Development and Characterization of a Rat Model of Nonpenetrating Liver Trauma

Abstract

The aim of this study was to develop and characterize a rodent model of liver trauma suitable for preclinical evaluation of new treatments and diagnostic technologies. Liver trauma was created by dropping a steel cylinder through a plastic tube onto the abdomen of supine, anesthetized rats. Internal hemorrhage in the absence of liver trauma was simulated by instilling fresh blood into the peritoneum. Platelet counts were elevated significantly after liver trauma but not simulated hemorrhage. Liver trauma and simulated internal hemorrhage both increased blood levels of the factor growth-regulated oncogene–Kupffer cell. Transcription of plasminogen activator inhibitor 1, heat shock protein 70, and suppressor of cytokine syntheses 3 was increased 77-, 22-, and 27-fold, respectively, 2 h after liver trauma but was unaltered by simulated internal hemorrhage. Levels returned to pretrauma levels by 24 h after trauma. Transcript levels for hypoxia-inducible transcription factor 1α were increased 2.8-fold at 24 h but not 2 h after trauma and were not affected by simulated hemorrhage. Production of heat shock protein 70 and inducible nitric oxide synthase in liver was limited to a penumbra surrounding areas of necrosis associated with trauma. The rat model described produces lesions similar to those that occur in humans after blunt trauma.

Blunt liver trauma is an important source of morbidity and mortality in the United States. Nonpenetrating trauma to the abdomen related to automobile accidents, falls, and industrial accidents can lead to the breakage of both small and large vessels within the liver.¹⁰ Blunt liver trauma also affects soldiers injured during wartime operations.⁶

Diagnosing and properly treating patients with hepatic injury is complicated. Traumatic liver injuries are evaluated in stable patients through the use of contrast-enhanced ultrasonography and focused assessment with sonography for trauma (FAST).^{6,10,19,20,22} Although 71% to 89% of all blunt liver injuries can be assessed nonsurgically with current diagnostic practices,²² patients who are hemodynamically unstable or have peritonitis must undergo either diagnostic peritoneal lavage or emergency laparotomy before the full extent of hepatic injury can be diagnosed.¹⁸ Injuries that must be diagnosed at locations that lack imaging technology present a greater challenge.

Progress in this area has been made principally through case series and case reports. Animal models are an ideal alternative for the study of liver injury and a conduit for the development of assessment and treatment practices. Swine have been the animal model of choice. To induce nonpenetrating liver trauma in swine, researchers have used diverse mechanical methods including firing a projectile from a modified nail gun or crossbow and striking with a blunt object such as a steel rod.²¹ Swine models of blunt, nonpenetrating liver trauma have been characterized in terms of their visual similarity to human injury.²¹

A rodent model for the study of blunt liver trauma has never been developed and characterized. Liver trauma resulting from other types of injury has been studied in both swine and rats. For example, reproducible high-grade liver trauma can be induced by opening the peritoneum and crushing a lobe of the liver with forceps.^2,19 Indeed, this model has been used recently to evaluate various methods of hemostasis and resuscitation; however, little has been done to characterize liver injury arising from blunt trauma at the molecular level. A comparative study of rat and human liver anatomy found that the lobes of the rat liver are equivalent to the sectors of the human liver, although they do not resemble one another in appearance.¹² Our goal was to develop a reproducible rodent model of blunt, nonpenetrating liver trauma that does not require custom equipment (for example, modified nail gun, crossbow) yet creates injury similar to that of humans. Our results demonstrate a reasonably simple method that reproducibly creates blunt, nonpenetrating liver trauma with features similar to those associated with blunt trauma in humans.

Materials and Methods

Creating blunt, nonpenetrating liver trauma in rats.

Male Sprague–Dawley rats (150 total; age, 7 to 9 wk; weight, 275 to 325 g; Charles River, Portage, MI) were used. According to vendor health reports, study rats were free of Helicobacter hepaticus, Pseudomonas aeruginosa, Mycoplasma pulmonis, Klebsiella pneumonia, Streptobacillus moniliformis, Kilham rat virus, H1 virus, rat parvovirus, Hantavirus, Sendai virus, lymphocytic choriomeningitis virus, reovirus, helminths, Giardia spp., ectoparasites, and Spironucleus spp. Rats were quarantined for 10 d on arrival and acclimated to the study environment for at least 3 d. Random rats from the same shipment were euthanized during quarantine and underwent gross visual examination and histology by a pathologist associated with our institution. Rats were housed individually in microisolation caging with commercial bedding (Sani-Chip, PJ Forest Products, Saddle Brook, NJ) in a temperature and humidity (30% to 70%)-controlled environment and had ad libidum access to water and a standard diet (Formula Diet 5008, Purina Lab Diet, Gray Summit, MO).

Surgical depth anesthesia was induced by inhalation of isoflurane in O₂ (60%:40%) by using a rodent anesthesia chamber (IMPAC,⁶ VetEquip, Pleasanton, CA). Anesthetized rats were positioned underneath a hollow ¾-in. PVC tube 50 or 100 cm in length, so that the end of the tube was positioned at the abdomen directly below the zyphoid process and costal cartilage. A steel punch (weight, 73.6 g; length, 8 mm) with a flat surface was dropped through the tube onto the anesthetized subject animal. Rats then received buprenorphine (0.3 mg/kg SC) to reduce discomfort and were allowed to recover from anesthesia in their home cages. In addition, rats received ad libidum gelatin cubes (Jell-O, Kraft Foods, Northfield, IL) containing 0.2 mg/kg buprenorphine for additional pain relief; these cubes were left in cages until consumed. At 2 or 24 h after trauma, rats were anesthetized deeply with isoflurane and then euthanized through exsanguination by cardiocentesis. All animal work was approved by the Brooks City-Base Institutional Animal Use and Care Committee and conducted in accordance with principles stated in Guide for the Care and Use of Laboratory Animals.¹¹

Treatment groups.

Rats were allocated into 1 of 4 possible treatment groups, n = 11 (unless otherwise stated): 1) sham (anesthesia) control; 2) hemorrhage control; 3) liver trauma induced by dropping the punch from 50 cm; and 4) liver trauma induced by dropping from 100 cm. Rats were euthanized and evaluated at 2 and 24 h after treatment (control or trauma).

Liver injury can lead to hemorrhage into the peritoneal cavity. Preliminary experiments enabled us to estimate the maximal amount of blood lost into the peritoneum after liver injury induced through our method as 4 mL (data not shown). To separate effects due to hemorrhage into the peritoneal cavity, 4 mL blood from the tail vein was placed into tubes containing citric acid buffer (30 mM citric acid, 75 mM trisodium citrate, 136 mM glucose; pH 6.4) and injected into the peritoneum of the same animal; these rats were designated as ‘hemorrhage controls’ and were euthanized at either 2 or 24 h after baseline blood collection.

To assess whether anesthesia affected cytokine production or gene expression, some rats underwent anesthesia only and were termed sham (anesthesia) controls. These rats were not injured but were euthanized for blood collection at 2 or 24 h after anesthesia. Samples from these rats were not used for evaluation of blood cell counts, as that evaluation required baseline data (see Evaluation of blood).

Visual assessment of injury.

Immediately after euthanasia of rats, laparotomy was performed and the liver exposed. The length and number of lacerations, number of lobes affected, lobe detachments, discoloration, and the presence of clots were noted. Based on these observations, a trauma score (1 to 4) was assigned by a member of the research team, who was blinded to both sample number and grouping. The trauma scale (Figure 1) was adapted from that previously published by the American Association for the Surgery of Trauma.^15,21

Figure 1.

Rat liver injury scale. ^aWhen accompanied by hematoma or discoloration of the same or a higher score, the overall injury score increases by 1. ^bVaried from white to light pink. Possibly an indication of hypoxia. ^cMust be present in injuries receiving scores of 3 or 4.

Evaluation of blood.

Samples of blood from cardiocentesis at euthanasia were collected in 5-mL nontreated glass vacuum-phlebotomy tubes, and serum was recovered for cytokine analysis. Blood was collected at baseline (0 h) and at euthanasia (2 or 24 h after liver trauma) and allowed to clot at room temperature. Serum was collected by centrifugation at 1400 × g for 15 min and stored at −80 °C until analysis. In addition, 1-mL samples of whole blood were collected in EDTA-containing vacuum-phlebotomy tubes and used for cell-count analysis.

Hematology was performed automatically (Cell-Dyn 3700, Abbott Diagnosis, Santa Clara, CA) and adjusted for rat cells. Neutrophils, monocytes, lymphocytes, RBC, and platelets were counted. Prior to analysis of experimental samples, blood obtained from an untreated rat was evaluated to ensure that the machine was adjusted correctly.

Cytokine levels in serum were determined by using a commercially available bead-based system (Luminex system, BioRad, Hercules, CA). A rat-specific multiplex kit for evaluation of 24 cytokines and chemokines (RCYT-80K- PMX24, Linco Research, St Charles, MO) was used according to the manufacturer's instructions. Concentrations of cytokines were estimated by using the inverse-function 5-paramter logistic regression applied to cytokine standards of known concentration, utilizing BioPlex Manager, version 4.0 (BioRad). Samples were analyzed according to height of pin drop and time after injury: anesthesia control 24 h, n = 7; anesthesia control 2 h, n = 9; hemorrhage control 24 h, n = 9; hemorrhage control 2 h, n = 6; 100-cm drop 24 h, n = 10; 100-cm drop 2 h, n = 10; 50-cm drop 24 h, n = 11; and 50-cm drop 2 h, n = 10. All samples were evaluated in duplicate.

Gene expression analysis.

For gene expression experiments, livers were excised in toto, each placed into a small plastic bag containing 10 mL RNAlater (Ambion, Austin, TX), and kneaded by hand for 5 min to disperse and mix liver tissue. Suspensions were stored at 20 °C until analysis could be conducted. Samples were not pooled for PCR analysis. Individual rat liver suspensions were rinsed twice in ice-cold filtered PBS, RNA was isolated by using TriZol (Invitrogen, Carlsbad, CA), and cDNA was synthesized from 500 ng RNA (cDNA iScript kit, BioRad) by incubating for 5 min at 25 °C, 30 min at 42 °C, and 5 min at 85 °C and holding at 4 °C. All cDNAs were prepped with, 100 nM of forward and reverse primers and IQ SYBR Green SuperMix (BioRad), denatured for 3 min at 95 °C, and amplified through 40 cycles of 30 s at 95 °C, 35 s at 55 °C, 30 s at 72 °C, followed by 1 min at 95 °C and 1 min at 55 °C. Primer sequences and accession numbers were: β-actin (NM_031144), 3′ GAA ATC GTG CGT GAC ATT AAA GAG 5′ and 3′ GCG GCA GTG GCC ATC TC 5′; hypoxia-inducible transcription factor 1α (HIF1α; NM_024359), 3′ AAG AAA CCG CCT ATG ACG TG 5′ and 3′ CCA CCT CTT TTT GCA AGC AT 5′; heat shock protein 70 (HSP70; NM_212504), 3′ CAA GAA TGC GCT CGA GTC CTA 5′ and 3′ GGA GAT GAC CTC CTG GCA CTT 5′; inducible nitric oxide synthase (iNOS; NM_012611), 3′ GGA GAG ATT TTT CAC GAC ACC C5′ and 3′ CCA TGC ATA ATT TGG ACT TGC A5′; plasminogen activator inhibitor 1 (PAI1; NM_012620), 3′ AAC CCA GGC CGA CTT CA 5′ and 3′ CAT GCG GGC TGA GAC TAG AAT 5′; and suppressor of cytokine synthases 3 (SOCS3; NM_053565), 3′ CTG CGC CTC AAG ACC TTC A 5′ and 3′ CGG TTA CGG CAC TCC AGT AGA 5′. Relative fold differences were calculated as 2^–(ΔΔCt), where ΔΔCt represents the cycle difference corrected for β-actin and the anesthesia control (that corresponding to the time at of euthanasia of experimental rat, namely 2 or 24 h after liver trauma). Individual rat data were averaged for calculation of 2^–(ΔΔCt). Samples were analyzed according to time after injury (because height of drop did not significantly alter mRNA expression [data not shown]): hemorrhage control 24 h, n = 10; hemorrhage control 2 h, n = 11; 24 h after injury, n = 10; and 2 h after injury, n = 11.

Distribution of HSP70 and iNOS in traumatized liver.

Representative samples of liver tissue from each trauma grade were prepared for histologic analysis by cutting a 3-mm section containing a trauma lesion and adjacent normal-appearing tissue, fixing it in 10% neutral buffered formalin, and processing into 5-µm paraffin-embedded sections, which were mounted on microscope slides (Superfrost Plus, Fisher, San Jose, CA) for immunostaining or staining with hematoxylin and eosin.

Sections were deparaffinized, rehydrated, and subjected to epitope retrieval with sodium citrate (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) for 40 min at 95 °C. Sections then were washed in Tris-buffered saline containing Tween 20 (TBST; 50 mM Tris.HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20) for 5 min, blocked for 1 h in 10% appropriate serum (HSP70, normal rabbit serum; iNOS, normal goat serum; GIBCO, Carlsbad, CA), and blocked overnight at room temperature with the following antibodies in 10% serum: iNOS (1:1000, catalog no. 160862, Cayman Chemical, Ann Arbor, MI), hypoxia-inducible transcription factor 1α (HIF1α; 1:50, catalog no. sc10790, Santa Cruz Biotechnology, Santa Cruz, CA), HSP70 (1:50, catalog no. sc1060, Santa Cruz Biotechnology). Sections were rinsed 3 times for 5 min each in TBST, endogenous peroxidase activity was blocked by incubating in 3% H₂O₂ in TBST for 10 min, and sections again were washed 3 times for 5 min each in TBST. Biotinylated secondary IgG antibody (1:250 in 10% serum) was added for 30 min at room temperature, followed by peroxidase-conjugated streptavidin (1:600) at room temperature for 30 min. Enzymatic reactions were performed by using diaminobenzidine (BioRad). All sections were counterstained with methyl green (0.5% in 0.1 M sodium acetate, pH 4.2).

Statistical analysis.

One-way ANOVA, performed utilizing SYSTAT software, version 9 (Systat Software, Chicago, IL), was used to determine whether treatment groups were statistically different at a P value of less than 0.05. When ANOVA was significant, Tukey posthoc tests were used to determine which groups differed from others. Paired t tests on differences between 0 h and 2 or 24 h after trauma or sham were made in order to detect changes in levels of platelets, neutrophils, monocytes, RBC, and lymphocytes.

Results

Grade of liver trauma induced by ballistic injury.

Liver trauma induced in rats according to the method we described was scored at 2 and 24 h after injury. The wounds in rats varied widely in appearance but exhibited several of the features common to those in humans, including a ‘cracked’ surface, blood clots adhering to the surface of the liver, and abnormal coloration due to either loss of blood flow or hematoma. Lobes typically affected were the left lateral, median, and caudate, among which the left lateral lobe sustained the most injury (Figure 2). At 2 h, the trauma score (mean ± 1 SD; n = 11 per group) due to dropping a weight from 100 cm (2.909 ± 0.944) was significantly (P < 0.001) greater than that caused by dropping a weight from 50 cm (1.182 ± 1.168). At 24 h, the relative injury produced by dropping from 50 cm increased relative to 2 h. Dropping the weight from different heights did not lead to a significant difference in trauma score at 24 h (100 cm, 2.700 ± 1.059; 50 cm, 2.636 ± 0.924).

Figure 2.

Induced liver injury. (A) Grade 4 liver injury involving caudate (not shown) and left lobes, 24 h after trauma. Extent of injury includes 6 cm² of tissue discoloration, a 4-cm² hematoma, and a single, laceration (crack; length, 4 cm). (B) Grade 4 liver injury involving caudate (not shown) and left lobes, 24 h after trauma. Extent of injury includes 4 cm² of tissue discoloration, a 1.5-cm² hematoma, and 2 lacerations totaling 5 cm in length. (C) Grade 4 liver injury of the left lobe only, 2 h after trauma. Liver excised, with lobes spread apart for visualization. Extent of injury includes 3 cm² of tissue discoloration, a 2-cm² hematoma, and 3 lacerations totaling in 9 cm in length. (D) Normal, uninjured liver tissue from anesthesia control. CL, caudate lobe; LL, left lobe; ML, median lobe; RL, right lobe.

Alterations in the population of peripheral blood cells.

Baseline (0 h) samples were collected from all animals (n = 28), except anesthesia control groups. Samples were analyzed according to the following groups: 2 h after injury, n = 6; 24 h after injury, n = 9; 24 h hemorrhage controls (blood control), n = 7; data beyond the range of detection for the automatic analyzer were excluded from analysis. Liver trauma led to increased levels of monocytes and platelets 2 h after dropping the weight from 100 cm (Table 1). By 24 h after trauma, platelets returned to preinjury levels, but neutrophils and monocytes remained increased. In contrast, the injection of fresh blood into the peritoneal cavity significantly (P < 0.05) decreased numbers of platelets but not the abundance of leukocytes. Neither liver trauma nor intraperitoneal blood altered the numbers of basophils, eosinophils, or lymphocytes (data not shown). RBC levels at both 2 and 24 h after injury significantly (P < 0.05) decreased by both trauma and injection of blood into the peritoneal cavity.

Table 1.

Hematology after liver trauma in rats

	Baseline (n = 28)	2 h after trauma (n = 6)	24 h after trauma (n = 9)	24 h after hemorrhage (n = 7)
Neutrophils (×103 cells/µL)	2.0 ± 0.18	5.4 ± 1.2 (0.001)	8.4 ± 3.7 (0.001)	6.1 ± 7.6 (NS)
Monocytes (×103 cells/µL)	0.99 ± 0.07	1.22 ± 0.29 (NS)	1.52 ± 0.56 (0.004)	1.09 ± 0.63 (NS)
Platelets (×103 cells/µL)	686 ± 35	823 ± 191 (0.006)	577 ± 203 (NS)	431 ± 161 (0.017)
RBC (×106 cells/µL)	7.78 ± 0.05	6.63 ± 0.91 (0.006)	6.95 ± 0.67 (0.007)	6.60 ± 0.89 (0.026)

NS, not significant

Data are given as mean ± 1 SD of difference between baseline (0 h) and time after 100 cm-drop trauma or simulated internal hemorrhage. P value (paired t tests) of comparison with baseline value is shown in parentheses.

Expression of cytokines in liver after trauma.

A 24-component cytokine and chemokine array was used to determine whether serum levels of these factors were altered specifically by liver trauma. ANOVA identified only growth regulated oncogene–Kupffer cell (Gro-KC) as significantly (P < 0.05) responsive to liver injury and blood in the peritoneal cavity. Trauma due to dropping the projectile from 100 cm caused Gro-KC levels to increase 9-fold compared with those of the anesthesia control (Figure 3). However, injection of blood into the abdominal cavity increased Gro-KC expression more than 5-fold compared with the anesthesia control, suggesting that blood in the peritoneal cavity is an important signal leading to this response. By 24 h, the changes in Gro-KC had diminished markedly, although the circulating levels of Gro-KC were still significantly (P < 0.05) higher for rats in the 100-cm drop group, compared with rats receiving intraperitoneal injection of blood (simulated hemorrhage) or rats receiving anesthesia alone.

Figure 3.

Gro-KC in blood. Levels for blood and 100 and 50 cm trauma at 2 h are increased relative to anesthesia control but do not differ from one another. Levels at 24 h are elevated after 100 cm trauma, but 50 cm trauma and blood are similar to the anesthesia control.

Changes in gene transcript levels in liver tissue after trauma.

The genes encoding PAI1, HSP70, HIF1α, and SOCS3 all showed increased levels of mRNA transcript after blunt, nonpenetrating liver trauma in rats. HIF1α transcripts were increased at 24 h but not 2 h after trauma compared with those of anesthesia-control rats. and was not affected by injection of blood (Figure 4). PAI1 transcript levels increased 77-fold by 2 h after trauma and subsequently declined from this peak to just 12-fold above the initial level by 24 h after injury; injection of blood into the peritoneal cavity had no effect on the transcript levels of PAI1. Transcript levels for SOCS3 were increased 26-fold and 7-fold after 2 h for rats receiving blunt injury and injection of blood, respectively; these levels returned to baseline within 24 h after injection of blood, whereas transcript levels remained elevated at 5-fold above baseline 24 h after blunt liver injury. Transcript levels of HSP70 in liver increased 22-fold 2 h after blunt injury and then declined to baseline by 24 h. Injection of blood into the peritoneal cavity had no effect on the mRNA levels of HSP70.

Figure 4.

Expression of selected genes in liver. *, Significantly (P < 0.05) greater than 1 relative to anesthesia controls.

Spatial distribution of iNOS and HSP70.

Immunohistochemistry was performed to assess the infiltration of iNOS, HIF1α, and HSP70 in injured liver tissue. Slides were evaluated visually for gross distribution of protein. iNOS and HSP70 both were localized at sites of injury, identified with hematoxylin and eosin staining, and borders between uninjured and injured tissue were clearly visible (Figure 5). There was no noteworthy distribution of HIF1α in any section of injured liver. Controls showed little or no iNOS, HIF1α, or HSP70.

Figure 5.

Production of HSP70 (left) and iNOS (right) in traumatized liver. Immunohistochemical staining of tissue harvested 24 h after blunt liver trauma. Enzymatic reaction was performed with diaminobenzidine; methyl green counterstain.

Discussion

This study established a reproducible method for inducing blunt, nonpenetrating trauma to rat liver in which the degree of injury can be varied reproducibly by changing the height through which the weight is dropped. Injury can be characterized visually by the number and length of lacerations, number of liver lobes affected, and changes in color at the site of the injury, as is done to evaluate human liver injury during laparotomy. In addition to this visual assessment of trauma, the injury to rat livers was verified by analysis of cytokine production, gene induction, and cellular analysis of blood.

Of all the cytokines examined, Gro-KC appears promising as a potential biomarker for liver trauma; however, further investigation is needed. Gro-KC is a rat homolog to human Groα, a chemokine belonging to the CXC family of chemokines. Gro proteins are secreted by Kupffer cells, the interstitial macrophages of the liver that act as its first line of defense.^13,17 Groα is believed to be responsible for the chemotaxis of neutrophils in response to injury. Gro-KC has been shown to increase chemotaxis in monocytes.¹⁷

Our observation that neutrophils and monocytes but not lymphocytes increase after liver trauma in rats is consistent with higher levels of Gro-KC in serum. Levels of platelets in blood were elevated at 2 h after liver trauma but returned to baseline levels by 24 h. In contrast, platelet count decreased after injection of blood into the peritoneum, as expected. The observed increase in platelet levels after trauma may be attributable to a phenomenon occurring in the liver. A possible explanation for this observation is increased expression of PAI1 in liver after trauma but not injection of blood into the peritoneal cavity. We observed significant increases (77- and 7-fold) increase in expression in PAI1 at 2 and 24 h, respectively, after trauma. PAI1 is a glycoprotein that inhibits both plasminogen activator and urokinase activator and contributes to decreased production of plasmin, which in turn causes inhibition of fibrinolysis. PAI1 can be found in low concentrations in plasma, is released from hepatocytes and endothelial cells, and may be released from platelets in an inactive form.^1,3,4,9,16 Decreased levels of plasmin support our observation of clot formation in the liver and increased mobilization of platelets from the bone marrow; however, proteins that support clot formation and platelet mobilization were not investigated in the present study.

Results reveal increased expression of HIF1α 24 h after liver injury. HIF1α is responsible for cell migration and differentiation, matrix metabolism, angiogenesis, and tissue repair. Increased HIF1α expression has been shown to have a direct relation to severity to fibrogenesis within the liver.⁵ PAI1 expression is increased 2 h after injury, and this increase may account for the increased HIF1α gene expression seen at 24 h, because PAI1 is responsible for increasing fibrogenesis. Hemorrhage controls showed no significant increase or decrease in HIF1α and PAI1 expression or platelet count.

Transcript levels for HSP70 were elevated at 2 h but returned to baseline by 24 h, and HSP70 transcript levels were not responsive to intraperitoneal injection of blood. HSP70 transcription is induced when cells respond to increases in Ca²⁺ generated by trauma or shock. HSP70 has been shown to protect hepatocytes against oxidative stress and apoptosis. In one study,¹⁴ HSP70 remained elevated for only 8 h or less after introduction of stress, a finding that may account for the return of HSP70 expression to baseline by 24 h after injury in the current study.

Hepatocytes and Kupffer cells release a number of cytokines and chemokines in response to injury and inflammation in the liver. In turn, SOCS3 and iNOS are released as regulators to this influx of activity. SOCS3 gene expression was significantly increased at 2 h and decreased by 24 h. SOCS3 acts as a negative regulator in response to inflammatory cytokine release, which is known to block cytokine signaling from signal transducers and activators (STATs).⁷ iNOS is responsible for the synthesis of nitric oxide, a free radical, which protects hepatocytes in acute inflammation and injury against apoptosis and mitochondrial damage and induces protective proteins. iNOS is expressed by Kupffer cells, endothelial cells, and hepatocytes and typically is expressed in resting cells. For this reason, gene expression data (data not shown) may not have been as informative as immunohistochemical evaluation of local infiltration of iNOS (Figure 5). Histologic evaluation supports this infiltration of iNOS-expressing cells found at the site of necrosis due to cessation of blood flow.⁸

In conclusion, we have established a reliable and reproducible method for inducing blunt force liver trauma in the rat. The method allows induction of trauma in a desired location, with severity of injury dependent on the distance that the projectile is dropped. Through investigation of cytokine and chemokine production, gene and protein expression, and immunohistochemical analysis, we have shown the utility of our model in detection of physiologic and inflammatory hallmarks associated with traumatic liver injury. Further studies of pathophysiologic and inflammatory mechanisms could be conducted using this rat model. The ease of managing and maintaining this model makes it ideal for study. Our rat model of liver trauma may be useful in future applications for both improving clinical diagnosis and treatment of traumatic liver injury and replacing the use of higher sentient species in liver trauma research.