Steatotic mice are particularly susceptible to hepatic ischemia/reperfusion injury compared with their lean littermates. We have previously demonstrated that livers of mice having a spontaneous mutation in the leptin gene (ob/ob), resulting in global obesity and liver steatosis, are ATP depleted, are endotoxin sensitive, and do not survive (I/R) injury. We hypothesize that administration of an anti-LPS monoclonal antibody (mAb) prior to initiation of I/R would be protective from that insult. Steatotic mice (ob/ob) were subjected to 15 min of ischemia via complete porta-hepatis occlusion and varying lengths of reperfusion with or without pre-treatment with an anti-LPS mAb. There was 14–31% survival of isotype matched control mAb treated ob/ob mice after 15 min of ischemia and 24 h of reperfusion. In contrast, 75–83% of ob/ob mice pre-treated with an anti-LPS mAb prior to initiation of I/R survived both ischemia and 24 h of reperfusion. Furthermore, there was a decrease in ALT and circulating endotoxin levels when treated with an anti-LPS mAb compared with control antibodies. Attenuation of the endotoxin load with anti-LPS mAb, prior to initiation of I/R, was cytoprotective and improved survival. Consequently, these studies might offer a solution to the problems associated with using steatotic livers in clinical transplantation.
Steatosis (hepatocyte lipid accumulation) is a primary factor in determining the usability of potential donor organs due to poor outcomes associated with fatty livers (1,2). Nearly one-third of cadaveric donor livers are discarded secondary to the fat content alone. The prevalence of asymptomatic steatosis has been identified and recognized as a determining factor in the prequalification criteria of living donors (3). In response to increasing demand, many transplant programs are accepting moderate to severely steatotic livers that, in the past, would have been rejected for transplantation (4).
In human transplantation the exposure of the liver to endotoxin (LPS) is thought to be a significant factor in the pathogenesis of ischemia/reperfusion (I/R) injury (5). Endotoxin is the key constituent in the outer membrane of Gram-negative bacteria and is responsible for many of the adverse effects associated with sepsis (6). Endotoxin is composed of three regions. The first is a highly antigenic, outer O-polysaccharide region, which is unique for each Gram-negative bacterial strain. The second is the core which is comprised of outer, middle and deep or inner subregions. The third is the Lipid-A portion, which is poorly antigenic, but is highly conserved between species and genera (6). There are multiple factors leading to the increase of circulating endotoxin levels; however, the primary cause of this increase during liver transplantation results from portal occlusion during the anhepatic phase of the operation (7–10).
The heterogenicity of endotoxin delivered to the liver as a consequence of portal occlusion is well known (11). Consequently, two monoclonal antibodies (mAb) whose antigen binding sites are directed against a common epitope of the LPS molecule were used. They have been shown to be endotoxin antagonists and have a high affinity to their binding domains (6,12). First, the anti-J5 antibody is derived from a mutant form of Escherichia coli 0111:B4 (J5) that lacks the enzyme UDP-galactose 4-epimerase resulting in inhibition of the bacteria's ability to incorporate exogenous galactose into its lipopolysaccharide structure. This mutant lacks the O-side chain and consists only of the lipid A and inner core sugars. Clinical trials involving the anti-J5 antibody which was given after the onset of sepsis were not as successful as hypothesized (13). A precise mechanistic explanation as to why the J5 antibody was unsuccessful is unknown, but the current hypothesis is that its failure was likely due to the timing of its administration. The second antibody investigated was against a common antigenic portion of the Deep Core/Lipid A (DCLA) region of the polysaccharide molecule (6). The anti-DCLA mAb was produced using a mixture of LPS from Salmonellaminnesota Re and E. coli J5 as the antigens for vaccination. By using such a mixture of LPS in the preparation of the anti-DCLA mAb, good cross reactivity was achieved with this antibody. Both the anti-J5 and anti-DCLA monoclonal antibodies binding regions include the common epitope of the antigenic deep core and the highly conserved lipid A regions (13). Therefore, these antibodies were selected for this study because the deep core/lipid A components of endotoxin are highly conserved across a wide range of Gram-negative bacteria, and it was hypothesized that both of these antibodies would yield a protective result against I/R injury. Previous studies have shown that these mAbs were able to decrease endotoxin activity, in vitro as well as in vivo, when the mAb and the LPS were administered together (14,15). Consequently, studies were initiated to determine whether or not these mAbs would be an effective therapy to block endotoxin injury associated with portal occlusion and hepatic I/R if they were administered prior to the ischemic insult. We hypothesized that through such a protocol of administration, the anti-LPS mAbs would protect steatotic livers from endotoxin-mediated injury.
Adult 6- to 8-week-old male ob/ob mice and their lean littermates were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Mice were housed in temperature- and light-controlled chambers on a 12 h:12 h light/dark cycle and provided with rodent chow and water ad libitum. Sterile monoclonal antibody (200 μg/mouse) in phosphate-buffered saline (PBS) was injected intravenously via tail vein injection 1 h prior to ischemic challenge. In situ warm hepatic ischemia with bowel congestion was initiated with administration of Nembutal anesthesia at 50 mg/kg body weight intraperitoneally (i.p) followed by a mid-line laparotomy. The portal triad was identified and mice were ischemicized using a pediatric vessel loop ligature for 15 min followed by reperfusion (16). Animals were sacrificed immediately or 24 h after reperfusion, and samples were collected. Animals were sacrificed by exsanguination and occipital-cervical subluxation according to approved IACUC protocols.
Purified anti-Deep Core/Lipid A (anti-DCLA) monoclonal antibody was developed and characterized by Dr Davin L. Dunn. Its isotype-matched control mAb IgG2b was harvested from a hybridoma culture (ATCC no. CRL-1770) and supplied by Dr Dunn. Anti-J5 monoclonal antibody was harvested from a hybridoma culture (ATCC no. HB-8297) grown in medium of RPMI-1640 (Biowhittaker Cambrex, Walkersville, MD, USA), supplemented with 10% heat-inactivated fetal bovine serum (FBS, Hyclone, Logan, UT, USA) (110 mL/L), 200 mM l-glutamine (11 mL/L), 100 mM sodium pyruvate (11 mL/L), 100× nonessential amino acids (11 mL/L), and 100 U penicillin/streptomycin (11 mL/L). The mAb was purified using a mAbTrap(tm) Kit (Amersham Pharmacia Biotech, Uppsala Sweden), and the final product was tested for endotoxin contamination using a LAL endotoxin assay (Associates of Cape Cod, Folmouth, MA, USA). An IgG1 antibody against Tri-Nitro Phenol (anti-TNP) mAb was used as an isotype-matched control for anti-J5. It was grown from a hybridoma (ATCC no. TIB-191) in medium of RPMI-1640, 10% heat-inactivated FBS(110 mL/L) 200 mM l-glutamine (11 mL/L), sodium bicarbonate (1.5 g/L), glucose (2.5 g/L), 10 mM HEPES (1.1 mL/L), 100 mM sodium pyruvate (11 mL/L), and 100 U penicillin/streptomycin (11 mL/L).
Serum alanine aminotransferase (ALT)
ALT was evaluated by a Synchron LX20 System (Bectman Coulter Inc., Fullerton, CA, USA) and expressed as IU/L (Clinical Laboratory Services, Medical University of South Carolina, SA, USA).
Concentrated sera were diluted 1:10 in LAL reagent grade water and boiled for 2 min to irreversibly denature serine proteases. Endotoxin analysis was performed according to manufacturer's instructions using the chromogenic endpoint assay (Associates of Cape Cod, Inc., Cat. no. C0180) (17–19).
Oil-Red-O (ORO) and hematoxylin and eosin (H&E) staining were performed as previously described (20,21). Slides were double-blinded and necrosis grading was quantitated as previously described on a scale from 0 to 3. Five high-powered fields per section were analyzed in relation to the central vein (22).
In situ TUNEL staining
Liver tissue sections were parafinized as instructed by the manufacturer of the ApopTag® kit (Chemicon International®, Temecula, CA, USA). Sections were deparafinized with xylene and rehydrated through serial ethanol, digested with proteinase K (10 μg/mL for 15 min at room temperature), and rinsed in PBS. Endogenous peroxidase activity was blocked by 2% H2O2 (5 min at room temperature). The liver sections were incubated with terminal deoxynucleotide transferase (TdT) and digoxigenin-dUTP at 37 °C for 1 h. The incorporated digoxigenin-dUTP was detected by peroxidase-conjugated antidigoxigenin antibody and signal developed by incubation with 3, 3′-diamino-benzidine (DAB) in the presence of H2O2.
Experimental results were analyzed for their significance (Student's t-test). Significance was established at the 95% confidence level (p < 0.05)
Anti-LPS mAbs improve animal survival
We have previously shown, using our warm ischemia model, that steatotic mice (ob/ob) are more sensitive to I/R than their lean counterparts (16). In order to ascertain whether or not LPS played a role in the decreased survival after I/R, we hypothesized that administration of an anti-LPS mAb prior to the ischemic insult would protect the steatotic liver from the effects of endotoxin, thus leading to increased animal survival. Consistent with our previous experiments using this animal model, mice that received the isotype-matched control mAbs, IgG2b or anti-TNP, as a pre-treatment 1 h prior to ischemia followed by 15 min of ischemia and 24 h of reperfusion had a survival of 14.2% and 31%, respectively (Figure 1). Alternatively, when the mice received either of the anti-LPS mAbs (anti-DCLA or anti-J5), the survival was significantly increased to 83.3% and 75%, respectively (p < 0.05). A slight, but not statistically significant difference, was observed in mice who received the control mAb (p = NS).
Histological analysis was performed to determine if the pre-treatment of ob/ob mice with anti-LPS mAbs was protective against LPS-induced I/R injury. All samples were compared with baseline mouse tissue (Figure 2A). Evaluation of the tissues revealed decreased centrilobular necrosis and a greater percentage of viable tissue with anti-DCLA mAb (Figure 2E) and anti-J5 mAb (Figure 2C) treatments as compared with animals receiving control mAbs (Figure 2B,D). Cell death was concentrated to zone 3 (central vein) in the antibody treatment groups and was more extensive in samples taken from the animals receiving the isotype-matched control mAbs. To further evaluate the histology, the H&E stained specimen's were analyzed using a previously described necrosis quantification method (22). Values ranged from 0 (no apoptosis) to 3 (confluent foci) using five high-powered fields per section. The results confirm the anti-LPS mAbs ability to significantly protect the liver from endotoxin-induced I/R injury compared with its isotype-matched control antibody (p < 0.05) (Table 1).
Table 1. Graded histology of liver H&E sections
Average necrotic index
Slides were double blinded using the scale as described by Neil et al. Slide scores ranged from 0 to 3. 0 = absent apoptosis, 1 = single apoptosis, 2 = small foci 2–3 cells, 3 = confluent foci. Values were averaged from all experiments.
*p < 0.05.
1 ± 0.0
IgG2b control mAb
2.6 ± 0.9
IgG1 control mAb
2.2 ± 0.25
1 ± 0.0*
1 ± 0.0*
To further characterize the ability of anti-LPS mAbs to protect steatotic livers from I/R, we next focused our attention on whether these mAbs were diminishing hepatocyte injury by assessing the level of ALT. At baseline, ob/ob mice have an ALT level of 100 ± 16/IU (Figure 3). Three hours after reperfusion, ALT levels were 1082 ± 399/IU vs. 510 ± 183.8/IU for control mAb and anti-LPS mAb, respectively. At 24 h, ALT levels observed in control ob/ob mice pre-treated with an isotype-matched control mAb followed by I/R, continued to be increased to 450 ± 25/IU. However, when the mice were pre-treated with either of the two anti-LPS mAbs, followed by I/R, the ALT levels returned to those routinely observed at baseline. Pre-treatment with anti-DCLA resulted in ALT levels that were significantly reduced to 85 ± 17/IU compared with the isotype-matched control-treated mice (p < 0.05) (Figure 3). Similar results were also observed in the mice treated with anti-J5 mAb vs. isotype-matched control mAb (data not shown).
Circulating endotoxin levels
For endotoxin to initiate its receptor ligand binding and signal a cascade of events, its lipid A portion must bind to LPS binding protein, which then binds to toll-like receptor 4 (TLR4) (23). To determine whether the protective effects of the administration of anti-LPS mAb were a consequence of preventing the lipid A portion of the LPS from reaching the receptor involved in signal amplification, we assessed serum endotoxin levels immediately after reperfusion. Circulating serum endotoxin levels were measured with and without treatment using an anti-LPS mAb and/or its isotype-matched control mAb. In steatotic mice, a low level of circulating endotoxin was observed at baseline prior to I/R (Figure 4). Upon introduction of the ischemic challenge followed by reperfusion there was a significant increase in serum endotoxin from 0.8 ± 0.18 EU/mL observed at baseline to 1.75 ± 0.06 EU/mL after I/R when a control antibody was administered prior to the insult. This rise in serum endotoxin was significantly attenuated in ob/ob mice pre-treated with anti-DCLA mAb 0.71 ± 0.33 EU/mL (p < 0.05). Similar results were also seen with the anti-J5 mAb (data not shown).
The mechanism of hepatic injury after I/R is controversial. It has been shown to be a consequence of necrosis, apoptosis or a mixture of both (24–26). Exposure to LPS has been shown to increase apoptosis in hepatocytes (27). Given our histopathologic results and to further examine the protective effect of anti-DCLA mAb pre-treatment, we assessed the liver tissue by investigating apoptosis. A TUNEL assay was used to evaluate liver hepatocyte apoptosis after I/R. Apoptosis was not observed in baseline tissues (Figure 5A). In tissue harvested from animals pre-treated with a control mAb prior to I/R, a significant increase in apoptotic bodies 15 ± 10.51 was observed. In contrast, in tissue collected from mice pre-treated prior to I/R with an anti-LPS mAb, a significant decrease (p < 0.05) in the degree of apoptosis was observed. On average in those, the numbers of apoptotic bodies per field, 1.56 ± 2.18, were very similar to the number observed from tissue harvested from baseline animals (Figure 5A). Apoptotic bodies were primarily located in the area outside of the necrotic injury.
As the prevalence of obesity in the general population increases, so will the incidence of hepatic steatosis (28). Consequently, the number of available livers for transplantation will potentially decrease while the number of patients on the waiting list will continue to grow, both in terms of the number of individuals awaiting a suitable graft and length of time on the list. Because of this, there is a growing need for therapeutic interventions that will increase the likelihood of success when challenged with use of steatotic livers in transplantation and thus favor the use of ‘expanded criteria’ organs (29,30).
We investigated a novel application in hepatic injury of a well-described therapy to block the effects of exposure to endotoxin. Historically, anti-LPS mAbs have been very successful in animal models in blocking sepsis and improving survival when administered prior to a septic insult (15,31–33). These results were so promising and sepsis has such a high mortality that two anti-LPS mAbs were developed for human use. These agents were studied in phase 3 and 4 clinical trials. However, they failed to provide a survival advantage and were never adopted for clinical use (13,34). Analysis of their failure showed that the lack of effectiveness was possibly due to the timing of administration of the mAb in relation to the septic insult and the inability to reverse the cascade of cytokine-mediated events that followed (34).
In clinical liver transplantation, endotoxin plays a central role after reperfusion, when circulating levels increase intra-operatively (35). Furthermore, circulating endotoxin is elevated in the setting of cirrhosis (36). There are numerous sources of endotoxin during a transplant. The major source is from translocation of bacteria to the portal venous system during the anhepatic phase of the operation secondary to portal occlusion (7–10,37). In the previous attempts to attenuate sepsis with monoclonal antibodies against LPS, antibodies were unable to reduce endotoxin because they were given after the septic insult (13,38,39). Because these events can be anticipated during a liver transplant and it is known that serum endotoxin levels will increase, patients can be pre-treated with an anti-LPS mAb prior to portal occlusion. Attenuating the levels of circulating endotoxin may thereby result in the protection of the organ from endotoxin-associated I/R injury.
To study the effectiveness of this therapy, animals were pre-treated with anti-LPS mAbs prior to I/R. This pre-treatment resulted in an increased survival rate compared with those animals treated with an isotype-matched control antibody. The survival data presented in this study show that the anti-LPS antibodies, which are effective in binding a wide variety of endotoxin in vitro secondary to their common epitopes, were protective in a lethal warm ischemia model. These results suggest that through the use of antibodies to the common Lipid A region that it was possible to neutralize the toxic effects of endotoxin-mediated damage secondary to I/R. It is likely that the protective effect associated with the antibodies resulted from the fact that the epitopes recognized the majority of the endotoxin associated with Gram-negative bacteria.
There are conflicting data concerning the efficacy of anti-LPS antibodies for the treatment of sepsis in humans. Most of the antibodies used in the human trials were directed against the core or lipid A region of the endotoxin molecule, and all achieved similar results (13,40). In each of these trials, they attempted to reverse the effects of sepsis by administering the antibodies after the onset of sepsis. In our study the animals were pre-treated with the antibodies prior to I/R. By administering the antibodies 1 h prior to the ischemic insult, it was possible to insure that the antibody was sufficiently perfused throughout the animal. Consequently, pre-treatment blocked the endotoxin from eliciting its inflammatory response after I/R. By blocking the effects of endotoxin through its sequestration with the mAb a dramatic improvement in animal survival as well as a decrease in liver injury was observed. Therefore, it was the pre-treatment of animals prior to I/R that was critical to the survival differences noted between anti-LPS and control-treated mice, as it is difficult to reverse the inflammatory cascade once it has begun.
Endotoxin is thought to activate hepatic macrophages which then release toxic substances resulting in hepatocyte injury, repair and fibrosis (41–43). Thus, by sequestering the endotoxin through its binding to the mAbs it was possible to prevent or significantly limit endotoxin from binding to Kupffer cells. This limited the activation of the inflammatory cascade resulting in the protection of the steatotic liver observed in this study. Indeed, our data strongly support this in that there was a decrease in serum ALT after I/R in mice treated with an anti-LPS mAb compared with treatment with an isotype-matched control. Reduced ALT levels were paralleled by a decreased cellular necrosis. Furthermore, it has been previously demonstrated that exposure to LPS resulted in an increase in apoptosis and necrosis in hepatocytes, especially in steatotic livers (44). This observation was confirmed in our control mice receiving I/R. Again, in animals receiving the appropriate anti-LPS mAbs prior to I/R there was a marked attenuation of apoptosis, as measured by TUNEL. Taken together, these results suggest this decrease in cellular injury is due to the sequestration of circulating endotoxin. This too was confirmed in that the circulating endotoxin was decreased in the mice treated with an anti-LPS antibody compared with the isotype-matched controls. Consequently, the anti-LPS antibodies used in this study appear to be binding the endotoxin resulting in blocking the interaction of LPS with the hepatocytes, thereby preventing the lethal cascade of signaling that follows. Therefore, the conclusion that can be drawn is that it is the pre-treatment of the animal prior to I/R with an anti-LPS mAb that is responsible for the muffling of the inflammatory cascade preventing and/or limiting injury to the steatotic hepatocyte.
In summary, the data presented in this study support the administration of anti-LPS mAbs prior to an ischemic insult in order to protect a steatotic liver from the effects of endotoxin-associated I/R injury. Both the anti-J5 and anti-DCLA, two different antibodies directed towards a common antigen, demonstrated protection in a lethal warm ischemia model. The anti-LPS mAbs conferred protection from endotoxin and its subsequent toxic cascade by sufficiently sequestering the endotoxin prior to its interaction with the steatotic hepatocyte. Further analysis of the mechanism associated with protection from endotoxin using these agents is warranted. Presently, the determination and understanding of the anti-LPS ability to interrupt the cascade of inflammatory cytokines after endotoxin exposure is being investigated. In addition, this intervention with anti-LPS mAbs has the potential to be an easy supplement to current clinical practices to help expand the use of steatotic donor livers in transplantation.
We would like to thank Margaret Romano in the MUSC Department of Pathology and Laboratory Medicine for her help in cutting and processing the samples.