Establishment and Characterisation of Two Canine Prostate Cancer Cell Lines with Stem Cell Marker Expression

TL;DR

Researchers created two new cell models from aggressive canine prostate cancer to study its biology and improve treatments for both dogs and humans.

Contribution

The novel contribution is the establishment and characterization of two canine prostate cancer cell lines with preserved stem cell markers.

Findings

The cell lines retained key epithelial markers and stem cell features from the original tumors.

Xenografts generated from one cell line recapitulated the histopathological features of the primary tumor.

Selective loss of certain stem-like populations was observed during in vitro culture.

Abstract

Prostate cancer in dogs is uncommon but very aggressive, and it is often diagnosed late when treatment options are limited. This makes it difficult to study and to improve outcomes for affected animals. Some cancer cells, often called “stem-like” cancer cells, are thought to play a key role in how cancers grow, spread, and resist treatment. In this study, we developed two new laboratory-grown cell models from naturally occurring prostate cancers in dogs. We carefully compared these cells with the original tumours they came from, and with tumours created in mice using the cells, to see how closely they matched. We found that the new models kept many important features of the original cancers, including markers linked to aggressive behaviour and treatment resistance, although some features were lost when the cells were grown in the laboratory. These results show that our new models are reliable tools for studying prostate cancer in dogs. In the future, they may help researchers to better understand how this disease develops and spreads and support the development of improved treatments for both dogs and people with prostate cancer.

Canine prostatic adenocarcinoma is a rare but highly aggressive cancer that is typically diagnosed at an advanced stage, due to the lack of effective screening methods and poor recognition of early lesions. Cancer stem cells are known to drive tumour progression and treatment resistance in human prostate cancer, but their role in naturally occurring canine disease remains poorly defined. A deeper understanding of the biology of canine prostatic adenocarcinoma is therefore essential to improve prognosis and to develop relevant comparative models. We established and comprehensively characterised two novel canine prostatic adenocarcinoma cell lines, Kodiak and Bobby, with detailed comparison to their tumours of origin and, for Kodiak, xenografts generated in immunodeficient mice. Both lines displayed variable epithelial morphology influenced by culture conditions, and Kodiak xenografts recapitulated key histopathological patterns of the primary tumour. Expression of the luminal epithelial marker CK8/18 and the basal marker CK14 was largely retained across tumour, cell line, and xenograft, whereas the basal markers CK5 and p63, and the urothelial marker UPIII, were diminished or lost during in vitro culture. Evaluation of cancer stem cell-associated markers showed consistent expression of CD44, Nanog, Oct3/4, and Sox2 in the original tumours and cell lines, while CD133, Nestin, and Trop2 were present in the tumours but absent in vitro, indicating selective loss of specific stem-like populations. Media-dependent plasticity was evident in the Bobby line. These models retain key epithelial and stemness features and provide robust platforms for translational prostate cancer research in dogs and humans.

Figures

Full text

1. Introduction

Dogs are one of the few species known to develop spontaneous prostate cancer, and when it arises it is typically aggressive and highly metastatic. Historical reports consistently show that canine prostatic adenocarcinoma in dogs is rare. The only large survey of canine prostatic carcinoma identified an extremely low rate—just 14 cases (0.09%) were diagnosed among more than 15,000 canine necropsy and histopathology cases submitted to the Washington Animal Disease Diagnostic Laboratory [1]. Similarly, studies on dogs with prostatic disease have reported that only 3.5 to 15.7% were diagnosed with prostatic carcinoma [2,3,4].

However, these numbers likely underestimate the true occurrence of canine prostatic adenocarcinoma. There is no established screening method for detecting subclinical disease, and preneoplastic lesions remain poorly characterised which makes early identification extremely difficult. Early or subtle clinical signs are easily missed and many dogs therefore present only once the disease is advanced, at which time metastasis is common. At diagnosis, pulmonary involvement has been documented in roughly 8–50% of cases, and local metastasis to lymph nodes or bone in about 15–72% [5]. Additionally, advanced cases may be misdiagnosed—skeletal metastases can mimic neurologic or orthopaedic conditions, and owners may decline extensive diagnostic testing when the prognosis appears poor [6]. Thus, improved diagnostic strategies are needed to determine the true incidence of prostatic adenocarcinoma in the canine population.

In humans, cancer stem cells (CSCs) are thought to underpin prostate cancer initiation, progression, therapeutic resistance, and metastatic dissemination [7]. Multiple CSC-enriched populations have been identified in prostate tumours and cell lines, with CD44 expression frequently noted as a characteristic of these populations [8,9,10,11,12,13]. Trop2 has also been implicated in prostate cancer self-renewal and drug resistance [14]. Other key pluripotency regulators—including Nanog, Oct3/4, and Sox2—have been shown to confer stem-like properties in prostate cancer. Nanog promotes clonogenic growth, tumorigenicity, and castration resistance [15,16]. Nanog^+^ prostate cancer cells form more colonies and larger xenograft tumours than their Nanog^−^ counterparts [15], while Nanog downregulation reduces tumour formation [16]. Sox2 contributes to CSC self-renewal, tumorigenicity, and resistance to therapy [17,18], and Oct3/4 supports tumorigenicity and drug resistance [18,19]. Notably, targeting cells overexpressing Sox2 and Oct3/4 diminishes tumour initiation, growth, and chemoresistance [18]. CD133 and Nestin—well-established markers of progenitor cells [20,21]—have also emerged as candidate prostate CSC markers. CD133, associated with stem cell multipotency, identifies a subset of prostate cancer cells with hallmark CSC features, including enhanced self-renewal, tumorigenic capacity, and increased proliferation [22,23]. Nestin has been postulated to be involved in human prostate cancer self-renewal, multipotency, and tumorigenicity [8]. In contrast, few studies have investigated CSC markers in dogs [24,25].

The spontaneous, aggressive, and late-presenting nature of prostate cancer in dogs makes the species a uniquely relevant comparative model for studying CSC-driven pathogenesis and for identifying mechanisms that may underline advanced, treatment-resistant human prostate cancer. To expand the tools available for studying stem cell-driven mechanisms in canine prostate cancer, we established two new cell lines from prostate tissues obtained during prostatectomy in dogs with naturally occurring disease. While the establishment of several canine prostatic carcinoma cell lines has been described in the scientific literature, a detailed comparison of cell phenotype and CSC marker expression between the cell lines and the tumours from which they were derived has not been reported for any of the cell lines. Therefore, it is not known how closely the cell lines recapitulate the characteristics of the original tumours. It is well recognised that tumour cells can undergo phenotypic and molecular changes during in vitro culture, adapting to artificial growth conditions and potentially losing aspects of the heterogeneity and biological features present in the primary tumour. Such culture-induced alterations may compromise the interpretation and translational relevance of experimental findings. Therefore, we performed a comprehensive evaluation of each cell line and its corresponding tumour, examining a luminal epithelial marker (CK8/18), basal epithelial markers (CK5, CK14, p63), a urothelial marker (UPIII), and critically, a panel of CSC-associated markers (CD44, CD133, Nanog, Nestin, Oct3/4, Sox2, Trop2). This level of characterisation enables a more accurate assessment of the representativeness of the cell lines and provides a foundation for their use in future studies exploring the biology of canine prostate cancer. Focusing on stem cell-related phenotypes provides a clearer picture of how well these cell lines recapitulate the CSC characteristics of the original tumours and creates an essential platform for future work aimed at understanding CSC biology, defining therapeutic vulnerabilities, and advancing treatment strategies for canine prostate cancer.

2. Materials and Methods

2.1. Establishment of Cell Lines

Two canine prostate cancer cell lines were generated from prostate tissues collected during prostatectomies performed at Brisbane Veterinary Specialist Centre, Australia. The first cell line, Kodiak, was derived from an eight-year-old castrated Husky. Immediately following excision, a ~1 cm^3^ section of grossly abnormal prostate tissue was placed in sterile phosphate-buffered saline (PBS) on ice for cell culture, while a separate portion was fixed in 10% neutral-buffered formalin for subsequent embedding in paraffin (FFPE—formalin-fixed, paraffin-embedded) so that histopathology and immunohistochemistry (IHC) could be performed. Histopathological examination of the resected prostate confirmed a prostatic adenocarcinoma of mixed morphology (cribriform with central necrosis and solid components), with a score of 10 according to the grading system proposed by Palmieri and Grieco [26]. The PBS-preserved sample was cut into very small fragments using scalpel blades. The tissue fragments were then evenly distributed across 25 cm^2^ cell culture flasks (Nunc™ EasYFlask™ Cell Culture Flasks, Thermo Fisher Scientific, Waltham, MA, USA) coated in Matrigel^®^ Basement Membrane Matrix (Corning, Corning, NY, USA) containing complete MesenCult^TM^ MSC Medium (MSC) (STEMCELL Technologies, Vancouver, BC, Canada) supplemented with penicillin–streptomycin and L-glutamine. After 48–72 h of incubation, residual tissue fragments were removed and media were refreshed. Cells were subsequently maintained with routine media changes every 2–3 days.

Primary cultures were initially established in 25 cm^2^ culture flasks. Cells were allowed to grow until approximately 70–80% confluence before passaging. At that stage, cultures were split at an approximate 1:2 ratio into fresh flasks. No predefined quantitative seeding density thresholds were applied; instead, passaging decisions were based on visual assessment of confluence and cellular morphology. Following several passages and once stable proliferative growth was observed, cultures were cryopreserved in medium containing 10% dimethyl sulfoxide and stored in liquid nitrogen.

The second cell line, Bobby, was established from the prostate sample of a ten-year-old castrated Border Collie. Histopathology similarly confirmed a prostatic adenocarcinoma of mixed morphology (cribriform with central necrosis and solid), with a score of 10 according to the grading system proposed by Palmieri and Grieco [26]. This line was developed using the same protocol applied to Kodiak.

Kodiak and Bobby cells that were initially grown in MSC were subsequently tested in complete PrEGM™ Prostate Epithelial Cell Growth Medium (PrEGM) (Lonza, Basel, Switzerland), formulated for normal human prostate epithelial cells. The Bobby line demonstrated moderate growth in this medium, whereas Kodiak exhibited poor viability.

Regular testing for Mycoplasma was performed on both cell lines and was consistently negative.

2.2. Xenografts

The Kodiak cell line was transduced with a lentivirus expressing firefly luciferase and then injected subcutaneously into five female non-obese diabetic/severe combined immunodeficiency mice. Once the mice developed a detectable bioluminescent signal, they were euthanised and the tumours were harvested. Part of the tumour tissue was formalin-fixed and paraffin-embedded (FFPE) for histopathology and immunohistochemistry.

2.3. Immunohistochemistry

IHC was performed on the FFPE samples of the Kodiak and Bobby prostate tumours and the Kodiak xenograft to characterise the expression of CK8/18, CK5, CK14, p63, UPIII, CD44, CD133, Nanog, Nestin, Oct3/4, Sox2 and Trop2. The specifications and dilutions of the primary antibodies, and the antigen retrieval methods, detection systems, and positive control tissues used are outlined in Table 1.

Negative control slides consisted of one slide where the primary antibody was replaced by PBS and a second slide where it was replaced with non-immune, species-matched IgG diluted in PBS to the same concentration as the primary antibody.

All slides were deparaffinised and rehydrated and then treated with 3% hydrogen peroxide in methanol for 45 min at room temperature to inhibit endogenous peroxidase activity. After rinsing in distilled water, antigen retrieval was performed. Slides were then washed in Tris-buffered saline (TBS) or TBS with Tween™ (TBS-T) (Thermo Fisher Scientific, Waltham, MA, USA) after a 20 min cooling period when heat-induced epitope retrieval (HIER) was used or immediately following Proteinase K digestion.

When the VECTASTAIN^®^ Elite^®^ ABC-HRP Kit (Vector Laboratories, Newark, CA, USA) was used, non-specific staining was reduced by incubating the slides in blocking solution—either 5% normal goat serum, 5% bovine serum albumin (BSA) or Animal-Free Blocker^®^ (Vector Laboratories, Newark, CA, USA)—for 30 min at room temperature. Avidin and biotin blocking was performed using an avidin and biotin blocking kit before the slides were incubated overnight at 4 °C either with negative control solution or the primary antibody diluted in PBS. Slides were then rinsed three times in TBS or TBS-T and incubated with a biotinylated secondary antibody diluted to 1:200 in PBS for 30 min at room temperature. The biotinylated secondary antibody used for these markers was goat anti-mouse (Goat Anti-Mouse IgG Antibody (H+L) Biotinylated, Vector Laboratories, Newark, CA, USA) for CK8/18, CK5, CK14, p63, UPIII, CD44 and Sox2; goat anti-rabbit (Goat Anti-Rabbit IgG Antibody (H+L) Biotinylated, Vector Laboratories, Newark, CA, USA) for Nanog and Trop2; and rabbit anti-goat (Rabbit Anti-Goat IgG Antibody (H+L) Biotinylated, Vector Laboratories, Newark, CA, USA) for Nestin and Oct3/4. The CD133 slides were incubated with a rabbit anti-rat antibody (Rabbit Anti-Rat IgG Antibody (H+L) Mouse Adsorbed Unconjugated, Vector Laboratories, Newark, CA, USA) at a 1:200 dilution for 30 min at room temperature, rinsed three times with TBS-T and then incubated with the biotinylated goat anti-rabbit antibody at a 1:200 dilution for 30 min at room temperature.

After incubation with the biotinylated antibody, slides were rinsed in TBS or TBS-T twice before being incubated with the ABC-peroxidase solution for 30 min at room temperature. The slides were then rinsed three times in TBS or TBS-T and incubated with a peroxidase substrate solution for five minutes at room temperature in the dark. Finally, slides were rinsed in distilled water, counterstained with haematoxylin, dehydrated and mounted.

When the ImmPRESS^®^ HRP Universal Polymer Kit (Vector Laboratories, Newark, CA, USA) was used, non-specific staining was reduced by incubating the slides in the kit-supplied blocking solution (2.5% normal horse serum) for 30 min at room temperature. Slides were then incubated overnight at 4 °C with either the primary antibody diluted in PBS or the negative control solution. After rinsing in TBS, the ImmPRESS^®^ Universal Polymer Reagent was applied for 30 min at room temperature. The slides were then rinsed in TBS and incubated with a peroxidase substrate solution for five minutes at room temperature in the dark. Finally, the slides were rinsed in distilled water, counterstained with haematoxylin, dehydrated and mounted.

Semi-quantitative assessment of marker expression was performed for each sample by a single evaluator by estimating the percentage of positive neoplastic epithelial cells in 10 fields at 40× magnification and then calculating the average. Expression was then classified as high (>50% cells positive), moderate (>10% to <50% cells positive), low (<10% cells positive) or no expression (no cells positive).

2.4. Immunofluorescence

Immunofluorescence (IF) for CK8/18, CK5, CK14, p63, UPIII, CD44, CD133, Nanog, Nestin, Oct3/4, Sox2 and Trop2 was performed on Kodiak and Bobby cells grown in MSC and Bobby cells grown in PrEGM. IF could not be performed on Kodiak cells grown in PrEGM due to the cell line’s poor growth in this medium. The Kodiak and Bobby cells grown in MSC were both passage 9 and the Bobby cells grown in PrEGM were passage 10.

Cells were cultured overnight on sterilised Matrigel^®^-coated coverslips. After removing the medium, cells were rinsed with room-temperature PBS and fixed for five minutes in methanol chilled to −20 °C. They were then washed three times with PBS pre-chilled to 4 °C and incubated in 1% BSA for one hour at room temperature to block non-specific binding. The cells were subsequently incubated overnight at 4 °C with either the primary antibodies diluted at 1:50 in 1% BSA or negative control solution. The primary antibodies were the same as those used to perform IHC. Three negative controls were included in each run: (1) replacement of both primary and secondary antibodies with 1% BSA, (2) replacement of only the primary antibody with 1% BSA, and (3) replacement of the primary antibody with non-immune, species-matched IgG. After the overnight incubation, the cells were washed three times with room temperature PBS and then incubated for one hour at room temperature in the dark with fluorescein-conjugated secondary antibodies diluted at 1:200 in 1% BSA. The secondary antibodies were horse anti-mouse (Horse Anti-Mouse IgG Antibody (H+L) Fluorescein, Vector Laboratories, Newark, CA, USA) for CK8/18, CK5, CK14, p63, UPIII, CD44 and Sox2; goat anti-rabbit (Goat Anti-Rabbit IgG Antibody (H+L) Fluorescein, Vector Laboratories, Newark, CA, USA) for Nanog and Trop2; rabbit anti-goat (Rabbit Anti-Goat IgG Antibody (H+L) Fluorescein, Vector Laboratories, Newark, CA, USA) for Nestin and Oct3/4; and rabbit anti-rat (Rabbit Anti-Rat IgG Antibody (H+L) Fluorescein, Vector Laboratories, Newark, CA, USA) for CD133. Cells were then washed three times with room-temperature PBS in the dark, rinsed with distilled water, and mounted on glass slides using a DAPI-containing mounting medium.

Semi-quantitative assessment of marker expression was performed for each sample by a single evaluator by estimating the percentage of positive cells in six fields at 20× magnification and then calculating the average. Expression was then classified as high (>50% cells positive), moderate (>10% to <50% cells positive), low (<10% cells positive) or no expression (no cells positive).

3. Results

3.1. Characteristic Features of Cell Lines and Xenograft Tumours

The Kodiak cell line grown in MSC produced cells that varied in morphology from spindle-like to epithelial-like, with large sheets of cells, small amounts of secretory material, and formation of gland-like structures. When cultured in PrEGM, the cells adopted a more epithelial-like morphology, retained their gland-like structures, and produced noticeably greater amounts of secreted material. However, overall growth was suboptimal in this medium. Instead of forming the extensive cell sheets observed in MSC, the cultures developed only small, dispersed clusters that did not exceed approximately 40% confluence.

The Bobby cell line grown in MSC produced clumps of epithelial-like cells surrounded by spindle-like cells. Although secretion was observed, no gland-like structures were present in MSC. In contrast, when cultured in PrEGM, the cells exhibited a broader morphological spectrum, ranging from spindle-like to epithelial-like forms. Secretion persisted and numerous gland-like structures were present, albeit less well-defined than those observed in the Kodiak cell line. The Bobby cell line demonstrated comparable growth in both MSC and PrEGM, forming extensive sheets of cells in both media. See Figure 1 for examples of the cell lines grown in the different media.

The histopathological features of the mouse xenograft tumours created using the Kodiak cell line closely resembled the original tumour (Figure 2). The xenograft tumours contained large clusters of neoplastic cells without a specific histological pattern, similar to the regions of the original tumour that demonstrated a solid histotype. The xenograft tumours also demonstrated areas of necrosis surrounded by multilayered rings of neoplastic cells, which resembled the sections of cribriform with central necrosis present in the original tumour.

3.2. Expression of Cell Phenotype Markers

For the Kodiak cell line cultured in MSC, the expression of CK8/18 and CK14 was consistent across the original tumour, the cell line and the mouse xenograft tumour, with high cytoplasmic expression of CK8/18, mainly in luminal cells, and low cytoplasmic expression of CK14, mainly in the basal cell population, observed in all three (Table 2; Figure 3). In contrast, p63 and UPIII expression differed between the cell line and the tissue samples. While p63 showed low nuclear expression in the original tumour, with a predominant distribution in the basal cells, and the xenograft, it was undetectable in the cell line. UPIII was absent in both the original tumour and xenograft but exhibited low cytoplasmic expression in the cell line. CK5 expression was low in the original tumour tissue but was not detected in either the cell line or the xenograft.

For the Bobby cell line, CK14 expression was comparable between the original tumour and the cell line cultured in both MSC and PrEGM, with moderate cytoplasmic CK14 levels observed across all three samples (Table 3; Figure 4). Cytoplasmic CK8/18 was highly expressed in the original tumour and in the cell line grown in MSC, but only moderately expressed in the cell line cultured in PrEGM. There was moderate expression of CK5 and low expression of p63 and UPIII in the original tumour but these markers were not detected in the cell line under either culture condition.

3.3. Expression of Stem Cell Markers

For the Kodiak cell line cultured in MSC, the expression of CD44, Nanog, Oct3/4, and Sox2 was consistent with that observed in the original tumour, with all four markers showing high expression in both the tumour and the cell line (Table 4; Figure 5). CD44, Nanog and Oct3/4 were expressed in the cytoplasm in both the original tumour and the cell line, whereas Sox2 expression was nuclear and cytoplasmic in the original tumour but only nuclear in the cell line. In contrast, CD133, Nestin, and Trop2 were highly expressed in the original tumour but were undetectable in the cell line.

For the Bobby cell line, CD44 was highly expressed in both the original tumour and the cell line cultured in MSC and PrEGM. Oct3/4 was highly expressed in the original tumour and in the MSC-cultured cell line but showed moderate expression in PrEGM. Sox2 exhibited moderate expression in the original tumour and in the PrEGM-cultured cell line, whereas it was highly expressed in the MSC-cultured cells (Figure 6). CD44 and Oct3/4 expression were cytoplasmic in both the original tumour and the cell line, while Sox2 expression was nuclear and cytoplasmic in the original tumour but just nuclear in the cell line grown in both MSC and PrEGM. In the original tumour, CD133 displayed low expression, and Nanog, Nestin, and Trop2 were highly expressed; however, none of these markers were detected in the cell line under either culture condition (Table 5).

4. Discussion

This study reports the successful establishment and comprehensive characterisation of two novel canine prostatic adenocarcinoma cell lines (Kodiak and Bobby), including confirmation of the tumorigenic capacity of one cell line through xenograft development in immunodeficient mice. The xenografts retained the histopathological morphology of the original tumour, demonstrating that the in vivo model preserves key structural features of the primary neoplasm and supporting its relevance as a biologically representative system.

Regarding the Kodiak cell line, comparative expression analyses across the primary tumour, cell line, and xenograft revealed that the epithelial differentiation markers CK8/18 and CK14 were largely conserved, with consistently high expression of the luminal marker CK8/18 and low expression of the basal marker CK14. In contrast, the basal epithelial markers p63 and CK5 and the urothelial marker UPIII showed differential expression. p63 demonstrated low expression in the original tumour and xenograft tumour but was not expressed in the cell line. UPIII was not expressed in the original tumour or xenograft tumour but showed low expression in the cell line. The expression of CK5 was low in the original tumour tissue but was absent in the cell line and xenograft tumour. These differences suggest that the in vivo xenograft microenvironment may more closely replicate the conditions of the original tumour than in vitro culture, thereby preserving certain phenotypic characteristics lost under cell culture conditions. Evaluation of stem cell-associated markers demonstrated partial retention of stemness attributes in the cell line. Expression of CD44, Sox2, Nanog and Oct3/4 was high in both the original tumour and cell line, but the lack of expression of CD133, Nestin and Trop2 in the cell line compared to the original tumour may indicate potential culture-dependent plasticity.

Considering the Bobby cell line, the expression of CK14 and CK8/18 was comparable between the primary tumour and the cell line, indicating preservation of key epithelial markers. In contrast, CK5, p63, and UPIII were detectable in the original tumour but absent in the cell line, suggesting alterations of basal and urothelial-related characteristics during adaptation to in vitro culture, and partial urothelial differentiation in the original tumour, which is not uncommon in canine prostate cancer [27]. Expression of the stem cell-associated markers CD44, Sox2 and Oct3/4 was maintained; however, reduced Sox2 and Oct3/4 expression was observed when Bobby cells were cultured in PrEGM compared to MSC medium. Notably, the lower Sox2 expression in PrEGM more closely matched the limited Sox2 expression seen in the original tumour tissue. PrEGM is optimized for the growth of normal human prostatic epithelial cells, whereas MSC is designed for human mesenchymal cells. Thus, the higher expression of Oct3/4 and Sox2 in MSC may have been due to the media promoting a more mesenchymal phenotype. This phenotypic plasticity may be consistent with epithelial-to-mesenchymal transition (EMT) processes and stem cell-like features, as tumours undergoing EMT may acquire progressive enhanced adaptability and stemness potential. CD133, Nanog, Nestin and Trop2 were present in the original tumour but lost in the cell line, consistent with clonal selection and diminished tumour heterogeneity following cell culture establishment.

The differences in the protein expression profiles between the original tumours, the derived cell lines, and xenograft models are not unexpected and can be attributed to several biological and methodological factors.

(1)Loss of tumour microenvironmental influences: Primary tumours exist within a complex network of stromal cells, immune cells, extracellular matrix and soluble factors that can profoundly influence tumour cell phenotype, and transcriptional programs. In vitro cell lines are grown in an artificial, nutrient-rich, two-dimensional culture system that lacks these regulatory signals, leading to shifts in proliferation rates, differentiation status, and stress-response pathways.(2)Clonal selection and adaptation: The establishment of a cell line involves selection for tumour cell clones that can survive and proliferate under culture conditions. These clones represent only a subset of the heterogeneity present in the original tumour and may exhibit different gene and expression patterns.(3)Host-specific xenograft effects: Xenografts are established in an immunocompromised host from a different species, and this may influence the immune interactions and stromal composition that can modulate the expression of cancer-associated genes.(4)Temporal genetic and epigenetic drift: Both cell lines and xenografts can accumulate genetic and epigenetic changes, thus diverging from the genetic profile of the original tumour.

The observation of markers being consistently expressed, downregulated in culture, or reappearing in xenografts provides insights into stem cell plasticity and the influence of microenvironmental context on marker preservation and expression.

Taken together, our findings indicate that CD44, Sox2, Oct3/4 and, to a lesser extent, Nanog, are the most robust and consistently expressed stem cell markers in canine prostatic adenocarcinoma. Conversely, CD133, Nestin and Trop2 were detected in the original tumours but were lost during in vitro cell line establishment, which may be associated with a more restricted or plastic stem-like population that requires specific in vivo-like conditions for maintenance. This supports the concept of stem cell marker heterogeneity in canine prostate cancer [7], where only a subset of markers remains stable outside of the native tumour microenvironment and, at the same time, highlights the limitations of defining tumour stem cell phenotypes solely on the basis of cell culture studies that may not reflect the complexity of the disease in vivo.

Our study does have some limitations. Immunohistochemistry and immunofluorescence differ in several aspects, including detection chemistry, fixation and antigen retrieval requirements, and imaging modalities, and thus the dissimilarities in marker expression seen between the two methods may be partially due to differences in detection sensitivity. Additionally, due to the limited availability of Kodiak xenograft tissue and the need to prioritize essential comparative analyses, we did not assess the expression of the stem cell markers in the Kodiak xenograft, and we were not able to create xenografts using the Bobby cell line. Nevertheless, our study represents the most comprehensive comparison of cell phenotype and CSC marker expression between original canine prostatic adenocarcinoma tissue, derived cell lines, and xenografts performed to date.

The establishment of more than 10 canine prostate cancer cell lines is currently reported in the scientific literature [28,29,30,31,32,33,34,35,36,37,38,39]. The first canine prostatic carcinoma cell line, CPA-1, was established from a 10-year-old Doberman with prostatic adenocarcinoma and showed consistent morphology between the primary tumour and xenografts in athymic mice; neither androgen nor oestrogen receptors were found in the primary tumour or the cell line [32]. Similarly, DPC-1, also established from a Doberman with prostatic adenocarcinoma, displayed comparable morphology across the primary tumour, mouse xenografts and an orthotopic allograft generated through the injection of the cell line into the prostate of an immunodeficient dog [31]. The third cell line, CT1258, established from a 10-year-old Briard, was likewise reported to maintain morphological similarity between the primary tumour and xenografts with high proliferative index and peritoneal effusion production, but lack of metastatic spread [28,35,36].

The cell lines Ace-1 [33], Leo [30], Probasco [29], CHP-1 [34], CHP-2 [39] and LuMa [38] have been shown to produce xenografts in athymic mice that are morphologically similar to the original tumours. PC1 and PC2, originating from a 10-year-old mixed-breed dog and an 11-year-old Poodle, respectively, produced tumour spheres that expressed stemness markers including Oct3/4, Nestin, Nanog, and CD44 [25]. Recently, the creation of another four canine prostatic adenocarcinoma cell lines has been reported, but their characteristics in comparison to the originating tumours were not described [37].

Thus, the previously published studies report only morphological similarity between primary tumours and xenografts with limited or no comparative molecular characterization between the original tumours and cell lines/xenografts. This lack of detailed characterisation limits the ability to determine how well these models replicate the biology of canine prostate cancer in vivo, which is crucial for interpreting experimental outcomes and selecting appropriate cell lines for specific research questions.

5. Conclusions

In conclusion, the present work provides a detailed phenotypic and molecular comparison performed between primary canine prostatic adenocarcinoma tissue, corresponding cell lines, and xenografts. By explicitly documenting both retained and altered features, this study enhances confidence in the biological relevance of the Kodiak and Bobby models and establishes a robust reference framework for their future application. Furthermore, these findings underscore the importance of rigorous validation of newly developed cancer models to ensure appropriate representation of disease mechanisms and to maximize translational value for both veterinary and human prostate cancer research.