Abstract
Cats with feline leukemia virus (FeLV)–associated enteritis (FAE), enteritis of other known viral etiology (parvovirus [PV], enteric coronavirus [CoV]), and enteritis of unknown etiology with histologic features similar to those of FAE and PV enteritis (EUE) and FeLV-negative and FeLV-positive cats without enterocyte alterations were examined. Amount and types of infiltrating leukocytes in the jejunum and activity and cellular constituents of mesenteric lymph nodes, spleen, and bone marrow were determined. PV and CoV infections were confirmed by immunohistologic demonstration of PV and CoV antigen, ultrastructural demonstration of viral particles in the intestinal content, and in situ hybridization for PV genome. FeLV infection was detected by immunohistology for gp70, p27, and p15E. Latent FeLV infection was excluded by polymerase chain reaction methods for exogenous FeLV DNA. Enterocyte lesions involved the crypts in cats with PV enteritis, FAE, and EUE and the villous tips in cats with CoV enteritis. Inflammatory infiltration was generally dominated by mononuclear cells and was moderate in the unaltered intestine and in cats with PV enteritis and marked in cats with FAE, CoV enteritis, and EUE. In cats with EUE, myeloid/histiocyte antigen-positive macrophages were relatively numerous, suggesting recruitment of peripheral blood monocytes. Lymphoid tissues were depleted in cats with PV enteritis and with EUE but were normal or hyperplastic in cats with FAE. Bone marrow activity was decreased in cats with PV enteritis; in cats with FAE or EUE and in FeLV-positive cats without enterocyte alterations, activity was slightly increased. In cats with FAE and PV enteritis, a T-cell-dominated response prevailed. EUE showed some parallels to human inflammatory bowel disease, indicating a potential harmful effect of infiltrating macrophages on the intestinal epithelium.
Keywords
Feline leukemia virus (FeLV)–associated enteritis (FAE) is one of the nonneoplastic conditions that may be seen in cats persistently infected with FeLV. 26,43–45 The small intestine in cats with FAE bears histopathologic features similar to those of cats with feline panleukopenia, but bone marrow depletion is missing. At the same time, parvovirus (PV) cannot be demonstrated and FeLV infection is present. 26,28,29,43 A study on the differential expression of FeLV proteins in the small intestine showed strong expression of FeLV envelope proteins gp70 and p15E in cats with FAE, suggesting a pathogenic role for these proteins or their precursors in the development of the disease. 26 This feature has also been proposed for FeLV-related feline acquired immunodeficiency syndrome (FeLV-FAIDS). 13,26,40,41
The purpose of the present study was to further characterize FAE by evaluating the number and type of inflammatory cells in the jejunum and by evaluation of lymphoid tissue and bone marrow activity and composition. We compared FAE with other relevant forms of enteritis in cats, such as PV enteritis and coronavirus (CoV) enteritis and enteritis with unknown etiology (EUE) but of similar morphology, by comparing infected cats with FeLV-negative and FeLV-positive cats without intestinal lesions.
Materials and Methods
Animals and tissue processing
The study was retrospectively performed on tissues from 74 routinely necropsied cats. Fourteen cats were immunohistologically positive for FeLV antigen and exhibited histopathologic alterations consistent with those associated with FAE (group I; Table 1). 26,43 For comparative examinations, four additional groups were established. Group II (n = 17) represented both immunohistologically FeLV-negative (n = 15) and FeLV-positive (n = 2) cats with enteritis due to other viral infections (PV enteritis: n = 15; CoV enteritis: n = 2; Table 2). Group III comprised immunohistologically FeLV-negative cats with intestinal alterations resembling those associated with FAE (EUE: n = 13; Table 3). FeLV-positive cats without enterocyte alterations composed group IV (n = 14; Table 4), and FeLV-negative cats without enterocyte alterations composed group V (n = 17; Table 5).
Cats with FeLV-associated enteritis (group I).
∗ F = female; M = male; MC = male, castrated.
Cats with coronavirus (CoV) or parvovirus (PV) enteritis (group II).
∗ Female; M = male; MC = male, castrated.
Immunohistologically FeLV-negative cats with intestinal alterations similar to those of cats with FAE but of unknown etiology (group III).
∗ F = female; M = male.
Immunohistologically FeLV-positive cats without enterocyte alterations (group IV).
∗ F = female; M = male.
Immunohistologically FeLV-negative cats without enterocyte alterations (group V).
∗ F = female; M = male.
Five randomly chosen sections of jejunum and one section of bone marrow from all cats and spleen and mesenteric lymph nodes from most cats were fixed in 10% nonbuffered formalin for 12–18 hours. One section each of jejunum, bone marrow, spleen, and mesenteric lymph nodes also was fixed in methanol. Specimens were embedded in paraffin, and 5-µm-thick sections were stained with hematoxylin and eosin (HE) and Giemsa. Specimens from the small intestine were further stained with Gram's stain (modified by Weigert) and the Warthin–Starry stain. 30,35
Scoring system for histopathologic alterations
Histopathologic alterations of the jejunum were evaluated as mild, moderate, marked, or severe according to the degree of change in the intestinal crypts and villi based on the amount of degenerating epithelial cells and the presence and degree of depletion of crypts, dilation of residual crypts, and villous atrophy in the different jejunal sections. The degree of infiltration by inflammatory cells in the mucosa was evaluated as mild, moderate, or marked on the basis of the amount of infiltrating cells in the different jejunal sections. Grading and counting were performed blindly and independently by two of the authors (A. Kipar and J. Kremendahl).
Immunohistology
Indirect peroxidase, peroxidase-antiperoxidase, and avidin–biotin–peroxidase complex methods were performed on both formalin- and methanol-fixed, paraffin-embedded tissues as previously described. 24–27 Cross-reacting antibodies for the demonstration of T cells (rabbit anti-human CD3), B cells (rat anti-mouse CD45R), macrophages and neutrophils (mouse anti-human myeloid/histiocyte antigen), and PV antigen (mouse anti-canine PV) and mouse monoclonal antibodies against FeLV gp70, p15E, and p27 proteins and feline CoV p56 protein were applied (Table 6).
Antibodies used for immunohistology.
∗ FeLV = feline leukemia virus; CoV = coronavirus; PV = parvovirus.
Sections for negative controls were incubated with normal rabbit serum for polyclonal antibodies, normal rat serum for rat monoclonal antibodies, and an irrelevant monoclonal antibody directed against chicken lymphocytes for mouse monoclonal antibodies. 12 Normal lymphoid tissues served as positive controls for leukocyte markers, and intestinal tissues from cats with panleukopenia and granulomatous-necrotizing lesions of feline infectious peritonitis served as positive controls for PV and CoV antigen, respectively. 24–26,54
Scoring system for immunohistology
The percentage of infiltrating inflammatory cells that stained positive with the various antibodies in relation to the total amount of leukocytes was evaluated semiquantitatively (counting of immunohistologically positive cells and of all leukocytes in several high-power fields) by light microscopy using the following reaction scoring system: ± (0–5% of total infiltrating cells), + (6–25% of total infiltrating cells), ++ (26–50% of total infiltrating cells), +++ (51–75% of total infiltrating cells), and ++++ (76–100% of total infiltrating cells).
In situ hybridization for the demonstration of PV genome
In situ hybridization (ISH) was performed on formalin-fixed jejunal specimens of selected cats (group I, cat Nos. 2, 4–7, 9–12, 14; group II, cat No. 1; group III, cat Nos. 1, 3, 10, 12; group V, cat No. 1) as previously described. 26,54 Canine and feline intestinal specimens and canine myocardial specimens that were immunohistologically positive for PV antigen and exhibited positive hybridization signals for PV genome served as positive controls. Negative controls were represented by specimens incubated with the biotinylated control probe pBR322.
Polymerase chain reaction for exogenous FeLV DNA
For the detection of exogenous FeLV DNA, a polymerase chain reaction (PCR) technique was applied to both methanol- and formalin-fixed tissue specimens of the jejunum and, when available, mesenteric lymph nodes, spleen, and bone marrow of selected cats (group I, cat Nos. 1, 4, 5, 9, 10, 12; group II, cat No. 1; group III, cat Nos. 1–13; group IV, cat No. 8; group V, cat Nos. 1, 5, 6) (Table 7), as previously described. 21 Primers targeting a 166-base pair (bp) segment of the FeLV U3 long terminal repeat region were applied: 5′-TTACTCAAGTATGTTCCCATG-3′ (sense) and 5′-CTGGGGAGCCTGGAGACTGCT-3′ (antisense). 9,21 PCR products from cats not showing any specific band at the 166-bp level were used as templates for a second PCR.
Results ∗ of immunohistology for FeLV, PV, and CoV antigen and PCR for exogenous FeLV DNA in cats with and without enteritis.
∗ Number positive/number examined.
† I = FeLV-associated enteritis; II = enteritis of other viral etiology (CoV, PV); III = FeLV-negative with enteritis of unknown etiology; IV = immunohistologically FeLV positive without enterocyte alterations; V = immunohistologically FeLV negative without enterocyte alterations.
Specificity of the PCR had been determined previously. 21 Determination of sensitivity showed that detection of FeLV genome is possible at about 3.2 fg, representing 246 copies of genome in 3 × 105 cells. Template DNA for positive controls was taken from previously tested formalin-fixed, paraffin-embedded tissues that consistently yielded intense appropriate bands with agarose gel electrophoresis and from selected cats from groups I and IV that were immunohistologically positive for FeLV (Table 7). Template DNA for negative controls was derived from uninfected cultured feline T cells, from cats immunohistologically negative for FeLV but without intestinal alterations (group V, cat Nos. 1, 5, 6), and from a cat with CoV enteritis (group II, cat No. 1) (Table 7). Reagent controls containing no DNA were included in each PCR.
Bacteriologic examinations
Intestinal and organ material from various cats (group I, cat Nos. 7, 11; group II, cat Nos. 1, 7; group III, cat No. 5) was cultured on blood agar containing 5% defibrinated sheep blood and water-blue metachrome-yellow lactose agar (E. Merck, Darmstadt, Germany) and incubated at 37 C for 24 and 48 hours.
Transmission electron microscopy
Feces or intestinal content from selected cats from group I (cat Nos. 4, 5), group II (cat No. 1), and group III (cat Nos. 2, 6, 12) were examined ultrastructurally for viral particles by means of the negative staining technique. 1,26
Results
Case histories
All animals were domestic shorthair cats. Cats in group I ranged in age from 4 months to 12 years, with an average of 4.5 years and a median of 3 years. Clinical signs included diarrhea (n = 6) and general symptoms such as weight loss and anemia (n = 4) (Table 1). Cats in group II ranged in age from 3 weeks to 4 years, with an average of 0.8 years and a median of 5 months. All cats had a clinical history of vomiting and/or diarrhea (Table 2). In cat Nos. II-1 and II-2, CoV enteritis was diagnosed. In the remaining cats (Nos. II-3–II-17) PV enteritis was diagnosed. Cat Nos. II-1–II-15 were immunohistologically negative for FeLV, and cat Nos. II-16 and II-17 were positive for FeLV. Cats in group III ranged in age from 9 weeks to 8 years, with an average of 2.8 years and a median of 1 year. Nine cats had gastrointestinal symptoms such as vomiting and/or diarrhea (Table 3). Cats in group IV ranged in age from 6 months to 13 years, with an average of 3.6 years and a median of 3 years. Cats in group V ranged in age from 4 months to 10 years, with an average of 3.9 years and a median of 2 years. Cats from groups IV and V had died with or had been euthanatized because of several clinical signs not related to the alimentary tract (Tables. 4, 5). Enterocyte alterations were not seen in cats in these two groups.
Histopathologic findings and immunohistologic characterization of infiltrating leukocytes in the jejunum
The degree of enterocyte alterations in the jejunum in groups I, II, and III and the amount of infiltrating cells in the mucosa in all cats are depicted in Table 8.
Degree of histologic alterations based on changes of epithelial cells and infiltration by inflammatory cells in the jejunum of cats with and without enteritis.
∗ I = FeLV-associated enteritis; II = enteritis of other viral etiology (CoV, PV); III = FeLV-negative with enteritis of unknown etiology; IV = immunohistologically FeLV positive without enterocyte alterations; V = immunohistologically FeLV negative without enterocyte alterations.
In most cats, lymphocytes represented the majority of infiltrating cells. Macrophages, plasma cells, neutrophils, and in some cases eosinophils were also observed. Inflammatory cells were generally diffusely distributed in the lamina propria. Occasionally, inflammatory cells accumulated predominantly in the upper or lower half of the villous mucosa. Percentages of T cells, B cells, and macrophages/neutrophils are shown in Fig. 4.
Special stains of the intestine did not provide evidence of bacterial, mycotic, or parasitic intestinal infections. In general, small numbers of gram-positive coccoid and rod-shaped bacteria were seen in the lumen, covering the villi without signs of bacterial adhesion or enterocyte invasion. In cat No. II-1, masses of gram-negative rod-shaped bacteria were observed in the intestinal lumen.
In all cats with PV enteritis where gut-associated lymphoid tissue was examined (10/15), moderate (4/10) or massive (6/10) lymphoid depletion was observed.
FAE (group I)
In most cats with FAE (8/14), changes were moderate (Table 8) and, similar to cats with PV enteritis, included degeneration of crypt enterocytes, depletion of crypts, dilation of residual crypts, and villous atrophy. 8,26,28,29,43 In addition to minimal changes of crypt enterocytes, cat No. I-4 exhibited markedly shortened villi and degeneration as well as sloughing of individual or small groups of enterocytes at villous tips.
In most cats (9/14), the mucosa was moderately infiltrated by mononuclear cells (Table 8), which were dominated by T cells (12/14: >50%) (Fig 1a). B cells and macrophages generally accounted for ≤25% of inflammatory cells and in most cases ≤5% (Figs. 1b, c, 4a).
Jejunum; cat No. I-8. FeLV-associated enteritis. Fig. 1a The majority of inflammatory cells in the mucosa are CD3-positive T cells. Peroxidase–antiperoxidase method. Bar = 50 µm. Fig. 1b Few CD45R-positive B cells (arrowhead) are seen. Bar = 50 µm. Inset: Higher magnification of CD45R-positive B cells. Avidin–biotin–peroxidase complex method. Bar = 20 µm. Fig. 1c A minority of infiltrating cells in the mucosa are myeloid/histiocyte antigen–positive macrophages (arrowheads). Peroxidase–antiperoxidase method, Papanicolaou's hematoxylin counterstain. Bar = 50 µm.
PV and CoV enteritis (group II)
In cats with CoV enteritis (cat Nos. II-1, II-2), shortened villi with denuded tips were observed as previously described. 25 In cat No. II-1, more severe changes were seen, including crypt dilation, enterocyte loss, hyperplasia, and the presence of enteroblasts indicative of regeneration. The mucosa was markedly infiltrated mainly by myeloid/histiocyte antigen-positive macrophages (>75%). In cat No. II-2, epithelial alterations were moderate and the inflammatory infiltration was mild and dominated by B cells.
In most cats with PV enteritis (cat Nos. II-3–II-17), intestinal alterations were severe (12/15; Table 8). Epithelial changes were restricted to the crypts and were identical to those described previously. 8,28,29 Severe villous atrophy also was observed. Mucosal infiltration was usually mild (13/15; Table 8) and dominated by T cells (11/15; Fig 2a). Macrophages and B cells were rare (Figs. 2b, c, 4b). However, in two cats that had shown a short clinical course (cat Nos. II-10, II-12), macrophages comprised >50% of the infiltrate.
Jejunum; cat No. II-13. Parvovirus enteritis Fig. 2a The majority of the few infiltrating cells are CD3-positive T cells. Peroxidase–antiperoxidase method. Bar = 50 µm. Fig. 2b Few CD45R-positive B cells (arrowheads) are found. Bar = 50 µm. Inset: Higher magnification of CD45R-positive B cells. Avidin–biotin–peroxidase complex method. Bar = 20 µm. Fig. 2c Single myeloid/histiocyte antigen-positive macrophages (arrowhead) are seen. Bar = 50 µm. Inset: Higher magnification of myeloid/histiocyte antigen-positive macrophage. Peroxidase–antiperoxidase method, Papanicolaou's hematoxylin counterstain. Bar = 20 µm.
Both FeLV-positive cats with PV enteritis showed severe intestinal alterations. Inflammatory infiltration was T cell dominated and marked in one cat and mild in the other cat.
EUE (group III)
Intestinal changes in cats in this group were comparable to those observed in cats with FAE and PV enteritis but generally were more severe than those in cats with FAE (Table 8).
As in FAE, inflammatory infiltration was usually moderate (9/13; Table 8) and consisted mainly of mononuclear cells. In three cats, the submucosa was also infiltrated. The percentage of T cells varied but in most cats was ≤25% (8/13; Fig 3a). B cells only surpassed 5% in one cat (Fig 3b). The number of macrophages and neutrophils varied but often was >25% (8/13; Figs. 3c, 4c). Frequently, a marked proportion of neutrophils was identified (8/13).
Jejunum; cat No. III-8. FeLV-negative cat with EUE. Fig. 3a Numerous CD3-positive T cells are found among infiltrating cells of the mucosa. Peroxidase–antiperoxidase method. Bar = 50 µm. Fig. 3b Single CD45R-positive B cells (arrowheads) are found. Bar = 50 µm. Inset: Higher magnification of CD45R-positive B cells. Avidin–biotin–peroxidase complex method. Bar = 20 µm. Fig. 3c Myeloid/histiocyte antigen-positive macrophages are numerous among infiltrating cells. Peroxidase–antiperoxidase method, Papanicolaou's hematoxylin counterstain. Bar = 50 µm.
Cats without intestinal epithelial cell alterations (groups IV and V)
In most cats in these groups, the mucosa exhibited mild mononuclear cell infiltration (group IV: 12/14; group V: 16/17; Table 8). The majority (≤75%) of infiltrating cells were T cells (group IV: 10/14; group V: 14/17). In FeLV-positive cats (group IV), B cells usually accounted for ≤25% (12/14), and macrophages and neutrophils ≤5% (12/14; Fig. 4d). The opposite situtation was observed in FeLV-negative cats (group V; Fig. 4e).
Proportions of CD3-positive T cells, CD45R-positive B cells, and myeloid/histiocyte antigen-positive macrophages among infiltrating cells in the mucosa of the feline jejunum. Group I = FAE; group II = enteritis with other known viral etiology (CoV enteritis, PV enteritis); group III = EUE; group IV = FeLV-positive cats without enterocyte alterations; group V = FeLV-negative cats without enterocyte alterations.
Immunohistologic demonstration of viral antigens (FeLV, CoV, PV), transmission electron microscopy for viral particles, and ISH for the demonstration of PV genome in the jejunum
Immunohistologic results are summarized in Table 7. FeLV antigen expression was generally observed in crypt epithelial cells and in infiltrating mononuclear cells in the mucosa. Staining for PV antigen was found in the cytoplasm, and ISH signals for PV genome were observed in nuclei of intact crypt enterocytes. CoV antigen was expressed by enterocytes at the villous tips. Negative controls for viral antigens, PV genome, and nonspecific staining did not show any reaction.
FAE (group I)
Staining for FeLV proteins was comparable to that previously described for FAE. 26 Gp70 was generally expressed most intensely, followed by p15E, whereas p27 staining was mostly faint. Inflammatory cells in the intestinal mucosa showed moderate to strong gp70 expression and faint to strong p15E expression but failed to stain for p27. 26
In cat No. I-4, CoV antigen was detected in single detached villous tip enterocytes, and CoV particles were identified in the intestinal content by electron microscopy, establishing the additional diagnosis of CoV enteritis.
PV antigen and genome were not detected in any cats in this group.
Cats with intestinal alterations of other viral etiology (group II)
CoV enteritis (cat Nos. II-1, II-2) was diagnosed based on CoV antigen expression in single villous tip enterocytes in both cats and in crypt enterocytes in cat No. II-1. 25 In cat No. II-1, CoV particles were detected in the feces by electron microscopy. Neither FeLV antigens nor PV antigen or genome (cat No. II-1) were detected in these cats.
In cats with PV enteritis (cat Nos. II-3–II-17), PV antigen was expressed by a variable number of intact crypt enterocytes. CoV antigen was not detected in any cats. FeLV antigens were moderately expressed by crypt enterocytes of cat Nos. II-16 and II-17. Gp70 and p27 were observed in a few crypt cells, and p15E was restricted to single cells. Gp70 was expressed by numerous and p15E by single inflammatory cells, which lacked p27 expression.
EUE (group III)
FeLV antigens, CoV antigen, and PV antigen and genome were not detected in any cats in this group. Viral particles were not identified by electron microscopy in the intestinal content of cat Nos. III-2, III-6, and III-12.
FeLV-positive cats without intestinal alterations (group IV)
FeLV antigen expression was consistently observed in these cats and the staining pattern was comparable to that previously described. 26,27 Crypt epithelial cells generally stained intensely for p27 compared with gp70 and p15E; expression of gp70 and p15E was occasionally lacking. Infiltrating cells generally stained strongly for gp70 and p15E but not for p27. 26
CoV and PV antigen and PV genome were not detected in any cats in this group.
FeLV-negative cats without intestinal epithelial cell alterations (group V)
FeLV antigens, CoV antigen, and PV antigen and genome were not detected in any cats in this group.
Histopathologic findings, cellular composition, and expression of FeLV antigens in spleen, mesenteric lymph nodes, and bone marrow
Spleen
Composition of the red pulp was generally comparable among the groups. In a few cats (Nos. I-7, I-13, III-3, IV-10, IV-13, V-15), moderate megakaryocytopoiesis was present. B-cell zones exhibited variable activity (Table 9). In cats with FAE (group I), they showed normal activity to moderate hyperplasia, whereas they exhibited mild to marked depletion in cats with PV or CoV enteritis (group II). Normal activity to marked depletion was observed in association with EUE (group III). In FeLV-positive cats without enterocyte alterations (group IV), B-cell activity was variable, but it was generally normal (11/17) in FeLV-negative cats without enterocyte alterations (group V).
B-cell activity in the white pulp of the spleen and in the lymphoid follicles of the mesenteric lymph nodes of cats with and without enteritis.
I = FeLV-associated enteritis; II = enteritis of other viral etiology (CoV, PV); III = FeLV-negative with enteritis of unknown etiology; IV = immunohistologically FeLV positive without enterocyte alterations; V = immunohistologically FeLV negative without enterocyte alterations.
T cells, B cells, and macrophages generally showed a normal distribution. T cells usually were up to 25%, except for the majority of cats with PV or CoV enteritis (group II; 9/17), where T cells accounted for ≤5%. B-cell proportions usually reached 25% or 50%, whereas macrophages never surpassed 25%.
Staining for FeLV antigens (groups I and IV; cat Nos. II-16, II-17) was present in mononuclear cells and dendritic cells of both the red and white pulp. Staining was generally strongest in group I cats, in which a variable number of leukocytes expressed all FeLV antigens tested. 26 Megakaryocytes stained positive for gp70 and p15E. In cat No. IV-14, the spleen was FeLV negative. In the remaining cats of group IV and in cat Nos. II-16 and II-17, expression of one FeLV antigen was occasionally lacking in some cell types.
Mesenteric lymph nodes
In FAE cats (group I), follicular activity was normal in most examined mesenteric lymph nodes (7/10), whereas follicular depletion (14/15) was observed in cats with PV enteritis (cat Nos. II-3–II-17; Table 9). In the one cat with CoV enteritis where mesenteric lymph nodes were available (cat No. II-2), moderate follicular hyperplasia was observed. Cats with EUE (group III) showed mild follicular depletion (2/3) or hyperplasia (1/3) (Table 9). In FeLV-positive cats without enterocyte alterations (group IV), follicular activity was variable, but activity was mostly normal (13/17) in FeLV-negative cats without enterocyte alterations (group V) (Table 9).
T-cell, B-cell, and macrophage distributions were similar in all groups. T cells usually accounted for ≤25%, B-cell numbers varied and reached 25%, 50%, or 75%, and macrophages only surpassed 5% in a few cases.
Staining for FeLV antigens was generally less intense than that in the spleen. In cats with FAE (group I), staining was often restricted to single faintly gp70- and occasionally p15E-positive follicular dendritic cells. In two FeLV-positive cats without enterocyte alterations (group IV), staining for FeLV antigens was not observed. In the remaining group IV cats and in both FeLV-positive cats with PV enteritis (cat Nos. II-16, II-17), all FeLV antigens were expressed in many cells.
Bone marrow
Bone marrow activity was variable. In the majority of cats, activity was moderate or marked and rated as physiological with respect to the cat's age (Table 10). In the FAE group (group I), four cats with marked activity were ≥2 years of age. Panmyelophthisis was only observed in three cats with PV enteritis (Table 10). Both cats with CoV enteritis (cat Nos. II-1, II-2) showed high activity. In the EUE group (group III), four cats with marked activity were adults (Table 10). In FeLV-positive cats without enterocyte alterations (group IV), four cats with marked and both cats with high bone marrow activity were adults (Table 10).
Activity of bone marrow of cats with and without enteritis.
I = FeLV-associated enteritis; II = enteritis of other viral etiology (CoV, PV); III = FeLV-negative with enteritis of unknown etiology; IV = immunohistologically FeLV positive without enterocyte alterations; V = immunohistologically FeLV negative without enterocyte alterations.
The distribution of CD3-positive cells, CD45R-positive cells, and myeloid/histiocyte antigen-positive cells was generally similar in all cats except those with enteritis of known viral etiology. Although CD3-positive and CD45R-positive cells mostly accounted for up to 25% each, myeloid/histiocyte antigen-positive cells, which were generally most numerous, often accounted for up to 75%. Most animals with enteritis of known viral etiology (9/17; group II) only had up to 5% myeloid/histiocyte antigen-positive cells.
In all FeLV-positive cats, FeLV gp70 was expressed most intensely, whereas either p27 or p15E was not detected in a few cats. Gp70 was expressed by megakaryocytes and, to a lesser extent, cells of the lymphoid and myeloid lineage. In cat No. IV-14, bone marrow was FeLV negative.
PCR for exogenous FeLV DNA
PCR results are listed in Table 7. In FeLV-positive cats, a specific band at the 166-bp level was detected after either both the first and second PCR runs or only the second PCR run. In negative cats, no specific band was detected after the first or the second PCR run.
Microbiology
Escherichia coli were isolated in high numbers (>200 colonies) from the small intestine of cat Nos. I-5, I-7, I-11, and II-1 and from mesenteric lymph nodes of cat Nos. I-7 and I-11. In cat No. II-2, moderate numbers (50–200 colonies) of E. coli and enterococci were isolated from small intestine, mesenteric lymph nodes, liver, and lungs.
Discussion
In this report, we describe both type and degree of enterocyte changes and inflammatory cell infiltration in the jejunum and lymphoid tissue and bone marrow activity in FAE and other relevant forms of enteritis in cats with similar morphologic features. Immunohistologic, molecular biologic, and electron microscopic methods were applied to establish a diagnosis of FAE, PV enteritis, CoV enteritis, or EUE.
Type and degree of intestinal alterations varied. In PV enteritis, FAE, and EUE, crypts had lesions characteristic of PV enteritis. 8,28,29,43 Lesions were usually least severe in FAE. In cats with CoV enteritis, degeneration and loss of villous tip enterocytes were observed. 25,39
In the morphologically normal jejunum regardless of the age of the cat and in cats with PV enteritis, mild to moderate T-cell–dominated infiltration of the mucosa was seen. Thus, in the jejunum as in the duodenum, a moderate T-cell infiltration is normal. 55 However, there are differences between experimental and natural PV enteritis. In the former, many neutrophils are found around and within crypts in the early phase of disease. 8 Therefore, cats naturally infected with PV either might not die in the early phase of the disease or might suffer from a lack of recruitable leukocytes due to granulocytopenia in the course of bone marrow hypoplasia and lymphoid depletion. 2,29 In cats with FAE, infiltration of the intestinal mucosa was generally more intense but also was T cell dominated. Together with the fact that the majority of T cells were CD8 positive at least in the normal feline colon, findings may indicate an FeLV-specific local cytotoxic T-cell response in FAE. 53 In FeLV-positive cats without enterocyte alterations, however, B cells were relatively numerous among infiltrating cells, suggesting a specific humoral response to FeLV antigens. In most cats with EUE, numerous myeloid/histiocyte antigen-positive macrophages were found among infiltrating cells. Additionally, a portion of infiltrating mononuclear cells failed to stain with T-cell, B-cell, and monocyte/histiocyte markers in the jejunal mucosa. In humans, resident macrophages are negative for the myeloid/histiocyte antigen (calprotectin), the synthesis of which cannot be induced in mature macrophages. 4,50 About 25% of infiltrating cells in the normal feline duodenum are cells with a histiocytic morphology. 55 Therefore, the number of macrophages among infiltrating cells in EUE might actually be even higher, and recruitment of peripheral blood monocytes in the course of mucosal inflammation is suspected. 50,55
Lymphoid tissue activity and bone marrow activity were generally normal in FeLV-negative cats without enterocyte alterations regardless of the other lesions these cats showed. FeLV-positive cats without enterocyte alterations showed variable lymphoid tissue activity and a tendency towards increased bone marrow activity, confirming that both lymphoid tissue hypoplasia and hyperplasia and bone marrow hyperplasia can develop as a result of FeLV infection. 17 In cats with FAE, however, bone marrow activity was normal and lymphoid tissue was either normal or hyperplastic. This phenomenon has also been observed in the initial phase of experimental infection with FeLV–felineAIDS (FAIDS) variants, which are (as natural variants) considered as a potential etiology of FAE. 19,26 Although viral replication can be anticipated in lymphoid tissues of cats with FeLV-FAIDS and FAE, there is no evidence of a cytopathic effect of FeLV on lymphoid and hematopoetic cells in either syndrome. 19 In cats with PV enteritis/feline panleukopenia, however, lymphoid tissue depletion is observed. 8,29 Many cats in the present study showed decreased bone marrow activity with depletion of myelomonocytic myeloid/histiocyte antigen-positive cells, a feature which is also seen in the early phase of experimental PV infection. 8 In a prolonged course of experimental PV enteritis and in cases of naturally acquired disease in our study, however, unaltered and even hyperplastic bone marrow was present. 2,6,29 This finding further indicates the variable course of natural PV enteritis. In cats with EUE, a tendency for lymphoid depletion and increased bone marrow activity was obvious, potentially reflecting the recruitment of inflammatory cells to the intestine.
In cats with FAE and EUE, PV infection was excluded by immunohistology and ISH. 26,42,54 Therefore, failure to confirm PV infection, as encountered in a natural PV enteritis outbreak in a colony of experimentally FeLV-infected specific-pathogen-free cats, is unlikely. 31,42,54 In general, FeLV and PV coinfections are rare. 44–46 In the two cats in our study with FeLV and PV coinfection, lesions were dominated by alterations typical of PV enteritis. The FeLV staining pattern corresponded to that of FeLV-positive cats without intestinal alterations. 26,27 In cats with FAE and in cats with PV/FeLV coinfection, lesions were distinct from the erosion of villous tips in the small intestine as they have been described in the “panleukopenia-like syndrome” (myeloblastopenia) in FeLV-infected cats. 16 Furthermore, the lymphoid depletion and severe granulocytic hypoplasia of the bone marrow described for this syndrome were also absent. 16
In our study, the lack of both typical lesions of villous enterocytes and CoV antigen expression excluded CoV enteritis in cats with FAE and EUE. 20,25,38 As observed in one cat, lesions typical of FAE and CoV enteritis may occur simultaneously. 25
FeLV infection was confirmed by immunohistology and by PCR for a 166-bp DNA fragment of the exogenous FeLV U3-LTR. 9,21 Latent FeLV infection was unlikely in all FeLV-negative cats, including all cats with EUE. 21 Occasional negative PCR results in cats immunohistologically positive for FeLV may be due to destruction of DNA by tissue processing factors. 21
In cats with EUE, infection with other enteric viruses, such as astrovirus, rotavirus, or torovirus, cannot be definitively excluded; ultrastructural examination of the intestinal content was not performed in all cats. However, these infections seem unlikely because they are generally rare and of minor pathologic significance. 15 Intestinal changes in cats with EUE do not suggest feline immunodeficiency virus (FIV) infection. However, intestinal lesions associated with FIV range from nonexistent to extensive mucosal inflammatory cell infiltration, mucosal erosions or ulceration, and transmural typhlitis. 3,7,36,51 Furthermore, lymphatic tissues and bone marrow do not show features indicative of FIV infection, such as large and often depleted secondary lymphoid follicles and a distinctly increased bone marrow cellularity. 3,6,7,51 However, because serology for FIV infection was not performed for cats with EUE, FIV infection cannot be definitively excluded.
Histopathology did not reveal involvement of bacterial, protozoal, or fungal pathogens in any cats. Bacterial adhesion or invasion of intestinal epithelial cells was not evident. Therefore, the numerous gram-negative rod-shaped bacteria observed histologically and the E. coli growing heavily in a few cultures cannot be considered pathogens. Abundant bacterial colonization with almost equal amounts of aerobes and anaerobes has been identified as normal in the small intestine of healthy cats. 37,52
In cats with EUE, exclusion of other causes of gastrointestinal inflammation renders the diagnosis of feline inflammatory bowel disease (IBD) acceptable. However, the cats in this group do not fulfill the criteria of lymphocytic-plasmacytic enteritis, the most common form of IBD in cats. 18,22,55 EUE is somewhat similar to human IBD, which is characterized by infiltration of the lower lamina propria by newly recruited calprotectin-positive histiocytes. 48,50 In human IBD, type IV hypersensitivity response to chronic antigenic challenge and dysregulation of cytokine secretion have been observed. 33,47 Calprotectin-positive macrophages seem to have a proinflammatory potential; they are primed for tumor necrosis factor α (TNF-α) and interleukin 1 production. 49 Furthermore, calprotectin has an antimicrobial effect in vitro and might therefore represent a defense mechanism against microbial invasion. 5,48 These findings suggest that the pathogenesis of lesions in EUE is similar to that in human IBD; in most cases of EUE, myeloid/histiocyte antigen (calprotectin)-positive macrophages dominated the inflammatory cell population in the jejunal mucosa. These macrophages may release cytotoxic substances such as reactive oxygen species and TNF-α, thereby inducing the observed crypt epithelial changes. 32,49
Footnotes
Acknowledgements
We are grateful to Dr. A. Waldvogel (Department of Veterinary Pathology, University of Bern, Switzerland) for providing the canine parvovirus DNA probe and to Prof. Dr. A. Pospischil (Department of Veterinary Pathology, University of Zurich, Switzerland) for providing the facilities for the parvovirus in situ hybridization. Furthermore, we thank Mrs. R. Weidenmann and Mrs. B. Trask for excellent technical assistance. A. Kipar was supported by a grant from the Deutsche Forschungsgemeinschaft (II B 8-KI 617/1-1); A. Kipar and J. Kremendahl were supported by the “Hochschulerneuerungsprogramm” of the State of Saxony and the Bundesministerium für Forschung und Technologie.