Intestinal neoplasia is rare and approximately accounts for 0.6% of canine tumors (19). Carcinomas are the most common large intestinal malignancy in dogs (28). Among the rectal tumors in dogs, carcinomas in situ (CiS) account for up to 25% (29). Dogs with colorectal CiS generally show a favorable prognosis after the surgical excision (29). However, depending on the type of surgery, the post-surgical recurrence of colorectal CiS is reported to occur up to 30% among affected dogs (2). Approximately, 14% of dogs with colorectal CiS may undergo malignant transformation to dysplastic carcinomas (27).

Regulatory T cells (Tregs) – a subset of CD4⁺ T cells characterized by transcriptional factor FOXP3 expression – are critical for immune homeostasis in normal and pathophysiological conditions, including cancer (10). The tumor-infiltrating Tregs (TI-Tregs) have been closely associated with poor prognosis due to anti-tumor immune surveillance in multiple types of canine cancers (6,23,24). The status of immune cell infiltration in canine colorectal cancers (CRC) has been poorly investigated. It is unclear how other immunosuppressive lymphocytes, such as Tregs, are infiltrated in canine CRC.

To our knowledge, this is the first known report of clinical evidence emphasizing the immunogenicity and immune-suppressive environment of canine colorectal CiS. Our case will be valuable in providing a rationale for basic research that dissects the immune environment for CiS for future development of cancer immunotherapy.

A 7-year-old castrated male poodle dog was referred to the Jeju National University Hospital with a 7-month history of hematochezia with blood clots. History evaluation revealed a loss of appetite, a 15% weight reduction, and intermittent rectal bleeding. Supportive medical care for the gastrointestinal symptoms, including famotidine (0.5 mg/kg, orally twice daily), amoxicillin-clavulanic acid (20 mg/kg, orally twice daily), metoclopramide (0.2 mg/kg, orally twice daily), metronidazole (7.5 mg/kg, orally twice daily), and digestive enzyme (1/4 T, orally twice daily) did not resolve the hematochezia. The physical examination revealed that the body condition score was six out of nine. Superficial lymph nodes were within normal limits. A digital rectal examination found multiple nodules protruding from the ventral mucosa of the rectum. The complete blood count (CBC) and serum chemistry parameters were within the normal limits. Aseptically obtained by cotton swabs, fecal samples were subject to real-time PCR to detect genes encoding enterotoxins of pathogens associated with canine diarrhea (Green vet, Yong-In, South Korea). The results of real-time PCR were present in Supplementary Data 1. The gene encoding enterotoxins of Clostridium perfringens was significantly amplified, suggesting tested positive. No significant findings were observed in chest radiographs and ultrasound examination. Abdominal ultrasonography revealed the splenic nodules, which were cytologically diagnosed with reactive nodular hyperplasia. Computed tomography (CT) results revealed contrast-enhanced, multifocal nodular lesions on the descending colon and rectum (Fig. 1A, B). Mild enlargement of the cecal and ileocolic lymph nodes was observed. Diagnostic imaging found no evidence of metastatic lesions in major organs. An endoscopy that was performed to examine the appearance of the lesion showed diffuse, multiple intraluminal nodules protruding from the intestine mucosa in the lower descending colon (Fig. 1C, D). The endoscopy-guided incisional biopsy was performed on the colorectal masses, which were subject to cytopathologic and histopathological examinations.

Figure 1. Computed tomography and gross lesion appearance of colorectal CiS in a dog. (A, B) CT of intestinal mass. Increased wall thickness and multiple nodules are present in the contrast enhancement at the level of the colon and rectum. (C) Endoscopy of the colon. Multiple nodules and mucosal swelling are present in the lumen of the descending colon. (D) Multiple nodules are present on the luminal side of the colorectal lesions resected.

The impression smears of the biopsied samples revealed clusters of cuboidal to columnar epithelial cells (Fig. 2A). No remarkable cellular and nuclear atypia were observed. Meanwhile, there were inflammatory cells that consisted of neutrophils, macrophages, and reactive lymphocytes (Fig. 2B). A differential diagnosis included inflammatory polyps and carcinoma with low malignancy. The rectal masses were fixed in 10% neutral-buffered formalin, processed, and embedded in paraffin wax. The three-micrometer sections were prepared and stained with hematoxylin and eosin (H&E). The histopathologic examination revealed a proliferation of neoplastic epithelial cells arranged in tubules, cords, and a few papillary fronds (Fig. 3A). Occasional multifocal squamous metaplasia with dyskeratosis was noted. The epithelial cells had variably distinct borders and were cuboidal to columnar, with abundant eosinophilic to amphophilic, often vacuolated cytoplasm. Nuclei were round to ovoid, with finely stippled chromatin and 1-2 prominent nucleoli (Fig. 3B). Within the surrounding stroma, numerous infiltrates of leukocytes, including neutrophils, lymphocytes, and plasma cells, were observed (Fig. 3C). Of note, there were multifocal areas of mineralized bone deposition where bone spicules were frequently and intimately associated with the neoplastic epithelial cells (Fig. 3D). The primary histopathologic diagnosis of colorectal mass was an inflamed neoplasm with ossification, favoring adenocarcinoma.

Figure 2. Fine needle aspiration of colorectal CiS in a dog. (A) A cluster of epithelial cells is exfoliated. Inflammatory cells are also seen. (B) Inflammatory cells, including numerous degenerative neutrophils with rod-shaped bacteria, occasional histiocytes, and small lymphocytes, are present. (C) Clusters of epithelial and spindloid cells with marked cytoplasmic basophilia are present. Note the infiltration of inflammatory cells, mainly consisting of neutrophils, adjacent to the cluster. (D) Mild to moderate anisokaryosis with nucleoli are present. The arrowhead indicates three plasma cells characterized by a perinuclear clear zone and a round, eccentric nucleus with coarse chromatin. (E) Multinucleated cells (arrows) and round-to-spindle cells with an eccentric nucleus are present. H&E stain, A: 100×, B: 400×, C: 200×, D, E: 400×. Scale bars = 100 μm (A), 20 μm (B), 50 μm (C), and 20 μm (D, E).

Figure 3. Histopathologic examination of colorectal CiS in a dog. (A-D) Representative histopathologic images of presurgical endoscopy-guided biopsy of colorectal mass. (A) Neoplastic cells are arranged in cords, tubules, and papillary patterns. (B) Cuboidal to columnar neoplastic cells with vacuolated cytoplasm. (C) Chronic-active inflammation is composed of lymphocytes, plasma cells, and neutrophils in the surrounding supportive stroma. (D) Mineralized bone is intimately associated with neoplastic focus (arrow). (E, F) Representative histopathologic images of surgically excised colorectal mass. (E) Papillary colorectal carcinoma in situ. Note the multifocal area of abundant mucinous materials with osseous and squamous metaplasia. (F) The dashed line refers to an enlarged section from panel E. Asterisk refers to ossification in all slides. H&E stain, A: 10×, Scale bar = 100 μm, B, C, D, F: 40×, Scale bars = 50 μm, E: 20×, Scale bar = 50 μm.

The patient was treated with amoxicillin-clavulanic acid (12.5 mg/kg, orally twice daily) and tranexamic acid (10 mg/kg, orally twice daily) to prevent Clostridium infection and rectal bleeding. The patient underwent a colectomy utilizing the Swenson pull-through procedure to remove the colorectal tumor (Fig. 1D). Regional lymphadenectomy was also performed on mesenteric, rectal, and enlarged medial iliac lymph nodes to examine metastasis. Cytologic features of the surgically excised colorectal masses in fine needle aspirates were similar to those observed in the imprint smear of the incisional biopsied sample, suggesting polypoid colitis, adenoma, and carcinoma (Fig. 2C, D). Admixed with epithelial cells, degenerative neutrophils, small lymphocytes, and plasma cells were often found (Fig. 2D, arrow). Occasionally, individualized round to oval cells with marked cytoplasmic basophilia and eccentric nucleus, multinucleated cells, and extracellular eosinophilic materials were found (Fig. 2E). Cytologic atypia was not found in the multinucleated cells.

The representative samples were subject to histopathologic examinations. The exophytic polypoid mass with multifocal erosion and ulcer was protruding into the intestinal lumen. Mass was composed of moderate to densely cellular, well-demarcated, unencapsulated neoplastic epithelial cells (Fig. 3E). Occasionally, the neoplastic cells within these papilliform structures marginally invaded the underlying lamina proprial cores. The histopathologic features of the neoplastic cells were consistent with those observed in pre-operative biopsied samples (Fig. 3F). The mitotic count was 2 to 6 per 2.37 mm² fields. Similar to the previous incisional biopsy results, approximately 25% of the carcinomatous foci were associated with osseous metaplasia (Fig. 3E, F, asterisk). The histologic diagnosis was multifocal, papillary colorectal carcinoma in situ with invasion into the lamina propria. Lymphofollicular hyperplasia of rectal and medial iliac lymph nodes was diagnosed with no evidence of metastasis.

Immune checkpoint inhibitors (ICIs), including anti-PD1 and PD-L1 antibodies, have tremendously changed the paradigm of cancer treatment in dogs (8,17). Canine colorectal cancers have been suggested to be a potential target for immune checkpoint inhibition (13). However, no clinical immunophenotypic evidence is available to show that canine colorectal cancers, including CiS, shape an immune suppressive environment. Thus, immunohistochemistry (IHC) and flow cytometric analysis were performed to characterize the immune infiltrates. IHC was performed in the serial H&E-stained sections, as previously described (11), using monoclonal mouse anti-human CD3 (LN10 clone; Biocare Medical, CA, USA). The IHC revealed mild to moderate infiltration of CD3⁺ T cells that were spatially distributed in the resident lamina propria (Fig. 4A), tumoral stroma, and neoplastic epithelial compartments (Fig. 4B, C). Paired flow cytometric analysis using the peripheral blood mononuclear cells (PBMC) and tumor-infiltrating leukocytes (TIL) was performed, as previously mentioned (10). The antibodies and reagents used in this study are available in Table 1. The flow cytometric immunophenotyping characterized and demonstrated various immune subsets, including CD8⁺ T, CD25⁺FOXP3⁺ regulatory T, and CD11b⁺ myeloid subsets (Fig. 4D-F). Compared to blood, the proportion of the TI-Tregs remarkably increased by more than 4-fold (Fig. 4H). When compared to blood, the ratio of CD8⁺ T to Tregs in the tumor decreased (Fig. 4I). The fluorescent intensity of CD11b remarkably increased in tumor-infiltrating myeloid cells, compared to that of circulating ones, despite the reduction in tumor-infiltrating myeloid proportion.

Figure 4. Immunophenotyping of colorectal CiS in a dog. (A-C) Immunohistochemistry. (A) Inflammatory T cells display strong membranous to cytoplasmic immunostaining of CD3. (B) T cells are present within the lamina propria, intratumoral stroma, and neoplastic and non-neoplastic epithelial compartments. (C) CD3⁺ T cells are infiltrated into the neoplastic epithelial compartment and associated with foci of osseous metaplasia (asterisk) within the intratumoral stroma and intraepithelial compartment. The dashed line refers to an enlarged section from the original panel. (D-G) Flow cytometry. Forward and side scatter density plots are used to exclude potential doublets. A viability dye is used to remove debris and dead cells. Following the selection of live singlets, gating strategies of Tregs (D), CD8⁺ T cells (E), and CD11b⁺ myeloid cells (F) are shown. (G) Comparison between peripheral and tumor-infiltrating immune subsets using contour plot and histogram. Figures refer to the percentage of peripheral (in black) and tumor-infiltrating (in orange) cells indicated by a rectangle or bar closed. (H) Note the preferential infiltration of tumor-infiltrating TI-Tregs compared to circulating ones. (I) CD8⁺ T/Treg ratio remarkably decreases in the tumor compared to PBMC. The immune cell ratio is calculated by dividing the percentage of CD8⁺ T cells by the percentage of CD4⁺CD25⁺FOXP3⁺ T cells among live CD45⁺CD3⁺ T cells. IHC stain, A: 5×, Scale bars = 200 μm, B: 20×, Scale bars = 50 μm, C: 40×, Scale bars = 50 μm.

Table 1 . Antibodies and reagents used in this study.

Antibody/reagent	Clone	Company	Dye	Dilution
CD45	YKIX716.13	Invitrogen	PE	1:100
CD45	YKIX716.13	Novus Biologicals	AF700	1:100
CD3	CA17.2A12	Bio-Rad Laboratories	FITC	1:100
CD4	YKIX302.9	Bio-Rad Laboratories	AF647	1:100
CD8	YCATE55.9	Invitrogen	PE	1:100
CD21	CA2.1D6	Bio-Rad Laboratories	AF488	1:100
CD25	P4A10	Invitrogen	PE-Cy7	1:100
FOXP3	FJK-16s	Invitrogen	PerCP-Cy5.5	1:50
CD11b	M1/70	BioLegend	PE-Dazzle 594	1:100
Fc Receptor Binding Inhibitor	Not applicable	Invitrogen	Not applicable	1:100
Fixable viability dye	Not applicable	eBioscience	EF780	1:8,000

The c-reactive protein (CRP) level increased (31 mg/dL; reference interval (RI), 0.1-1.0 mg/dL) a day after the surgery, but it returned to the normal limits within 12 days postoperatively. The dog showed temporary, intermittent rectal bleeding, which was resolved by the treatment of lactulose (6 mL/day, orally twice daily), amoxicillin-clavulanic acid (12.5 mg/kg, orally twice daily), and metronidazole (15 mg/kg, orally twice daily) for 50 days. The rectal examination has confirmed no remarkable findings. The CBC results were within the normal limits, except for the presence of non-anemic reticulocytosis at 151.1 × 10⁶/μL (RI, 10-110 × 10⁶/μL). The clinical staging is TisN0M0. CT results have revealed no evidence of recurrence and metastasis in the dog. As of 200 days after the surgery, no abnormal clinical signs, including complications, have been found in the patient. The patient is currently under careful follow-up to monitor for local recurrence and development of tumor-associated clinical signs.

This paper reports the novel and unusual aspects of canine colorectal CiS. First and foremost, this is the first study to show spatial interaction between lymphocytes and neoplastic cells within canine colorectal CiS and demonstrate preferential infiltration of Tregs within the tumor. Our case is academically valuable, suggesting that canine colorectal CiS is highly immunogenic and shapes an immune-suppressive environment. Next, we report a rare manifestation – ossification induced by canine colorectal CiS. As far as the authors are aware, only a case of canine colon cystadenocarcinoma has described osseous metaplasia (4). We believe that the clinicopathologic and immunophenotypic results presented in this case will be valuable for providing a rationale for future works to map the immune landscape of canine CRC at the single-cell level.

In our case, reactive hyperplasia of rectal and medial iliac lymph nodes was confirmed. This is interpreted as antigenic stimulation following local nodal drainage of tumor antigens derived from colorectal CiS (12). In the same context, histopathologic examination found a proportional increase in tumor-infiltrating lymphocytes during the progression of colorectal CiS. Importantly, the CD3⁺ T cells were distributed within the neoplastic intraepithelial compartment within CiS. Therefore, we strongly propose that canine colorectal CiS is highly immunogenic and promotes adaptive immunity. By leveraging paired flow cytometric analysis of PBMC and TIL (6), we demonstrated a preferential infiltration of Tregs within canine colorectal CiS. IHC that demonstrates FOXP3⁺ TI-Tregs would further consolidate our findings. The enrichment of Tregs is associated with enhanced recruitment of circulating Tregs to the tumor and/or induction of clonal expansion of TI-Tregs (12). The former could be supported by previous studies that showed the preferential recruitment of peripheral Tregs to urothelial and prostate carcinomas in dogs (15,16). Although clonal expansion is likely to occur in dogs with carcinomas (1,23), the latter needs to be demonstrated by T cell receptor repertoire analysis. Given that Tregs have been demonstrated to be a strong inhibitor of anti-tumor immunity in multiple cancer types in dogs (15,16,24), colorectal CiS-infiltrating Tregs, in our case, will likely be functionally immune suppressive. Importantly, our case corroborates the transcriptional inference that canine CRC could be immune suppressive possibly due to infiltration of Tregs (13). Meanwhile, contrary to our notion, a previous study suggested that canine colorectal adenocarcinomas induce poor anti-cancer immunity due to a low expression of cancer-associated antigens (5). Indeed, Herstad et al. (5) drew the conclusion based on the observation that CD3⁺ T cells were spatially confined to the lamina propria. Currently, despite the cases of T cell tumor exclusion, our case is more likely to favor the immune suppressive status of human and canine CRC (13). Thus, future studies are of particular interest to investigate how distinct subsets of functionally unique T cells, including Tregs, affect colorectal carcinogenesis and clinical outcomes in dogs.

In our case, ossification was interpreted as metaplasia that occurred in association with cancer-mediated tissue injury, followed by an infiltration of inflammatory cells and downstream signaling sequelae among predominantly resident cells of mesenchymal origin and neoplastic cells (20). Although a single case cannot evaluate or determine the highly complicated pathophysiology of the ossification (20), we suggest potential etiological factors involved in the ossification of canine colorectal CiS. For instance, neoplastic cells and CD3⁺ T cells were found to be spatially adjacent to the ossification in the niche of tumor stroma. This finding may suggest that reciprocal interactions between the tumor microenvironment components may play a role. The pathogenesis of ossification suggested in human CRC supports our notion (14). In dogs, osteoinductive proteins produced by tumor cells, such as transforming growth factor β (TGF-β), are suggested to mediate the ossification (19). TGF-β is critical for the infiltration and clonal expansion of TI-Tregs in dogs (26). Lymphocytes are required for heterotopic bone formation and propagation (9). Currently, very little is known about the phenomenon of ossification, a rare manifestation with a reported case of canine colon cystadenocarcinoma (4). Future studies need to elucidate the clinical relevance or molecular pathogenesis of ossification in the context of the immune environment of canine alimentary carcinomas.

The potential factors for evaluating the prognosis of canine CRC may include multiplicity of masses, completeness of margins, recurrence, and occurrence of complications after the excision procedure (2,18,22,27). Although the prognosis evaluation period is far from concluding, we discuss key prognostic factors that may assist practitioners in guiding clinical decisions of the dog with colorectal CiS. In our case, a tumor-free margin could not be completely obtained. This was due to the inherent inaccessibility of the lesions and the need to preserve intrinsic colorectal functions. The achievement of complete margins at the histopathologic level is considered to reduce the recurrence and tumor-related death in canine CRC (2). However, there is still a lack of sufficient data to support that margin completeness is critical for predicting the prognosis of canine colorectal CiS. For example, a 5-year overall survival probability of 23 dogs with colorectal CiS was high (80%), despite 6 cases with incomplete margins (2). Likewise, three dogs with completely excised colorectal CiS could undergo recurrence but are likely to have a good prognosis, showing an average survival time of 851 postoperative days (2). The overall prognosis of canine colorectal CiS is considered to be favorable with a median survival time of 1,006 days (21,22). Currently, the most convincing evidence suggesting a poor prognosis might be an occurrence of malignant transformation in dogs with rectal tumors, including CiS (2,3,22,27). Meanwhile, although the postoperative occurrence of the complications might increase the recurrence rate of canine rectal tumors, non-permanent, minor complications are unlikely to be responsible for tumor-related death (2). Thus, practitioners should focus on careful follow-up to monitor tumor-associated clinical signs, local recurrence, and ultimately malignant transformation of colorectal tumors in dogs.

In 29 dogs with small intestinal adenocarcinomas, adjuvant chemotherapy did not improve the survival time of the affected dogs (25). Clinical benefits of adjuvant chemotherapy on canine gastrointestinal tumors remain elusive. Future studies might be of interest to determine the beneficial effect of adjuvant chemotherapy on the prognosis of canine intestinal tumors with or without margin completeness.

Clinical trials using ICIs that target PD1/PD-L1 blockade on canine cancers have shown some promising results (8,17). However, the overall response rate is low or highly variable depending on cancer type. A dog with colorectal adenocarcinoma did not respond to the anti-PD1 antibody (7). These results might implicate the presence of other immune checkpoints expressed on immune subsets within the immune suppressive tumor microenvironment. Thus, future studies that can dissect the immune landscape of canine CRC at the single-cell level will be an unmet need for characterizing functionally unique immune subsets, such as CD8⁺ T, Tregs, and CD11b⁺ myeloid-derived suppressive cells as demonstrated in our case.

Our case is the first study to reveal immune suppressive canine colorectal CiS, highlighting the study rationale to map the immune environment of the tumor at a single-cell level.

The authors express sincere appreciation to all the pathologists from IDEXX Laboratories involved with the pathology and immunohistochemistry reports used in this study. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE) (#RS-2023-00241779).

The authors have no conflicting interests.