The study of laminitis in horses has been hampered by the inability to serially characterize the progression of histologic changes and molecular events in the laminae of an individual affected horse. Histologically, stretching of the lamellae, dermal edema, hemorrhage, changes to the basal cells, presence of WBCs in the dermis, and signs of basement membrane detachment have been observed in specimens collected from horses euthanized during different stages of acute laminitis.1–4 Molecular events are similarly altered during the course of disease in experimentally induced laminitis.5–7 However, substantial variation from horse to horse evaluated within the same time frame has confounded attempts to characterize the sequence of early events without euthanasia of large numbers of experimental horses at various time points. Serial sampling of an individual horse would allow study of the patterns of events in early laminitis that may proceed at different rates among horses.
Laminar biopsy specimens have been collected through the hoof wall in standing animals.8–10 In cattle, tissues have been harvested from the sole-heel junction.9,10 Single laminar biopsy specimens have been collected in standing horses by use of a technique whereby the hoof wall was thinned with a drill and the biopsy specimen was obtained by cutting a square of tissue with a scalpel blade.8 The effects of collection of serial laminar biopsy specimens in horses have not been reported.
In laminitis, skin distant from laminae has molecular, biochemical, and leukocyte migration changes similar to laminae.11–13 Events in the skin appear to parallel events occurring in the relatively inaccessible lamellar tissue in experimental models.12 Results of a recent study14 indicate a lack of superoxide dismutase in equine laminar tissues, suggesting a tissue-specific susceptibility to damage by inflammatory leukocytes during laminitis. These recent findings support continued investigation into the similarities and differences between these closely related tissues.
We hypothesized that multiple biopsy specimens could be harvested from a single hoof during a period of 24 hours for morphologic examination and RNA recovery without inducing undue pain or causing morphologic changes in adjacent tissues that would be biopsied at later times. The purpose of the study reported here was to determine the feasibility of performing serial laminar and skin biopsies on an individual horse and whether sampling affected adjacent tissues.
Materials and Methods
Six healthy adult horses (body weight, 440 to 545 kg) with no preexisting lameness were used in accordance with a protocol approved by an institutional animal care and use committee. Horses were selected from the Equine Health Studies Program herd. Horses were sedated with xylazine hydrochloridea (125 mg) and acepromazineb (5 mg) administered IV. Detomidinec (5 mg) was administered IV if necessary for additional sedation. Horses were restrained in stocks during the biopsy procedures. The metacarpophalangeal (fetlock) and proximal phalangeal (pastern) areas were clipped of hair and aseptically prepared. Medial and lateral abaxial sesamoid digital nerve blocks were performed with mepivacaine hydrochlorided at each collection time point. An elastic bandagee was tightly applied from the midpastern region to above the fetlock joint to provide hemostasis.
Laminar specimens were collected from the dorsum of the hoof at 3 cm and 5 cm distal to the coronary band at approximately 3-cm intervals parallel and equidistant from the coronary band (Figure 1). A 1-cmdiameter area of the hoof wall at each biopsy site was thinned with a drillf until approximately 1 mm of stratum medium remained, as determined by evaluation of movement in response to pressure applied with a hemostat. A disposable 6-mm-diameter biopsy punchg was used to cut into the laminae to the level of the distal phalanx. Care was taken during biopsy to ensure the distal phalanx was not damaged with the biopsy punch (Figure 2). The biopsy specimen core was elevated with a probe and removed with a scalpel.h The ease of biopsy specimen removal from the site was recorded. Biopsy sites were packed with sterile, dry gauze and secured with tape. Biopsy sites were bandaged until the biopsy sites were dry. Horses were evaluated prior to each specimen collection and daily for signs of lameness by use of the grading scale recommended by the American Association of Equine Practitioners (score, 0 to 5). Phenylbutazonei (2 g once daily) was administered after the 24-hour specimen collection. After cornification of the hoof biopsy sites, an epoxy polymerj was applied to the sites. The horses were maintained on pasture and monitored and received regular hoof trimming until the hoof biopsy sites healed completely. Full-thickness skin specimens were collected 1 cm and 3 cm proximal to the coronary band with a disposable 6-mm-diameter biopsy punch. Skin biopsy sites were sutured with 2-0 nylon in a single cruciate pattern.
Illustration of the locations of biopsy specimens obtained from the digit of horses. Hashed line indicates the sagittal plane of the digit. (Illustration provided by Michael Broussard).
Citation: American Journal of Veterinary Research 70, 6; 10.2460/ajvr.70.6.697
Photograph of a longitudinal cross-section of a horse's foot. Notice locations where biopsy specimens were obtained. Lines through the hoof wall indicate areas removed with a rotary drill. Black boxes represent laminar tissue removed with a 6-mm-diameter biopsy punch. Care was taken to not disturb the periosteum of the distal phalanx (P3); P1 and P2 indicate the proximal and middle phalanx, respectively.
Citation: American Journal of Veterinary Research 70, 6; 10.2460/ajvr.70.6.697
Study design—Laminar and skin biopsy specimens were harvested at 0, 6, 12, and 24 hours. Two biopsy specimens of each type were collected at each time interval. One biopsy specimen was placed in neutral-buffered 10% formalin for routine histologic processing. The other biopsy specimen was immediately preserved at −70°C in a monophase solution of phenol and guanidine thiocyanatek until processing for extraction of RNA.
Histologic evaluation—Tissue specimens were embedded in paraffin, sectioned, and stained with H&E and periodic acid–Schiff stains. Lamellar structures were examined by an experienced veterinary pathologist masked to sample time point and evaluated for inflammatory changes and basement membrane characteristics.1 Skin structures were evaluated similarly for inflammatory changes.
Molecular techniques—Total cellular RNA was isolated via chloroform extraction as per the manufacturers' instructions. The RNA was suspended in 0.1% diethyl pyrocarbonate in water,l analyzed via agarose gel electrophoresis, and quantitated via spectrophotometry. All RNA samples (1.2 μg) were reverse transcribed to cDNA by use of murine Moloney leukemia virus reverse transcriptasem at 40°C for 1 hour.15 This cDNA was used as template and quantified via real-time reverse transcriptase PCR assay by use of a sequence detection system.n To determine concentrations of mRNA of the housekeeping genes β-gus and β-actin in laminar biopsy tissues, 2 primers and a probe for each gene were used. The probes were labeled with a reporter dye (6-carboxyfluoresceine) and a quencher dye.o β-Gus primer and probe sequences were provided.p The β-actin primers and probe were designed from the National Center for Biotechnology Information published gene sequence (accession No. AF035774 1,128-bp mRNA linear MAM 07-APR-2000; definition, Equus caballus β-actin mRNA; complete coding sequences) by use of software.q The primers and probe for equine COX-2 were designed from the National Center for Biotechnology Information published gene sequence (accession No. AB041771 1,267-bp mRNA linear MAM 20-APR-2000; definition, Equus caballus COX-2 mRNA; partial coding sequences) by use of software.
A relative standard curve was constructed from equine laminar tissue by serially diluting (1:1) mRNA, then reverse transcribing it to cDNA. The standard curve for COX-2 expression was constructed from liver specimens from horses with acute laminitis secondary to gastrointestinal tract disease. The highest point on the standard curve containing 1,200 ng of total RNA was designated as a relative value of 1,024 units. The first dilution containing 600 ng of total RNA was designated as a relative value of 512 units. The dilutions were continued until 11 samples were prepared, with the lowest sample designated as a relative value of 1 unit. Cycle threshold values were obtained for β-actin, β-gus, and COX-2 in all tissue specimens as well as for the standard curve. Linear regressions that used the standard curves were carried out for the β-actin, β-gus, and COX-2 cycle threshold values as described.16,17 The interpolated values for β-actin, β-gus, and COX-2 mRNA concentrations were determined by use of the derived linear regression. Values for β-actin and β-gus were used to standardize the values for COX-2.
Data analysis—Histologic data are reported descriptively. Normality of the gene expression data (expressed as copy units) was checked by use of the univariate procedure of a software program.r The ANOVA mixed procedure was used to determine significant (P ≤ 0.05) differences in β-actin, β-gus, or COX-2 mRNA concentrations over study time points. If significant differences were observed, post hoc comparisons were made by use of the Tukey test to maintain type I error at 0.05. The COX-2 gene expression copy unit values were compared with baseline COX-2 values obtained at time = 0 hours.
Results
Adequate laminar and skin biopsy specimens were obtained from all horses for histologic and molecular biological analysis. During the biopsy procedures, medial and lateral abaxial sesamoid digital nerve blocks provided adequate analgesia, and horses did not have signs of pain during specimen collection. Application of the elastic bandage to control hemorrhage was required to maintain sight of the laminar biopsy specimen during retrieval from the hoof wall. The degree of lameness prior to each sampling period during the first 24 hours ranged from grade 0 to 1 for all horses. Five of the 6 horses had intermittent grade 1 lameness at subsequent daily evaluations for 5 days following biopsy. Phenylbutazone (1 g), administered on the basis of assessment of lameness each day following biopsy, alleviated signs of lameness in the 5 horses. One horse had a grade 1 lameness for 3 weeks. This horse had poor hoof quality prior to initiation of the study but fully recovered from lameness induced from the procedure and did not have any further lameness for 12 months after completion of the study.
The defects in the hoof wall at the biopsy sites dried and cornified within a week after collection. The hoof wall grew out without long-term ill effects or cracking of the hoof wall adjacent to the biopsy site for 12 months. After conclusion of the study, horses were returned to the Equine Health Studies Program herd.
Histologic evaluation—Full-thickness biopsy specimens were obtained during all time points for skin and laminae. There was no evidence of progressive inflammation or damage at adjacent laminar tissue or skin collected at subsequent time intervals as a result of the procedure. In 4 laminar specimens, damage to the tissues occurred because of tissue handling during removal from the biopsy site. This problem became less of an issue as experience was gained with the technique. Proper orientation of the laminar biopsy specimens during paraffin embedding required careful attention; a specimen obtained at time = 6 hours from 1 horse became unusable because of improper orientation during embedding and subsequent sectioning. As with biopsy specimen collection, experience with the embedding of the specimens resolved this issue.
Histologic examination of the laminar tissues with H&E and periodic acid–Schiff stains revealed there was no increase during the 24-hour period with respect to signs of inflammation or basement membrane damage in the laminar tissues when the biopsy specimens were easily harvested (Figure 3). The degree of difficulty in removing the laminar tissue from the biopsy site seemed to be related to the degree of inflammation or structural damage. The histologic findings for laminar biopsy specimens that were difficult to remove were characterized by torn segments of tissue, mild endothelial swelling, mild edema, and the presence of WBCs (Figure 4). The morphology of laminar tissue differed from dorsal to abaxial locations with slight caudally directed bends in the tips of laminae present abaxially.
Photomicrograph of the laminae of a horse's digit. Notice normal histologic appearance after biopsy specimen collection. H&E stain; bar = 150 μm.
Citation: American Journal of Veterinary Research 70, 6; 10.2460/ajvr.70.6.697
Photomicrograph of the laminae of a horse's digit. Notice moderate inflammatory infiltrate located between and near the tips of the secondary epidermal laminae after biopsy specimen collection. H&E stain; bar = 150 μm.
Citation: American Journal of Veterinary Research 70, 6; 10.2460/ajvr.70.6.697
No progressive inflammation or damage in the dermal or epidermal layers was detected in skin biopsy specimens. In 3 horses, mild edema and mild endothelial cell rounding were present in the superficial layers of the skin at all time points, whereas dermal tissues from these 3 horses were histologically normal during the 24-hour period.
Molecular techniques—Mean ± SD amount of total RNA obtained from the laminae biopsy specimens was 369 ± 274 μg/mL. Mean ± SD amount of total RNA obtained from the skin was 332 ± 380 μg/mL. The RNA extracted from the specimens was of high quality as judged on the basis of gel electrophoresis that revealed clear, distinct bands. Laminar and skin housekeeping gene data for β-actin and β-gus had normal distributions. Throughout the course of the 24-hour sampling period, there were no significant differences in expression of the 2 housekeeping genes in the laminar specimen β-actin (P = 0.3) and β-gus (P = 0.150) or in the skin specimen β-actin (P = 0.227) and β-gus (P = 0.150) over time. Gene expression of COX-2 was present in the standard curve samples, but no or low expression was found in the biopsy specimens (Table 1). There was no significant effect of time on COX-2 expression in the laminar or skin biopsy specimens. Compared with baseline specimens, expression of COX-2 differed by factors of 1.61, 0.079, and 0.655 for 6-, 12-, and 24-hour laminar specimens, respectively. In skin specimens, expression of COX-2 differed by factors of 2.55, 0.60, and 0.03 for 6-, 12-, and 24-hour specimens, respectively.
Mean ± SEM copy units for gene expression of COX-2 in laminae and skin from biopsy specimens collected at 4 time points during a 24-hour period in horses. Values for laminae and skin did not change significantly (P = 0.302 and P = 0.351, respectively) during the study.
| Time of biopsy (h) | Laminae | Skin |
|---|---|---|
| 0 | 3.83 ± 1.80 | 2.50 ± 1.12 |
| 6 | 6.17 ± 2.85 | 5.58 ± 2.86 |
| 12 | 0.38 ± 0.19 | 1.25 ± 0.67 |
| 24 | 2.51 ± 1.53 | 0.06 ± 0.03 |
Discussion
The laminar biopsy collection method used in the present study allowed for consistent collection of tissue in conscious sedated horses. Each biopsy specimen was of sufficient quality and quantity for histologic evaluation and gene expression analysis.
Medial and lateral abaxial sesamoid digital nerve blocks provided adequate analgesia during the tissue collection procedures, and lameness was not detected immediately after dissipation of regional anesthesia following each biopsy. The procedure as described did not induce signs of pain, with mild lameness detected after 24 hours in some horses; however, this was alleviated for at least 24 hours after the administration of phenylbutazone. One horse with poor hoof quality initially had mild lameness that persisted for 3 weeks, suggesting that hoof quality may be an important factor in selection of subjects for use of this technique.
An important finding of this study was that serial sampling did not result in detectable progressive inflammation or damage to the adjacent tissues as judged on the basis of histologic evaluation and COX-2 gene expression. Physical damage to the laminar tissues occurred 4 times and was associated with extraction of the biopsy specimen. As skill with the technique was developed, physical distortion was minimized and histologic specimens were adequate to assess morphology.
Proper orientation of the laminar biopsy specimens during paraffin embedding was difficult and required re-embedding of several specimens early in the study. This problem was eliminated by the addition of a foam insert in the specimen cassette to maintain correct orientation during paraffin embedding. Laminar morphology varied from dorsal to abaxial specimens. Douglas and Thomason18 described equine laminar histologic morphology in various locations and found that the laminae in the region of the quarters are shaped and oriented differently than the laminae in the dorsal aspect of the foot. The laminae in the dorsum of the foot are spaced closer and oriented 90° relative to the hoof wall, which is believed to position the laminae along the primary line of force in a clinically normal horse. The laminae in the quarter regions have different morphology because the hoof wall encounters different forces than along the dorsum of the foot. The authors suggest that these morphologic differences should be considered when examining specimens from various regions of the foot if morphology is critical to the study.
Histologic examination of the skin did not reveal progressive inflammation or damage in the dermal or epidermal layers caused by the serial biopsy procedure. Skin (including chestnut and ergot) alterations during the prodromal stages of laminitis has been hypothesized by researchers to parallel alterations in the laminae.13,19 Experimental laminitis models have been studied in efforts to pinpoint potential similarities between laminar tissues and skin that may reflect common inflammatory, vascular, enzymatic, and mechanical changes ultimately leading to laminar structural failure.14 Results of a recent study14 suggest that the laminar tissues may in fact have different physiologic features from the skin, making them more susceptible to damage from inflammatory leukocytes than other tissues. Similarities or differences between the skin and laminar tissues support the continued study of both tissues. The site of skin specimen collection for the study reported here was selected to be as close to the hoof as possible, in hopes of reflecting anatomic-physiologic regional changes. Mild edema and mild endothelial cell rounding were detected in the superficial layer of the skin in a few of the specimens and may represent response to regional anesthesia or tourniquet application. These changes did not worsen during the study and were only detected in a small number of specimens.
Total RNA recovered from both types of biopsy specimens was of high quality and of sufficient quantity to allow for production of adequate cDNA for PCR studies. Real-time quantitative PCR is a sensitive technique to determine gene expression alterations associated with various disease conditions. Normalization of the specimens by use of at least 1 reference gene (termed the housekeeping gene) is an important step in the early stages of gene expression analysis. Validation of the housekeeping gene is essential to ensure that the gene is not differentially expressed because of experimental protocols or animal age, breed, or disease state, to name a few of the possible influential variables. We selected 2 housekeeping genes, β-gus and β-actin, because of their use in gene expression studies20–25 of equine tissues including laminae and skin. Our findings indicated the stability of these 2 genes throughout the serial biopsy procedure and support their continued use in future studies that use the biopsy technique.
Cyclooxygenase-2 is the inducible form of the COX enzymes and is often considered central to the inflammatory process, although many cytokines contribute to inflammation.26–28 Expression of COX-2 was absent or low in the biopsy specimens from the laminae or skin in the present study, suggesting low constitutive concentrations of COX-2 in these tissues; throughout the study, significant increases in COX-2 expression were not found. Although constitutive concentrations of COX-2 have been identified in the CNS and kidneys of rats, other tissues have little or no expression of COX-2 under normal conditions.29 Studies29 in other species have also found that 100% or 200% changes in expression of COX-2 are not physiologically important and that much larger changes in expression more accurately reflect an inflammatory response. Our positive controls expressed COX-2 in a concentration-dependent manner, and the housekeeping gene data were of good quality for the laminar and skin biopsy specimens; therefore, the little to no COX-2 expression in the biopsy specimens should be accurate.
Adequate laminar and skin biopsy specimens can be serially obtained in conscious sedated horses for histologic and molecular studies. The procedure does not cause undue signs of pain or long-term adverse effects. Laminar specimens collected serially do not appear to be adversely affected or have evidence of biopsy-induced inflammation secondary to previous adjacent biopsy specimen collection. Thus, this technique should be considered a potential tool for the study of sequential pathophysiologic events in experimental laminitis models and naturally acquired disease.
ABBREVIATIONS
| β-Gus | β-Glucuronidase |
| COX | Cyclooxygenase |
AnaSed, 100 mg/mL, Lloyd Laboratories, Shenandoah, Iowa.
PromAce, 10 mg/mL, Fort Dodge Animal Health, Fort Dodge, Iowa.
Dormosedan, 10 mg/mL, Pfizer Animal Health, Exton, Pa.
Carbocaine, 2% (20 mg/mL), Pfizer Animal Health, Exton, Pa.
Esmark Bandage, McKesson Medical-Surgical Corp, Richmond, Va.
Dremel Lithium-Ion cordless drill, Dremel, Racine, Mich.
Miltex Dermal Biopsy Punch, 6 mm diameter, Miltex Inc, Bethpage, NY.
BD Beaver Mini-Blade Blades, 8D Ophthalmic Systems, Waltham, Mass.
Phenylbutazone tablets, The Butler Co, Dublin, Ohio.
Equi-Thane “Super-Fast” Instant Shoe, Vettec Hoof Care Products, Oxnard, Calif.
Tri-Reagent, Schering-Plough Animal Health, Kenilworth, NJ.
Mo Bio Laboratories Inc, Carlsbad, Calif.
Invitrogen, Carlsbad, Calif.
TaqMan, Applied Biosystems, Foster City, Calif.
BHQ-1, Biosearch Technologies, Novato, Calif.
Gluck Equine Research Center, University of Kentucky, Lexington, Ky.
Applied Biosystems, Foster City, Calif.
SAS, version 9.1, SAS Institute Inc, Raleigh, NC.
References
- 1.↑
Pollitt CC. Basement membrane pathology: a feature of acute equine laminitis. Equine Vet J 1996;28:38–46.
- 2.
Galey FD, Whiteley HE, Goetz TE, et al. Black walnut (Juglans nigra) toxicosis: a model for equine laminitis. J Comp Pathol 1991;104:313–326.
- 3.
French KR, Pollitt CC. Equine laminitis: loss of hemidesmosomes in hoof secondary epidermal lamellae correlates to dose in an oligofructose induction model: an ultrastructural study. Equine Vet J 2004;36:230–235.
- 4.
Garner HE, Hahn AW, Salem C, et al. Cardiac output, left ventricular ejection rate, plasma volume, and heart rate changes in equine laminitis-hypertension. Am J Vet Res 1977;38:725–729.
- 5.
Blikslager AT, Yin C, Cochran AM, et al. Cyclooxygenase expression in the early stages of equine laminitis: a cytologic study. J Vet Intern Med 2006;20:1191–1196.
- 6.
Fontaine GL, Belknap JK, Allen D, et al. Expression of interleukin-1B in the digital laminae of horses in the prodromal stage of experimentally induced laminitis. Am J Vet Res 2001;62:714–720.
- 7.
Kyaw-Tanner M, Pollitt CC. Equine laminitis: increased transcription of matrix metalloproteinase-2 (MMP-2) occurs during the developmental phase. Equine Vet J 2004;36:221–225.
- 8.↑
Morgan SJ, Hood DM, Wagner IP, et al. Submural histopathologic changes attributable to peracute laminitis in horses. Am J Vet Res 2003;64:829–834.
- 9.
Singh SS, Ward WR, Murray RD. Technique of hoof biopsy in cattle. Vet Rec 1993;133:190–191.
- 10.
Singh SS, Murray RD, Ward WR. Gross and histopathological study of endotoxin-induced hoof lesions in cattle. J Comp Pathol 1994;110:103–115.
- 11.
Black SJ, Lunn DP, Yin C, et al. Leukocyte emigration in the early stages of laminitis. Vet Immunol Immunopathol 2005;109:161–166.
- 12.↑
Johnson PJ, Ganjam VK, Slight SH, et al. Tissue-specific dysregulation of cortisol metabolism in equine laminitis. Equine Vet J 2004;36:41–45.
- 13.
Wattle OS. Cytokeratins of the matrices of the chestnut (torus carpeus) and periople in horses with acute laminitis. Am J Vet Res 2001;62:425–432.
- 14.↑
Loftus JP, Belknap JK, Stankiewicz KM, et al. Laminar xanthine oxidase, superoxide dismutase and catalase activities in the prodromal stage of black-walnut induced equine laminitis. Equine Vet J 2007;39:48–53.
- 15.↑
Chirgwin SR, Nowling JM, Coleman SU, et al. Effect of immunostimulatory oligodeoxynucleotides on host responses and the establishment of Brugia pahangi in Mongolian gerbils (Meriones unguiculatus). J Parasitol 2003;89:483–489.
- 16.
Leutenegger CM, Mislin CN, Sigrist B, et al. Quantitative real-time PCR for the measurement of feline cytokine mRNA. Vet Immunol Immunopathol 1999;71:291–305.
- 17.
Sequence Detection System user bulletin #2. Foster City, Calif: Applied Biosystems, 2006.
- 18.↑
Douglas JE, Thomason JJ. Shape, orientation and spacing of the primary epidermal laminae in the hooves of neonatal and adult horses (Equus caballus). Cells Tissues Organs 2000;166:304–318.
- 19.
Riggs LM, Franck T, Moore JN, et al. Neutrophil myeloperoxidase measurements in plasma, laminar tissue, and skin of horses given black walnut extract. Am J Vet Res 2007;68:81–86.
- 20.
Quinlivan M, Nelly M, Prendergast M, et al. Pro-inflammatory and antiviral cytokine expression in vaccinated and unvaccinated horses exposed to equine influenza virus. Vaccine 2007;25:7056–7064.
- 21.
Bogaert L, Van Poucke M, De Baere C, et al. Selection of a set of reliable reference genes for quantitative real-time PCR in normal equine skin and in equine sarcoids. BMC Biotechnol 2006;6:24–30.
- 22.
Waguespack RW, Cochran A, Belknap JK. Expression of the cyclooxygenase isoforms in the prodromal stage of black walnut-induced laminitis in horses. Am J Vet Res 2004;65:1724–1729.
- 23.
Waguespack RW, Kemppainen RJ, Cochran A, et al. Increased expression of MAIL, a cytokine-associated nuclear protein, in the prodromal stage of black walnut-induced laminitis. Equine Vet J 2004;36:285–291.
- 24.
Loftus JP, Belknap JK, Black SJ. Matrix metalloproteinase-9 in laminae of black walnut extract treated horses correlates with neutrophil abundance. Vet Immunol Immunopathol 2006;113:267–276.
- 25.
Vick MM, Murphy BA, Sessions DR, et al. Effects of systemic inflammation on insulin sensitivity in horses and inflammatory cytokine expression in adipose tissue. Am J Vet Res 2008;69:130–139.
- 26.
Smith WL, Garavito RM, DeWitt DL. Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2. J Biol Chem 1996;271:33157–33160.
- 27.
Dubois RN, Abramson SB, Crofford L, et al. Cyclooxygenase in biology and disease. FASEB J 1998;12:1063–1073.
- 28.
Stack E, DuBois RN. Regulation of cyclo-oxygenase-2. Best Pract Res Clin Gastroenterol 2001;15:787–800.
- 29.↑
Botting RM. Inhibitors of cyclooxygenases: mechanisms, selectivity and uses. J Physiol Pharmacol 2006;57(suppl 5):113–124.
