Molecular biology has been crucial in refuting a central dogma in equine laminitis and in organ injury in human sepsis, stating that laminar/organ injury was due to decreased blood flow, and not due to inflammatory injury. Many physiology studies indicated a decreased vascular supply or blood flow to the digital laminae at different clinical stages of the disease. Because routine histology (i.e., H&E staining) showed the lack of leukocytes, it was concluded that laminar inflammation was not playing a role in the etiology of laminitis. The first study to challenge this dogma involved ISH, using equine-specific 35S-labeled riboprobes for interleukin-1 beta (IL-1β) in paraffin-embedded tissues from a model of laminitis (the black walnut extract [BWE] model), in which the digits were perfused with 10% formalin immediately following euthanasia. The ISH study demonstrated IL-1β-positive cells in the laminar dermis located outside of the laminar venules which were markedly positive, whereas no IL-1β signal was present in the laminae from control animals (Figure 16.1a). Further, studies using CD13 (a marker of myeloid leukocytes) demonstrated a similar cellular pattern for emigrating leukocytes, indicating that the IL-1β–positive cells are emigrating neutrophils and possibly monocytes (Black et al., 2006). More recently, we have used non-isotopic ISH methodology for IL-6 to demonstrate a similar pattern indicating that leukocytes are the source of laminar IL-6 (Figure 16.1b), and to demonstrate that an important chemokine, CXCL1, is expressed in laminar epithelium (discussed below, Figure 16.1c).
There are limited studies of laminar gene expression prior to the invention of qRT-PCR. This was partially because only a few equine investigators had training in molecular biology, but also because of the limited amount of RNA that could be obtained from laminar samples for Northern hybridization (Rodgerson et al., 2001; Kyaw-Tanner & Pollitt, 2004). Quantitative RT-PCR was rapidly incorporated into laminitis research starting in 2004, when it was used both for the assessment of laminar mRNA concentrations of inflammatory proteins (Waguespack et al., 2004b), and for the assessment of laminar mRNA concentrations for a protease proposed to be involved in breakdown of the laminar adhesion, matrix metalloprotease-2 (MMP-2) (Kyaw-Tanner & Pollitt, 2004). In the first study, mRNA differential display, a genomic screening technique commonly used prior to the introduction of microarray technology, was employed to assess differential gene expression in affected laminae of BWE-treated horses compared to control laminae (Waguespack et al., 2004b). In this study, the investigators did not screen pooled samples due to the variability of response in individual horses in the laminitis models, but screened 3 pairs each of laminar samples from BWE-treated and control animals (Figure 16.2). Due to the variability in using the outbred equine population, the investigators assessed genes that were differentially regulated in ≥2 of the 3 pairs of samples (Figure 16.2) (Waguespack et al., 2004b). Interestingly, although this was an unbiased screening (i.e., that could have resulted in cloning of an equine homologue of a gene important for toenail growth), the first differentially expressed mRNA that was cloned from the gel and sequenced was MAIL (molecule with ankyrin repeats induced by lipopolysaccharide, LPS) (Waguespack et al., 2004b). This molecule is directly related to inflammatory signaling and was originally discovered in a survey of differentially regulated genes in the brain of mice administered with LPS (Haruta et al., 2001). MAIL is now known as one of the IκB proteins, IκB zeta (IκB-ξ). This protein is induced by numerous bacterial products/TLR ligands and by IL-1β, but not by TNF-α (Yamamoto et al., 2004; Motoyama et al., 2005). IκB-ξ is essential for the up-regulation of specific cytokines including IL-6 and IL-12. In the same study in which MAIL/IκB-ξ was found by differential mRNA display in affected equine laminae, qRT-PCR was used to demonstrate a 4-fold increase in IκB-ξ, 30-fold increase in IL-1β, and 160-fold increase in IL-6 mRNA concentrations in the laminae of BWE-treated animals at a developmental stage in the BWE model of laminitis versus controls (Waguespack et al., 2004b). In later studies, qRT-PCR has documented marked increases in laminar mRNA concentrations of cytokines (IL-1β, IL-6, IL-12) and chemokines (CXCL8/IL-8, CXCL1/Groα) in the developmental and lameness stages of both the BWE model and the carbohydrate overload (CHO) model of laminitis (Belknap et al., 2007; Loftus et al., 207; Leise et al., 2011). Interestingly, TNF-α has never been reported to be increased in any model of laminitis (Belknap et al., 2007; Loftus et al., 2007; Leise et al., 2011). Laminar COX-2 was also demonstrated to be up-regulated, originally by qRT-PCR (Waguespack et al., 2004a; Blikslager et al., 2006) and later by immunochemical techniques (Blikslager et al., 2006). These studies contrasted the differences between the BWE and CHO models: cytokine mRNA concentrations increase at early developmental time points (as early as 1.5 hours post BWE administration) in the BWE model (Loftus et al., 2007), whereas cytokine gene expression did not increase until the onset of clinical signs of lameness in the CHO model (Leise et al., 2011). Also, qRT-PCR was recently used to demonstrate that the increased incidence of laminitis in the forelimbs is not due to any difference between the front and hind feet regarding physiological or cellular events affecting cell signaling, but rather due to differences in weight bearing (Leise et al., 2009). Because digital laminae appears to be the “target organ” in the horse with sepsis, versus visceral organs like liver and lung in the human, we used qRT-PCR to compare inflammatory events between visceral organs and the laminae in horses. Although we did find increases in pulmonary and hepatic cytokine and chemokine mRNA concentrations (up to 83-fold increase in pulmonary IL-6 mRNA concentrations), the increases were not as high as those observed in the laminae in the same animals (up to 553-fold increase in laminar IL-6 mRNA concentration) (Loftus et al., 2007). The results are consistent with the clinical reality that the septic equine patient is much more likely to undergo clinically apparent laminar injury than significant pulmonary or hepatic dysfunction. However, the increased inflammatory gene expression in the liver and lungs suggests that, if we were able to maintain horses with severe state of sepsis in intensive care like it is done with human patients (equine patients have usually been euthanized at this point due to the inability to maintain a recumbent adult horse with severe sepsis or due to severity of laminitis), we would probably observe a similar visceral organ dysfunction as occurs in the septic human patient. There are no reports regarding the use of functional genomics in the study of laminar injury related to equine metabolic syndrome.
Similarly to sepsis-related organ injury in other species (Belknap et al., 2009), leukocyte emigration appears to be the source of inflammatory cytokines in laminar injury in laminitis. Therefore, two events essential for leukocyte adhesion and extravasation were assessed by qRT-PCR: endothelial adhesion molecule expression and laminar chemokine (cytokines chemotactic for leukocytes) expression. It was found that two important adhesion molecules, ICAM-1 and E-selectin, had a similar pattern of expression at the onset of leukocyte extravasation in the two laminitis models. The two molecules undergo peak expression at the 1.5 hour time point in the BWE model (Loftus et al., 2007), but not until the onset of lameness (approximately 24–36 hours) in the CHO model (Leise et al., 2009). Chemokines, reported to be essential for leukocyte activation and migration, demonstrated a somewhat similar pattern peaking very early in the BWE model (Loftus et al., 2007) and peaking at the onset of lameness in the CHO model, but undergoing some increases at the developmental stage (approximately 12 hours post CHO administration) in the CHO model. The chemokines that have been examined include CXCL1/Gro-α and CXCL8/IL-8. Both were found to undergo a marked increase (140–160-fold) at the 1.5 hour time point in the BWE model, decreasing to more moderate increases (approximately 20-fold) at later developmental stages (Loftus et al., 2007). As leukocytes are thought to migrate on a chemokine gradient (toward higher chemokine concentrations), we were interested to determine if laminar epithelial cells express CXCL1, a chemokine commonly expressed by epithelial cells in different organs. Using, non-isotopic ISH, we found CXCL1 expression in the laminar basal epithelial cells of the epidermal laminae, in endothelium, and also in cells that appear morphologically identical to those we stain with the macrophage marker CD163 in the secondary dermal laminae (Figure 16.1c) (Faleiros et al., 2009). This work demonstrates that there might be a chemokine gradient that induces leukocytes to migrate toward the point of failure in laminitis – the laminar dermal/epidermal interface.
The possibility that dysadhesion of the basal epithelial cells from the underlying basement membrane and dermis may occur due to degradation of matrix proteins by proteases, including matrix metalloproteases, led to the assessment of laminar mRNA concentrations for MMP-2, MMP-9, and, more recently, for aggrecanase ADAMTS-4. Quantitative RT-PCR results for laminar MMP-2 and MMP-9 expression in laminitis models are not consistent with reports of either minor increases or no change (Kyaw-Tanner and Pollitt, 2004; Loftus et al., 2006; Loftus et al., 2007). Most likely, this is because a great deal of the regulation of MMPs is post-translational. There are zymography data showing increases in MMP-2 activation in the CHO laminitis model, thus indicating that this MMP may play a role in laminar injury (Loftus et al., 2009). Laminar mRNA concentrations of the aggrecanase ADAMTS-4 undergo consistent increases in both BWE and CHO models, and in acute clinical cases of laminitis (Coyne et al., 2009). This increase correlates with an increase at the laminar protein level (personal communication, Samuel J. Black, University of Massachusetts, Amherst).
Whole Genome Transcriptional Profiling
Whole-genome transcriptional profiling techniques in the past included incredibly work-intensive techniques including mRNA differential display (discussed above) and subtractive cloning techniques that required a great deal of effort to determine a small number of differentially regulated genes. More recently, microarray analyses have come to the forefront of human and veterinary research due to optimized formats that allow the screening of thousands of genes at one time. Although these genome-wide screening techniques have been derided at times as “fishing expeditions” or “non-hypothesis driven research”, they are extremely important in complex disease processes, such as laminitis in which the small number of laboratories throughout the world investigating this disease are unlikely to discover all of the different signaling mechanisms important to the disease process. Presently, only two publications on laminitis include the use of microarray technology. These studies were performed using a bovine oligonucleotide microarray chip (over 15,000 transcripts) and a custom equine cDNA microarray (over 3,000 genes) to assess gene expression in the early stages of laminitis in the BWE and CHO models of laminitis (Budak et al., 2009; Noschka et al., 2009). Using the bovine microarray chip, the investigators reported up-regulation of 155 genes (none down-regulated), with genes associated with “pro-inflammatory biochemical or cellular processes” and “protein degradation/turnover” being most prevalent (Budak et al., 2009). However, because a large number of genes were considered as up-regulated at less than a 2-fold increase, and the microarray had not been validated on equine tissues, these results require further validation by qRT-PCR. The investigators did perform qRT-PCR on 3 genes that showed a 4–6-fold increase in microarray signal intensity: the mRNA concentrations for all three had a 1.7–11.6-fold increase by qRT-PCR.
The custom equine cDNA microarray in the other study represents approximately 10% of the equine transcriptome and was designed using ESTs obtained from several equine cDNA libraries of leukocyte origin, and cDNA fragments of inflammatory genes provided by other investigators (Noschka et al., 2009). Thus, this microarray is largely dedicated to inflammatory signaling. The investigators assessed laminar signaling at three time points in the BWE model including two developmental time points (1.5 and approximately 3 hours [onset of leucopenia] post BWE administration), and the onset of first signs of laminitis (Obel Grade 1 defined as weight shifting between forelimbs and bounding digital pulses [approx 10–12 hours post BWE administration). The number of differentially regulated (DR) transcripts detected by the microarray increased temporally: with only 14 DR genes at 1.5 hours (8 genes up-regulated and 6 genes down-regulated), 22 DR transcripts at the 3 hour time point (17 up- and 5 down-regulated at each time point), and 62 DR transcripts at the onset of Obel Grade 1 laminitis (55 up- and 7 down-regulated). Four genes consisting of chemokines (CCL7/MCP-3, MCP-1, CXCL10/IP-10) and an inflammation-related acute phase protein (serum amyloid A protein, SAA) were up-regulated at all three time points in the laminar tissue. The discovery of up-regulation of SAA proteins, acute-phase proteins induced by IL-6 mediated activation on STAT3 in other tissues (Nerstedt et al., 2010), is in good agreement with other recent findings showing that IL-6 is the most highly up-regulated inflammatory gene in equine laminitis (Loftus et al., 2007; Leise et al., 2011), and that laminar STAT3 is activated in the two laminitis models (Leise et al., 2010). Quantitative assessment by qRT-PCR of the expression of SOD2 and MCP-3 showed that the two transcripts undergo a much greater increase (up to 68-fold for SOD2 and 404-fold for MCP-3) than calculated from the microarray data (Noschka et al., 2009). Furthermore, the authors stated that they would have detected DR in genes previously reported to undergo significant up-regulation in BWE laminitis, such as IL-1β (up to 50-fold increase) (Loftus et al., 2007) and COX-2 (32-fold increase) (Blikslager et al., 2006), if they lowered the stringency of analysis. These results indicate that the microarray may not be the most sensitive method of detecting differentially regulated genes, and that more attention needs to be paid to genes with small increases in the expression (i.e., need to be assessed quantitatively by qRT-PCR). In addition to the increased number of inflammatory transcripts at the onset of OG1 laminitis, there is also an increase in the number of transcripts encoding anti-inflammatory proteins including Mn-superoxide disumutase (Mn-SOD/SOD2), elafin, and TIMP-1, which may play a protective role to limit laminar injury.
It is anticipated that future microarray studies will take advantage of the availability of higher fidelity equine-specific microarrays, advanced software allowing to organize the DR genes into different signaling cascades, and cutting-edge techniques, such as laser capture microdissection (LCM), where investigators can dissect individual cell types from cryosections for qRT-PCR and microarray analysis. There is an ongoing study in the author’s laboratory where the laminar basal epithelial cell, the cell that dysadheres from the underlying laminar dermis, resulting in laminar failure in laminitis, is being obtained by LCM from serial laminar biopsies taken prior to induction of laminitis (control), and at developmental and lameness time points. The cells will be subjected for microarrary analysis combined with gene networking to assess the signaling mechanisms occurring early in the disease process.
Contribution of Genomics to the Therapy of Laminitis
Functional genomics has recently been used to assess the efficacy of two different therapies being used in laminitis to block inflammatory signaling (Van Eps et al., 2010; Williams et al., 2010). In one, a constant-rate infusion of intravenous lidocaine, which has been studied and used for its purported anti-inflammatory properties, was assessed in the BWE model of laminitis. Quantitative RT-PCR analysis demonstrated that a lidocaine CRI did not change expression of cytokines or COX-2, whereas it actually induced an increase in the expression of the adhesion molecule E selectin indicating endothelial activation in the laminar vasculature by the lidocaine infusion (Williams et al., 2010). In distinction to these results, qRT-PCR was recently used to assess the effect of local digital hypothermia (termed “cryotherapy”) on inflammatory signaling in the oligofructose model of laminitis. Local hypothermia/cryotherapy has come to the forefront of therapies for laminitis due to recent studies indicating clinical efficacy of the treatment when the horse’s limbs are maintained in ice water following intragastric administration of a laminitis-inducing solution of oligofructose (Van Eps & Pollitt, 2009). In the study, animals were administered oligofructose and then had one forelimb placed and maintained in ice water (treatment foot) while the other foot was maintained at ambient temperature (control foot); this model is very powerful for statistical analysis of qRT-PCR data due to the paired nature of the comparisons. In that work, local hypothermia was demonstrated to markedly decrease laminar mRNA concentrations of inflammatory mediators including cytokines, chemokines, COX-2, and endothelial adhesion molecules (Van Eps et al., 2010). Due to the ability of qRT-PCR to easily provide data from the paired samples from individual horses (vs. the usual use of pooled samples in former techniques such as Northern hybridization), we were able to discover the efficacy of hypothermia in effectively decreasing inflammatory mediator gene expression as much as 1,000-fold in individual horses (Figure 16.3).