations were achieved at 24 h with this delivery system, local anesthetic plasma concentrations from mice were confirmed by high performance liquid chromatography measurements at the Nathan Kline Institute (Orangeburg, NY) as described previously (table 1).12 

Six hours after pump implantation (the time interval is required for pump priming), the mice were reanesthetized with intraperitoneal pentobarbital and were allowed to spontaneously breathe room air on an electric heating pad under a warming light. CLP was performed as described by others.13Briefly, the cecum was isolated through a 1-cm midline incision and the distal (0.5 cm) portion of the cecum below the ileocecal valve (to avoid bowel obstruction) was ligated with a 4–0 silk suture. The cecum was then punctured through and along its antimesenteric border (double puncture) with a 20-gauge needle and a small amount of stool was extruded through the puncture site. After instillation of 0.5 ml normal saline into the peritoneal cavity, the abdomen was closed in two layers. The mice remained on the heating pad until they recovered from the anesthetic. Twenty-four hours after CLP, mice were killed with an overdose of intraperitoneal pentobarbital, and specimens (plasma and kidneys) were collected. The left kidney was used for histologic evaluation and the right kidney was snap frozen in liquid nitrogen for all in vitro  studies described (e.g. , DNA laddering, semiquantitative reverse transcriptase polymerase chain reaction, and immunoblotting).

To determine the effect that local anesthetic infusions would have on mortality from CLP-induced septic peritonitis, continuous subcutaneous infusion of lidocaine and bupivacaine was administered to mice for 7 days until the pump reservoir was exhausted. Survival studies were performed with infusions of varying concentrations of lidocaine (5% and 10%) and bupivacaine (1% and 2%) to determine if dose dependency existed. All mice had free access to water and food and were frequently observed by dedicated research personnel to determine 7-day survival statistics. All severely moribund animals were euthanized with an overdose injection of anesthetic in adherence with our animal protocol. To examine if the manner of administration played an important role in survival, a separate small cohort of mice (n = 8) were bolused with lidocaine every 8 h totaling the dose they would have received with the continuous osmotic pump infusion. Specifically, a 0.8 mg subcutaneous lidocaine bolus every 8 h for 24 h was equivalent to an infusion of 10% lidocaine at 1 μl/h over 24 h.

Renal function was assessed by measuring plasma creatinine 24 h after CLP by a colorimetric method based on the Jaffe reaction.14Hepatic function 24 h after CLP was assessed by measuring plasma, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) concentrations using a commercially available colorimetric method (Sigma, St. Louis, MO).

Murine plasma tumor necrosis factor (TNF)-α and keratinocyte-derived chemokine (KC) concentrations taken 24 h after CLP were measured using commercially available ELISA kits according to manufacturer instructions (ALPCO, Windham, NH and R&D Systems, Minneapolis, MN, respectively).

Renal inflammation 24 h after CLP was determined by measurement of renal cortical myeloperoxidase activity (a marker of leukocyte infiltration) and immunoblotting for intracellular adhesion molecule (ICAM)-1 as described previously.12In addition, measurements of mRNA encoding for the markers of inflammation (KC, macrophage inflammatory protein [MIP]-2, ICAM-1, monocyte chemoattractant protein [MCP]-1), TNF-α, regulated on activation normal T-cell expressed and secreted, interferon-induced protein 10, and glyceraldehyde 3-phosphate dehydrogenase (control for RNA quantity) in renal cortices with semiquantitative reverse transcriptase polymerase chain reaction were also performed as described previously.12For each experiment, we also performed semiquantitative reverse transcriptase polymerase chain reaction under conditions yielding linear results for glyceraldehyde 3-phosphate dehydrogenase to confirm equal RNA input. Primers were designed based on published GenBank sequences (National Institutes of Health database, Bethesda, MD) for mice (table 2).

Both necrosis and apoptosis was assessed by histologic evaluation. Renal apoptosis was also evaluated by DNA laddering as well as by in situ  Terminal Deoxynucleotidyl Transferase Biotin-dUTP Nick End Labeling assay Assay (TUNEL).

Explanted kidneys were bisected along the long axis and fixed in 10% formalin solution overnight. After automated dehydration through a graded alcohol series, transverse kidney slices were embedded in paraffin, sectioned at 5 μm and stained with hematoxylin and eosin. Morphological assessment was performed by an experienced renal pathologist (SHN) who was blinded to the treatments that each animal had received. An established grading scale (0–4) of necrotic injury to the proximal tubules was used for the histopathological assessment of damage as outlined by Jablonski et al.  15For apoptotic assessment the pathologist examined hematoxylin and eosin stained mouse kidney sections under light microscopy for apoptotic bodies. As previously described, the degree of renal tubular apoptosis was quantified by counting the number of apoptotic bodies in proximal tubules in the corticomedullary area of the kidney.16Twenty-five to fifty tubules were counted per field from each treatment group and at least six fields were examined per section. Results are representative of at least four separate experiments.

To assess potential differences in DNA laddering (indicative of apoptosis), extracted DNA (Wizard; Promega, Madison, WI) from dissected renal cortices taken at 24 h after CLP were subjected to electrophoresis for 6 h at 70 V in a 2.0% agarose gel in Tris-acetate-EDTA buffer. Gels were stained with ethidium bromide and photographed under ultraviolet illumination. DNA ladder markers (100 bp) were used as a reference for the analysis of internucleosomal DNA fragmentation.

To further evaluate potential differences in renal apoptosis, TUNEL assay was performed on formalin fixed mice kidney sections obtained at 24 h after CLP. Briefly, sections were deparaffinized and rehydrated. In situ  labeling of fragmented DNA was performed with TUNEL (green fluorescence) using a commercial in situ  cell death detection kit (Roche, Indianapolis, IN) according to manufacturer instructions. To visualize the total number of cells in the field, sections were also concomitantly stained with a red fluorescent monomeric cyanine nucleic acid stain (Molecular Probes, Eugene, OR).

One-way analysis of variance was used to compare mean values across multiple treatment groups with a Dunnett post hoc  multiple comparison test. The ordinal values of the Jablonski scale were analyzed by the Kruskal-Wallis nonparametric test with Dunn posttest comparison between groups. Survival statistics were compared with a Kaplan-Meier curve and log-rank test. In all cases, a probability value less than 0.05 was taken to indicate significance.

Unless otherwise specified, all reagents were purchased from Sigma (St. Louis, MO). Protein content was determined with the Pierce Chemical (Rockford, IL) bicinchoninic acid protein assay reagent using bovine serum albumin as a standard.

We initially measured the effects of chronic local anesthetic infusion on mortality from CLP-induced septic peritonitis. As demonstrated in figure 1, infusion of lidocaine and bupivacaine significantly protected mice against CLP-induced mortality. Both lidocaine infusion groups (5% and 10%) significantly reduced mortality to 0% (n = 20) and 3% (n = 25), respectively, at 24 h compared with the saline vehicle group (24 h mortal-ity = 49%, n = 31; P < 0.05). Log-rank analysis demonstrated that a significant level of protection was conferred by lidocaine throughout the study. Similar results were observed in the mice that received continuous subcutaneous infusions of bupivacaine. CLP-induced mortality with 1% and 2% bupivacaine infusion was 7% (n = 22) and 14% (n = 21, respectively) at 24 h, and a significant degree of protection compared with saline vehicle persisted throughout the study (P < 0.05). The 2% bupivacaine infusion significantly improved survival versus  the 1% bupivacaine infusion (P < 0.05). However, we chose to use the 10% lidocaine and 1% bupivacaine treatments in the remaining studies because these more accurately reflect clinically relevant doses. We also found the manner of administration had an impact on survival benefit. We observed no benefit in mortality with bolus administration (every 8 h for 24 h) of 10% lidocaine (data not shown).

The development of organ dysfunction and injury as a complication of sepsis is associated with increased mortality. We examined the degree of renal dysfunction by measuring plasma creatinine and the degree of hepatic dysfunction by measuring ALT and AST 24 h after CLP. In this model of sepsis, plasma creatinine was significantly increased in saline vehicle treated mice (1.2 ± 0.10 mg/dl, n = 13) compared with sham surgery controls (0.50 ± 0.10 mg/dl, n = 8) (table 3). Mice treated with 10% lidocaine or 1% bupivacaine infusion had significantly lower plasma creatinine at 24 h (0.50 ± 0.10 mg/dl, n = 10 and 0.50 ± 0.10 mg/dl, n = 5; P < 0.001 respectively) compared with mice treated with saline vehicle alone Therefore, local anesthetic infusion alleviates CLP-induced renal injury. Similarly, 24 h after CLP operation AST and ALT concentrations were significantly increased in saline vehicle treated mice (AST, 349 ± 15 U/ml, n = 15 and ALT, 142 ± 10 U/ml, n = 12) compared with sham surgery controls (AST, 91 ± 11 U/ml, n = 8 and ALT, 20 ± 2 U/ml, n = 8; table 3). At 24 h after CLP, the lidocaine and bupivacaine treated groups also demonstrated lower AST (134 ± 13 U/ml, n = 7 and 169 ± 28 U/ml, n = 4 respectively; P < 0.001) and ALT (78 ± 7 U/ml, n = 7 and 95 ± 23 U/ml, n = 4 respectively; P < 0.001) compared with mice receiving vehicle alone. This data also demonstrates that local anesthetic infusion ameliorates CLP-induced hepatic injury in mice.

TNF-α concentrations were undetectable in most sham surgery control mice (n = 5), indicating that baseline concentrations were less than 18.2 pg/ml (the sensitivity of our ELISA). In contrast, plasma TNF-α concentrations in mice undergoing CLP with saline vehicle was significantly increased (1626 ± 553 pg/ml, n = 6) in comparison with the sham and the local anesthetic groups. Infusion of lidocaine or bupivacaine resulted in undetectable plasma TNF-α concentrations, indicating that local anesthetics reduced TNF-α concentrations after CLP-induced sepsis to concentrations indistinguishable from sham controls (table 3). KC plasma concentrations at 24 h after operation were significantly increased in mice undergoing CLP with saline vehicle (21375 ± 5353 pg/ml, n = 5) compared with concentrations observed in the sham group (173 ± 19 pg/ml, n = 4). Lidocaine and bupivacaine significantly reduced KC plasma concentrations to 1346 ± 150 pg/ml, n = 7 and 3292 ± 770 pg/ml, n = 4, respectively (table 3).

Myeloperoxidase is an enzyme present in leukocytes and is an index of tissue leukocyte infiltration.17CLP increased myeloperoxidase activity (2.10 ± 0.24 ΔOD/min/mg protein, n = 9) in the renal cortex of mice compared with sham controls (0.02 ± 0.02 ΔOD/min/mg protein, n = 7). However, mice receiving 10% lidocaine or 1% bupivacaine infusion demonstrated significantly reduced myeloperoxidase concentrations (0.57 ± 0.11 ΔOD/min/mg protein, n = 3 and 0.31 ± 0.21 ΔOD/min/mg protein, n = 5, respectively) as compared with mice treated with vehicle alone (table 3).

Immunoblotting for ICAM-1 protein from dissected renal cortices taken 24 h after CLP revealed that ICAM-1 protein expression was significantly higher in the CLP group receiving saline vehicle (263 ± 49% of sham, n = 6) versus  the lidocaine and bupivacaine treated groups (91 ± 16% of sham, n = 6 and 98 ± 55% of sham, n = 5; respectively) (table 3). In a separate set of experiments, saline-treated mice subjected to CLP demonstrated increased gene expression of KC, MIP-2, ICAM-1, MCP-1, TNF-α, regulated upon activation normal T-cell expressed and secreted, and interferon-induced protein 10 compared with sham controls. However, infusion of 10% lidocaine or 1% bupivacaine resulted in reduced mRNA quantities of KC, MIP-2, ICAM-1, and MCP-1 (fig. 2). mRNA quantity of TNF-α, regulated upon activation normal T-cell expressed and secreted, and interferon-induced protein 10 were not significantly different between treatment groups (data not shown). This occurred despite a decrease in systemic plasma concentrations of TNF-α with local anesthetic infusion as described above. This is consistent with other published findings that suggest suppression of TNF-α can occur at a posttranscriptional level.18 

Although no appreciable renal necrosis was observed 24 h after CLP-induced sepsis (Jablonski scores were 0 for all groups), CLP-induced sepsis in mice receiving saline vehicle resulted in a significant increase in renal cortical apoptosis. However, local anesthetic infusion decreased the magnitude of renal apoptosis after CLP, as indicated by quantification of apoptotic bodies, the appearance of DNA laddering, and TUNEL positive cells.

There was a minimal occurrence of renal apoptotic bodies seen in sham operated mice (0.023 ± 0.022 apoptotic bodies/tubule). CLP resulted in a significantly higher number of renal apoptotic bodies (0.320 ± 0.117 apoptotic bodies/tubule; P < 0.05). In contrast, infusion of 10% lidocaine (0.018 ± 0.008 apoptotic bodies/tubule, n = 6) or 1% bupivacaine (0.006 ± 0.004 apoptotic bodies/tubule, n = 4) resulted in decreased renal apoptosis at 24 h after CLP (P < 0.05).

Similar results were observed when renal cortices were assessed for DNA laddering (fig. 3). Sham-operated mice displayed little to no DNA laddering. Samples from mice subjected to CLP and saline infusion showed obvious DNA laddering composed of DNA fragments in multiples of 180 basepairs, whereas samples from mice receiving local anesthetic infusion demonstrated minimal to no evidence of DNA laddering. Figure 3is representative of three separate experiments.

Furthermore, we failed to detect TUNEL positive corticomedullary kidney cells in sections from sham-operated mice (fig. 4). Consistent with our above findings, mice treated with saline infusion undergoing CLP showed increased TUNEL positive staining in the corticomedullary junction. However, mice receiving local anesthetic infusion and subjected to CLP showed reduced TUNEL-positive cells in the corticomedullary junction after 24 h. Figure 4is representative of three separate experiments.

The major findings of the current study are that chronic infusions of lidocaine or bupivacaine improve mortality, provide significant protection from acute renal and hepatic dysfunction, and attenuate the hyperinflammatory response and renal apoptosis associated with murine septic peritonitis induced by CLP.

Sepsis remains a major health challenge facing contemporary medicine. Despite advances in antibiotic, hemodynamic, and ventilatory support, sepsis remains responsible for approximately 215,000 deaths in the United States each year.1Given the high mortality rate associated with sepsis (which ranges between 30% and 70%), therapeutic interventions are primarily judged by their impact on survival.19 

To more accurately recapitulate the complex immunology of human sepsis, we adopted the cecal ligation and double puncture model that produces a clinical picture that more accurately reflects human sepsis than animal models that utilize endotoxin or bacteria.3,20This model induces septic peritonitis and has been shown to resemble human sepsis with respect to cytokine generation and progression to multiorgan failure and response to certain therapeutic interventions.21Another advantage of this model is the ability to manipulate the magnitude of sepsis. In preliminary studies, we compared survival responses to graded injuries from 18-gauge, 20-gauge, and 22-gauge CLP. We chose to employ a 20-gauge needle to induce a lethal degree of sepsis with a high predilection for fulminant multiple organ failure in the absence of any interventions. Our mortality rates correspond well with other published survival studies using this genetic strain in CLP.22We deliberately chose this severe model of sepsis to elucidate whether local anesthetic infusion could improve survival, even in the face of a known lethal injury. By 48 h after CLP, we observed a profound level of mortality (78%) and morbidity in the majority of mice receiving saline vehicle alone. They uniformly exhibited piloerection and conjunctival secretions, were disheveled in appearance, remained sedentary, and abstained from food and drink. No mouse from this treatment group survived beyond 3 days. In contrast, mice receiving local anesthetic infusion were more robust, mobile, and resumed drinking and eating by 4 days, and a significant number survived past 7 days.

It is well known that mortality in sepsis is strongly influenced by the development of organ injury and dysfunction.1In concordance, our study showed chronic local anesthetic infusion not only decreased CLP-induced mortality, it significantly reduced the magnitude of CLP-induced renal and hepatic injury. Local anesthetics were able to decrease both creatinine and ALT/AST, suggesting that the protection conferred by local anesthetics was global and not restricted to a particular organ. Creatinine concentrations were not statistically different between sham and local anesthetic groups. Taken together, chronic local anesthetic infusion provided global protection from organ injury and dysfunction resulting from severe sepsis.

The fact that global antiinflammatory regimens have shown some degree of success in human sepsis lends credence to interventions predicated upon antiinflammatory modulation.5These beneficial human sepsis studies utilizing insulin, low dose steroids, and activated protein C all share antiinflammatory characteristics.23–25Similarly, there exists ample evidence of the potent antiinflammatory effects of local anesthetics. In a rabbit model of lung injury, Takao et al.  demonstrated that lidocaine infusion decreased various chemotactic factors and resulted in less neutrophil infiltration.26In a separate study lidocaine and bupivacaine inhibited the release of inflammatory mediators leukotriene B4 and interleukin-1 from human neutrophils and monocytes.27Several studies have also demonstrated that local anesthetics, in addition to possessing antithrombotic effects, can inhibit inflammatory cell migration and proinflammatory cytokine generation.28–30On the basis of these studies, we hypothesized that modulation of initial inflammatory responses after septic insult with local anesthetics would lead to improved mortality rates as well as hepatic and renal protection, as inflammation is a major component of cell death associated with renal and hepatic injury in sepsis. In accordance with other published findings, significant inflammatory processes are elicited during CLP-induced sepsis in our saline vehicle group (shown by considerable elevations in proinflammatory plasma cytokines, renal mRNA expression, myeloperoxidase activity, and ICAM-1 protein expression). Although no mechanistic data are provided in this study, we show that local anesthetic treatment reduced inflammatory responses after CLP in mice. We speculate that the protective effects of local anesthetics occur by limiting inflammatory cell (macrophage, neutrophil, lymphocyte) activation or by attenuating inflammation on renal and hepatic parenchymal cells.

We first examined plasma concentrations of proinflammatory cytokines (TNF-α and KC) to assess the general state of inflammation in our mice. TNF-α is a prototypic example of a proinflammatory cytokine that plays an important role in propagating a host of secondary inflammatory cascades. As such, it has been shown in a variety of studies to mediate many of the immunopathologic features seen in sepsis. It also has been implicated to directly cause acute renal failure in sepsis by acting on the TNF-1 receptor31and TNF-α has also been shown to cause multiorgan failure and lethality in sepsis.32We demonstrated that local anesthetic infusion resulted in a dramatic decrease in circulating concentrations of TNF-α 24 h after CLP. Local anesthetics mitigated TNF-α plasma concentrations to undetectable concentrations (comparable to sham controls), and this may in part account for the decreased renal injury we observed in local anesthetic treated mice. Murine KC is a proinflammatory alpha chemokine that putatively represents the functional homologue of human interleukin-8. As such, KC not only serves as a potent leukocyte attractant and activator but its overexpression in experimental peritonitis has also been associated with liver enzyme abnormalities and increased early mortality.33We also demonstrated a reduction in plasma KC concentrations with local anesthetic infusion, which may have served to prevent overamplification of proinflammatory responses mediated by KC. These results illustrate that local anesthetic infusion attenuates the overexuberant production of proinflammatory cytokines induced during severe sepsis.

To further examine our hypothesis that local anesthetics decrease inflammation, we then focused on gene expression in the kidney to examine if this phenomenon also occurred at the organ level. Previous studies have demonstrated that up-regulation of chemokines within the kidney accompanies the acute renal failure resulting from polymicrobial sepsis.34We found reduced gene transcription of KC, MIP-2, ICAM-1, and MCP-1 in mice receiving local anesthetic infusion. In addition to the systemic plasma reductions in KC described above, we also observed renal KC gene expression was attenuated with local anesthetic infusions. Similar to KC, MIP-2 plays a pivotal role in leukocyte recruitment. Previous studies have shown a positive correlation with blunting MIP-2 overexpression and increased survival in sepsis models.35Similarly, we observed substantial elevations in KC and MIP-2 mRNA concentrations with CLP, and a significant attenuation of this overexpression in KC and MIP-2 mRNA concentrations was seen in those mice treated with chronic infusion of lidocaine and bupivacaine. ICAM-1 is a proinflammatory adhesion molecule that participates in leukocyte recruitment and is up-regulated during acute inflammatory responses. MCP-1 is a chemokine also involved in the pathogenesis of sepsis that has been shown by other authors to indirectly contribute to renal injury.36It plays an important role in the elicitation of both macrophages and neutrophils during sepsis. We also show that chronic local anesthetic infusion results in a decrease in MCP-1 gene expression in lidocaine and bupivacaine treated mice. We did not observe significant differences in gene expression of TNF-α, regulated on activation normal T-cell expressed and secreted, or interferon-induced protein 10. We did see a decrease in systemic plasma concentrations of TNF-α with local anesthetic infusion, suggesting that inhibition of TNF-α expression may occur at a posttranscriptional level.

Renal apoptosis is often an important component in the development of acute renal failure. In renal ischemia and reperfusion injury models, it has been demonstrated that increases in apoptosis were directly associated with increased renal inflammation.37Daemen et al.  also demonstrated that blocking apoptosis prevented renal inflammation.38Evidence is emerging that renal apoptosis also plays a significant role in the development of acute renal failure in sepsis.39,40In the current study, we observed increased levels of apoptosis in renal tubule cells 24 h after CLP and a concomitant increase in renal dysfunction. In contrast, mice treated with local anesthetic infusion demonstrated decreased renal apoptosis and significantly improved renal function. Whether the decrease in renal apoptosis was mediated by local anesthetic suppressing proinflammatory mediators or by a direct effect on apoptotic pathways remains to be elucidated. We conclude from our data that in this sepsis model local anesthetic infusion is associated with decreased apoptosis and improved renal function after CLP.

A great deal of valuable work and attention has recently been dedicated to the immunosuppressive and hypoinflammatory phase that follows the initial hyperinflammatory phase of sepsis.20There is a growing body of evidence to suggest that uncontrolled, hyperexuberant cytokine regulation, as occurs in sepsis, can lead to apoptosis in a variety of cell types and subsequently manifest as immunosuppression. The fact that immunosuppression may arise secondary to the uncontrolled hyperinflammatory phase leads to the question of whether attenuation of the hyperinflammatory phase would aid in preventing subsequent immunosuppression. We agree that once a generalized state of immunosuppression is achieved, it logically follows that further immunosuppressive or antiinflammatory modulation would be deleterious. However, as already mentioned, there are three clinical studies that have demonstrated a benefit from antiinflammatory modulation in sepsis. We believe the benefits we have demonstrated in this study likely follow similar antiinflammatory mechanisms.

In conclusion, we demonstrate that chronic infusion of local anesthetics provides protection from CLP-induced mortality and acute organ dysfunction. As the pathogenesis of organ dysfunction in sepsis is largely mediated by an imbalanced, uncontrolled inflammatory response, we believe that local anesthetics improve organ function after septic insult by attenuating this hyperinflammatory process. This attenuation is global in nature, exerting its effect on a variety of proinflammatory molecules, and culminates in a reduction in organ injury and thereby protects from progression to multiple organ dysfunction. Given the protective benefit on survival and organ dysfunction observed with chronic local anesthetic infusion, our findings could have important therapeutic implications for patients in early sepsis. Although it is premature to advocate intravenous anesthetics in septic patients, this intervention may someday prove beneficial for patients in early sepsis. Intense work is now being performed to elucidate the cell types involved in mediating the protective effects of local anesthetics and the cellular signaling mechanisms involved in protection.