Epithelia were evaluated for the type of keratinization: non-keratinized, orthokeratinized (no nuclei in stratum corneum) and parakeratinized (cells with pyknotic nuclei in stratum corneum). In addition, the presence/absence of rete pegs was recorded.

SE and JE were each evaluated for the presence of locally or diffusely distributed leukocytic infiltration (LI) and leukocyte cell type, using 400x total-magnification.

The degree of LI was scored using 200x total-magnification within the complete SE and JE. For statistical reasons, the scores were summarized in the binominal categories of either “low” or “high”:

Low LI = 0: no leukocytic infiltration; 1: 1–2 leukocytes (mild);High LI = 2: 3–10 leukocytes (moderate); 3: > 10 leukocytes (severe).

The attached area of the JE on the cementum surface was scored as 0: no attachment; 1: attachment present with 1-5 cell layers; 2: attachment present with > 5 cell layers. For statistical reasons the scores 1 and 2 were combined to create the binominal categories of either “no attachment” or “attachment present”.

The distribution of leukocytes in the LP and PDL was recorded concerning localization (e.g., close to the gingival sulcus, the cementum, the oral gingival epithelium or the alveolar bone) and categorized as not present, multifocal or multifocal to coalescing.

The leukocyte cell types were evaluated in 400x total magnification.

The degree of LI was scored using 100x total-magnification within the complete LP. The following score was applied:

Low LI = 1: 0–50 leukocytes (mild);

High LI = 2: 51–100 leukocytes (moderate); 3: > 100 leukocytes (severe).

Eosinophilic infiltration was recorded separately as present or not present.

The distribution and localization of blood vessels in the lamina propria and the PDL was recorded and grouped either as homogenous (physiologic) or inhomogeneous, their quantity was scored using one optic field in 100x total-magnification:

Low numbers of blood vessels = 0–5 blood vessels;

High numbers of blood vessels = ≥ 6 blood vessels.

The collagen fiber arrangement was evaluated in the LP and PDL as regular or irregular (fibers running in criss-cross manner with irregular distances between connective tissue cells and matrix) in HE stain.

Higher mineralized cementum from “rest periods” in cementum formation was visible as parallel incremental lines as described by Fiorellini et al. (29). Evaluation of the alignment of incremental lines was recorded as regular or irregular.

The presence and localization of plaque (bluish deposition on cemental surface) and/or caries [brownish discoloration of the bluish plaque matrix and loss of calcified tissue forming flake or flask-like lesions, were defined as peripheral caries (30)] was recorded as present or not present, and classified as supragingival (above the sulcus gingivalis) or gingival (within the gingival sulcus).

A parakeratinized stratified squamous epithelium was found in 99.1% (115/116) of the PPID and in 79.1% (91/115) of the control specimens. Only one PPID sample (0.9%) featured a non-keratinized stratified squamous epithelium (108 palatinal), whereas in controls, 20.9% of the specimens showed an orthokeratinized stratified squamous epithelium. Rete pegs were visible in all specimens and no blood vessels or leukocyte infiltrates were found in the oral gingival epithelium in either group. No significant differences between the groups were found.

In the PPID group, 50.9% of the specimens showed a keratinized stratified squamous epithelium, while in the control group 44.1% were keratinized (p = 0.3). None of the specimens showed orthokeratinization. In both groups, the coronal aspect of the SE was usually keratinized, and the apical aspect showed a non-keratinized stratified squamous epithelium. Rete pegs were visible in all PPID samples and in 97.4% (112/115) of the control specimens.

In both groups, approximately 15% of specimens showed no leukocytes within their SE. For statistical analysis, the degrees of LI in SE and LP were combined to leukocytic infiltration of the gingiva (LIG) and are displayed in the LP section. In both groups, the cell type of the LI in the SE consisted predominantly of lymphocytes, followed by neutrophils. Plasma cells were present only occasionally.

In both groups, the JE was classified as a non-keratinized stratified squamous epithelium. After controlling for age, rete pegs were present in 5.2% (5/97) of PPID specimens and 12.4% (14/113) of controls (Figure 3). Neither group nor age showed a significant influence on the presence of rete pegs. LI were present in 81.8% (81/99) of PPID samples and in 84.2% (96/114) of controls. Neither group (p = 0.35) nor age (p = 0.054) showed an influence on the degree of leukocytic infiltration in the JE. Like in the SE, the predominant cell types were lymphocytes, followed by neutrophils and plasma cells. An intact attachment of the JE to the tooth was not more often seen in the control group (96.5%) than in the PPID group (92.8%; p = 0.64).

The LP contained homogenously distributed blood vessels and a subepithelial capillary plexus beneath the JE and SE in every specimen. The number of blood vessels in the LP did not differ between the groups. Perivascular leukocyte infiltration was seen in all specimens of both groups in the gingiva. The PPID group showed more often (87.7%) a high degree of leukocytic infiltration of the gingiva (LIG) than the controls (79.3%), but there were no statistically significant differences, even if the adjustment for a mean age of 24.4 years was applied. However, the distribution of LI in the lamina propria showed a significant difference between both groups (p = 0.002). In the PPID group, 65.4% of the specimens showed multifocal LI and 34.6% multifocal to coalescing LI. Controls showed multifocal LI in 92.8% and multifocal to coalescing LI in only 7.2% (Figure 4). When comparing the groups at the mean age of 24.4 years, PPID affected animals showed a higher chance (OR: 6.83) for a multifocal to coalescing LI distribution in the LP than controls. The leukocyte cell populations consisted predominantly of lymphocytes and plasma cells. The third most common cell population in both groups in the LP were eosinophils. Eosinophils were observed in both groups as focal or multifocal infiltrations, lying close to the gingival sulcus or the cementum. Neither the group nor age influenced significantly the presence of eosinophils in the LP. The PPID group showed slightly fewer eosinophilic infiltrations (45.8%), than the controls (50.7%). Macrophages and neutrophils were never seen as the dominant cell population but were present, especially if there were macroscopically noted periodontal pockets close to the histological sampling location.

Blood vessels in the PDL of both groups were observed in every specimen, lying either close to the alveolar bone or close to the cementum in a homogenous distribution. Most of the specimens (65.8% in PPID and 75.8% in controls) showed high blood vessel numbers with no significant difference. Age but not group, had a major effect (p < 0.0001) on the amount of LI in the PDL. Animals older than 24.4 years had a higher chance (OR: 1.24) for high amounts of LI in the PDL (Figure 5). In the PPID group, 13.3% (15/113) showed no LI in the PDL, 65.5% showed multifocal LI and in 21.2%, a multifocal to coalescing distribution of LI in the PDL was noted. In the control group, 33.6% (38/113) showed no LI, 54% showed a multifocal distribution of LI and 12.4% a multifocal to coalescing LI in the PDL. Even after mathematically controlling for age, no significant effect on the distribution of LI in the PDL between groups was calculated. However, if the age decreases by one year, the chance to show none or multifocal LI in the PDL increases (OR: 1.1; p = 0.05). Leukocyte cell populations consisted predominantly of lymphocytes and plasma cells in both groups with some apoptotic lymphocytes. Eosinophils were seen more frequently in the PPID group (10.6%) than in the controls (2.7%), but neither group nor age showed a significant effect on the presence of eosinophils in the PDL (Figure 6). Neutrophils and macrophages were never the dominant cell population, but single such cells were observed in individual samples from both groups. There was no significant difference between the two groups concerning the presence of an irregular collagen fiber network (PPID 5.6%; controls 2.6%).

Irregular cemental structures (Figure 7) were found in 15.5% of the PPID group and in 21.3% of the control samples. Within each group, an increase of 1 year in age results in a higher chance for irregular cemental structures (PPID OR: 1.66; control OR: 1.15). The interaction of age and group was significant (p = 0.005) in this parameter. Plaque was present within the gingival sulcus in 41.2% of the PPID and in 31.9% of the control samples, but there were no significant differences between the groups or age. Caries was found almost exclusively above the gingival sulcus and was recorded in 48.9% of PPID and 42.6% of control specimens. The interaction between age and group was significant with regard to the presence of caries (p = 0.007). In the (older) PPID group, an increase in age by 1 year even yields to a lower chance for caries (OR: 0.75).

The predominant presence of a parakeratinized stratified epithelium in the oral gingiva of horses is in line with findings documented for brachydont species like dogs, cats and humans (41–43). The higher amount of orthokeratinized epithelia in controls might reflect a higher mechanical irritation in these sites (8, 44, 45). The types of epithelia within the oral gingiva, sulcular epithelium and junctional epithelium were not significantly different between PPID and controls. Diffusely distributed LI in the gingiva are already documented in a wide range of species and they are claimed to be a normal finding in the gingiva (8, 9, 46, 47). The JE shows unique features compared with other oral epithelia. On one hand, it forms an epithelial barrier against the oral cavity, but permits the passage of extracellular fluid, inflammatory cells and other components of the immunologic host defense system. On the other hand, the function of the JE is to provide a stable connection with the teeth, which is essential for a healthy periodontium (3). Macroscopically horses affected by PPID showed a higher number of gingival sulci of ≥ 1 mm than healthy controls (p = 0.004) (28). Our findings here show a higher but not significant number of PPID affected horses (7/97) with a smaller attachment zone or missing contact than controls (4/113). The remodeling and eruption of teeth in hypsodont species requires ongoing cellular renewal and activity in order to keep the JE attached and the gingival sulcus < 1 mm. A gingival sulcus deeper than 1 mm or a loss of attachment could be an expression of a poorer dental status, predisposing for food and bacterial entrapment in the gingival sulcus. In this study we could not show a difference between PPID affected and healthy controls concerning the gingival epithelia.

The collagen fiber arrangement in the LP showed no difference between groups and an irregular fiber network was seen in 7% (8/114) of PPID and 5.2% (6/116) of controls. However, collagen fiber arrangement was evaluated exclusively in HE stained specimens and not in collagen fiber specialized staining methods. Histological evaluations of tendons from PPID affected horses (n = 4) showed a reduced longitudinal arrangement of collagen fibers in the suspensory ligament and an accumulation of proteoglycans between suspensory ligament fibers in comparison to two groups of healthy control horses (48). Likewise, in humans with adrenal hypercortisolism, alterations in collagen and epidermal thinning were observed in histological sections from skin with no evidence of leukocyte infiltration (49). In human dentistry, the normal lamina propria is described as collagenous tissue that usually shows no inflammatory infiltrates except in the area of the gingival sulcus (3, 5, 50). The main finding of our study is that PPID affected horses showed significantly (p = 0.002) more multifocal to coalescing LI than controls in the LP. Cox et al. (9) described that leukocyte infiltration is a physiological aspect of the equine gingiva mainly in the superficial lamina propria of the free gingiva, while the deep lamina propria contained mild perivascular infiltrations of lymphocytes and plasma cells. In humans, risk factors for severe forms of periodontal disease with LI in LP and even PDL are diseases that influence immune function (e.g., diabetes mellitus, HIV, cushing's syndrome) or lead to exaggerated immune responses (21, 34, 51, 52). In our study, the higher presence of multifocal to coalescing LI in the PPID group could show the onset of immune dysregulation within the LP. The infiltrations consisted mostly of lymphocytes and plasma cells, consistent with a chronic inflammatory/immunological reaction. Interestingly, eosinophils represented the third most common leukocyte cell population within the infiltrates of the lamina propria and in the PDL. The role and function of eosinophils in the periodontium is currently not fully understood. The presence of eosinophils might play a role in gingival protection and might be a transient expression of remodeling rather than a pathological finding (10). Eosinophils beneath the gingival epithelium were reported as a result of parasitic infestation of the gingiva by larvae of botflies (53). In rats with experimentally induced periodontal disease, eosinophils were identified in the lamina propria and are thought to play a role in cell and collagen destruction (9, 54). Our findings of eosinophils are in line with the findings of Cox et al. (9) who found eosinophils “irrespective of the presence or degree of periodontal disease” in the lamina propria and submucosa of their specimens. In contrast, Steinfort et al. (8) found no eosinophils in specimens of horses free from dental and periodontal disorders. Therefore, we assume that the often clustered eosinophils found in this study are a sign of enhanced immunological activity in the gingiva. However, there were no statistical differences detected between groups or related to age in our older probands.

There are few studies in equine dentistry that evaluated the subgingival periodontal ligament and their results are consistent with those of our control group (9, 55). We could not detect a significant difference in the distribution of LI between the groups but 21.2% (24/113) of PPID affected horses showed a multifocal to coalescing distribution of LI in the PDL, while in the control group only 12.4% (14/113) showed this distribution of LI. The reason why only a part of the PPID group in our study showed a higher amount of multifocally to coalescing distributed LI could be the varying amounts of POMC-derived peptides and their multiple systemic effects. PPID is a slowly developing and progressing neurodegenerative disease. Therefore, effects from this disease might be balanced by physiological regulations of the body and its immune system. Recent papers on PPID affected horses showed that they suffer from an impaired immune function (38, 56). Their neutrophil function is decreased resulting in a low oxidative burst capacity and decreased adhesion. Interestingly, an unexpected finding in the study of McFarlane et al. (56) was the increased chemotaxis in healthy aged horses, which is contrary to humans showing an age-related loss of chemotactic ability. On the other hand, age-related inflammation leads to an up-regulation of chemotaxis. In PPID, the age-induced higher LI might be counteracted by immunosuppressant POMC-peptides (e.g., α-MSH and CLIP). This could lead to a multifocal to coalescing high leukocyte infiltration without obvious signs of tissue demarcation, like we have shown in the periodontium.

On interdental gingiva sites, the PPID group showed more than four times more irregularities in cemental structures than controls (p = 0.02). This might still be a normal reaction due to remodeling capacities of the cellular tertiary cementum in hypsodont species (57), possibly caused by trauma or infections, which lead to cemental deposition in resorptive and reparative lesions (2, 58). Irregular cemental structure could also be an age related occurrence (35) as geriatric horses have shorter teeth, which is accompanied by higher tooth mobility. This is counter-regulated by a feature of senile equine dentition: the formation of new cementum replacing and compensating the missing tooth substance (35, 57, 58). However, in our probands, age had no significant influence (p = 0.79) on the presence of cemental irregularities. Therefore, we speculate, that PPID might interfere with the building of new cementum.

Nearly all samples in both groups showed plaque within the gingival sulcus or below the gingival epithelium level in interdental locations. In dental locations, caries was found almost exclusively above the gingival sulcus. The nutrition of geriatric horses often involves feed with simpler carbohydrate content (e.g., senior mash with higher starch vs. cellulose content), therefore the high amount of soluble sugars with their cariogenic potential may explain these findings. However, information about diet was not available for the probands. There was no higher prevalence of peripheral caries or plaque in PPID horses. This contradicts the proposal of Simon and Herold (59), who hypothesized that older, but especially PPID affected horses will develop plaque and caries due to immunosuppressive mechanisms.