Chronic inflammatory enteropathies (CIEs) are characterized by the clinical characteristics of chronic persistent or recurrent gastrointestinal signs, including vomiting, diarrhea, weight loss, and abdominal pain (Jergens and Heilmann, 2022). The pathogenesis of the disease is complex and involves genetics and environmental factors, resulting in an exacerbated and perpetuated intestinal mucosal immune response (Jergens and Heilmann, 2022). Canine CIE is currently subclassified based on the retrospective assessment of the treatment response and disease remission into either food-responsive enteropathy (FRE), steroid- or immunosuppressant-responsive enteropathy (SRE/IRE), and non-responsive enteropathy (Jergens and Heilmann, 2022). These entities of CIE are histopathologically similar, and the assessment of disease severity is currently primarily based on grading schemes evaluating the clinical and clinicopathological data (Allenspach et al., 2007; Jergens et al., 2010). A special subcategory of CIE is protein-losing enteropathy (PLE) resulting from marked inflammatory infiltration.

Endoscopy and histopathologic assessments are vital for documenting the disease (Washabau et al., 2010; Slovak et al., 2015). These diagnostics are not routinely employed as a monitoring tool, especially as the histologic resolution of duodenal lesions in CIEs was absent in most dogs responding during the induction phase of treatment (Garcia-Sancho et al., 2007; Schreiner et al., 2008). However, the role of mucosal healing (deep remission) has not been extensively investigated in canine CIEs (Jergens and Heilmann, 2022).

Canine CIE shares many similarities with chronic inflammatory enteropathies or inflammatory bowel diseases (IBDs) in human patients (Jergens and Simpson, 2012), comprising Crohn’s disease (CD), ulcerative colitis (UC), and indeterminate inflammatory bowel disease (Beniwal and Harrell, 2010; Ashton and Beattie, 2023). Despite these overlapping disease characteristics, several diagnostic and management features appear to be unique to either canine CIE or human IBD, including the disease location and distribution, immunological signatures such as the cytokine profiles of T helper cells (Heilmann and Allenspach, 2017), and long-term risks (Jergens and Simpson, 2012). However, canine CIE—despite presenting a good spontaneous disease model for human IBD—is generally less studied than human IBD.

The intestinal tract has the basic structure of a membranous–muscular tube. The luminal mucosa comprises an epithelium overlaying a basement membrane and a loose connective tissue layer (lamina propria mucosae). The underlying lamina muscularis mucosae consist of smooth muscle cells that facilitate the motility of the intestinal mucosa. In the small intestine, mucosal enlargement results from mucosal folds with characteristic finger-like villi of approximately 0.5–1.6 mm in length that protrude into the intestinal lumen and tubular crypts that extend into the intestinal wall (Liebich, 2010; Kummer and Welsch, 2018).

Villi are generally absent on the mucosa throughout the large intestine, but the large intestinal mucosa contains small folds and deep tubular crypts at a high density. The villi and crypts are covered by a single layer of columnar epithelium, mainly consisting of absorptive enterocytes, goblet cells, and few endocrine (Paneth) cells. Enterocytes are characterized by an apical brush border consisting of microvilli. These microvilli have an approximate width of 80 nm, height of 1.0–1.4 µm, and are covered by a 0.3–0.5-µm glycocalyx layer (Marin et al., 1983). However, the abundance of microvilli in the epithelial cells of the large intestine is significantly lower than that in the small intestine (Washabau, 2013). Enterocytes have an oval-shaped nucleus that is located in the basal ⅔ compartment of the cell and is overlaid by the Golgi apparatus, lysosomes, mitochondria, and rough and smooth endoplasmic reticulum (ER). Goblet cells ensure a continuous merocrine release of a glycoprotein- and glycolipid-rich cytoprotective mucin layer. These cells have a narrow base with a basal nucleus and organelles composed of a well-developed ER and Golgi apparatus. The apical part of the goblet cells is typically dilated and filled with mucin droplets, which are membrane-bound and 1–2 µm in diameter.

Epithelial cells are attached to the basal membrane via hemidesmosomes and are firmly linked together by an apical junctional complex composed of—from the apical to basal region—tight junctions (TJs) (zonula occludens), adherens junctions (AJs) (zonula adherens), and desmosomes (macula adherens), providing adhesive and other mechanical properties that seal and restrict the exchange of substances across the paracellular space (Toner, 1968; Demling et al., 1969; Roda et al., 2010; Henderson et al., 2011; Buckley and Turner, 2018).

The submucosa separates the lamina muscularis mucosae from the muscularis propria and contains loose connective tissue with blood vessels, lymphatics, adipose tissue, and scattered immune cells, as well as the submucosal plexus (Meissner plexus), a nerve plexus that controls the secretion and motility of the inner intestinal wall layers. The muscularis propria consists of the inner circular and outer longitudinal layers of smooth muscle cells and a few interstitial cells of Cajal that have an intestinal pacemaker function. Located between these two muscle layers is the myenteric plexus (Auerbach plexus) that regulates the peristalsis of the muscularis layer. The muscularis propria in most areas of the intestinal tract is covered by a single layer of flat epithelium, the serosa, which covers an underlying thin layer of connective tissue (Hoyle and Burnstock, 1989; Ibba-Manneschi et al., 1995; Liebich, 2010; Kummer and Welsch, 2018).

Ultrastructural lesions can be expected to have a higher sensitivity of detection than structural changes revealed upon light microscopy and may even precede or remain after the resolution of the latter (Sbarbati et al., 2003). However, data on ultrastructural lesions in canine CIE are limited. Abnormalities of the brush border and mitochondrial lesions (cristeolysis and swelling) in endoscopic biopsies of the duodenum have been described in the food-responsive phenotype of canine CIE (Table 1) (Walker et al., 2013). Furthermore, these lesions improved upon clinical remission, showing a reduction in the enterocyte intermicrovillar space and increased microvillus height after 6 weeks of dietary intervention using a hydrolyzed protein diet (Walker et al., 2013). However, potential lesions in the ileum and/or colon have not been evaluated in treatment-naïve dogs with CIE or in those dogs undergoing sequential treatment options (Jergens and Heilmann, 2022), warranting further research into morphometric changes at the subcellular level.

As a breed-specific disease entity separate from canine CIE, gluten-sensitive enteropathy in Irish setters was also revealed to produce microvillus lesions, including a reduction in the size and number, irregularities, and vesiculation.

Differential mucosal abnormalities in IBD (Table 2) included a preserved mucosal integrity with the loss of regular polygonal units and increased mucous production in patients with CD but sloughing and mucosal disintegration with decreased mucous production in UC patients (Trabucchi et al., 1986). Interestingly, these UC lesions could even be observed in endoscopically normal areas of the colon (Trabucchi et al., 1986; Bertini et al., 1998), making electron microscopy (EM) a sensitive tool to detect mucosal lesions and potentially deep remission. In other studies of human IBD, which focused on either CD or UC, the ultrastructural lesions evaluated included several alterations affecting the different structural units of the intestinal wall (Table 2) (Aluwihare, 1971; Ranlov et al., 1972; Dvorak et al., 1979; Myllarniemi and Nickels, 1980; Rickert and Carter, 1980; Lewis et al., 1984; Nyhlin and Stenling, 1984; Dvorak and Silen, 1985; Shields et al., 1985; Nagel et al., 1995; Bertini et al., 1998; Fratila and Craciun, 2010; Zhao et al., 2021; Zhou et al., 2023). While most lesions overlapped between childhood and adult cases of IBDs, some lesions presumed to reflect the chronicity of the disease process (e.g., villus bridging and goblet cell reduction) were primarily observed in adult IBD (Bertini et al., 1998) or were suspected to be specific for childhood IBD (Lewis et al., 1984). As an important differential diagnosis of IBD in humans, irritable bowel disease can also produce ultrastructural lesions (Miglietta et al., 2021). In addition, some overlapping (e.g., microvillus reduction or loss and intermicrovillus space enlargement) and also unique features (e.g., increased mucous layer and pseudomembrane coating of the epithelium) exist compared to chronic infectious diarrhea of bacterial (Fagundes-Neto et al., 2000) or protozoal origin (Poley and Rosenfield, 1982), non-infectious etiologies such as the irritable bowel syndrome (IBS)-like disorders including celiac disease (Shiner and Birbeck, 1961; Miglietta et al., 2021), or even systemic conditions with the potential to affect the integrity of the gastrointestinal barrier, such as an experimental sepsis model (Obermuller et al., 2020).

The intestinal epithelial cell ultrastructure was also altered in experimental rodent models of IBD (Pfeiffer et al., 1997; Tian et al., 2003; Bou-Fersen et al., 2008; Bochimoto et al., 2019). The colonic ultrastructure in a 2,4,6-trinitrobenzene sulfonic acid (TNBSA)-induced IBD/colitis model, for example, revealed a mixed picture of deformed intestinal crypt areas with high cell migration rates and a more regular structure with low cell migration rates in the remaining intestinal crypts (Bochimoto et al., 2019). The depletion of goblet cell mucin, stacked to curled Golgi apparatus in absorptive cells, and remnants of the ER support that ER stress plays an important role in the pathogenesis of IBD (Bochimoto et al., 2019).

Some characteristics, such as axonal degeneration or necrosis (Dvorak and Silen, 1985), might be an important feature and pathogenic correlation in IBD but require tissue biopsies that extend at least to the level of the intestinal submucosa. However, this depth is usually not reached with routine endoscopic biopsies (Day et al., 2008; Willard et al., 2010).

Lesion evaluation in human IBD and experimental animal models has traditionally focused on the enterocyte and typically includes the evaluation of the lesions that reflect changes in mitochondrial size and integrity, cytoplasmic injury, microvillus mass, and disruption of tight junctions (Mughal and Filipe, 1992). However, other structures of the mucosa might be equally important to be examined. Transmission EM (TEM) or scanning EM (SEM) could present a useful adjunct tool to evaluate the structures and detect lesions or patterns of tissue regeneration that remain undetected during routine histology (Laschi et al., 1987; Schattenfroh et al., 1994; Bertini et al., 1998). EM is an important tool in medical research and diagnostics, and this technique is usually accessible in most universities and research laboratories.

Very little information about ultrastructural lesions is available for canine CIEs (Walker et al., 2013). Morphological characteristics yet to be investigated in canine CIEs using TEM and/or SEM include any features and particularly goblet cells (and their subcellular mucin droplets) in the colon, any structural lesions in the ileum, mucosal structures other than the enterocyte, and a number of other enterocytic subcellular structures (e.g., ER, Golgi complexes, desmosomes, nuclei, and nucleoli) and cellular lesions including 1) junctional complexes (JCs) focused on TJs, AJs, and desmosomes; 2) autophagic bodies; 3) evidence of apoptosis or necrosis; 4) lipid/chylomicron droplets; 5) other particles (including viral or phage structures); and 6) other lesions shown in human IBD (Table 2). In addition, ultrastructural changes in the response to treatment (or lack thereof) other than dietary intervention—including immunomodulatory treatment or alternative therapeutic options such as pre-/pro-/synbiotics, fecal microbiota transplantation, cholestyramine as bile acid sequestrant, or stem cell therapy (Jergens and Heilmann, 2022)—remain to be studied. In this regard, synbiotic treatment decreases the dispersion and size variation of microvilli and the disruption of enterocytes in a dog with CIE (Figure 1) (Sahoo et al., 2022). Given the number and complexity of possible ultrastructural alterations (Table 2), their detection, grading, and association with other patient and disease characteristics might benefit from using a machine learning algorithm. Deep machine learning might also allow researchers to identify and compare structural lesions along the gastrointestinal tract, follow and integrate longitudinal changes over a more extended time, and utilize an unsupervised convoluted neural network approach to identify currently underestimated or even undetected lesions (Syed and Stidham, 2020; Javaid et al., 2022; Zand et al., 2022). Despite the tedious preparative steps involved in electron microscopy (Laschi et al., 1987), this method shows potential to investigate and integrate still unknown or undetected aspects of canine CIE, especially in the context of its pathogenesis, diagnosis, and response to treatment. A better characterization and deeper understanding of the ultrastructural mucosal changes in canine CIE will ultimately lead to a better definition of the similarities and differences between human IBD and canine CIE and, thus, will shed more light on the suitability of the dog as a spontaneous animal model for human IBD.