• Open Access

In Vivo and In Vitro Evidence of the Involvement of CXCL1, a Keratinocyte-Derived Chemokine, in Equine Laminitis


  • Portions of this manuscript were presented at the Veterinary Emergency and Critical Care Society Annual Meeting, Phoenix, AZ, 2008.

Corresponding author: Dr James K. Belknap, Department of Veterinary Clinical Sciences, The Ohio State University, 440 VMAB, 1900 Coffey Road, Columbus, OH 43210; e-mail: belknap.16@osu.edu


Background: C-X-C motif ligand 1 (CXCL1) is an important chemokine of epithelial origin in rodents and humans.

Objectives: To assess in vivo and in vitro the regulation of CXCL1 in equine laminitis.

Animals: Twenty adult horses.

Methods: Real-time quantitative polymerase chain reaction (PCR) was used to assess expression of CXCL1 in samples of laminae, liver, skin, and lung from the black walnut extract (BWE) model of laminitis, and in cultured equine epithelial cells (EpCs). Tissue was obtained from control animals (CON, n = 5), and at 1.5 hours (early time point [ETP] group, n = 5), at the onset of leukopenia (developmental time point [DTP] group, n = 5), and at the onset of lameness (LAM group, n = 5) after BWE administration. EpCs were exposed to Toll-like/Nod receptor ligands, oxidative stress agents, and reduced atmospheric oxygen (3%). In situ PCR was used to localize the laminar cell types undergoing CXCL1 mRNA expression.

Results: Increases in laminar CXCL1 mRNA concentrations occurred in the ETP (163-fold [P= .0001]) and DTP groups (21-fold [P= .005]). Smaller increases in CXCL1 expression occurred in other tissues and organs. In cultured EpCs, increases (P < .05) in CXCL1 mRNA concentration occurred after exposure to lipopolysaccharide (LPS [28-fold]), xanthine/xanthine oxidase (3.5-fold), and H2O2 (2-fold). Hypoxia enhanced the LPS-induced increase in CXCL1 mRNA (P= .007). CXCL1 gene expression was localized to laminar EpCs, endothelial cells, and emigrating leukocytes.

Conclusion and Clinical Importance: These findings indicate that CXCL1 plays an early and possibly initiating role in neutrophil accumulation in the BWE laminitis model, and that laminar keratinocytes are an important source of this chemokine. New therapies using chemokine receptor antagonists may be indicated.


black walnut extract


complementary DNA


cytokine-induced-neutrophil chemoattractant


control group


C-X-C motif ligand


developmental time point


3-residue motif


epithelial cells


early time point


growth-related oncogene




keratinocyte-derived chemokine


onset of lameness time point




lipoteichoic acid


melanoma growth stimulatory activity


nitroblue tetrazolium and bromochloroindolyl phosphate


pathogen-associated molecular pattern protein


polymerase chain reaction




polymorphonuclear cell

RT in situ PCR

reverse transcriptase in situ PCR


real-time quantitative PCR


systemic inflammatory response syndrome

Recent studies have demonstrated that local laminar and systemic inflammation similar to that observed in systemic inflammatory response syndrome (SIRS) occurs in experimental models of laminitis.1–5 Local laminar events that mimic events leading to organ failure in humans with sepsis include endothelial activation, cytokine and chemokine gene expression, and, importantly, leukocyte extravasation and infiltration throughout the laminar tissue.1–5 Important differences exist, however, between equine laminar tissue and the target organs that fail in humans with sepsis. One of the primary differences is that, although epithelial cells (EpC) are injured in numerous target organs in humans with sepsis including lung, kidney, and intestinal tract,6 the laminae are the only tissue in which the primary cell at the point of failure is the epidermal EpC. Chemokines (ie, cytokines that are chemotactic for leukocytes) play a major role in tissue injury in inflammatory diseases such as sepsis because they contribute to leukocyte extravasation and activation.7 Whereas C-X-C motif ligand 8 (CXCL8) (IL-8) is the most studied cytokine in organ injury in sepsis in humans and animals, the present study sought to assess a chemokine mainly expressed by epidermal EpCs, CXCL1.

Originally termed growth-related oncogene-α (GROα) or human melanoma growth stimulatory activity protein, CXCL1 was first isolated and characterized in late 1980s as a consequence of its stimulatory action on human melanoma cells.8 GROα was also recently reported to have an important role inducing inflammatory cell migration into tissues, and was included in the large multifunctional family of cytokines termed chemokines.7–9 By the arrangement of the N-terminal cysteines and the presence of a 3-residue motif (ELR) in its molecule, GROα was classified as an ELR + CXC chemokine and named CXCL1.7–9 The cytokines from the subgroup ELR + CXC can attract neutrophils that express the receptors CXCR1 and CXCR2, and include the well-studied CXCL8 (commonly termed interleukin 8 [IL-8]). The receptor CXCR1 binds predominantly IL-8, whereas CXCR2 binds both IL-8 and CXCL1. In addition to neutrophils, CXCL1 and its closely related chemokines (CXCL2 [GROβ] and CXCL3 [GROγ]) have been implicated as chemotactic or activating factors for basophils, eosinophils, monocytes, smooth muscle cells, and lymphocytes.7,9

In the epidermis, CXCL1 has been reported to play an important role in the recruitment of inflammatory cells in the wound healing process after injury; it also plays a role in angiogenesis and re-epithelialization of wounds.10–12 CXCL1 can be produced not only by endothelial and inflammatory cells,10 but also by keratinocytes.13 In mice, the structurally homologous protein is termed keratinocyte-derived chemokine (KC).14 Increased expression of CXCL1 has been reported in inflammatory skin diseases such as psoriasis,15 and is suggested to play a role in the neutrophil recruitment and epidermal hyperproliferation processes characteristic of this disease.15 In horses, the only report regarding CXCL1 is a study documenting CXCL1 gene expression by EpCs isolated from the bronchoalveolar lavage fluid of horses with recurrent airway obstruction.16

In laboratory animals, models of SIRS associated with sepsis have shown that the analogue rodent proteins of CXLC1 (KC) and CXCL8, namely cytokine-induced-neutrophil chemoattractant 1, play a fundamental role in tissue injury associated with polymorphonuclear cell (PMN) recruitment into different types of tissue.17,18 For instance, treatment with a CXCR2 inhibitor prevented intestinal and pulmonary neutrophil influx and increased vascular permeability induced by intestinal ischemia and reperfusion in rats,17 and PMN recruitment was abolished after lipopolysaccharide (LPS) inhalation in knockout mice lacking CXCR2.18 Furthermore, KC played a central role in nonbacterial SIRS and remote neutrophil accumulation in a study in which administration of anti-KC antibody before induced trauma and hemorrhage ameliorated neutrophil infiltration and edema formation in lung and liver.19

Considering the role of CXCL1 in the numerous types of leukocyte-mediated inflammation discussed above, and its production by the keratinocyte, a central cell type in laminar pathophysiology, we hypothesized that CXCL1 is overexpressed in laminae and other tissues at different time points in the black walnut extract (BWE) laminitis model. We also investigated the effect of sepsis-related stimuli thought to play a role in laminitis (exposure to pathogen-associated molecular pattern proteins [PAMPs], oxidative stress, and hypoxia) on CXCL1 gene expression by cultured equine keratinocytes.

Materials and Methods

In Vivo Study

Frozen (−80°C) archived laminae, skin, lung, and liver tissue samples from previous studies were used to assess inflammatory events at specific time points during the developmental stage of the BWE model.3–5 Laminar sections were used from 4 groups (5 horses each): control group (CON), early developmental time point group (EDT, 1.5 hours after BWE administration), developmental time point group (DTP, onset of leukopenia, approximately 3–4 hours after BWE), and onset of lameness group (LAM, approximately 10–12 hours after BWE). Skin, liver, and lung samples were only available from CON, DTP, and LAM groups. All animal protocols were approved by the Animal Care and Use Committees of Ohio State University or Auburn University. The methodology for laminitis induction was as previously described.4 Briefly, extract was made for administration to horses by soaking 2 g of black walnut heartwood shavings/kg of body weight of the horses in 6 L of deionized water. Six liters of BWE (principal horses) or water (control horses) were administered by nasogastric tube. Anesthesia was induced after BWE administration at 1.5 hours in the ETP group, at the onset of leukopenia (30% decrease in leukocyte count occurring approximately 3–4 hours postadministration) in the DTP group and at the onset of Obel grade 1 lameness (approximately 10–12 hours) in LAM group. In the CON group, anesthesia was induced 3 hours after water administration. Forefeet were rapidly removed by disarticulation of the metacarpophalangeal joint after placement of a tourniquet and 1.5-cm-thick sagittal sections of the digit were cut with a band saw. Laminae were rapidly dissected from the hoof and third phalanx, and samples were immediately frozen in liquid nitrogen and stored at −80°C or immediately placed in neutral-buffered 10% formalin for 24 hours, placed in 70% ethanol, and subsequently embedded in paraffin. Each horse was subjected to euthanasia with 20 mg/kg (IV) of a solution containing pentobarbital sodium (390 mg/mL) and phenytoin sodium (50 mg/mL) after laminar specimen collection.

Isolation and Culture of Equine Keratinocytes

Two horses subjected to euthanasia for reasons other than laminitis or systemic illness were used to collect sections of skin from the shoulder region to establish cultures of equine primary keratinocytes as previously described.20 Briefly, fresh skin biopsies were microdissected to completely remove subcutaneous tissues in order to avoid contamination by fibroblasts. The remaining epidermis was minced and trypsinized (0.25% in Hank's buffered saline solution) for 4 hours at 37°C. Cells were placed in T-25 flasks coated with collagen type Ia and cultured in epithelial cell culture medium (Epi-CMb) supplemented with 2% fetal bovine serum (the same lot of commercial serum was used for all experiments), epithelial cell growth supplement,c and penicillin and streptomycin at 37°C with 5% CO2. EpCs, at third passage, were plated in 12-well cell culture plates, at 2 × 104 cells per well. Immunofluorescence staining for cytokeratin (Cytokeratin A1/A3d) was performed in the same passaged EpCs cultured in chamber slides on collagen type I matrix to verify pure culture of keratinocytes.

Stimulation of Keratinocytes

All experiments were performed after the EpCs reached confluency using the same commercial epithelial medium, supplemented with 2% calf serum from the same batch. Cells from both horses were independently stimulated by different substances and conditions in triplicate.

In the TLR ligand experiment, keratinocytes were exposed to concentrations of PAMPs used previously in other keratinocyte studies21–25 including 3 concentrations (100 ng/mL, 500 ng/mL, and 5 μg/mL) of LPS,e 4 concentrations (100 ng/mL, 500 ng/mL, 5 μg/mL, and 10 μg/mL) of peptidoglycan (PGN),f 2 concentrations (1 and 10 μg/mL) of lipoteichoic acid (LTA),g 3 concentrations (100 ng/mL, 500 ng/mL, and 5 μg/mL) of flagellin,h and 1 concentration (1 μM) of unmethylated CpG DNA.i The keratinocytes were incubated with the ligands at 37°C for 4 and 24 hours. Medium (supplemented as described above) was only placed in 4 wells with keratinocytes as a negative control.

To determine the keratinocyte response to oxidative stress, the skin EpCs were incubated at 37°C for 1 and 4 hours with 3 concentrations of H2O2 (50 μM, 200  μM, and 1 mM) and 2 concentrations of xanthine/xanthine oxidase (X/XO, 100 μM/3 mU, 200 μM/30 mU). LPS ([5 μg/mL] positive control) and medium only (negative control) were used in the same conditions.

Nitroblue tetrazolium (NBT) dye reduction was used to confirm superoxide radical production by X/XO, using 100 μM/3 mU and 200 μM/30 mU X/XO in EpC culture containing 100 or 250 μM NBT for 1 and 4 hours. Spectrophotometricj analysis at 560 nm for NBT activity was obtained for both concentrations and at both time points.

To determine the effects of hypoxia, keratinocytes were exposed to 2 concentrations of LPS (100 ng/mL and 5 μg/mL) and incubated in normoxic (20% O2) or hypoxic (3% O2)26–28 environment for 4 and 24 hours. In both environment conditions, medium only served as a negative control.

RNA Isolation, cDNA Synthesis, and Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR) Procedure

From homogenized laminae, lungs, and liver tissue samples, total RNA was extracted by use of a commercial phenol/guanidium isothiocyanate protocolk followed by mRNA isolation.l PolyA mRNA was used to make complementary DNA (cDNA) for each horse by reverse transcription.m Total RNA was extracted with trizol from homogenized skin samples. From the cultured keratinocytes, total RNA was extracted from each sample (triplicate samples were extracted individually) with an RNA isolation kit (Absolutelyn). cDNA was then made for each sample by reverse transcription (Retroscripto). The cDNA was frozen at −20°C and stored until used for RT-qPCR analysis.

RT-qPCR was performed with a thermocycler (LightCyclerp) and quantified with external standards with the fluorescent format for SYBR Green I dye as previously described.29,30 Primers for CXCL1 and the housekeeping genes (β-actin, β-2 microglobulin, glyceraldehyde-3 phosphate dehydrogenase, and TATA-box binding protein) were designed from equine specific sequences by computer programs as previously described.29 Amplified cDNA fragments of each gene were inserted into a vector and the vectors linearized for the use of templates to generate a standard curve for the RT-qPCR reaction.29

Reverse Transcriptase (RT) In Situ PCR

Five-micrometer sections from paraffin-embedded samples of the in vivo study (3 for each group) were deparaffinized and subjected to an in situ PCR hybridization protocol for localization of cells expressing RNA for CXCL1 using a protocol previously described.31,32 Briefly, optimal protease digestion time was determined using nonspecific incorporation of the reporter nucleotide digoxigenin dUTP. Optimal protease digestion was followed by overnight incubation in RNase-free DNase and 1 step RT/PCR using the rTth system and digoxigenin dUTP. The chromogen was nitroblue tetrazolium and bromochloroindolyl phosphate (NBT/BCIP) with nuclear Fast Red as the counterstain. Controls for the GROα cDNA signal included use of tissues known to be negative as well as positive for the target as determined by RT-qPCR, omission of the primers and no DNase in situ amplification control, as previously described.31,32 The cells positive for CXCL1 expression stained dark blue compared with the negative ones that stained pale red.

Data Analysis

As previously reported, RT-qPCR data from the 4 housekeeping genes were evaluated by a computer software program (geNormq) to determine which 2 genes received the best acceptable score to be used to create a normalization factor for each tissue.3

Data were expressed by fold-increase compared with the mean value from control samples and subjected to one-way analysis of variance, followed by the Student-Newman-Keuls test. When normal distribution could not be verified, data were log transformed in order to achieve normality. In the hypoxia study, log transformation of the data was not sufficient to produce normality, and the Kruskal-Wallis test was used followed by a Mann-Whitney test to compare the means by a specific software program (Sigma Stat 3.5r). Differences were considered significant when P < .05.


In Vivo Study

RT-qPCR. In the equine laminae, mean (min to max) 163-fold (30–485-fold, P= .0001), and 21-fold (1.5–57-fold, P= .005) increases in laminar CXCL1 mRNA concentrations were observed in the ETP and DTP groups, respectively, reaching values similar to basal at the LAM time point (Fig 1). In liver, lungs, and skin of the same animals, increases in CXCL1 expression also were documented at lower magnitude (Fig 1). In liver and skin, the mean increases (P= .001) peaked at 1.5 hours (23-fold [2–61-fold] and 7.5-fold [3–12-fold], respectively), but, in lung, the highest mean value (12.5-fold [3–24-fold]) occurred at the 3 hour time point.

Figure 1.

 Fold increase changes (compared with the control group) of gene expression of CXCL1 in the laminae, skin, lung, and liver of horses subjected to the black walnut extract (BWE) laminitis model. Groups (5 horses each): control, early time point (ETP, 1.5 hours after BWE), developmental time point (DTP, 3–4 hours after BWE), and lameness time point (LAM, 10–12 hours, after BWE). In the same graph, groups that share the same letter do not differ from each other (P < .05).

RT In Situ PCR. No signal was seen in the treated horses if the CXCL1 primers were omitted and rare or no positive cells were seen in the control lamina (Fig 2A), where most of the cells stained pale red with Fast Red. At the ETP and DTP, CXCL1-positive cells (dark blue stained) in the laminae included the majority of basal EpCs, many endothelial cells, many intravascular leukocytes (Fig 2B), and few interstitial cells presumed to be extravascular leukocytes because of their morphology and spatial orientation (Fig 2C, see gray arrows and inset for CD163-positive macrophage with similar morphology to CXCL1-positive cells).

Figure 2.

 Photomicrographs of laminae from horses with black walnut extract-induced laminitis, showing the RNA expression of C-X-C motif ligand 1 (purple stain) detected by reverse transcriptase in situ polymerase chain reaction hybridization in the horses of the control group (A) and those in the early and developmental time points (B, C). Nitroblue tetrazolium and bromochloroindolyl phosphate with nuclear Fast Red as counterstain; bar = 40 μm. Note the primary dermal (PDL) and the primary (PEL) and secondary (SEL) epidermal laminae in control group; rare positive cells were detected (A). At the early time point and developmental time point, most of the keratinocytes (B and C, gray arrows), many endothelial cells (B, black arrows), many intravascular leukocytes (B, open arrow), and few interstitial cells (C, red arrows) were positive. Note a positive interstitial cell infiltrated in the secondary dermal laminae (C, inset) that have morphology similar to a macrophage.

In Vitro study

In cultured equine skin epidermal EpCs, increases (P < .001) in CXCL1 mRNA expression were observed in cells exposed to the highest concentration (5 μg/mL) of LPS after 4 and 24 hours (Fig 3). Exposure to PAMPs, most commonly associated with Gram-positive organisms (PGN and LTA), did not evoke a response, whereas flagellin (500 ng/mL) induced a mild increase (1.5-fold, P < .001).

Figure 3.

 Fold increase changes (compared with the control group) of gene expression of C-X-C motif ligand 1 in primary cultivated equine keratinocytes obtained from the skin of normal horses after 4 and 24 hours of exposure to TLR ligands. #In the same graph, differ from the control group (P < .05).

Mild increases (P < .001) were observed after 1 and 4 hours exposure to X/XO (200 μM/30 M, 2.6–3.5-fold) and after 4 hours exposure to H2O2 (1 mM, 2-fold) in the oxidative stress study (Fig 4). In the hypoxia study, decreasing the oxygen concentration to 3% was not sufficient to increase CXCL1 expression (Fig 5). However, cells exposed to the 5 μg/mL of LPS for 4 hours under hypoxic conditions exhibited higher expression compared with those incubated with same dose of LPS under normal oxygen conditions (8.3-fold versus 12.3-fold, P= .007).

Figure 4.

 Fold increase changes (compared with the control group) of gene expression of C-X-C motif ligand 1 in primary cultivated equine keratinocytes obtained from the skin of normal horses, after 1 and 4 hours of exposure to xanthine/xanthine oxidase (X/XO) or hydrogen peroxide. #In the same graph, differ from the control group (P < .05).

Figure 5.

 Fold increase changes (compared with the control group) of gene expression of C-X-C motif ligand 1 in primary cultivated equine keratinocytes obtained from the skin of normal horses, after 1 and 4 hours of exposure to lipopolysaccharide (LPS) in normoxic (20%) or hypoxic (3%) environmental conditions. #In the same graph, differ from the corresponding (normoxic or hypoxic) control (P < .05).


Migration of neutrophils from blood vessels into tissues is a pivotal step in remote organ injury after an initial septic or traumatic injury in human beings.33 Similar leukocyte emigration into the equine laminae has been described in the early stages of laminitis,1 and the events leading to leukocyte emigration and activation are a potential target for treatment of this disease process. Research on inflammation using animal models consistently has documented that neutrophils cross the basement membrane of blood vessels and migrate into tissues to induce tissue injury (and subsequent organ dysfunction) by the release of reactive oxygen (and nitrogen) species and proteases.34 In horses, the main cause of structural failure in laminitis is believed to be the dysadhesion of laminar basal epidermal EpCs from the underlying basement membrane (BM). This dysadhesion may occur due to degradation of the BM itself,35,36 inflammatory injury to the EpCs leading to possible failure of adhesion of the EpC to the BM or both.37 Emigrated leukocytes are likely to play an injurious role in these events occurring at the epithelial-dermal interface. Chemokines play a central role in leukocyte activation, adhesion to the endothelium, and extravasation into the interstitium of the tissues.33 Additionally, leukocytes are reported to follow a “chemokine gradient” in some reports, migrating toward a higher concentration of chemokines.33 As these leukocytes have been recently reported to migrate to the dermal-epidermal interface in affected laminae,38 chemokine expression by EpCs may play a central role in stimulating migration of leukocytes to the point of failure in the affected digit. Although we have reported marked increases in laminar mRNA concentrations of the well-described neutrophil chemokine CXCL8 (IL-8) in the developmental stages of BWE-induced laminitis,3,4 we recently have determined, by in situ hybridization, that CXCL8 (IL-8) is expressed in dermal cells (endothelial cells and presumptive extravascular leukocytes), with minimal to no expression in laminar EpCs (J.K. Belknap et al, unpublished data). We therefore were interested in CXCL1 because of its potent neutrophil chemotactic properties and a consistent expression pattern in EpCs in other studies.10–13

The profile of CXCL1 gene expression in the laminae of BWE-treated horses is somewhat similar to the expression pattern for CXCL8/IL-8, the chemokine to which neutrophil emigration is frequently attributed in SIRS/sepsis studies.33 Previous studies from our group showed that IL-8 mRNA concentrations were increased at all time points examined in the BWE model, with expression peaking at the ETP (142-fold mean increase),4 but remaining increased at DTP (20-fold mean increase), and sustained at LAM time point (10-fold mean increase).3 CXCL1 exhibited a similar overall pattern of expression peaking at the ETP (163-fold mean [30–485-fold]) dropping to 21-fold mean increase (1.5–57-fold) at DTP. However, expression decreased more precipitously for CXCL1, with mRNA concentrations returning to control values at the LAM (Fig 1). It is difficult to speculate about such discrepancies because the source of those cytokines has not been completely defined in SIRS models. However, studies conducted in wound healing by in situ hybridization techniques indicated that after an insult, CXCL1 is one of the first chemokines produced by endothelial cells and pericytes, and that IL-8 is produced mainly by arriving inflammatory cells that accumulated later in the tissue.30 Such an explanation (ie, that CXCL1 is one of the first chemokines produced by resident cells initiating the attraction of neutrophils) also is possible in laminitis because keratinocytes were a major source of CXCL1 but not IL-8. However, because both CXCL1 and IL-8 experience similar high expression levels at the ETP assessed, we cannot determine if CXCL1 expression precedes IL-8 expression. Interestingly, whereas it has been suggested that inflammatory cells are a source of IL-8 but not CXCL1 in other species, our findings indicate that equine inflammatory cells express both chemokines. Both CXCL1 and IL-8 peak before the peak in neutrophil emigration, which occurs at the DTP in the BWE model.3,4,38

Apparently, CXCL1 and IL-8 have synchronous actions attracting inflammatory cells from the blood to the tissue.10 This chemokinesis is reported to be mediated by the receptors CXCR1 and CXCR2. It is believed that CXCR2 binding by CXCL1, IL-8 and other ELR + CXC chemokines occurs first causing the initial activation of adherent, emigrating leukocytes.10 With increasing numbers of inflammatory cells in the tissue, more IL-8 is produced locally, which results in a desensitization and downregulation of this receptor. However, because IL-8 also binds CXCR1 on neutrophils, there is a secondary response attracting more neutrophils.10 Studies with cultured human neutrophils demonstrated that CXCL1 pretreatment enhances the IL-8-induced neutrophil release of calcium, whereas IL-8 pretreatment abolishes the CXCL1 response.39 By this line of reasoning, the CXCR1 activation by IL-8 at DTP and LAM time points may explain the sustained increase of leukocytes despite a return of CXCL1 expression to baseline by the LAM time point.

Interestingly, CXCL1 has other actions not associated with its chemokine effects, including induction of changes in the local circulation resulting in damage to the surrounding tissue.7,40 CXCL1 has been shown to indirectly impair endothelium-dependent vasorelaxation in porcine coronary arteries by inducing overproduction of superoxide anion and downregulation of endothelial nitric oxide synthase.7 Furthermore, in a rodent model in which mice were inoculated with an adenovirus modified to result in overexpression of the rodent CXCL1 analogue (KC), overexpression of inflammatory genes including IL-1β and calprotectin in the liver was reported to be responsible for inflammatory injury and ultimately hepatic necrosis.40 Even considering differences among species, organs, and models, such roles for CXCL1 may be important in equine laminitis because circulatory dysfunction,41,42 overexpression of IL-1β,3 and intense calprotectin expression by the epidermis38 have been demonstrated in laminar tissue from BWE-treated horses at time points after the peak of CXCL1 presently described.

The concomitant increased gene expression of CXCL1 in skin, liver, and lungs during ETP and DTP corroborate our previous findings that the effects of oral administration of BWE are not restricted to the hoof laminae or integumentary system5 and confirm that BWE induces a systemic response comparable to the SIRS described in humans and laboratory animal models associated with sepsis and severe tissue trauma.19 As previously described for IL-1β, IL-6, and IL-8,5 increased CXCL1 expression in other organs and tissues was much smaller in magnitude than that reported in laminae at the same time points after BWE administration. Altogether, these findings highlight the susceptibility of laminar tissue to systemic inflammation (and possibly sepsis) and are consistent with laminitis being the most commonly reported complication in equine acute abdomen and endotoxemia or sepsis. The organ that had the most similar pattern of CXCL1 expression with the laminae was the liver, which exhibited a peak at the ETP (Fig 1). This finding may be related to the fact that, in normal horses, laminae and liver are reported to have very few resident neutrophils compared with skin and lungs,1,5 and, therefore, may have evolved with less protective mechanisms to address the injurious actions that neutrophils (ie, proteases) may have on the host cells when activated. This similarity between hepatic and laminar inflammatory responses is supported by the reported hepatic dysfunction in a clinical study of intestinal disease in horses likely to be associated with sepsis.43

The CXCL1 response from the cultured equine skin keratinocytes to the Toll-like/Nod receptor agonists was similar to the IL-6 and IL-8 responses we previously reported,20 demonstrating that equine keratinocytes undergo ELR + CXC chemokine gene expression when exposed to PAMPs associated most commonly with Gram-negative sepsis (LPS and flagellin) and no response to PAMPs associated more with Gram-positive sepsis (LTA and PGN). Although significant, the increases in CXCL1 gene expression in cultured EpCs (Figs 3–5) were much less than those observed in vivo (Fig 1), indicating that the cell phenotype changes that occur with cell culture impact the magnitude of response of the keratinocytes. It is also possible that either the magnitude of response of the equine skin keratinocyte used in these studies is not the same as that of the laminar epidermal EpC or that there are other cell-derived cytokines in vivo in the laminae that play an inhibitory role at the tissue level. The laminar EpC was not used in this study due to inability to maintain pure laminar EpC cultures (ie, without fibroblast contamination).

The BWE model differs from carbohydrate overload models due to the fact that events occur much more rapidly, and the laminae of BWE-treated horses rarely undergo failure.44 The mechanism by which BWE induces a systemic inflammatory response and laminitis in horses is not well understood. BWE-induced disseminated inflammation may follow bacterial product absorption due to a reported decreased mucosal resistance in large colon.45,46 However, as rapid as the response occurs in the BWE model (peak responses at the earliest time point examined [1.5 hours post-BWE administration in the ETP]), it is also possible that a PAMP-like molecule is present in BWE that is absorbed by the proximal GI tract and binds to receptors resulting in systemic inflammation. If this is true, it is likely that the lack of structural failure of the laminae in this model is due to the same reason that the endotoxin bolus rodent model for human sepsis commonly does not result in organ failure (ie, the inflammatory response is too transient compared with the sustained systemic inflammatory response in the CHO models in horses and cecal ligation and puncture models in laboratory rodents).47 However, due to the fact that black walnut is a clinical cause of laminitis, and we have recently found similar inflammatory events to occur in the early stages of a CHO model of laminitis (J.K. Belknap et al, unpublished data), the BWE model is likely to be of value for deciphering early events in the disease process. Finally, preliminary results have been presented indicating that BWE may result in oxidative events in equine serum.48 Regardless of the initiating event, these findings indicate that keratinocyte-derived CXCL1 expression may contribute to leukocyte accumulation in the laminae and other tissues of horses administered BWE, and is likely to play a role in laminitis in horses suffering by diseases associated with endotoxemia.49

The in vitro experiments also showed that hypoxia can enhance the CXCL1 response to LPS, and that keratinocytes undergo increased CXCL1 expression when exposed to oxidative agents. In human and rodent cell culture models, oxidative stress has been extensively studied in skin keratinocytes exposed to ultraviolet radiation and other agents that cause cellular injury,50 and hypoxia has been shown to increase cell mobility51 and prolong life span of keratinocytes.52 In the present study, those conditions were chosen based on previously reported evidence that BWE disturbs blood flow in the developmental stages of laminitis.41,42 Because hypoxia and oxidative stress are expected events during and after tissue hypoperfusion, and assuming that the laminar keratinocytes have the same capability to produce CXCL1, our findings do not exclude a possible role of the vascular events exacerbating induction of CXCL1 expression by epidermal laminar tissue in horses treated with BWE. We must therefore consider that the marked increase in inflammatory response in the laminar tissue in comparison with other organs exposed to the same circulating molecules is suggestive of an isolated event such as ischemia occurring only in the digit.

In conclusion, in vivo results indicate that CXCL1 may play a central and initiating role in the previously described neutrophil accumulation in the early stages of equine laminitis induced by BWE. The in vitro results demonstrated that equine keratinocytes could be an important source of this chemokine when exposed to conditions that have been reported as relevant in the developmental phase of equine laminitis such as presence of bacterial products, hypoxia, and oxidative stress. These results, in light of recent reports on laminar IL-8 expression, indicate that chemokines are likely to play a critical role in early pathologic events in laminitis, and that recently introduced chemokine receptor inhibitors should be investigated as possible treatment options for septic horses at risk of laminitis.


aBovine collagen type 1, BD Biosciences, Billerica, MA

bEpi-CM, ScienCell Research Laboratories, Carlsbad, CA

cCatalog 4152, ScienCell Research Laboratories

dCytokeratin A1/A3 clone antibody, Dako, Carpinteria, CA

eLipopolysaccharide E. Coli, Invitrogen, Carlsbad, CA

fPeptidoglycan S. aureus, Invitrogen

gLipoteichoic acid S. aureus, Sigma-Aldrich, St. Louis, MO

hFlagellin Salmonella typhimurium, InvivoGen, San Diego, CA

iUnmethylated CpG DNA, InvivoGen

jSpectra Max M2, Molecular Devices, Sunnyvale, CA

kGibco BRL, Life Technologies, Grand Island, NY

lmRNA Isolation Kit, Roche Molecular Biochemical, Indianapolis, IN

mTranscriptor Reverse Transcriptase, Roche Molecular Biochemical

nAbsolutely RNA Miniprep; Stratagende Inc, La Jolla, CA

oRetroscript; Ambion Inc, Austin, TX

pLightCycler, Roche Molecular Biochemical

qgeNorm, Ghent University; Ghent, Belgium

rSigma Stat 3.5, Systat Software Inc, San Jose, CA


This work was supported by grants from USDA/CSREES (#2007-35204-18563) and by Conselho Nacional de Desenvolvimento Científico e Tecnológico (Faleiros' scholarship).