Kaiser, Vicki L; Sifri, Ziad C; Dikdan, George S; Berezina, Tamara; Zaets, Sergey; Lu, Qi; Xu, Da-Zhong; Deitch, Edwin A
It has been recognized for over 25 years that the development of a dysregulated inflammatory response as well as acute lung injury and multiple organ failure (MOF) complicates the recovery of patients with severe trauma (1). In spite of this recognition, advances in medical care have done little to reduce the incidence or mortality rate of trauma-induced acute lung injury or MODS. This failure of medical therapy appears to be related, at least in part, to an incomplete understanding of the mechanisms by which trauma-hemorrhage leads to acute organ dysfunction. Although several hypotheses have been proposed to explain the pathogenesis of trauma-induced systemic inflammation and organ injury (2), until recently the mediators, mechanisms, and sources of the factors promoting acute organ dysfunction after trauma-hemorrhage have not been fully clarified. However, recent studies in rodents and primates indicate that gut-derived factors contained in intestinal lymph, but not portal blood, are key factors in the pathogenesis of acute lung injury, bone marrow dysfunction, neutrophil activation, and red blood cell dysfunction in animals subjected to trauma-hemorrhagic shock (T/HS) and that these injurious effects involve endothelial cell as well as neutrophil activation (3, 4). This conclusion is based on the following two sets of observations. First, ligation of the efferent mesenteric lymphatic duct exiting the gut, which prevents intestinal lymph from reaching the systemic circulation, also prevents T/HS-induced lung injury, bone marrow dysfunction, endothelial cell injury, upregulation of endothelial cell adhesion molecule expression, and neutrophil activation, as well as red blood cell dysfunction (3-10). Second, that the in vitro effects of T/HS lymph, but not lymph collected from sham-shocked animals, recreates the endothelial cell and neutrophil changes observed in vivo (3, 4, 7, 11-13). These studies clearly indicate that the source of the factors leading to T/HS-induced acute inflammation and lung and bone marrow dysfunction originate in the ischemic gut.
Having identified the gut as the source of factors leading to acute organ dysfunction and a systemic inflammatory response, determination of the identity of these gut-derived factors is the next critical question to address. Based on previous studies, we know that the biologic properties of T/HS lymph are not due to translocating bacteria or endotoxin because these lymph samples are sterile and do not contain measurable levels of endotoxin (13). Likewise, the biologic activity of T/HS lymph does not appear to be cytokine-mediated or due to other potentially toxic factors such as xanthine oxidase (13, 14). Yet until the identity of these factors is determined, the development of potential therapeutic options will be limited. Consequently, because the endothelium is recognized as being a central effector of the immunoinflammatory response, and endothelial activation/injury contributes to organ failure (1, 15), in the current study, we have focused on identifying the putative mediators contained in T/HS lymph that leads to endothelial cell injury. The results of this study indicate that protein and lipid factors toxic for endothelial cells are present in T/HS lymph.
MATERIALS AND METHODS
Male Sprague-Dawley rats (pathogen free; Charles River Laboratory, Wilmington, MA) were housed in barrier-sustained conditions under 12-h light/dark cycles and were allowed free access to water and chow (Teklad 22/5 rodent diet W-8640; Harlan Teklad, Madison, WI). Rat maintenance was performed in accordance with the “Guide for the Care and Use of Laboratory Animals,” and all procedures were approved by the New Jersey Medical School Animal Care Committee.
T/HS model and lymph collection
Male rats (300-350 g) were subjected to trauma (laparotomy) and hemorrhagic shock or trauma-sham shock (T/SS) as previously described (4). Briefly, rats were anesthetized with pentobarbital (50 mg/kg, i.p.) and the femoral artery was isolated and cannulated with polyethylene (PE-50) tubing containing heparinized saline. The catheter was attached to a blood pressure recorder, which was used to continuously monitor the animals' blood pressure during the shock period. The jugular vein was cannulated, followed by a laparotomy to cannulate the mesenteric lymphatic duct for lymph collection. The abdomen was subsequently closed, and after a 1-h stabilization period, the animals were subjected to HS or SS.
In the T/HS group, blood was withdrawn from the internal jugular to a mean arterial pressure (MAP) of 30 to 40 mmHg and were maintained at this level for 90 min by withdrawing or reinfusing shed blood as required to maintain this blood pressure. At the end of the shock period, the animals were resuscitated by reinfusing all shed blood. Mesenteric lymph samples were collected in sterile tuberculin syringes on ice in hourly increments for 6 h postshock. The T/SS rats were anesthetized and cannulated as described above and lymph was collected, but no blood was withdrawn or infused.
Lymph collected at each time point was centrifuged (500g for 20 min at 4°C) to remove all cellular elements and obtain humoral cell-free lymph after which the lymph samples were frozen at -80°C.
Solid phase extraction (SPE)
Frozen lymph samples from T/HS and T/SS rats were thawed and subjected to microcentrifugation (13,000 rpm at 4°C for 5 min) to pellet the material that had precipitated during storage. Approximately 0.5 mL of lymph was diluted with an equal volume of sterile 1× phosphate-buffered saline (PBS) and was applied to a 500-mg Strata reversed-phase C18-E column (Phenomenex, Torrance, CA). The lymph was eluted at a gravitational flow rate at room temperature. The eluted samples were immediately placed on ice. Lymph was separated using a step-wise gradient of acetonitrile (ACN) in dH2O (0%-100% ACN, 10% increments, 5 mL each ACN concentration). The material for each elution was collected in a 50-mL polypropylene conical tube that was then dried by vacuum centrifugation to remove the ACN. Pellets were resuspended in 100 μL of sterile dH2O and 10 μL was used to test for endothelial cell toxicity (10% [v/v] final dilution).
Anion exchange chromatography
Frozen lymph samples were thawed and subjected to microcentrifugation (13,000 rpm at 4°C for 5 min) to remove material that had precipitated during storage. Approximately 0.5 mL of lymph was diluted with an equal volume of 25 mM NaCl in 20 mM Tris buffer (pH 7.2) and was applied to a 5-mL HiTrapQ HP anion exchange column (Pharmacia, Piscataway, NJ). The lymph was then eluted using a step-wise gradient of NaCl in 20 mM Tris buffer (pH 7.2) as follows: 25 mM NaCl (6 mL), 100 mM NaCl (10 mL), 200 mM NaCl (10 mL), 400 mM NaCl (10 mL), 2 M NaCl (15 mL), and 25 mM NaCl (30 mL) for column re-equilibration. Material was eluted at a flow rate of 2 mL/min using a peristaltic infusion pump (BioChem Technology, Malvern, PA) and was collected as 1-mL fractions. Absorbance peaks were obtained by placing an aliquot from each fraction (100-200 μL) into a 96-well UV microtiter plate (Falcon, Becton Dickinson, Franklin Lakes, NJ) and reading at a wavelength of 280 nm (SpectraMax 384 Plus; Molecular Dynamics, Sunnyvale, CA). The appropriate elution buffers were used as blanks.
A modified step-wise gradient was also used to separate lymph samples. Lymph was separated as described above with the following modifications: Lymph was eluted from a 1-mL HiTrapQ HP anion exchange column (Pharmacia) at a flow rate of 0.75 to 1 mL/min and was collected as 0.5-mL fractions. The gradient consisted of NaCl in 20 mM NaPO4 as follows: 25 mM NaCl (4 mL), 100 mM NaCl (4 mL), 200 mM NaCl (4 mL), 225 mM NaCl (4 mL), 250 mM NaCl (4 mL), 300 mM NaCl (4 mL), 500 mM NaCl (4 mL), 2 M NaCl (5 mL), and 25 mM NaCl (4 mL) for column re-equilibration.
Gel electrophoresis, electroelution, and protein assays
Precast Tris-CL polyacrylamide gels (7.5% and 12%), electrophoresis buffer reagents, sample buffer, and gel stains were purchased from Bio-Rad (Hercules, CA). Electrophoresis under native or denaturing conditions was performed as per the manufacturer's instructions. Bands were visualized by staining gels with Amido black or Coomassie Brilliant Blue (Bio-Rad) for 20 min under gentle agitation and were destained in several changes of 40% methanol/10% acetic acid. Gels stained by ProtoBlue (National Diagnostics, Atlanta, GA) were performed as per the manufacturer's instructions.
Protein electroelution was performed using a elecroeluter (model 422; Bio-Rad). Protein bands were excised from native gels and eluted in native buffer as per the manufacturer's instructions. Protein was desalted using sterile MilliQ dH2O and were concentrated using Ultrafree Biomax protein concentrators according to the manufacturer's instructions (10K NMWL, 0.5 mL; Millipore, Bedford, MA).
Total protein concentration was measured using the Bio-Rad protein assay following the manufacturer's standard protocol for microtiter plates.
Cell viability assay
Human umbilical vein endothelial cells (HUVECs; Biowhittaker, Walkersville, MD) were used to test whole lymph and fractions obtained from SPE and ion exchange chromatography (IEX) separations of whole lymph. Viability of the HUVECs was determined using a mitochondrial cell viability assay (MTT; Sigma, St. Louis, MO). Cells were seeded at 2 × 104/well on a 96-well tissue culture plate (Falcon), and were grown to confluence at 37°C/5%CO2 in endothelial cell basal media (Biowhittaker). The media was supplemented as per the manufacturer's instructions with fetal bovine serum, bovine brain extract, heparin, endothelial cell growth medium, L-glutamine, hydrocortisol, and antibiotic-antimycotic solution (gentamicin and amphotericin B). Only cells of low passage (3-5) were used for assays.
Toxicity was determined by incubating cells with media and whole lymph (5%, v/v) or lymph fractions (10%, v/v) in a final volume of 100 μL. Cells incubated with media alone served as a negative control. After an 18-h incubation, the supernatant was removed and replaced with 90 μL of basal media without phenol red and 10 μL of MTT (5 mg/mL) and was then incubated for an additional 3 h. Formazan crystals produced by viable cells were then solubilized as per the manufacturer's instructions. The plates were read at 570 nm, and viability was calculated as a percentage of the cells incubated with media alone.
Because separation of lymph by anion exchange results in a large number of fractions (80+), samples tested for HUVEC toxicity were obtained by pooling, concentrating, and then desalting the fractions constituting each peak(OD280). Pooled fractions were concentrated to approximately 75 μL, desalted against sterile dH2O (Ultrafree Biomax protein concentrators, 10K NMWL; Millipore), and then tested on HUVECs at 10% (v/v; final concentration) as described above.
In vitro albumin glycosylation
in vitro glycosylation of rat serum albumin (RSA; Sigma) was performed following the procedure of Day et al. (16). Briefly, RSA was incubated with β-D-glucose in 200 mM NaPO4 (pH 7.4) at 37°C for up to 6 weeks (50 mg/mL RSA and 1.4 M β-D-glucose, final concentrations).
Mass spectrometry was performed at the Biomolecular Resource Facility Mass Spectrometry Laboratory at the University of Texas Medical Branch (Galveston, TX).
Data were analyzed using Sigma-Plot 3.0 and comparisons were evaluated using the Student's t test). They were considered significant at P < 0.05.
Lymph samples were collected from animals subjected to T/HS or T/SS as described in “Materials and Methods.” Lymph samples from different time points after shock were then tested for toxicity to HUVECs using the standard MTT test, and a representative sample of these measurements is shown in Table 1. Consistent with previous studies (13, 17), lymph collected 1-, 2- and 3-h post-T/HS showed maximal toxicity to HUVECs as reflected in the low viabilities (<20%). These low viabilities are in sharp contrast to the high viabilities of the HUVECs incubated with lymph collected from T/SS animals. Therefore, lymph collected from single animals at 1-, 2 and 3-h postshock was pooled for subsequent separation as “T/HS” lymph. Likewise, the corresponding lymph fractions from sham animals were pooled for subsequent separation as “T/SS” lymph.
Reversed-phase SPE, which separates components based upon their hydrophobic properties, and not their molecular weights, was used to fractionate lymph samples with the objective of eluting factors toxic to endothelial cells into discreet fractions. Lymph samples applied to each column were eluted in a stepwise gradient using increasing concentrations of an organic solvent (ACN). In this elution scheme, the more hydrophilic components elute in low ACN concentrations and the more hydrophobic (lipophilic) components elute in higher ACN concentrations. Lymph collected from six T/HS and six T/SS animals was separated in this manner and the collected fractions were tested for endothelial cell toxicity. The collected fractions were dried by vacuum centrifugation and resuspended in 100 μL of dH2O. These fractions (10 μL) were tested for endothelial cell toxicity (10% [v/v] final concentration). As shown in Figure 1, only fractions from separations of T/HS lymph yielded maximal HUVEC toxicity, which shows that the SPE methodology does not introduce toxic factors. Viabilities that were comparable with unseparated T/HS lymph (<20%) were detected in all (6/6) 40% ACN fractions and most (5/6) 50% ACN fractions. These toxic fractions were further analyzed in an attempt to characterize their components.
Because protein was consistently detected in the T/HS SPE 40%, this fraction was further analyzed to identify the protein components. Protein components from the toxic fractions were determined by PAGE analysis. Only the amount of material that yielded HUVEC toxicity (10 μL) was used for analysis. This was done as a second form of screening based upon the rationale that only proteins present in detectable amounts in a “toxic” aliquot were physiologically relevant. PAGE analysis consistently detected protein in the T/HS 40% ACN fraction (data not shown). No protein bands were detected in the SPE 50% ACN fraction even when the entire eluted sample was used for electrophoresis (data not shown). In addition, bands of lower molecular weight were not detected on a 4% to 20% gradient gel run under native conditions nor did subsequent silver staining of this gel reveal any additional bands to those detected by Coomassie stain. In fact, total protein measurements of the T/HS SPE fractions (Table 2) show that most of the protein elutes in the initial wash with dH2O (i.e., 0% ACN), which shows they are very hydrophilic. These data show that the SPE separation method immediately excludes a majority of the proteins in lymph as candidate toxic factors. Only a small fraction (approximately 0.10 mg) of the total protein applied to the column (approximately 10 mg) is present in the fraction having detectable toxicity to HUVECs (SPE 40%). In addition to the PAGE analysis and protein measurements of the SPE 50% fraction, the 50% SPE T/HS lymph fractions were analyzed by liquid chromatography (LC)/mass spectrometry (MS). Five peaks were detected by LC. Mass spectral analysis of the material composing these peaks showed them to be of low molecular weight (303-561 Da). Each of these entities was then subjected to fragmentation analysis (MS/MS).This analysis showed that the fragments produced had mass repeats of 14 Da, which is indicative of the sequential breaking of CH2 groups found in fatty acid chains. Additionally, the collision energy needed to fragment the material in the SPE 50% ACN fraction was higher than that needed to fragment a peptide bond but consistent with breakage of the C-C bond in fatty acid carbon chains. Therefore, these data suggest that the HUVEC cytotoxic factor(s) in the 50% SPE shock lymph fraction are lipids. These analyses therefore suggest that two separate factors, a protein (SPE 40%) and a lipid (SPE 50%), might be responsible for T/HS lymph-induced HUVEC toxicity. Further analysis focused on the T/HS SPE 40% ACN fraction to investigate the possibility of a toxic protein factor.
Analysis of 40% SPE fraction
The SPE 40% fractions from T/HS and T/SS separations were compared by native PAGE to identify possible differences in protein band migrations. Native conditions were used because protein migration is not based solely on molecular weight but also by charge. This is a more powerful method of analysis because small weight modifications to proteins may impart notable charge differences in the tertiary state of proteins. Additionally, this method allows for the visualization of protein aggregates, which are destroyed under denaturing and reducing conditions. Figure 2 shows the analysis of a T/HS and a T/SS SPE 40% fraction by PAGE under native conditions. As can be seen in this figure, there is a clear difference in the appearance of bands that comigrate in the T/HS and T/SS groups (compare bands T/HS-1 with T/SS-1). To determine if T/HS-1 or any of the other T/HS bands present in the gel were responsible for the detected endothelial cell toxicity, all visualized bands depicted by arrows (Fig. 2), were excised from the gel, electroeluted, and tested for HUVEC toxicity. No endothelial cell toxicity was detected with any of the T/SS bands or with T/HS bands 2 and 4 (HUVEC viability 88%-100%). However, the viability of HUVECs incubated with T/HS lymph bands 1 and 3 was reduced (30% and 36% viability, respectively).
Additional toxicity testing was done using bands electroeluted from other toxic T/HS SPE 40% fractions. Bands that correspond to the T/HS band 3 have been excised and electroeluted from gels of two additional toxic T/HS SPE 40% samples and both yielded reduced HUVEC viabilities (35% and 14%). The bands from these same samples that corresponded to T/HS-1 band did not reduce HUVEC viabilities (77% and 79%).
To further identify the bands, additional aliquots of the T/HS and T/SS samples were electrophoresed on a duplicate gel under conditions identical to the first analysis. Bands T/HS-1, -3, and T/SS-1 were excised, subjected to in-gel trypsin digestion, and the fragments were analyzed by MS. All three bands were identified as RSA (T/HS-1 and T/SS-1) or RSA multimers (T/HS-3). The presence of RSA multimers is also supported by PAGE analysis of SPE 40% samples under denaturing and reducing conditions. When disulfide linkages are disrupted under reducing conditions, the high molecular weight bands observed under native conditions are no longer present and the monomer RSA band is the most predominant band (data not shown). The appearance (i.e., smearing) of the T/HS-1 band is indicative of protein glycosylation and therefore T/HS-1 and T/SS-1 bands were subjected to carbohydrate analysis. The presence of glucose was detected for T/HS-1 and T/SS-1 bands. Because it was not determined that an equal amount of each eluted protein was sent for analysis, a comparison between the relative amounts of glucose between the T/HS and T/SS samples could not be made. However, because proteins migrate by size and charge under native conditions, the faster migration of the T/HS-1 band shows this RSA species is more negatively charged, and therefore may be more glycosylated than its T/SS-1 counterpart. These results indicate that T/HS lymph contains a glycosylated form of RSA (mod-RSA), which may contribute to T/HS lymph-induced endothelial cell toxicity.
Although RSA is not glycosylated before secretion (18), it has been shown that albumins can be glycosylated by a nonenzymatic process and has been shown to be present in rat plasma at approximately 7% (16). To test whether this type of albumin glycosylation could be a factor in the observed HUVEC toxicity, RSA was glycosylated in vitro (gly-RSA) and was tested for HUVEC toxicity. No HUVEC toxicity was detected in two separate measurements with RSA or gly-RSA at final concentrations of 5, 4, 3, 2, or 1 mg/mL (data not shown), which are protein concentrations far in excess of that for the toxic 40% ACN fraction. It is also clear from PAGE analysis (Fig. 3) that the migration of the gly-RSA is different from the mod-RSA species present in the T/HS SPE toxic fractions. In addition, each of the major bands appearing in Figure 3 were electroeluted from the gel and analyzed by MS. The measured masses for each species confirm that the RSA species is different from native RSA or gly-RSA (Fig. 3). Therefore, the mod-RSA present in T/HS lymph may contain a modification in addition to glycosylation.
Two-dimensional gel electrophoresis was also used to compare a toxic T/HS and nontoxic T/SS SPE fractions using the sensitive dye SYPRO Ruby. Unlike the analysis above where only 10 μL of each fraction was analyzed, an increased amount of material (200 μg of protein; 15-25 μL) was analyzed. Although three proteins, an albumin fragment, actin, and vitamin D-binding protein, were only detected in the T/HS fractions, these differences were not detected in a two-dimensional comparison of unfractionated T/HS and T/SS lymph (data not shown). In addition, the vitamin D-binding protein was found in T/HS and T/SS in high abundance in another SPE fraction that was not toxic to endothelial cells.
Because PAGE analysis indicates that the mod-RSA species is more negatively charged than native albumin, an IEX was used as an alternate separation technique that should separate the mod-RSA from the native RSA and confirm the SPE results. Lymph specimens from four rats subjected to T/HS and three rats subjected to T/SS were separated by anion exchange chromatography. The elution profiles of all T/HS and T/SS lymph samples were similar. A representative profile from a sham and a shock lymph anion separation is shown in Figure 4. Comparisons of the two profiles show that the number and distribution of the peaks of each elution profile are identical for the sham and shock lymph separations. Each profile contains six peaks (labeled A-F), and the highest absorbance readings occur in almost identical fractions.
Despite the similarities, there are notable differences between the elution profiles from the two lymph groups that were consistent in all samples tested. First, the E peak was the most prominent peak in the T/HS lymph elution profile, which is in sharp contrast to its level in the T/SS lymph profile. This result suggests there is a marked increase in or a modification of a pre-existing T/SS lymph factor(s) or that a new factor not contained in T/SS lymph is generated under shock conditions. A second difference between the T/HS and T/SS lymph samples is the increased prominence of the C peak in the T/HS lymph profile when compared with the T/SS lymph profile. Although qualitative in nature, these profiles show the existence of consistent and clear differences between lymph collected from T/SS and T/HS animals.
The fractions comprising each sham and shock anion peak were pooled and tested for HUVEC toxicity to identify the peaks in which the toxic factors were eluting. These measurements showed that significant HUVEC toxicity, as reflected by low viabilities, was detected in the fractions associated with the D and E peaks of the T/HS lymph samples, but not the T/SS lymph samples (Fig. 5). The prominence of the E peak in the shock lymph profile clearly indicates that lymph-mediated endothelial cell toxicity is caused by a change in the lymph composition during shock. However, the similarity of the sham and shock lymph D peaks does not readily account for the detected difference in HUVEC toxicity.
Because of the detected HUVEC toxicity, the material eluting under the D and E peaks was further analyzed in an attempt to identify lymph components that contribute to toxicity and/or the peak absorbance differences observed in the elution profiles. Despite the marked “E” peak OD280 values in the anion profiles of T/HS lymph separation, very little protein was found in these fractions (data not shown). Although absorbance at 280 nm is used as a guideline for protein presence, it is based on the assumption that most proteins contain aromatic amino acids (e.g., tryptophan) that are responsible for the detected absorbance at this wavelength. The peptide bond itself absorbs at a wavelength of 215 to 220 nm, along with many other substances, making it unreliable for protein measurements unless it is known a very pure protein is present. Therefore, the high absorbance of the E peak may reflect the presence of some other species in lymph that happens to absorb at 280 nm. Similar to analysis of the SPE 50% fraction, only masses <800 Da were detected (data not shown).
A native PAGE analysis was performed to analyze the proteins present in the anion exchange “D” peak fractions from T/HS and T/SS separations and compared with native rat albumin and a toxic T/HS SPE fraction. As can be clearly seen in Figure 6, the albumin species in the anion T/HS D peak comigrates with the albumin species in the T/HS SPE fraction but migrates differently from the native albumin and that in the T/SS D peak. Because many protein bands were detected by gel electrophoresis in the “D” peak fractions, the NaCl gradient was therefore modified in an attempt to further separate the proteins that eluted in “D” peak fractions and therefore reduce the number of proteins to be analyzed in this fraction having detectable HUVEC toxicity.
The fractions comprising the toxic peak from two T/HS and the corresponding nontoxic peaks from two T/SS lymph separations using this modified anion separation were analyzed by gel electrophoresis (Fig. 7). Two bands were detected in the toxic peak fractions obtained from T/HS separations that were not present in the T/SS separations (arrows). Both bands were excised, subjected to trypsin digestion, and then analyzed by mass spectrometry. The lower band was identified as RSA. The upper band could not be identified with any certainty. Therefore, these anion exchange separation results are consistent with the SPE separation results. Both separation methods indicate that a lipid component (SPE 50% and anion exchange E peak) and a protein component (SPE 40% and anion exchange D peak) contribute to T/HS lymph-induced endothelial cell toxicity. Furthermore, rat albumin has been identified in both toxic protein factors from both separation methods. Characterization of the albumin species from the T/HS lymph SPE isolations show that this rat albumin species is modified from the native species as shown by the difference in gel migration under native conditions, by the presence of glucose by carbohydrate analysis, and by mass difference by MS. Although the RSA species in the toxic anion peak (Fig. 6) was not further characterized to confirm the presence of glucose, control experiments with native and glycosylated rat albumin support the presence of a glycosylated form of albumin in this fraction. In these control experiments (data not shown), glycosylated rat albumin prepared in vitro, eluted in the fractions that correlate to the anion D peak (T/HS toxic fraction, Fig. 4), whereas commercially purchased native RSA eluted in the fractions that correlate to the anion C peak (T/HS nontoxic fraction, Fig. 4).
Although the relationship between trauma-hemorrhage and acute pulmonary dysfunction as well as systemic inflammation has been recognized for many years, the pathogenesis of T/HS-induced pulmonary microvascular dysfunction is not well understood. Nevertheless, it was clear that hemorrhagic shock and major trauma are associated with intestinal ischemia, loss of gut barrier function, and subsequent bacterial translocation (20-22). In fact, in the early 1990s, human studies first began documenting that gut permeability was increased shortly after major trauma with some of the more recent studies showing a direct correlation between the extent of the trauma and the magnitude of the increase in gut permeability (23-26). Our work and the work of others showing that factors contained in mesenteric lymph, rather than the portal blood, are an important link between gut ischemia and distant organ dysfunction (27) and provide a unique opportunity to isolate and identify the factors involved in the pathogenesis of the gut-induced systemic inflammatory response, acute lung injury, and MODS. Because the endothelium plays a central role in these trauma-hemorrhage-induced systemic responses, we have focused our initial efforts on isolating the factors in T/HS mesenteric lymph that are toxic to endothelial cells. Our earlier studies have shown that this toxic effect is due to the humoral, but not the cellular, components of T/HS lymph and that it is not related to bacteria, endotoxin, or cytokines contained within the lymph (13, 14). The current study has extended these findings by documenting that T/HS lymph collected over the first 3 h after the end of the 90-min T/HS period contains protein and lipid factors that manifest endothelial cell toxicity.
We have shown by two independent separation methods that fractions obtained from the separation of T/HS lymph having detectable endothelial cell toxicity contain mod-RSA. Analysis by PAGE shows the RSA protein in the toxic T/HS SPE 40% fractions has different electrophoretic properties than those of commercially purchased native RSA or the corresponding fractions from T/SS lymph separations. Also, the detection of glucose by carbohydrate analysis of the mod-RSA protein indicates that the protein is glycosylated. In addition, reduced HUVEC viability detected after incubation with an electroeluted mod-RSA protein band implicates mod-RSA in HUVEC toxicity.
It is unlikely that the mod-RSA species isolated from T/HS lymph is an artifact formed during the isolation process or during storage. First, the mod-RSA found in separations of T/HS lymph is not present in T/SS lymph separations. Second, PAGE analysis of fractions obtained from a SPE of 20 mg of commercially purchased RSA showed that RSA eluted from the SPE column in the initial dH2O wash and no bands consistent with a modified form of RSA were detected in any of the other SPE fractions (V. Kaiser, unpublished data). Third, T/HS lymph samples are stored at -80°C and do not increase in toxicity over time (13). Finally, the addition of exogenous bovine serum albumin (BSA) did not increase the toxicity of the lymph, which would be expected if albumin were acting as a substrate for a modifying enzyme already present in lymph. In fact, addition of unmodified BSA to HUVECs exposed to shock lymph is protective against T/HS lymph-induced endothelial toxicity (28). Therefore, we consider this mod-RSA to be a real component of shock lymph.
Electrophoretic and carbohydrate analysis indicate that the mod-RSA from T/HS lymph is glycosylated despite studies showing that RSA is not modified before secretion (18). Glycosylation of albumin can result from a nonenzymatic process, and this modified form of albumin is thought to contribute to the pathology of certain disease states, such as diabetes (19). Thus, we glycosylated RSA (gly-RSA) in vitro to investigate its toxic properties. However, this form of gly-RSA had no HUVEC toxicity and its electrophoretic properties and mass did not match that of the mod-RSA isolated from rat shock lymph. Additionally, because the mass difference obtained between native RSA and mod-RSA (500 Da) is not evenly divisible by the weight of one or more glucose modifications (162Da/glucose), the data suggest other modifications are also present. Comparisons of the enzymatic digests between RSA and mod-RSA by MS will be needed to further identify the type of modification and the amino acid residue(s) involved in this modification.
Because T/HS lymph-induced endothelial activity is detected shortly after shock (1-3 h) and protein modification is a rapid way to create proteins with bioactivity, the notion of protein modification as a mechanism of converting nontoxic to toxic molecules is consistent with the rapid appearance of toxicity. Because albumin is the most abundant protein present in lymph and plasma, it is readily available as a substrate for modification. In fact, a recent study shows the modification of human serum albumin (HSA) by neutrophil-derived reactive oxygen intermediates and the ability of this modified HSA to then act as a proinflammatory mediator (29). In addition, modification of albumin is also a marker of myocardial ischemia (30) and is thought to result from reactive oxygen species produced from ischemic tissue. Alternatively, because serum albumin is a transport molecule, factors bound to the albumin species present in T/HS lymph may be directly responsible for endothelial cell toxicity. However, the elution of the T/HS albumin species in 40% ACN would favor the dissociation of bound material. More rigorous experiments are needed to fully address this issue.
The mechanisms by which the toxic factors in T/HS lymph are generated in the ischemic gut remain to be determined; however, there is increasing evidence to implicate pancreatic enzymes as an important element in this process. This notion is most recently supported by a series of studies from the laboratory of Schmid-Schonbein showing that pancreatic factors are involved in the generation of gut-derived factors that contribute to leukocyte activation, hemodynamic instability, and organ injury in a superior mesenteric artery occlusion shock model (31-33). Our recent studies documenting that the administration of pancreatic protease inhibitors into the lumen of the gut, but not systemically, limits T/HS-induced lung injury and systemic inflammation (34), as does pancreatic duct ligation (35), supports this notion that pancreatic enzymes in the lumen of gut act on the ischemic intestine to generate biologically active factors. The Schmid-Schonbein investigative group has also recently found that the pancreas induces biologically active aqueous and lipid factors, although they have not identified what these factors are (36).
Identification of the toxic factors in lymph will make it possible to study their production, regulation, entry into the lymph, and effects on different tissues. Elucidation of each of these mechanisms creates several possible levels of intervention. Inhibition of the formation of the toxic factor may offer one possible protective mechanism. Alternatively, antibodies that recognize specific modifications on toxic proteins would inhibit their mode of action. Additionally, inhibiting the toxic factors action on target tissues is another level of intervention that might limit the damage produced by these factors to prevent MODS.
The identification of the lipid-based components is more complex because multiple factors elute in a single fraction. We have attempted to further separate these factors into individual fractions. However, no toxicity was detected with any single fraction, although toxicity could be regained by mixing several of the fractions together (V. Kaiser, unpublished data). This result suggests that more than one of these lipid factors is needed for toxicity. Therefore, modifications to our purification scheme are necessary to further investigate the lipid species that are responsible for endothelial cell toxicity. However, the observation that T/HS lymph contains biologically active lipid factors is supported by the studies showing that lipid factors are involved in T/HS-induced activation of neutrophils (37).
In conclusion, our data implicates protein and lipid factors in T/HS lymph-induced endothelial cell toxicity. Furthermore, one of these factors may be a modified form of RSA or a multimer of RSA or modified RSA. Further investigations to determine the RSA modification, the identity of the lipid factors, as well as the genesis of these factors and their mechanisms of toxicity are needed to understand and attenuate shock lymph-mediated inflammatory and cellular injury that leads to MODS.
The authors thank Dr. Yan for two-dimensional gel electrophoresis and Dr. G. Diamond for critical reading of the manuscript.
1. Deitch EA: Multiple organ failure: pathophysiology and potential future therapy. Ann Surg 216:117-134, 1992.
2. Deitch EA, Sambol JT: The gut origin hypothesis of MODS. In Deitch EA, Vincent J-L, Windsor A (eds.): Sepsis and Multiple Organ Dysfunction: A Multidisciplinary Approach. Philadelphia, PA: W.B. Saunders Company, 2002, pp 105-116.
3. Deitch EA, Forsythe R, Anjaria D, Livingston DH, Lu Q, Xu DZ, Redl H: The role of lymph factors in lung injury, bone marrow suppression and endothelial cell dysfunction in a primate model of trauma-hemorrhagic shock. Shock 22:221-228, 2004.
4. Magnotti LJ, Upperman JS, Xu DZ, Lu Q, Deitch EA: Gut-derived mesenteric lymph but not portal blood increases endothelial cell permeability and potentiates lung injury following hemorrhagic shock. Ann Surg 228:518-527, 1998.
5. Xu DZ, Lu Q, Adams CA, Issekutz AC, Deitch EA: Trauma-associated shock-induced upregulation of endothelial cell adhesion molecules is blunted by mesenteric lymph duct ligation. Crit Care Med 32:760-765, 2004.
6. Adams JM, Hauser CJ, Adams CA, Xu DZ, Livingston DH, Deitch EA: Entry of gut lymph into the circulation primes rat neutrophil respiratory burst in hemorrhagic shock. Crit Care Med 29:2194-2198, 2001.
7. Zallen G, Moore EE, Johnson JL, Tamura DY, Ciesla DJ, Silliman CC: Posthemorrhagic shock mesenteric lymph primes circulating neutrophils and provokes lung injury. J Surg Res 83:83-88, 1999.
8. Zaets SB, Berezina TL, Caruso J, Xu DZ, Deitch EA, Machiedo GW: Mesenteric lymph duct ligation prevents shock-induced red blood cell deformability and shape changes. J Surg Res 109:51-56, 2003.
9. Anjaria DJ, Rameshwar P, Deitch EA, Xu DZ, Adams CA, Forsythe RM, Sambol JT, Hauser CJ, Livingston DH: Hematopoietic failure after hemorrhagic shock is mediated partially through mesenteric lymph. Crit Care Med 29:1780-1785, 2001.
10. Lu Q, Xu DZ, Davidson MT, Hasko G, Deitch EA: Hemorrhagic shock induces endothelial cell apoptosis, which is mediated by factors contained in mesenteric lymph. Crit Care Med 32:2464-2470, 2004.
11. Caruso JM, Feketeova E, Dayal SD, Hauser CJ, Deitch EA: Factors in intestinal lymph after shock increase neutrophil adhesion molecule expression and pulmonary leukosequestration. J Trauma 55:727-733, 2003.
12. Dayal SD, Hasko G, Lu Q, Xu DZ, Caruso JM, Sambol JT, Deitch EA: Trauma-hemorrhagic shock mesenteric lymph upregulates adhesion molecule expression and IL-6 production in human umbilical vein endothelial cells. Shock 17:491-495, 2002.
13. Adams CA, Xu DZ, Lu Q, Deitch EA: Factors larger than 100 kD in post-hemorrhagic shock mesenteric lymph are toxic for endothelial cells. Surgery 129:351-362, 2001.
14. Davidson MT, Deitch EA, Osband A, Feketeova E, Nemeth ZH, Hasko G, Xu DZ: A study of the biologic activity of trauma-hemorrhagic shock mesenteric lymph over time and the relative role of cytokines. Surgery 136(1):32-41, 2004.
15. Vane JR, Anggard EE, Botting RM: Regulatory functions of the vascular endothelium. N Engl J Med 323:27-36, 1990.
16. Day JF, Thromburg RW, Thorpe SR, Baynes JW: Nonenzymatic glucosylation of rat albumin. Studies in vitro and in vivo. J Biol Chem 254:9394-9400, 1979.
17. Deitch EA, Adams CA, Lu Q, Xu DZ: Mesenteric lymph from rats subjected to trauma-hemorrhagic shock are injurious to rat pulmonary microvascular endothelial cells as well as human umbilical vein endothelial cells. Shock 16:290-293, 2001.
18. Peters T Jr, Reed RG: The biosynthesis of rat serum albumin. J Biol Chem 256:3156-3163, 1980.
19. Rumble JR, Cooper ME, Soulis T, Cox Z, Wu L, Youssef S, Jack S, Jenums G, Gilbert RS: Vascular hypertrophy in experimental diabetes. Role of advanced glycation and products. J Clin Invest 99:11016-11027, 1997.
20. Deitch EA, Bridges W, Baker J, Ma J, Grisham M, Granger DN, Specian R, Berg R: Hemorrhagic shock-induced bacterial translocation is reduced by xanthine oxidase inhibition or inactivation. Surgery 104:191-198, 1988.
21. Ramzy PI, Wolf SE, Irtun O, Hart DW, Thompson JC, Herndon DN: Gut epithelial apoptosis after severe burn: effects of gut hypoperfusion. J Am Coll Surg 190:281-287, 2000.
22. Biffl WL, Moore EE: Splanchnic ischemia/reperfusion and multiple organ failure. Br J Anaesthesia 77:59-70, 1996.
23. Faries PL, Simon RJ, Martella AT: Intestinal permeability correlates with severity of injury in trauma patients. J Trauma 44:1031-1036, 1998.
24. Langkamp-Henken B, Donovan TB, Pare LM: Increased intestinal permeability following blunt and penetrating trauma. Crit Care Med 23:660-664, 1995.
25. Deitch EA: Intestinal permeability is increased in burn patients shortly after injury. Surgery 107:411-416, 1990.
26. Ryan CM, Yarmush ML, Burke JF, Tompkins RG: Increased gut permeability early after burns correlates with the extent of burn injury. Crit Care Med 20:1508-1512, 1992.
27. Deitch EA: Role of the gut lymphatic system in multiple organ failure. Curr Opin Crit Care 7:92-98, 2001.
28. Osband AJ, Deitch EA, Hauser CJ, Lu Q, Zaets S, Berezina T, Macheido GW, Rajwani KK, Xu DZ: Albumin protects against gut-induced lung injury in vitro and in vivo. Ann Surg 240:331-339, 2004.
29. Kormoczi GF, Wolfel UM, Rosenkranz AR, Horl WH, Oberbauer R, Zlabinger GJ: Serum proteins modified by neutrophil-derived oxidants as mediators of neutrophil stimulation. J Immunol 167:451-460, 2001.
30. Bar-Or D, Lau E, Winkler JV: A novel assay for cobalt-albumin binding and its potential as a marker for myocardial ischemia: a preliminary report. J Emerg Med 19:311-315, 2000.
31. Kistler EB, Hugli TE, Schmid-Schonbein GW: The pancreas as a source of cardiovascular cell activaing factors. Microcirculation 7:183-192, 2000.
32. Mitsuoka H, Kistler EB, Schmid-Schonbein GW: Generation of in vivo activating factors in the ischemic intestine by pancreatic enzymes. Proc Natl Acad Sci USA 97:1772-1777, 2000.
33. Mitsuoka H, Kistler EB, Schmid-Schonbein GW: Protease inhibition in the intestinal lumen: attenuation of systemic inflammation and early indicators of multiple organ failure in shock. Shock 17:205-209, 2002.
34. Deitch EA, Shi HP, Lu Q, Feketeova E, Xu da Z: Serine proteases are involved in the pathogenesis of trauma-hemorrhagic shock-induced gut and lung injury. Shock 19:452-456, 2003.
35. Cohen DB, Magnotti LJ, Lu Q, Xu da Z, Berezina TL, Zaets SB, Alvarez C, Machiedo G, Deitch EA: Pancreatic duct ligation reduces lung injury following trauma and hemorrhagic shock. Ann Surg 240:885-891, 2004.
36. Kramp WJ, Waldo S, Schmid-Schonbein GW, Hoyt D, Coimbra R, Hugli TE: Characterization of two classes of pancreatic shock factors: functional differences exhibited by hydrophilic and hydrophobic shock factors. Shock 20:356-362, 2003.
37. Gonzalez RJ, Moore EE, Biffl WL, Ciesla DJ, Silliman CC: The lipid fraction of post-hemorrhagic shock mesenteric lymph inhibits neutrophil apoptosis and enhances cytotoxic potential. Shock 14:404-408, 2000.