Focal intra-colon cooling reduces organ injury and systemic inflammation after REBOA management of lethal hemorrhage in rats

Resuscitative endovascular balloon occlusion of the aorta (REBOA) is a lifesaving maneuver for the management of lethal torso hemorrhage. However, its prolonged use leads to distal organ ischemia–reperfusion injury (IRI) and systemic inflammatory response syndrome (SIRS). The objective of this study is to investigate the blood-based biomarkers of IRI and SIRS and the efficacy of direct intestinal cooling in the prevention of IRI and SIRS. A rat lethal hemorrhage model was produced by bleeding 50% of the total blood volume. A balloon catheter was inserted into the aorta for the implementation of REBOA. A novel TransRectal Intra-Colon (TRIC) device was placed in the descending colon and activated from 10 min after the bleeding to maintain the intra-colon temperature at 37 °C (TRIC37°C group) or 12 °C (TRIC12°C group) for 270 min. The upper body temperature was maintained at as close to 37 °C as possible in both groups. Blood samples were collected before hemorrhage and after REBOA. The organ injury biomarkers and inflammatory cytokines were evaluated by ELISA method. Blood based organ injury biomarkers (endotoxin, creatinine, AST, FABP1/L-FABP, cardiac troponin I, and FABP2/I-FABP) were all drastically increased in TRIC37°C group after REBOA. TRIC12°C significantly downregulated these increased organ injury biomarkers. Plasma levels of pro-inflammatory cytokines TNF-α, IL-1b, and IL-17F were also drastically increased in TRIC37°C group after REBOA. TRIC12°C significantly downregulated the pro-inflammatory cytokines. In contrast, TRIC12°C significantly upregulated the levels of anti-inflammatory cytokines IL-4 and IL-10 after REBOA. Amazingly, the mortality rate was 100% in TRIC37°C group whereas 0% in TRIC12°C group after REBOA. Directly cooling the intestine offered exceptional protection of the abdominal organs from IRI and SIRS, switched from a harmful pro-inflammatory to a reparative anti-inflammatory response, and mitigated mortality in the rat model of REBOA management of lethal hemorrhage.

Animal model. Figure 1 shows the flowchart and timeline of key events of this study. Male Sprague Dawley rats about 450 g were obtained from Charles River (Boston, USA). They were randomly pre-assigned identification numbers before the procedures. They were intubated, ventilated, and anesthetized throughout the procedure period. All survival surgery was performed by using aseptic techniques. Small incisions were made for inserting: (i) the inferior vena cava catheter line via the jugular vein in the neck for withdrawing blood; (ii) the right axillary arterial catheter for recording the proximal mean arterial pressure (MAP); (iii) the tail arterial catheter for recording the distal MAP; and (iv) a 2F Fogarty balloon catheter (Edwards Life Sciences, Irvine, USA) via the right femoral artery into the descending thoracic aorta for the implementation of REBOA. A TransRectal Intra-Colon (TRIC) temperature management device (see Supplement Materials and Methods) was inserted via the rectum into the descending colon into approximately 10 cm from the anus. A temperature probe was inserted into the descending colon to indicate the "intra-colon temperature. " Another temperature probe was inserted into the esophagus to indicate the upper body temperature. A final temperature probe was inserted into the bladder via the urinary tract to measure the bladder temperature to represent the abdominal cavity temperature. The rat hemorrhage model was produced by withdrawing about 50% of the estimated total blood volume (EBV) via the inferior vena cava catheter (Supplement Materials and Methods). Animals were bled over a 30 min period in an exponential manner based on the method described by Park et al. 14 . The 30-min blood withdrawing period was based on our pilot study which showed that withdrawing 50% of the EBV from the jugular-inferior vena cava catheter could take about 20-30 min.
The complete REBOA was confirmed by the difference between proximal and distal pressure. The 50% of EBV hemorrhage led to both the proximal and distal blood pressure to a 25-30 mmHg range (Supplemental Tables 3  and 4). Inflation of the REBOA balloon led to a sharp increase in the proximal MAP from 25 to > 90 mmHg while a sharp drop of the distal MAP from 25 to about 10-20 mmHg in all groups. The results are consistent with the pressure changes in other complete REBOA animal models such as the swine model 14 .
The TRIC device was activated at 10 min after blood withdrawal. Ice-cold water was circulated through the TRIC device to reduce the colonic wall temperature to 12ºC in the TRIC12ºC groups. Water warmed to 36-37ºC was circulated in a same manner to maintain a colonic wall temperature of 37ºC in the TRIC37ºC group. The upper body temperature was maintained via a warmed pad and an overhead lamp as close to 37 °C as possible in all groups. Following the 30-min hemorrhage period, the REBOA balloon was inflated to occlude the descending aorta for a period of either 25 or 30 min. At the end of the REBOA phase, the withdrawn blood was returned, and the balloon was deflated to start the post-REBOA reperfusion phase. During the post-REBOA phase, the TRIC device remained active for a total of 270 min.
In the TRIC12°C group, rewarming was initiated at 180 min after REBOA, which was done by gradually increasing the temperature of the circulating water in the TRIC catheter to achieve a rewarming rate 0.15 °C/min as measured by the esophageal temperature until a target range of 30 °C was reached. The TRIC device was then removed, anesthesia discontinued, and the rats were returned to their cages. In the TRIC37°C group, the TRIC device was activated in the same manner and for the same period as those in TRIC12°C group but circulated with 37 °C water during the entire period.
The intra-colon and esophageal temperature (Supplemental Table 1 and 2), as well as the proximal and distal MAPs (Supplemental Table 3 and 4) were continuously recorded throughout the experimental period by Powerlab 16-channel data acquisition system (ADInstruments, Mountain View, CA, USA). The arterial blood gas (ABG) (pH, pCO2, pO2, electrolytes, lactate, base deficit, bicarbonate, glucose, and potassium) were measured via an ABL90 FLEX blood gas analyzer (Radiometer, USA). The prothrombin time (PT) and international normalized ratio (INR) was measured via the Coag-Sense PT/INR Professional System according to manufacturer instructions (CoaguSense, Inc. Fremont, CA, USA).
Blood samples were collected before blood withdrawal (pre-hemorrhage phase) at 0 min and then at 90 and 180 min after the end of REBOA deployment (referred to as the post-REBOA phase or post-REBOA hereafter). All animals were included for analysis except for one rat that died due to an unanticipated anesthesia event and was later replaced with another rat. The animal inclusion and exclusion criteria were predefined and not determined on a post-hoc basis.

Experimental groups.
The key events of the study are shown in Fig. 1  Statistical analysis. Data are expressed as mean ± standard error of the mean (SEM). One-way ANOVA followed by Tukey post-hoc test for statistical analysis of values of tissue injury biomarkers and cytokines. The log-rank test for survival rate analysis. Chi-Square test for mortality rate. GraphPad Prism version 7.00 for Windows was used (GraphPad Software, La Jolla, California, USA). p < 0.05 was considered a statistically significant difference. Figure 2A shows that, relative to the pre-hemorrhage level (0), the plasma FABP2 levels were significantly increased by about 4.5-fold at 90 min and more than fivefold at 180 min post-REBOA in TRIC37°C rats. The focal intestinal cooling (25 min REBOA + TRIC12°C) www.nature.com/scientificreports/ significantly reduced the plasma FABP2 levels at both 90 (about 44.92%) and 180 (about 51.55%) min post-REBOA, compared to the same time points of the TRIC37°C group (25 min REBOA + TRIC37°C) ( Fig. 2A). An increase in REBOA duration by 5 min (30 min REBOA + TRIC12°C) insignificantly increased the plasma FABP2 level, compared to that of the 25 min REBOA + TRIC12°C group ( Fig. 2A). On the contrary, the infusion of 0 °C HEXTED (30 min REBOA + TRIC12°C + 0 °C Hex) significantly decreased the blood FABP2 level by about 32.36%, relative to that of the 30 min REBOA + TRIC12°C group.

TRIC12°C reduces REBOA-induced intestinal injury.
TRIC12°C Reduces REBOA-induced Sepsis. Figure 2B shows that, relative to the pre-hemorrhage level (0), the plasma endotoxin levels were significantly increased by about fourfold at 90 min and more than 4.5-fold at 180 min post-REBOA in TRIC37°C rats. The focal intestinal cooling (25 min REBOA + TRIC12°C) significantly reduced the plasma endotoxin levels at both 90 min and 180 min post-REBOA by about 50%, compared to the same time points of the TRIC37°C group (25 min REBOA + TRIC37°C) (Fig. 2B). An increase in REBOA duration by 5 min (30 min REBOA + TRIC12°C) significantly increased in the plasma endotoxin level by about 22.49% (p < 0.01), compared to that of the 25 min REBOA + TRIC12°C group (Fig. 2B). The infusion of icecold HEXTED (30 min REBOA + TRIC12°C + 0 °C Hex) further decreased the blood endotoxin level by about 31.45% (p < 0.01), relative to that of the 30 min REBOA + TRIC12°C group.
TRIC12°C reduces REBOA-induced organ injury. Figure 3 shows the plasma concentrations of the organ injury biomarkers. Figure 3A,B reveal that, relative to the pre-hemorrhage level (0 min), the plasma AST and FABP1 levels were significantly increased at 90 (fourfold) and 180 (sixfold) min post-REBOA in 25 min REBOA + TRIC37°C rats. In comparison, the AST and FABP1 levels in 25 min REBOA + TRIC12°C group were significantly reduced by > 50% at both 90 min and 180 min post REBOA relative to the same time points of the 25 min REBOA + TRIC37°C group. When the REBOA duration was increased from 25 to 30 min (30 min REBOA + TRIC12°C group), there was no significant additional increase in either the AST or FABP1. However, the infusion of ice-cold HEXTED (30 min REBOA + TRIC12°C + 0 °C Hex) further decreased the blood AST or FABP1 levels by about 36.8% (p < 0.01), relative to that of the 30 min REBOA + TRIC12°C group. Figure 3C shows the levels of creatinine. There was a 2.5-fold increase in the creatinine in the 25 min REBOA + TRIC37°C group at 90 min and a threefold increase at 180 min in the post-REBOA phase, relative to the creatinine level at the pre-hemorrhage level (0 min). When TRIC12°C was applied, there was non-significant decrease in the creatinine level by about 34.41% (p > 0.05) at 90-min and a remarkable decrease by about 44.8% (p < 0.01) at 180-min post-REBOA time point. The level of creatinine was significantly higher at 180 min post-REBOA in 30 min REBOA + TRIC12°C group than that in 25 min REBOA + TRIC12°C group. The level of creatinine was significantly lower at 180 min post-REBOA in 30 min REBOA + TRIC12°C + 0 °C Hex group than that in 30 min REBOA + TRIC12°C group. Figure 3D illustrates the concentration of serum cTnI. Relative to the pre-hemorrhage level (0 min) the levels of cTnI were significantly increased at both 90 and 180 min post-REBOA in all four groups. There were no statistical differences in the levels of cTnI among the four experimental groups, although the levels of cTnI in TRIC12°C groups tended to be lower than in TRIC37°C group.

TRIC12°C increases anti-inflammatory cytokines.
TRIC12°C reduces mortality. The survival rate was compared among the experimental groups used in this study (Fig. 6). All rats died within 24 h in the 25 min of REBOA + TRIC37°C group (Fig. 6A). In contrast, all rats survived until the endpoint in the 25 min of REBOA + TRIC12°C group (Fig. 6A). All rats died within 2 days in the 30 min of REBOA + TRIC12°C group, while all rats survived to the endpoint in the 30 min of REBOA + TRIC12°C + 0 °C Hex group (Fig. 6B). Therefore, the mortality rate was 100% for 25 min   (Fig. 7A,B). However, the PT/INR value tended to be persistently higher in TRIC12°C groups than in the TRIC37°C group, although the changes were insignificant (Fig. 7A,B). The insignificant changes in PT/INR values between cooling and non-cooling groups were likely because that the study was not powered to detect the significant difference.  Fig. 2

Discussion
This study demonstrates that after 25 min REBOA management of 50% bleeding, all normothermic rats died (25 min REBOA + TRIC37°C group). Strikingly, none of the rats undergoing cooling by TRIC (25 min REBOA + TRIC12°C group) died. This difference was reflected in the plasma levels of organ injury biomarkers and pro-inflammatory cytokines, which were all drastically elevated in the normothermic group (25 min REBOA + TRIC37°C). In comparison, direct TRIC cooling the intestines at 12 °C significantly reduces the plasma levels of the organ injury biomarkers and the injurious pro-inflammatory cytokines. TRIC12°C also upregulated the plasma levels of anti-inflammatory cytokines. Furthermore, TRIC12°C combined with a quick arterial infusion of ice-cold HEXTEND solution (TRIC12°C + 0 °C HEX) offered an additional and staggering reduction of the key organ injury biomarkers and injurious pro-inflammatory cytokines. These results demonstrate that both TRIC12°C and TRIC12°C + 0 °C HEX provide a robust protection of abdominal organs from IRI and SIRS.
Post-REBOA mortality. This study showed that directly cooling the intestine offered exceptional protection against mortality in the rat model of REBOA management of lethal hemorrhage. The results are consistent with a recent study from our laboratory, which showed that TRIC12°C reduced mortality after 20 min of REBOA management of (50% EBV) lethal hemorrhage 16 . The power analysis of the mortality of the previous study between cooling and non-cooling groups showed that 3 rats per experimental group were required in this study.

Post-REBOA organ injury biomarkers.
The changes in the plasma levels of the abdominal organ injury biomarkers are consistent with the several histopathological injuries to the intestines, liver, and kidneys during Post-REBOA pro-inflammatory cytokine storm and SIRS. SIRS result from a surge in the release of pro-inflammatory cytokines. The term cytokine storm occurs when there is a persistent and uncontrolled release of these pro-inflammatory cytokines. This release of cytokines can occur due to severe tissue injury or response to a profound infection or foreign entity 17 . TNF-α, IL-1β, and IL-17F represent a typical array of pro-inflammatory cytokines. The plasma levels of the pro-inflammatory cytokines are significantly higher in TRIC37°C group than in the TRIC12°C groups after REBOA. However, within the TRIC12°C groups, the plasma levels of the proinflammatory cytokines were all relatively low, even though all the rats in the 30 min REBOA + TRIC12°C group died. This may suggest that, although the cooling inhibited the upregulation of pro-inflammatory cytokines, factors other than pro-inflammatory cytokines may also play a vital role in the survival. Our study provides evidence to suggest that both abdominal organ injuries and pro-inflammatory response contribute to post-REBOA mortality.
TRIC activates the anti-inflammatory response after REBOA. This study further shows that directly cooling the intestine activated an anti-inflammatory response during the post-REBOA phase. This finding is consistent with reports that hypothermia can elicit a powerful anti-inflammatory reaction 18,19 . Previous results have shown that hypothermia upregulates the blood levels of IL-4 and IL-10 and can inhibit neutrophil aggregation and production of nitric oxide during endotoxemia 18 . Although IL-4 and IL-10 are key anti-inflammatory cytokines downregulating SIRS, we found that TRIC12°C increased in the plasma level of IL-10 more significantly than that of IL-4 (p < 0.05). A plasma IL-10 level higher than 108 pg/ml seemed to correlate with 100% survival (p < 0.05). Furthermore, the ratios of TNF-α/IL-10 < 0.8, IL-1β/IL-10 < 1.2, and IL-17F/IL-10 < 1.5 seemed to also correlate with 100% animal survival in our model.  www.nature.com/scientificreports/ TRIC may be a better therapeutic hypothermia modality. The present study provides evidence that utilizing TRIC for cooling the intestine offers exceptional protection against fatal IRI and SIRS. There are currently two main therapeutic temperature management techniques: (i) external skin cooling and (ii) endovascular cooling. Both cooling techniques are highly inefficient, as cooling down an average-sized 65 kg human (65 kg × 65 mL blood/kg body weight = 4 L of human blood and its supplied tissue) from 37 to 30 °C would take several hours at a cooling rate about 2-4 °C/hour 12,20 . By contrast, the TRIC device can cool down the gut temperature from 37 to 10 °C in about 2-5 min because of its direct contact. This localized cooling also creates a temperature gradient in the surrounding organs while maintaining a tolerable upper body temperature range above 28 °C and avoids low temperature-induced fatal cardiorespiratory failure 21,22 . Therefore, TRIC cooling is not a systemic hypothermia, rather than a focal cooling technique to achieve the optimal cooling temperature (TRIC12°C) for the maximal protection of the intestines during the post-REBOA phase. This study demonstrated that the level of endotoxin and FABP2/I-FABP were drastically increased during the post-REBOA phase, indicative of severe intestinal injury and the likely development of intraabdominal sepsis. In addition to sepsis, the intestines are equipped with the largest pool of macrophages and represent more than 70% of the entire immune system in the body 23,24 . Therefore, intestinal injury during the post-REBOA phase activates immune cells and macrophages, which propagates systemic inflammation resulting in SIRS. As a result, intestinal injury has been considered the "origin" or "motor" of the SIRS after lethal hemorrhage 25 . TRIC12°C significantly reduced the intestinal IRI as indicated by the lower plasma levels of endotoxins and FABP2/I-FABP. Therefore. cooling the intestines may equal to cooling the immune organ for the management of SIRS after REBOA management of lethal hemorrhage.
Post-REBOA cold perfusion shows further protection. The cold perfusion may represent the most effective measure for preserving transplant organs outside the body 26 . However, it may not be clinically feasible to perform deep hypothermia in most living patients, as the vital organs, particularly the heart and lungs, require warm oxygenated blood to function properly 21,22 . Once the REBOA is deployed, it blocks the blood flow to the distal abdominal organs and provides an opportunity to perform the cold perfusion of the distal abdominal organs inside the body while having insignificant impact on the heart and lungs. In early organ transplantation studies, simple surface cooling and cold perfusion with heparinized blood were used to preserve transplant organs for transport 26 . It was later realized that the flushing out the blood cellular components with cold physiological electrolyte solutions followed by the static cold storage offered considerably better organ preservation 26 . Our study showed, in conjunction with TRIC focal cooling, the quick arterial infusion of ice-cold Hextend solution into the abdominal organs offered greater protection against post-REBOA fatal IRI and SIRS without adverse effects to the heart or lungs.
Limitation of this study. Several limitations exist in this study. A limitation of the present study may be that the lethal hemorrhagic shock model was produced by bleeding 50% of the EBV without a primary traumatic injury. However, the implementation of REBOA is likely to stop the torso hemorrhage from the primary injury, thus reducing the risk of the enhanced ongoing bleeding caused by intra-colon cooling. Furthermore, some clinical situations may require the use of REBOA management of significant bleeding without a primary traumatic injury, such as aortic aneurysm injury and repair, or the unanticipated and uncontrolled bleeding during abdominal surgery. Therefore, the lethal hemorrhagic model (without severe tissue trauma) was often used in swine REBOA studies 14 .
There is an inverse relationship between temperature and prothrombin time (PT) 27 . As shown in Fig. 7, the PT/INR were moderately increased after REBOA, which was likely due to the use of heparin for the surgical preparation of the animal model. However, the increases in the PT/INR seemed somewhat more in TRIC12°C than in TRIC37°C group, which could be a result of TRIC cooling. The insignificant difference in PT/INR between TRIC12°C and TRIC37°C groups is likely because that the study was not powered to detect the difference. However, the PT/INR differences between the two groups appeared not as dramatic as what was expected at the intestinal 12 °C condition 27 . The smaller PT/INR differences between TRIC12°C and TRIC37°C groups are likely because, during implementation of REBOA, most of the body blood volume was circulating in the "warmer" upper body. As a result, TRIC intra-colon cooling may have a limited impact on the systemic coagulation cascade. Furthermore, implementation of REBOA might alleviate the cooling-induced bleeding risk in the abdominal organs. Nevertheless, coagulopathy after injury is complex. PT/INR may not sufficiently represent the derangements. Further investigation of broader coagulation parameters is needed to understand the effects of TRIC cooling on the potential coagulopathy after injury.
Another issue may be whether TRIC device can be used in humans as efficiently in rats. In a currently ongoing study of direct gut cooling in a swine model, we have demonstrated that TRIC directly cooling the gut can quickly reduce the intra-colon temperature from 37 to 12 °C within 10 min and bladder from 37 °C to about 25 °C within 30-40 min in a 45 kg naïve swine, while maintaining the esophageal upper body temperature in the tolerable range.
It also should be pointed out that TRIC cooling was initiated prior to the REBOA implementation in this study. Further animal study may be needed to investigate the scenario of the simultaneous implementation of both REBOA and TRIC cooling for mimicking patients who present in extremis and thus require prompt implementation of REBOA.
Additionally, the timescales of major biological events are different between rats and humans. For example, an average life expectancy is ~ 27 times shorter while the basal metabolic rate is ~ 6.4 times faster in rats than in humans 28 . However, the rat model is easily manageable, can be controlled in virtually identical experimental conditions and economical, while having the cardiovascular, organ injury biomarkers, and inflammatory response www.nature.com/scientificreports/ systems similar to those of humans. Therefore, the rat model may be suitable for the proof-of-concept studies. Significant efforts are needed for potential translating the proof-of-concept TRIC protection from the small animal model to the large animal model, and eventually to human applications.

Potential mechanisms of anti-inflammatory effects of directly cooling intestines. A critical
question may be why fast and deep cooling gut is more effective in reducing SIRS than cooling other organs in the body. In broad terms, the gut is comprised of three entities: the intestinal lumen commensal flora, the epithelium, and the mucosal immune system 29 . The release of intestinal luminal bacteria and toxins into the circulation due to gut IRI propagates systemic inflammation, and thus the gut is considered the most dangerous abdominal organ, and "origin" of SIRS under many severe medical conditions [29][30][31][32][33][34] . This may be consistent with the fact that SIRS can be induced by a long list of intestinal diseases and by all "leaky" intestinal injuries 30 . In order to prevent invasion of the intestinal microorganisms and toxins during transport of nutrients, the gut is equipped with the largest pool of immune cells representing more than 70% of the entire immune system and the largest population of macrophages in the body 29,30 . A recent study from our laboratory showed that the intestine was the most susceptible organ to IRI after implementation of REBOA 35 . The post-REBOA histopathological damage to the intestines was also observed in the present study (Supplemental Fig. 1). Consistently, the plasma I-FABP and endotoxin were drastically increased during the post-REBOA phase (Fig. 2). Evidence suggests that deep cooling the gut may be equivalent to cooling the major immune organ and thus is highly effective in lessening life-threaten SIRS after REBOA management of hemorrhage.