Portions of this manuscript were presented at the 2009 American College of Veterinary Internal Medicine Forum, Montreal, CA, 2009. Horses were housed and samples were collected at University of Missouri, Columbia, MO and laboratory analysis were conducted at Ohio State University, Columbus, OH.
Corresponding author: Dr J.K. Belknap, Department of Veterinary Clinical Sciences, Ohio State University, 601 Vernon L Tharp St, Columbus, OH 43210; e-mail: email@example.com.
Background: While there is evidence of laminar leukocyte infiltration in black walnut extract (BWE)-induced laminitis, there is no such evidence for carbohydrate overload (CHO) laminitis.
Objective: To assess presence of leukocytes and signs of epidermal stress/injury in the laminar tissue from horses with CHO-induced laminitis.
Animals: Twenty-four adult horses.
Methods: Immunohistochemistry for myeloid cell markers calprotectin (CP) and monocyte-specific marker (CD163) was performed on laminar sections obtained from 2 groups of horses in the CHO model: the developmental time point (DTP) group (n = 6) and the onset of lameness (LAM) group (n = 6), and a control (CON) group (n = 8).
Results: DTP was characterized by an increase in CP+ leukocytes (7.8-fold increase versus CON, P < .001), and LAM time point was characterized by a more marked increase in laminar CP+ (108.5-fold, P < .001) and mild increase in CD163+ (1.9-fold, P= .007) cell counts. Increased CP epidermal signal (indicating epidermal stress or injury) occurred consistently at the LAM time point, although histological evidence of basement membrane (BM) detachment was minor, only being present in 3/6 horses.
Conclusions and Clinical Relevance: Maximal laminar leukocyte infiltration and epithelial stress occurred at the onset of lameness in the CHO model showing a different temporal pattern from the BWE model, where maximal leukocyte infiltration clearly precedes epithelial stress. Leukocyte infiltration before major histological changes in the CHO model indicates that leukocyte infiltration can be a cause of and not a reaction to BM degradation and structural failure.
Leukocytes play a central role in inflammatory injury of tissues and organs in multiple diseases ranging from degenerative joint disease to the systemic inflammatory response and associated organ failure observed in human sepsis.1–5 Equine laminitis involves a similar remote tissue injury (remote from the source of sepsis) in horses suffering from bacterial sepsis as occurs in the target organs such as the lungs and kidney in human sepsis. Because leukocytes play such an important deleterious role in inflammatory injury in other disease states, the role of leukocytes in pathogenesis of laminitis has been a focus of laminitis research for the past decade.6–8
The 2 primary model systems to study equine laminitis incorporate the nasogastric administration of either black walnut extract (BWE) or a “carbohydrate overload” (CHO) in which either a corn starch-wood flour gruel mixture or oligofructose is administered.9,10 The BWE model is not only characterized by a rapid onset of systemic inflammation as seen both in human sepsis and recently reported in an equine model of intestinal obstruction,11–14 but also by a rapid local laminar inflammation.6,15 However, the BWE model results in less severe clinical signs and minimal laminar pathologic changes when compared with the CHO models, which undergo a similar severity of laminar injury as the clinical case of laminitis.11
Although leukocyte emigration into the laminar interstitium has been well-established in the BWE model,15–18 the role of leukocytes in laminar failure in the CHO model has been questioned because of the fact that, although laminar neutrophils can be easily identified in BWE-treated animals, leukocytes are not readily apparent in laminar tissues in the early stages of laminitis in the CHO model with routine immunostaining (ie, hematoxylin and eosin [H&E] staining).19 This lack of histological evidence of inflammatory cells in the 1st stages of naturally occurring and CHO-induced laminitis led to the postulation that equine laminitis is not inflammatory but basically a degenerative disease.20 Additionally, while a marked decrease in circulating white blood cell counts, one of the criteria used for the diagnosis of systemic inflammatory response syndrome, occurs in the developmental period in the BWE model, the same decrease in circulating leukocyte counts has not been reported at any developmental/prodromal time point in the CHO laminitis model.21,22
Thus, because of both (1) the important role of leukocytes in tissue injury in multiple inflammatory diseases, and (2) the similar degree of laminar injury in the CHO model as seen in clinical laminitis cases, we were interested in assessing the presence within the laminar tissue of different types of leukocytes in the traditional CHO model of laminitis. Because of the reported lack of leukocyte accumulation before laminar basement membrane (BM) injury, we also used a grading scale similar to that established by Pollitt19 to assess laminar morphology/injury in order to establish the temporal relationship of leukocyte emigration into the laminar tissue with BM injury in the CHO model. Because of the fact that neutrophils are not readily observed histologically in the CHO model, we used antibodies that are not only effective against neutrophils, but highly effective against different phenotypes of monocytes.
Material and Methods
Immunohistochemistry (IHC) was performed on paraffin-embedded laminar samples obtained from the following groups using the traditional CHO model (corn starch/wood flour) of laminitis: a control (CON) group (samples harvested 24 hours postnasogastric administration of water, n = 8), the developmental time point (DTP) group (onset of fever, 10–20 hours post-CHO administration, n = 8), and the onset of lameness (LAM) group (20–46 hours post-CHO administration, n = 8). Routine H&E and periodic acid-Schiff (PAS) histological stain techniques were performed in order to assess morphology of laminar keratinocytes and BM. IHC was performed for calprotectin (CP) in order to identify neutrophils, classically activated monocytes/macrophages and stressed keratinocytes.18 IHC was also performed for CD163 for further identification of laminar monocytes/macrophages (which may not be positive for CP). All image analysis was performed with the observer blinded to the group to which the laminar sections belonged.
CHO-Induced Laminitis and Sample Collection
Twenty-four clinically normal horses with a median body weight of 421 kg (ranging from 341 to 524 kg,) and a median age of 5 years (ranging from 3 to 12 years old) were used. All horses were determined to be healthy (body condition scores of 4–6) and free of digital disease by physical and lameness examinations and radiographic evaluation of the forelimb distal phalanges. Before the experiment, the horses were quarantined for 2 weeks at the University of Missouri, housed in stalls, and fed a regular ration of free choice hay. Horses were randomly allotted in 3 groups (each with 8 subjects), the CON group, the developmental time-point group (DTP), and the onset of LAM group.
A carbohydrate gruel consisting of 85% cornstarch and 15% wood flour (17.6 g/kg body weight) was administered to each horse in the DTP and LAM groups by nasogastric tube as described previously.21,22 Six liters of deionized water was administered via nasogastric tube to each horse in the CON group. All horses received complete physical examinations (consisting of rectal temperature, heart rate, respiratory rate, abdominal sounds, digital pulses, evaluation with hoof testers) and gait evaluation (horses were walked and trotted in straight lines and walked in circles in each direction immediately after nasogastric tube passage, and at 2-hour intervals after administration of either carbohydrate or water). Blood was taken from jugular catheter at 0, 2, 4, 8, 12, and 16 hours in all horses, and every 4 hours after 20 hours in the LAM group until onset of Obel grade 1 (OG1) lameness. Anesthesia was induced in each horse of the DTP group at onset of fever (≥102°F), and in horses from the LAM group at the onset of lameness, indicated by shifting leg, lameness when standing, and demonstrating a short, stilted gait at trot. Control horses were anesthetized at 24 hours after administration of water.
Although classically a 20% drop in central venous pressure (CVP) is often used to determine the developmental time point in the CHO model,23 fever was used instead because of the fact that (1) CVP can be variable with poor repeatability between multiple observers, and (2) fever, a much more repeatable and accurate variable to measure, was found to occur at approximately the same time as the drop in CVP (reported previously between 10 and 16 hours)23 upon review of previous CHO protocols at University of Missouri (P. Johnson, unpublished data).
Anesthesia was induced with a combination of xylazine (1 mg/kg of body weight IV) and ketamine (2.2 mg/kg of body weight IV). Horses were placed in lateral recumbency and anesthesia was maintained with isoflurane. While under general anesthesia, the feet were removed and processed one at a time (after placement of a tourniquet) by disarticulation of the metacarpophalangeal joint, and 1 cm thick sagittal sections of the digit were cut with a band saw. The laminae were quickly dissected from the hoof and 3rd phalanx, cut into sections (approximately 1 cm × 3 mm × 3 mm), and immediately immersed in neutral-buffered 10% formalin for 24 hours, followed by placement in 70% ethanol until paraffin embedding occurred. For each horse, 3 samples were embedded together in the same paraffin block. Once the tissues were collected, each horse was subjected to euthanasia with 0.22 mL/kg (IV) of a solution containing pentobarbital sodium (390 mg/mL) and phenytoin sodium (50 mg/mL). All protocols were approved by the Institutional Animal Care and Use Committees of The Ohio State University and University of Missouri.
Formalin-fixed tissues were embedded in paraffin, sectioned at 5-μm thickness and stained with routine H&E and PAS techniques. H&E-stained sections were used to evaluate laminar cells and structures, especially laminar keratinocytes. PAS staining was used to evaluate abnormalities associated with the laminar BM. The histopathological grading system previously described by Pollitt19 was modified (presented in Table 1) including 3 new grades (1–3) in order to better assess the progressive changes that occurred in the laminar morphology during early stages (before disseminated BM collapse). In this new system, the former 1–3 grades are represented for grades 4–6. All sections were blindly evaluated and a grade from 0 (normal) to 6 (advanced BM detachment from keratinocytes) was given for each section based on signs of keratinocyte and BM degradation/injury.
Table 1. Grading system to evaluate histopathological changes in equine laminitis, modified from Pollitt.19
Basal layer keratinocytes with elliptical nuclei, at right angle to BM. Secondary epidermal laminae (SEL) with rounded tips. The apex of the secondary dermal laminae (SDL) is close to the keratinized axis
Up to 50% of the basal layer keratinocyte nuclei are rounded, but no other evident changes
Most of the basal layer keratinocyte nuclei are rounded, but no other evident changes
Rounded keratinocyte nuclei, and up to 50% of the SEL tips are elongated presenting a pointed tip and BM detachment from basal layer cells
4 (Pollitt's grade 1)
Rounded keratinocyte nuclei, most of the SEL tips elongated presenting a pointed tip and BM detachment from basal layer cells
5 (Pollitt's grade 2)
Further elongation of the SEL and increased distance from the apex of the SDL to keratinized axis characterizing the progression of the BM detachment
6 (Pollitt's grade 3)
Most of the SEL is separated from the SDL; there is breakdown of the BM and cellular disorganization in the SEL
The paraffin-embedded samples were sectioned at 5-μm thickness for general immunostaining and at 3-μm thickness when serial sections were stained with CP and with CD163. Immunohistochemical staining using a universal alkaline phosphatase red detection technique was performed in all sections for detection of CP. These procedures were performed with the assistance of an automated slide processing system.a The sections were deparaffinized, treated with protease solution (0.5 protease U/mL,b 4 minutes), then incubated with a mouse monoclonal anti-human CP antibodyc (1 : 250, 37°C, 96 minutes). The immunoreactivity was observed using a detection kit containing 1 alkaline phosphatase-conjugated secondary antibodyd following the instructions of the manufacturer, and hematoxylin was used for the counterstain. In the serial sections, diaminobenzidine (DAB) was used as a chromogen. Sections of spleen and lung (from CON horses in current study) were stained following the same protocol as positive controls as previously described.18,24 Human anti-CD68e primary antibody (an antibody that does not recognize an equine epitope) was used in the same protocol in the negative control sections.
For CD163 imunohistochemistry, the avidin-biotin complex method was performed with the assistance of an automated slide processing system.f After removal of the paraffin, sections were pretreated with an antigen retrieval solutiong under controlled heat and pressure (125°C, 20 minutes using a decloaking chamberh). After endogenous peroxide quenching (hydrogen peroxide 3%, 5 minutes) and incubation with serum-free protein block,i sections were incubated initially with CD163 primary antibodyj (37°C, 30 minutes) and subsequently with the biotinylated secondary antibodyk (22°C, 30 minutes) and the avidin-horse radish peroxidase complex (22°C, 30 minutes). Color was developed with chromagen (DAB, 5 minutes) and hematoxylin was used for the counterstain. Sections of lung were stained following the same protocol as the positive controls. Inactivated CD163 primary antibody was used in the same protocol in the negative control sections.
All slides were examined by light microscopy in order to identify the presence and location of stained cells. Whole slide digital images (WSIs) were obtained from stained sections using an automated scanning robotl with a 40 × magnification at a spatial sampling period of 0.2 μm per pixel. WSIs were assessed using a software programm to randomly capture images from each section, which were transformed into jpg files. For CP stained sections, fields with a 10 × magnification (randomly selected 5 from the sublaminar dermal area and 5 from the laminar area) were used, and, for CD163 sections, fields with a 20 × magnification (10 from the dermal area and 10 from the laminar area) were used. CP+ and CD163+ leukocytes were manually counted in each image and the mean number of cells/mm2 for each section was determined with the assistance of a software for image analysesn; the operator was blinded to the origin of the section (ie, control versus principal).
During the counting process, the positive cells in the laminar region were classified in accordance with their localization: primary dermal laminae, secondary dermal laminae (SDL), and epidermal laminae (EL, primary, and secondary EL were counted together because the low cellular number in this layer). In the deep dermis (laminar corium), CP+ leukocytes were classified as intravascular (defined as being within the limits of the endothelium), perivascular (within or immediately around the vessel walls) and in the dermal tissue spaces (the connective tissue between perivascular interstitium and BM of EL; denoted “tissue” in graphs) as previously described.18 The CD163+ cells were classified only as intravascular (within the limits of the endothelium), or perivascular (inside the connective tissue around the vessel walls) as previously described.24
A grading system (0–5) was used to score the IHC signal of CP in EL of each sample as previously described.18 The sections were blindly scored from 0 to 5 : 0) no signal, (1) spots of localized signal, (2) presence of areas of diffuse signal, but <25% of the EL; (3) presence of areas of diffuse signal in 25–50% of the EL; (4) presence of areas of diffuse signal in 50–75% of the EL; and (5) intense and diffuse signal in more than 75% of the area of the EL. As each slide assessed had 3 sections of laminae, the sections were individually scored and a mean grade was used for raw statistical analysis.
Parametric data were presented as mean ± SEM. The statistical analyses were performed using specific software.o Parametric data were analyzed by one-way ANOVA; log transformation was used for parameters that did not exhibit normal distribution. Nonparametric data were analyzed by Kruskal-Wallis test. Posthoc comparisons were made using the Student-Neuman-Keuls test to compare the means. Significance was set at P < .05 for all tests.
Six of 8 horses in the DTP group developed fever (≥102°F) between 10 and 20 hours after CHO administration. Six of 8 horses in the LAM group developed signs of OG1 lameness, which occurred between 20 and 48 hours post-CHO administration. The 2 horses that did not develop fever and the 2 that did not develop lameness were considered nonresponders from the DTP and LAM groups, respectively, and were excluded from statistical analyses.
Although changes in keratinocyte morphology (rounded nuclei in basal cells) were found in laminar sections from the CON and DTP groups, there was no sign of BM abnormality in any horse in these groups. In the LAM group, 3/6 horses had signs of BM abnormality; with 2 grade 3 (up to 50% of the SEL tips elongated presenting a pointed tip and BM separation from basal layer cells) and 1 grade 4 (majority of SEL tips with the same abnormality). Grades for epidermal changes in LAM group were greater (P= .004) than the CON and DTP groups (Fig 1), which were not different from each other (P= .37).
CP IHC: Laminar Epidermal Staining
Laminar sections from all horses in the CON and DTP groups exhibited some degree of CP staining (grades 1–3, Fig 1) in laminar epidermis (especially suprabasal keratinocytes of the SEL, Fig 2), but there was no difference in staining between these 2 groups. CP epidermal stain score was greatly increased (P= .003, Fig 1) in the LAM group compared with other groups. The CP signal was diffuse in the EL in the LAM group, with both basal and suprabasal cells staining (Fig 2). Interestingly, there was variation in the CP epidermal stain in sections from the same individual in 4 of the 6 “responders” in the LAM group. As an example, in 1 horse one of the 3 sections was graded 0 whereas the others were graded 4 and 5 (Fig 3).
CP IHC: Laminar Leukocyte Staining
In CON horses, rare CP+ leukocytes were found in laminar region (0.4 ± 0.3 cells/mm2) and deep dermis (0.4 ± 0.1 cells/mm2). In the DTP group, there was a discrete but significant (P < .001) increase in CP+ leukocytes counts in laminar region (primary and SDL and EL combined, 3.1 ± 2.1 cells/mm2) and deep dermis (5.4 ± 2.2 cells/mm2) compared with CON group (Fig 4). In the LAM group, counts of CP+ leukocytes in laminar region (43.4 ± 15.7 cells/mm2) and deep dermis (72.7 ± 20.3 cells/mm2) were greatly increased (P < .001) compared with the CON (108.5- and 181.2-fold, respectively) and DTP (14- and 13.5-fold, respectively). When primary and secondary laminae were counted separately, no changes in CP+ leukocytes occurred at the DTP (compared with CON), but increased CP counts occurred at LAM time point in primary and secondary dermal layers of the laminae and also in all areas of the deep dermis (intravascular, perivascular, and interstitial spaces, Figs 5 and 6).
A large number of CD163+ cells were present in the laminar region (98.9 ± 14.9 cells/mm2) and deep dermis (207 ± 27.6 cells/mm2) in the clinically normal horses (CON group). No significant changes were observed in the DTP (Fig 4) in either the laminar region (104 ± 16.1 cells/mm2) or deep dermis (190 ± 27.3 cells/mm2). However, in the LAM group, counts of CD163+ monocytes/macrophages in laminar region (185 ± 21.3 cells/mm2) and deep dermis (417 ± 60.7 cells/mm2) were increased (P < .01) compared with CON (1.9- and 2.0-fold, respectively) and DTP (1.8- and 2.2-fold, respectively) groups. In the stratified counts, increases in CD163+ counts occurred in primary laminar (P= .005) but not in the secondary laminar dermal layers of laminar region in the LAM group (Fig 5). In deep dermis, increases in the CD163+ cells were observed at vascular (P= .03) and perivascular (P= .01) regions (Fig 6).
Serial sections of laminar tissue stained with CP and CD163 antibodies demonstrated that some monocytes/macrophages were positive for both CP and CD163, whereas many cells were only CD163+ (Fig 7).
In this study, we have documented that similar to the BWE model,18,24 there is an increase in number of CP+ and CD163+ leukocytes in laminitis induced by CHO before major changes in structure of the laminae. Despite this similarity, there were differences in leukocyte recruitment into the laminae between the 2 models with respect to both type and timing of infiltration. The 2 markers used in this study, CP and CD163, are likely to identify the majority of myeloid cells in the interstitium, with CP marking most neutrophils and activated macrophages,25 and CD163 detecting macrophages of the type 2/“anti-inflammatory” phenotype24 and possibly also other activated macrophages in the horse.24,26,27
Normal digital laminae contained few CP+ cells but a moderate number of CD163+ cells, indicating the presence of a relatively large pool of tissue macrophages. Whereas BWE administration induced significant increases in laminar CD163+ cells counts during early developmental stages that returned to basal values at the onset of lameness,24 CHO administration only induced an increase in CD163+ cells at the onset of lameness. In addition, while BWE administration induced a massive increase in laminar CP+ cell numbers at 2 developmental time points (53-fold increase at 1.5 hours post-BWE administration and 151-fold at 3–4 hours),18 only a modest 8-fold increase in CP+ cells was detected at the developmental stage of laminitis following CHO administration, which was marked by the onset of fever in the current study. However, at the onset of lameness, similar numbers of CP+ cells are detected in the laminae in both the BWE and CHO models, reflecting a decreasing number of CP+ cells in the BWE model and an increasing number of CP+ cells in the CHO model (59.6 ± 11.6 cells/mm2 in the BWE model,18 and 43.3 ± 6.4 cells/mm2 in current CHO model). The increase in CP+ cells might be because of an extravasation of CP+ cells into the laminar interstitium, a change in phenotype of CD163+ macrophages to an activated CP+ phenotype, or both. The fact that few cells express both CD163 and CP proteins could be due to the possibility that these cells only represent a transient, intermediate phase as the cell changes phenotype. The increase in both CP+ and CD163+ cells at the onset of lameness indicates that there is an influx of both types of cells; the CD163+ cell count would have been expected to decrease if the marked increase in CP+ cells at the LAM time point were only from activation of CD163+ tissue macrophages to a more activated CP+ phenotype. Overall, whereas the myeloid inflammatory cell influx is early but transient in the BWE model, it occurs more slowly and might be more sustained in the CHO model.
The temporal relationship of leukocyte influx and injury to the epidermal/dermal laminar interface, the main point of failure in laminitis, might indicate whether the leukocytes play an initiating role in injury to the interface, or only respond secondarily to injury caused by other events.19,28,29 Two indicators are used to score laminar pathology, namely, structural abnormalities of the laminar basal epithelial cells (the cell type present at the site of laminar dysadhesion) and separation of the BM from the basal epithelial cells accompanied by its stretching into the dermal laminae presumably pulled there by attached collagen fibers.19 Using these indicators of tissue injury, we found that, although marked leukocyte infiltration was detected in all horses that developed OG1 lameness, only 50% of these animals had any histologic changes in the BM. Furthermore, in the 3 animals in which changes to the dermal/epidermal interface were detected, the histologic changes were mild. The development of OG1 lameness accompanied by infiltrating leukocytes in the laminae of horses that showed no signs of BM pathology, which was observed in 3 of 6 horses in the present study, is consistent with leukocyte immigration preceding and perhaps causing damage to the BM and basal epithelial cells. However, in the present study, BM degradation was assessed only by light microscopic histological evaluation; consequently, our findings do not exclude the possibility that minor changes in the BM or other components of the extracellular matrix not detectable by light microscopy occur before the increase in CP+ cells as has recently been suggested.30 If so, these changes may play a role in tissue macrophage activation and/or recruitment of circulating leukocytes out of the vessel and toward the damaged BM.
To further examine the relationship between leukocyte infiltration and pathological changes in the epidermal/dermal interface, we assayed CP expression by the laminar basal epithelial cells of healthy and CHO treated animals. In a similar temporal pattern as that previously described for the BWE model, intense CP staining of basal epithelial cells only occurred at the onset of lameness, again following the increase in CP+ cells and coincident with CD163+ cell increase. These results suggest that the infiltration of CD163+ leukocytes can play a role in laminar basal epithelial cell injury/stress and, or, that the same stimulus induces both CD163+ leukocyte infiltration and laminar interface injury.
In conclusion, leukocyte infiltration in the CHO model occurs before severe histological changes consistent with the possibility that it can be a cause of and not a reaction to major BM degradation and structural failure. Additionally, the increased concentration of macrophages/monocytes at the onset of lameness in the CHO model indicates that these cells can cause, directly or indirectly, the structural failure commonly seen in CHO-induced laminitis.
a Discovery XT Systemb, Ventana Medical Systems, Tucson, AZ
b Protease 1, Ventana Medical Systems
c MAC387, Abcam, Cambridge, MA
d Ultra View Red, Ventana Medical Systems
e HAM56, Enzo Life Sciences, Plymouth Meeting, PA
f Dako Autostainer, Dako, Carpinteria, CA
g Target Retrieval Solution, Dako
h Decloaking chamber, Biocare Medical, Concord, CA
i Serum-Free Protein Block, Dako
j AM-3K, CosmoBio, Tokyo, Japan
k Biotinylated anti mouse IgG, Vector Laboratories, Burlingame, CA
l T2 Scan Scope, Aperio Technologies, Vista, CA
m Image Scope, Aperio Technologies
n Image J, National Institute of Health, Bethesda, MD
o Sigma Stat 3.5, Systat Software Inc, San Jose, CA
Supported by grants from USDA/CSREES (#2007-35204-18563) and by Conselho Nacional de Desenvolvimento Científico e Tecnológico (Faleiros' scholarship). We are thankful to Ventana Medical System, which provided the detection reagents for immunohistochemistry as a gift.