CXCL1 is upregulated during the development of ileus resulting in decreased intestinal contractile activity

Although the development of ileus is widespread and negatively impacts patient outcomes, the mechanism by which ileus develops remains unclear. The purpose of our study was to examine the contribution of myogenic mechanisms to postoperative ileus development and the involvement of inflammation in mediating intestinal smooth muscle dysfunction.


| INTRODUC TI ON
Ileus, or depressed intestinal contractile activity, is manifested clinically as feeding intolerance, one of the most common postoperative complications. Grocott reported that 20% of orthopedic surgery patients, 92% of general surgery patients, and 51% of urology surgery patients experienced gastrointestinal postoperative complications. 1 Virtually, all abdominal surgery patients experience some degree of ileus. In general surgery patients, gastrointestinal complications were the leading reason for hospital readmission. 2 Ileus is also a common complication in trauma patients. Approximately 33% of moderate to severe trauma patients develop feeding intolerance, and 25% develop documented ileus. 3 Ileus significantly increases patient care costs by necessitating the treatment of resultant complications, feeding parenteral nutrition, and prolonging hospital stays. 4 More importantly, ileus causes significant patient discomfort and increases the risk of septic, pulmonary, and thromboembolic complications, and malnutrition. 5,6 Despite the widespread and negative impact of ileus on patient outcomes, the mechanism by which ileus develops remains unclear.
Ileus development is complex, with a number of contributing factors.
Undoubtedly, opioid use contributes to ileus in surgical patients.
However, in a trauma patient study, morphine equivalents were not significantly different between patients who developed feeding intolerance and those who did not, 3 indicating that other factors contribute to slowed gastrointestinal motility. Intestinal edema resulting from both excessive fluid use and/or inflammation also inhibits intestinal contractile activity, 7-10 and in recent years, inflammation has been shown to contribute to ileus development. [11][12][13][14][15] A combination of factors, including inflammation, edema development, and opioid use, is likely to contribute to the development of ileus; however, the mechanisms by which these factors induce ileus may be different.
The effects of inflammation, in particular, are poorly understood.
Inflammation affects the enteric nervous system (ENS) to suppress intestinal motility; however, studies suggest that intestinal smooth muscle is also affected by inflammation. 11,12 In abdominal surgery patients, intestinal manipulation is thought to activate resident macrophages in the intestinal muscle layers resulting in cytokine and chemokine release and recruitment of leukocytes. Inflammation results in contractile dysfunction, not only in the manipulated sections of the bowel, but in most of the small and large intestine. Systemic inflammation in trauma and general surgery patients may also contribute to the development of ileus.
While the effects of intestinal edema on the ENS are unclear, we have shown that intestinal edema alone, in the absence of inflammation, inhibits intestinal contractility by downregulating smooth muscle myosin light chain (MLC) phosphorylation. 10 Edema development results in increased intestinal wall stress. 16 Increased stretch of intestinal smooth muscle cells has been shown to down-regulate MLC phosphorylation. 17 Thus, edema-induced mechanotransduction in the intestinal smooth muscle layers is one mechanism by which edema causes smooth muscle dysfunction. Little is known about the interaction between mechanotransduction and inflammatory mediator signaling in smooth muscle. Furthermore, the effects of increasing intestinal wall stress during edema development on macrophage activation are unclear.
While our investigations and others suggest that smooth muscle dysfunction contributes to the development of ileus, the drugs available to treat ileus target the ENS. We postulate that if intestinal smooth muscle is dysfunctional, drugs targeting the ENS will have limited effectiveness. Thus, the purpose of our study was to examine the contribution of myogenic mechanisms to the development of postoperative ileus, and the involvement of inflammation in mediating smooth muscle dysfunction.

| Animal model
A gut manipulation (GM) model of postoperative ileus, described by Bauer, was utilized. 18 Male Sprague Dawley rats weighing between 250 and 350 g were used for all experiments. All procedures were approved by the University of Texas Medical School Care and Use Committee and are consistent with the NIH "Guide for the Care and Use of Laboratory Animals". Intestines were exteriorized via an abdominal incision in anesthetized rats. The small intestine was gently compressed using a rolling motion between two sterile cotton applicators, in a proximal to distal direction. This procedure was repeated 3 times. After returning intestines to the abdominal cavity, the incision was closed. Sham animals underwent the same procedure with a laparotomy but no exteriorization or manipulation of the intestines. Naïve animals were not subjected to any procedures.
Animals were sacrificed 0, 1, 2, 4, 12, and 24 hours after surgery, and the distal small intestine was collected for functional and biochemical analyses.
In a separate set of animals, rats were treated with the CXCR2 antagonist, SB265610 (3 mg/kg) or vehicle (3% DMSO) via intraperitoneal injection, immediately after gut manipulation or sham surgery. Animals were sacrificed 24 hours after surgery, and intestinal segments were collected.

Key Points
• Spontaneous and agonist-induced intestinal contractile activities and intestinal transit are decreased after gut manipulation.
• Macrophages release CXCL1 in response to mechanical stress.
• CXCL1 also increased in intestinal smooth muscle after gut manipulation in a rodent model and in the circulation of trauma patients who develop ileus.
• These data suggest that CXCL1, released from macrophages during intestinal wall stress, can suppress intestinal contractile activity.

| Intestinal contractile activity
Contractile activity was measured 0, 1, 2, 4, 12, and 24 hours after surgical manipulation in the distal small intestine, as described in previous publications. 7,10,17 After equilibration, 10 minutes of basal contractile activity data was recorded. A subset of intestinal strips was treated with tetrodotoxin (TTX, 0.3 μmol/L added to the bath), and another 10 minutes of contractile activity was recorded.
Agonist-induced contractile activity was measured by adding increasing concentrations of carbachol to the organ bath in 5-minute intervals. Total contractile activity was calculated as the area under the curve. Basal tone was defined as the average minimum of the contraction cycle. Amplitude was calculated as average cycle height.
All force development was normalized to tissue cross-sectional area. Measurements were performed on two separate intestinal strips and averaged.
For measurement of CXCL1 effects on contractile activity, intestinal sections were equilibrated for 30 minutes. After recording 5 minutes of baseline data, 200 ng/mL of CXCL1 (Biolegend, San Diego, CA) was added to the organ bath chamber and 10 minutes of data was recorded. After CXCL1 addition, a carbachol dose-response curve was generated as described above.
For determining the effects of conditioned media (see macrophage conditions below), animals were subjected to sham surgery only and 6 hours later tissue was collected for measurement of contractile activity. Equilibration and baseline measurements were as described above. Tissue was treated with media only, or media collected from macrophages after cyclical stretching. Media was added in a 1:10 V:V dilution. Measurements were normalized to baseline.
An aliquot of conditioned media was pretreated for 5 minutes with 30 μmol/L oxyhemoglobin before adding to the organ bath chamber in a subset of experiments.

| Measurement of MLC phosphorylation by Western blotting
The mucosa was removed immediately after collection of intestines, and smooth muscle tissue lysates were generated and subjected to Western blotting as described previously. 17 Antibodies used were MLC Ser19 (Cell Signaling) and total MLC. ImageJ 20,21 was used to quantify luminescence intensities. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to control for protein loading amounts.

| Intestinal tissue water
Wet to dry weight ratios were measured in the mid-section of the small intestine. Intestinal samples were weighed immediately after collection. After drying in a 65°C oven, samples were weighed again. Wet to dry weight ratio was calculated as [(wet weight) − (dry weight)]/dry weight.

| Cytokine array
Cytokine and chemokine levels in the rat ileal smooth muscle and conditioned medium were measured using a rat or human cytokine array (R&D Systems) following the manufacturer's directions. In

| Human primary cell isolation and culture
Disease-free small intestinal tissue was collected from organ donor

| Treatment of hISMC with conditioned THP-1 media
THP-1 cells are an immortalized human monocyte-like cell line.
THP-1 cells (ATCC) were seeded onto Flexcell plates and 1 hour later activated with phorbol-12-myristate-13-acetate (PMA, 12.5 ng/mL). Twenty-four hours after PMA treatment, cells were subjected to either CCS or ECS protocols. Media, termed "conditioned media," was collected immediately after the 4 hours stretch protocol. Conditioned media was either used to treat intestinal strips, as described above, or added to hISMC, which were previously seeded onto Flexcell plates. Two hours after pretreatment with conditioned media, hISMC were subjected to either the CCS or ECS for 4 hours

| Statistics
Analysis of Variance (ANOVA) and t tests were used to compare groups where appropriate. A P value <.05 was considered significant.
When ANOVA was significant, Fishers LSD was used for post hoc analysis. Data are shown as the mean ± standard error of the mean.  The carbachol dose-response curve at 12 hours after surgery is shown in Figure 1C. Agonist-induced contractile activity was significantly reduced in the GM group compared with the SHAM group, from 10 −8 M to 10 −6 M carbachol concentrations (P = .05, 0.007, and 2 × 10 −5 for 10 −8 , 10 −7, and 10 −6 M carbachol). The carbachol doseresponse curve 24 hours after surgery looked similar to the 12-hour curve (data not shown). Carbachol-induced contractions were not different at the 1, 2 and 4 hours time points (data not shown). The effects of tetrodotoxin were measured at the highest carbachol dose (10 −6 M) 12 hours after surgery, as shown in Figure 1D. Tetrodotoxin reduced agonist-induced contractile activity in the Sham group; however, agonist-induced contractile activity was significantly reduced in the GM group compared with the Sham group with or without tetrodotoxin (P = 2 × 10 −5 and 0.02 for GM vs SHAM in VEH and TTX treated groups; P = 9 × 10 −4 , SHAM TTX vs VEH).

| Contractile dysfunction
Relaxation after agonist-induced contraction was measured 4 hours after induction of ileus (at the point of maximal differences in tone).
Relaxation was significantly suppressed in the GM group compared with the Sham group, as shown in Figure 1E.
Intestinal transit, measured in a subset of animals 12 hours after surgery, is shown in Figure 1F. The geometric center was significantly decreased in the GM group compared with the Sham group (5.57 ± 0.57 vs 3.24 ± 0.68, P = .039).

| Changes in MLC phosphorylation and edema
MLC phosphorylation, the rate-limiting step for smooth muscle con-  Figure 2C). In addition, wet to dry weight ratios increased significantly from 4 to 24 hours after surgery in the GM group compared with the 0 time point.

| Effects of conditioned media from macrophages on contractile activity
We determined the effects of mechanical stimulation of macrophages on contractile activity and cytokine production. We collected "conditioned media" from macrophages subjected to CCS or ECS, mimicking intestinal wall mechanical stimuli under normal or edematous conditions, respectively. As shown in Figure 3A, conditioned media had no significant effects on contractile activity of intestinal sections collected from naïve rats. In contrast, conditioned media collected from macrophages subjected to ECS, but not CCS, inhibited the intestinal contractile activity in Sham animals (subjected to laparotomy only) ( Figure 3B). The effects of conditioned media collected from macrophages subjected to CCS showed no differences compared with media alone. Intestinal contractile activity was still inhibited by ECS conditioned media, after oxyhemoglobin treatment to oxidize nitric oxide, suggesting that nitric oxide was not causing the inhibition of contractile activity ( Figure 3B).
Primary hISMC were treated with conditioned media from macrophages subjected to either CCS or ECS ( Figure 3C). When hISMC were subjected to CCS, the conditioned media did not affect MLC phosphorylation. When primary hISMC were subjected to ECS, conditioned media from macrophages subjected to ECS significantly inhibited MLC phosphorylation compared with conditioned media from macrophages subjected to CCS.
Conditioned media from macrophages subjected to CCS or ECS were analyzed for cytokine content. In Figure 3D,3, cytokine levels in media are shown relative to quiescent cells. There are relatively small changes in the media of macrophages subjected to CCS. In contrast, several cytokines increase significantly in the media from macrophages subjected to ECS, including CXCL1, which was increased almost 3-fold.

| Cytokine production in vivo
To determine if in vivo changes in cytokines were similar to changes in macrophages subjected to mechanical stress in vitro, we collected intestinal tissue after gut manipulation or sham surgery and measured cytokine levels in the smooth muscle layers.
Data are shown as a ratio of GM to SHAM. As shown in Figure 4A, CXCL1 was increased 8-fold within 1 hour after gut manipulation. The increased CXCL1 levels were sustained until at least 24 hours after gut manipulation. LIX (CXCL5) was also upregulated in smooth muscle tissue after gut manipulation in an early and sustained manner. At later time points, IL-1β and TNFα were upregulated at 4 and 12 hours, respectively.
The upregulation of CXCL1 was confirmed by ELISA. As shown in Figure 4B, CXCL1 was significantly increased within 1 hours after GM compared with the Sham group. We also determined circulating CXCL1 levels in severely injured trauma patients 48-72 hours after hospital admittance. Plasma CXCL1 levels were significantly upregulated in trauma patients who developed ileus compared to trauma patients who did not develop ileus (P = .004) ( Figure 4C).

| Effects of CXCL1 on contractile activity
The effects of CXCL1 (200 ng/mL) on intestinal contractile activity were measured in an organ bath in animals subjected to sham surgery only ( Figure 5A). CXCL1 induced a significant decrease in contraction amplitude and a significant increase in tone, compared with VEH treatment (0.1% bovine serum albumin in phosphate-buffered saline). Agonist-induced contractile activity was significantly decreased after treatment with CXCL1 ( Figure 5B).
To further explore the effects of CXCL1 on contractile activity, we measured contractile activity after either sham surgery or gut manipulation in animals treated with a CXCR2 antagonist (SB265610, 3 mg/mL) or vehicle (3% DMSO). As shown in Figure 5C, treatment with the CXCR2 antagonist did not ameliorate the decreased contraction amplitude induced by gut manipulation. In contrast, inhibition of CXCR2 abrogated the decreased agonist-induced contractile activity induced by gut manipulation ( Figure 5D).

| D ISCUSS I ON
Our data suggest that gut manipulation-induced ileus is mediated, at    Figure 2. These data support the hypothesis that smooth muscle dysfunction contributes to ileus, and, therefore, drugs targeting the ENS or CNS will be ineffective in improving gastrointestinal motility in trauma patients.
The smooth muscle dysfunction observed in our study agrees with Farro et al, in which impaired smooth muscle contractility after gut manipulation was also observed. 24  * * in the absence of any treatments. We observed a higher tone in the gut manipulation group compared with the sham group. This significantly altered tone was preserved after tetrodotoxin treatment, indicating that the changes in tone were myogenic, not neurogenic.
In fact, at 24 hours after surgery, tone was significantly reduced in both the SHAM and GM groups in the presence of tetrodotoxin ( Figure 1B), indicating that the ENS had a positive effect on tone, and the differences in tone due to myogenic effects were unmasked with TTX treatment. Farro et al demonstrate impaired relaxation of intestinal smooth muscle after induction of ileus. 24 The increased tone in our study may be due to impaired relaxation. Figure 1E shows that relaxation after agonist-induced contractions is indeed suppressed in the ileus group compared with the Sham group, supporting the idea that the ability of smooth muscle to both contract and relax is impaired.  24 We also observed increased inflammation in intestinal smooth muscle after induction of ileus. In addition to increased cytokine production, intestinal edema developed within 4 hours after surgery ( Figure 2C). Edema development in the absence of vascular changes or hemodilution is indicative of inflammation.

Ileus in trauma patients
We have shown previously that edema itself, in the absence of gut manipulation, can decrease intestinal contractile activity. 7,10 Cox et al have shown that edema also results in increased mechanical stress in the intestinal wall. 16 This increased stress decreases MLC phosphorylation in primary intestinal smooth muscle cells. 17 We speculated that increased mechanical stress may also affect the resident macrophages in the intestinal wall. Thus, we subjected activated macrophages to either control cyclical stretch (CCS), as the macrophages would experience under physiological conditions, or edema cyclical stretch (ECS), which intestinal macrophages would experience after edema development. We collected the "conditioned" media after stretching the macrophages. Interestingly, conditioned media from macrophages subjected to CCS or ECS did not affect intestinal contractile activity of intestinal strips collected from naïve animals. In contrast, conditioned media collected from macrophages subjected to ECS significantly decreased contractile activity of tissue sections after surgical stress (Figure 3). Pretreatment of the conditioned media with oxyhemoglobin had no effect; thus, the inhibitory effect of the conditioned media on contractile activity was not due to nitric oxide release from the macrophages. Furthermore, treatment of primary hISMC with conditioned media had similar results; conditioned media from macrophages subjected to ECS inhibited MLC phosphorylation in ISMC subjected to ECS, but not CCS. We conclude from these data, that when macrophages are subjected to mechanical stress, such as they would experience during intestinal edema development or gut manipulation, the macrophage secretome inhibits intestinal contractile activity. Furthermore, the macrophage secretome did not affect naive tissue, but only affected tissue already subjected to stress.
To identify the substance secreted from macrophages that inhibited intestinal contractile activity, we used a cytokine array to determine changes in a panel of cytokines in the conditioned media.
In macrophages subjected to CCS, there were no significant changes in cytokines compared with quiescent cells. In media collected after macrophages were subjected to ECS, CXCL1 (GroA) was upregulated ( Figure 3D,3). CXCL1 was also the predominant cytokine upregulated in vivo in intestinal smooth muscle after gut manipulation ( Figure 4A). CXCL1 was upregulated within 1 hours after gut manipulation, and the upregulation was sustained for up to 24 hours ( Figure 4F). Of note, CXCL1 was also upregulated in trauma patients who developed ileus versus trauma patients who did not develop ileus ( Figure 4G). CXCL1 was upregulated very early after gut manipulation and was likely increased before leukocyte infiltration. Thus, CXCL1 was likely secreted by resident macrophages in the intestinal smooth muscle layers. The early release of CXCL1 is in agreement with the study by Farro, in which CXCL1 was significantly increased within the first 1.5 hours after gut manipulation. 24 However, unlike the Farro study, we did not see early increases in TNFα, IL-1β, or IL-1α. In our study, IL-1β increased 4 hours after gut manipulation and TNFα increased 12 hours after gut manipulation. These differences could have arisen for a number of different reasons. Firstly, Farro was measuring mRNA levels and we measure actual cytokine levels.
Secondly, we measured the ratio of cytokines in GM to Sham. Thus, IL-1β and TNFα could have increased in both groups due to surgical stress. In the Farro manuscript, the cytokines are shown in the gut manipulation group only; the cytokine levels in the laparotomy only group are not shown.
According to Figure 4, CXCL1 was the first cytokine to increase after gut manipulation, and CXCL1 was, for the most part, the most increased cytokine. Thus, we investigated the effects of CXCL1 on intestinal contractile activity ( Figure 5). When intestinal tissues from sham animals (laparotomy only) were treated with CXCL1, contraction amplitude decreased slightly but significantly and tone increased significantly compared with vehicle treatment ( Figure 5A).
CXCL1 substantially decreased agonist-induced contractile activity ( Figure 5B). These effects of CXCL1 on intestinal contractile activity are similar to the effects of gut manipulation on intestinal contractile activity. If animals were treated with a CXCR2 antagonist immediately after gut manipulation, spontaneous contractile activity was unaffected ( Figure 5C). In contrast, CXCR2 antagonism prevented decreases in agonist-induced contractile activity after gut manipulation ( Figure 5D). One drawback of the study is that the CXCR2 antagonist has off target effects which may affect intestinal contractile activity.
CXCL1 is a member of the C-X-C chemokine family. CXCL1 is upregulated in many inflammatory processes, often in response to nuclear factor Kβ (NF-kβ) or CCAAT-enhancer-binding proteinsβ (C/EBPβ). The mechanism by which CXCL1 is increased in intestinal smooth muscle is unclear. However, CXCL1 was increased in the conditioned media of cyclically stretched macrophages subjected to increased stretch ( Figure 3E). These data suggest that CXCL1 is secreted by macrophages (or other cells types in the intestinal smooth muscle) in response to mechanical stimuli. We have shown previously that NF-kβ which can upregulate CXCL1, is increased in edematous intestinal smooth muscle (ie tissue that is experiencing increased mechanical stretch) resulting in decreased intestinal motility. 25,26 Thus, NF-kB may upregulate CXCL1 in response to increased stretch.
The mechanism by which CXCL1 can affect smooth muscle contractility is unclear. CXCL1 acts as a full agonist at the CXCR2 receptor. CXCR2 activation triggers a number of downstream pathways, including Akt signaling and PAK1 signaling. 27,28 We have shown previously that increased PAK1 activation can inhibit intestinal contractile activity via decreased MYPT1 phosphorylation. 17 Thus, CXCL1 may signal through PAK1 to decrease intestinal smooth muscle contractility. Interestingly, in vascular smooth muscle, Akt signaling was shown to increase smooth muscle tone in a manner that was uncoupled to MLC phosphorylation. 29,30 This may explain the disparate results concerning decreased contraction amplitude versus increased tone ( Figure 1).

In summary, both spontaneous and agonist-induced intestinal
contractile activities were decreased after gut manipulation. The decreased contractile activity resulted in decreased intestinal transit.
Resident macrophages are a likely source of cytokine release early in the development of contractile dysfunction. Thus, we subjected macrophages to mechanical stress and showed that CXCL1 was released. CXCL1 also increased in intestinal smooth muscle after gut manipulation in a rodent model and in the circulation of trauma patients who develop ileus. CXCL1 decreased agonist-induced contractile activity; the suppression of agonist-induced contractile activity could be blocked with a CXCR2 antagonist. Taken together, these data suggest that CXCL1, released from macrophages during intestinal wall stress, can suppress intestinal contractile activity. CXCL1 is a potential target for treating decreased contractile activity in trauma and surgical patients.

CO N FLI C T S O F I NTE R E S T
No competing interests declared.

AUTH O R CO NTR I B UTI O N S
KU, TD, CW, and CC contributed to study design and concept; TD, DB, AS, and KU contributed to acquisition of data; KU, TD, and DB contributed to analysis and interpretation of data; KU and TD con-