Suppressive effects on metastasis and tumor growth by the knockout of chemokine receptors in a canine cutaneous lymphoma model

Abstract

Canine epitheliotropic cutaneous T-cell lymphoma (ECTCL) is a malignant neoplasm exhibiting various skin lesions. Metastasis to the lymph nodes and distant organs contributes to the poor prognosis of ECTCL; however, the underlying molecular mechanisms remain unclear. In the present study, the roles of chemokine receptors, such as CC chemokine receptor (CCR) 4 and CCR7, in the migration of tumor cells were examined using a canine cutaneous lymphoma model. Three mouse groups were prepared and xenografted with wild-type (WT), CCR4 knockout (KO), or CCR7KO canine ECTCL (EO-1) cells. The proportion of EO-1 cells in tissues and blood was significantly lower in the CCR4KO and CCR7KO groups than in the WT group. Only the iliac lymph node, a sentinel lymph node for the xenograft site, was enlarged with the infiltration of EO-1 cells in all groups. The size of the iliac lymph node was smaller in the KO groups than in the WT group. All mice developed a subcutaneous nodule at the xenograft site, which was smaller in the KO groups. In the cell proliferation assay, chemokine ligands stimulated an increase in EO-1 cells in the WT group, but not in the KO groups. These results suggest that CCR4 or CCR7 plays a vital role in the initial migration to sentinel lymph nodes and proliferation in a canine cutaneous lymphoma model.

INTRODUCTION

Canine lymphoma is classified as multicentric, mediastinal, gastrointestinal, hepatic, renal, ocular, nervous system, and cutaneous lymphoma [65]. Cutaneous lymphoma is further divided into epitheliotropic or non-epitheliotropic based on the histopathological localization of neoplastic lymphocytes. Dogs with epitheliotropic cutaneous T-cell lymphoma (ECTCL) show various skin lesions, including erythema, plaques, scaling, ulcers, and nodules, that vary depending on the individual patient [6, 16, 17]. Although the prognosis is generally poor due to the low efficacy of chemotherapy, some patients exhibit long-term survival [6]. Its clinical, histopathological, and immunohistochemical features have been investigated; however, the factors differentiating skin lesions and prognosis remains unclear. In human ECTCL, incidence, prognosis, and recommended treatments have been described for more than 10 subtypes [61]. Whereas canine ECTCL is classified into only three subtypes: mycosis fungoides, pagetoid reticulosis, and Sézary syndrome [16]. The pathogenesis of canine ECTCL needs to be elucidated in more detail for further classification and the establishment of novel treatments. In dogs with ECTCL, metastasis to the lymph nodes (LNs) and distant organs has also been observed at disease diagnosis or with disease progression [40, 43]. Definitive causes of death have not been identified in canine ECTCL; however, systemic metastasis may be a crucial factor because more than 50% of deaths from cancer are caused by metastasis [13]. The molecular mechanisms responsible for systemic metastasis need to be investigated.

Cancer metastasis consists of a multi-stage process that includes the migration and proliferation of tumor cells [51]. Since cell migration is regulated by interactions between chemokines and chemokine receptors [31], numerous studies have examined the roles of chemokines in cancer metastasis. An initial study revealed that CXC chemokine receptor (CXCR) 4 was highly expressed in human breast cancer. Furthermore, CXC chemokine ligand 12, the ligand for CXCR4, was preferentially expressed in metastatic organs, such as the lungs and LNs, suggesting the involvement of chemokines in organ-specific metastasis. In murine xenografted models, a CXCR4-blocking antibody inhibited lung and LN metastasis. Other chemokine receptors, including CC chemokine receptor (CCR) 6, CCR9, and CXCR3, were also shown to be involved in the metastasis of colorectal cancer [19], melanoma [49], and colon cancer [28], respectively. A study on dogs demonstrated that the transcription levels of CXCR3, CXCR4, CXCR7, and CCR9 were significantly higher in metastatic mammary carcinoma than in non-metastatic mammary carcinoma [1]. Canine cell lines of leukemia and osteosarcoma also expressed functional CCR9 [38] and CXCR4 [15]; however, the relationship between each chemokine receptor and organ-specific metastasis remains unclear in dogs.

Cell proliferation is induced by various molecules, including growth factors, colony-stimulating factors, interleukins, tumor necrosis factors, and chemokines [3]. The roles of chemokines in cell proliferation have been intensively investigated in tumors. In human glioblastoma cell lines expressing CCR5, the stimulation of CC chemokine ligand (CCL) 5 induced cell proliferation with an increase in intracellular phosphor-Akt (p-Akt) [66]. Since the knockdown of CCR5 decreased cell proliferation and p-Akt in the presence of CCL5, the Akt pathway was most likely involved in cell proliferation induced by CCL5 and CCR5. Similarly, tumor growth was suppressed by the inhibition of other chemokine receptors, such as CXCR3 in hepatocellular carcinoma [48], CXCR4 in breast cancer [21], and CCR10 in glioma [7]. These findings suggest that chemokines play an essential role in tumor cell proliferation, which has not yet been examined in dogs.

A previous study on canine ECTCL showed that the transcription levels of CCR4 and CCR7 were significantly higher in skin lesions than in normal skin [8]. A canine ECTCL cell line also expressed CCR4 [20] and CCR7 [26], exhibiting chemotactic activity to their ligands. In human and murine studies, activated keratinocytes and dendritic cells produced the ligands for CCR4, such as CCL17 [57, 58] and CCL22 [59], which promoted the skin tropism of CCR4-positive cells [63]. Murine CCR7 facilitated cell migration from peripheral tissues to LNs via lymphatic vessels at which CCL19 and CCL21, ligands for CCR7, were abundant [4, 11]. These findings indicate that CCR4 and CCR7 are involved in epitheliotropism and LN metastasis, respectively, in canine ECTCL. In mice xenografted with human breast cancer cell lines, CCR4-overexpressing cells formed larger primary tumors and developed more frequent lung metastases compared to wild-type (WT) control cells [30]. CCL21 enhanced in vitro dose-dependent cell proliferation in human lung and bladder cancer cells [42, 62], which was inhibited by the knockdown of CCR7 or a CCR7-blocking antibody. Therefore, the roles of CCR4 and CCR7 in cell proliferation also need to be investigated in canine ECTCL.

The cell line EO-1 has been established for canine ECTCL [20], and xenografted mice developed both skin lesions and systemic metastasis, similar to dogs with ECTCL [22]. In the present study, we examined the pathophysiological roles of CCR4 and CCR7 in tumor cell migration and proliferation in canine ECTCL.

MATERIALS AND METHODS

Cell line

The canine ECTCL cell line EO-1 was cultured in 75 cm² flasks with RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, and 1% penicillin/streptomycin at 37°C in a humidified atmosphere of 5% CO₂. The culture medium was changed three times each week. To validate the EO-1 cell line, its doubling time, morphology, and surface markers, such as CD3, CD8, and TCRγδ, were confirmed as previously reported [20].

Knockout (KO) of chemokine receptors in EO-1

The CRISPR/Cas9 system was used for the KO of the CCR4 or CCR7 genes in EO-1 cells. The sequences of guide RNAs (gRNAs) for canine CCR4 and CCR7 were designed using CHOPCHOP [32] and CRISPR direct [45]. gRNAs were as follows: CCR4: 5′-TAAGCCTTGCACCAAAGAAG-3′ and CCR7: 5′-AGATACGTACCTGCTCAACC-3′. gRNAs were inserted into the plasmid pGuide-it ZsGreen vector (Takara Bio Inc.; Kusatsu, Japan). The plasmid vector was then transfected into EO-1 cells by electroporation using the Neon Transfection System (Thermo Fisher Scientific; Waltham, MA, USA). Transfected EO-1 cells were sorted based on green fluorescence by Cell Sorter SH800S (Sony; Tokyo, Japan) and re-cultured. After a further incubation for three weeks, to confirm the loss of CCR4 or CCR7 expression in gene-edited cells, flow cytometric analysis and sorting were conducted. In brief, CCR4 expression was assessed using a mouse anti-human CCR4 monoclonal antibody (Clone:1G1, BD Biosciences; Franklin Lakes, NJ, USA) as the primary antibody, followed by a PE-conjugated goat anti-mouse IgG secondary antibody (Bio Legend; San Diego, CA, USA). For CCR7 detection, a cCCL19-hIgG-Fc fusion protein was employed as described previously [26], followed by an APC-conjugated goat anti-human IgG-Fc secondary antibody (Bio Legend). In both assays, isotype control staining was used to define background fluorescence. Cells exhibiting fluorescence intensities equivalent to isotype controls were gated as CCR4-knockout (KO) or CCR7KO cells and were sorted using Cell Sorter SH800S (SONY). Finally, the proportion of KO cells reached more than 99% (Supplementary data 1). Regarding the features of EO-1 cells described above, no changes were detected in CCR4KO or CCR7KO EO-1 cells.

To evaluate the efficacy of KO on cell migration, a chemotaxis assay was conducted in a 24-transwell plate with a 5-µm pore insert (Corning Inc.; Corning, NY, USA). In brief, mouse recombinant CCL17 (Bio Legend) and CCL19 (Prospec; Ness-Ziona, Israel) were diluted in RPMI 1640 from 0.1 to 500 nmol/L, and 600 µL of the diluted solution was placed in the lower chamber. WT, CCR4KO, and CCR7KO EO-1 cells were resuspended in RPMI 1640 at 1 × 10⁷ cells/mL, and 100 µL of each cell suspension was added to the upper chamber. After an incubation at 37°C for 3 hr under 5% CO₂, the number of migrated cells in the lower chamber was counted. Neither CCR4KO nor CCR7KO EO-1 cells showed active chemotaxis to each chemokine ligand (Supplementary data 2). The assay was independently performed twice.

Xenograft mice

Seven-week-old female NOD.Cg-Prkdc^scid/J (NOD-SCID) mice (formerly designated NOD.CB17-Prkdc^scid/J) were obtained from The Jackson Laboratory Japan, Inc. (Yokohama, Japan). Mice were maintained under specified pathogen-free conditions and fed autoclaved food and water. As an adaptation period, mice were observed for one week before xenograft implantation. All procedures were approved by the Institutional Animal Care and Use Committee of our university (Approval #2020-254).

Mice were divided into three groups (WT, CCR4KO, and CCR7KO). A total of 1 × 10⁸ EO-1 cells were subcutaneously injected into the left thigh of four mice per group under anesthesia with the inhalation of isoflurane (Pfizer, Tokyo, Japan). The general condition of mice was monitored daily with measurements of body weight every 3 days. Humane endpoints were set to include signs of marked changes in locomotion and posture, difficulties in accessing or ingesting food and drink, weight loss >15%, signs of pain (grimace scale >1) [33]. Twenty-eight days after xenograft implantation, WT mice reached humane endpoints due to tumor progression and were euthanized. Although CCR4KO and CCR7KO mice did not exhibit humane endpoint criteria, they were also euthanized on day 28 to allow for synchronized tissue collection and comparison across groups. Tissue samples from skin (above the injection site), LNs, the spleen, lung, liver, kidney, heart, and bone marrow were fixed in 10% neutral buffered formalin and embedded in paraffin for immunohistochemistry (IHC). Peripheral blood was collected in a heparin-containing syringe by cardiac puncture. Subcutaneous nodular sizes (longest diameter) were measured only at the endpoint after euthanasia and skin incision in order to obtain accurate and consistent measurements. The weight of nodules was not collected.

This experiment of xenograft mice was not conducted repeatedly in the view of the policy for reducing the use of experimental animals. A mouse model xenografted with the canine ECTCL cell line (EO-1) was established in the previous research [22], which ensured the reproducibility of the WT group in the present study. The number of mice per group was also determined referred to the previous studies using xenograft mice, in which multiple times of experiments were not required to obtain reliable results drawing a conclusion [12, 34, 52, 56].

IHC

Paraffin-embedded tissue sections (thickness of 3 µm) were pretreated in antigen retrieval solution (Dako; Glostrup, Denmark) by autoclaving at 121°C for 15 min followed by an incubation with either rabbit anti-human CD3 (C7930, Sigma) [22, 44] or normal rabbit IgG (Fujifilm Wako). After an overnight incubation at 4°C, sections were treated with a horseradish peroxidase-labeled polymer (Dako). The chromogen 3,3’-diaminobenzidine (DAB, Dako) was used for the visualization of immunolabelling, followed by counterstaining with hematoxylin. Slides were observed with an optical microscope (DM2500, Leica; Wetzlar, Germany).

To assess the potential contribution of host-derived CD3-positive cells, additional immunohistochemistry was performed on archived, tumor-untransplanted NOD-SCID mice at 12 weeks of age. Lung, spleen, and liver tissues were examined, and only sporadic CD3-positive cells were observed, with an area fraction below 0.1% (Supplementary data 3). Therefore, the influence of host-derived CD3-positive cells on the quantification of EO-1 cell infiltration was considered negligible, and CD3 immunoreactivity was used as a marker for evaluating EO-1 distribution in tissues.

Quantitative evaluation of EO-1 cells in tissues

The distribution of EO-1 cells in each tissue was quantified using ImageJ Fiji software ver. 2.3.0 [10, 50] as previously described [18, 25, 30]. In brief, hematoxylin (non-EO-1 cell region) and DAB staining (EO-1 cell region) were separated from IHC images by a color deconvolution procedure. The area of each stain was then measured. The area of infiltrated EO-1 cells was shown as a percentage of the DAB area in the hematoxylin area. Although cell counting is often used for discrete round cell populations, EO-1 cells do not exhibit typical round morphology [20]; thus, the area-based semi-quantitative method was selected for its established utility and reproducibility. This method is widely used for IHC quantification in lymphoid and immune cell–rich tissues, particularly when analyzing broad tissue sections using high-throughput image workflows [10, 50].

Quantitative evaluation of EO-1 cells in peripheral blood

The white blood cells of mice were isolated from 100 µL of peripheral blood by lysing buffer (BD Biosciences), followed by an incubation with FITC-conjugated mouse anti-canine CD3 (clone CA17.2A12, Bio-Rad Laboratories; Hercules, CA, USA) in washing buffer at 4°C for 30 min. An appropriate isotype control (FITC-conjugated mouse IgG1, clone MOPC-21, Bio Legend) was used. The percentage of EO-1 cells, which were positive for CD3, in white blood cells was evaluated by a flow cytometer (FACS Canto II; BD Biosciences). This CD3 antibody is specific for canine CD3 and is not known to cross-react with murine CD3 under our experimental conditions [41]. Therefore, the CD3-positive cells detected in this assay were considered to be EO-1-derived.

Chemokine transcription profiles in tissues

Chemokine transcription levels of various organs in the WT, CCR4KO, and CCR7KO groups were analyzed by real-time PCR. Correlations between chemokine transcription and the extent of EO-1 cell infiltration were also evaluated. In detail, lung, spleen, skin, liver, kidney, heart, iliac lymph node (LN), and blood were immediately submerged in RNAlater (QIAGEN, Venlo, Netherlands) after sampling and stored at −20°C until extraction of total RNA. The total RNA of tissue was extracted using the RNeasy Mini Kit (QIAGEN). Genomic DNA was removed with a TURBO DNA-free Kit (Applied Biosystems, Foster, CA, USA). Complementary DNA was synthesized from RNA using the PrimeScript^TM RT Reagent Kit (Takara Bio Inc.). The transcription levels of chemokine ligands (CCL17, CCL19, CCL21, and CCL22) were quantified by two-step RT-PCR (Thermal Cycler Dice Real Time System TP800; Takara Bio Inc.) using SYBR Premix Ex Taq^TM II (Takara Bio Inc.). Two reference genes, SDHA and RPL27, were selected by GeNorm and Bestkeeper among four candidate reference genes (GUSB, HMBS). Each primer sequence is listed in the table (Supplementary data 4), and was confirmed not to amplify canine EO-1 genes. PCR amplification consisted of an initial activation step (95°C for 10 sec) followed by 40 cycles of denaturation (95°C for 5 sec) and annealing/extension (60°C for 30 sec), and a dissociation step (95°C for 15 sec, 60°C for 30 sec, 95°C for 15 sec). Each PCR run included no-RT controls, which show no amplification. The ΔCt value was calculated by subtracting the geometric mean of the Ct values of the two reference genes from the Ct value of each chemokine ligand. All samples were assayed in duplicate, and the mean ΔCt was used for further analysis. Relative transcription levels were expressed as 2^-∆Ct, representing the n-fold difference relative to the mean of the two reference genes.

Cell proliferation assay

Cell proliferation was examined using Cell Counting Kit-8 (CCK-8, Fujifilm Wako) as previously reported [62]. In brief, WT, CCR4KO, and CCR7KO EO-1 cells were seeded on 96-well plates at 2 × 10⁴ cells/100 µL and treated with 100 ng/mL of recombinant murine CCL17, CCL19, CCL21 (PEPRO TECH; Cranbury, NJ, USA), or CCL22 (PEPRO TECH) for 0, 24, 48, or 72 hr. After the treatment, a coloring reagent of CCK-8 was added to each well according to the manufacturer’s instructions and incubated at 37°C for 2 hr. The optical density value of each well was measured using a microplate reader (Bio-Rad) at a wavelength of 450 nm. Cell proliferation at each time point was shown as a percentage relative to 0 hr. The assay was independently repeated three times.

Transcription analysis of CCR4 downstream signaling in EO-1 cells

To investigate CCR4 downstream signaling, real-time PCR was performed to compare the mRNA levels of downstream signaling molecules between CCR4-positive EO-1 cells and CCR4-negative canine lymphoma cell lines. Three canine lymphoma cell lines were used: EO-1 (CCR4⁺), UL-1 (CCR4⁻), and CLK (CCR4⁻) [56]. Total RNA was extracted from each cell line using the RNeasy Mini Kit according to the manufacturer’s instructions. Genomic DNA was removed with the TURBO DNA-free Kit, and RNA samples were stored at −80°C until use. Complementary DNA (cDNA) was synthesized from total RNA using the PrimeScript™ RT Reagent Kit (Takara Bio Inc.). Gene transcription was quantified by two-step RT-PCR using the Thermal Cycler Dice Real Time System III (Takara Bio Inc.) with primers listed in Supplementary data 5 and TB Green Premix Ex Taq™ II (Takara Bio Inc.). Ribosomal protein L13A (RPL13A), ribosomal protein L32 (RPL32), and β2-microglobulin (B2M) served as reference genes [26]. The cycling conditions were as follows: initial denaturation at 95°C for 10 sec, followed by 40 cycles of denaturation at 95°C for 5 sec and annealing/extension at 60°C for 30 sec, and a dissociation step (95°C for 15 sec, 60°C for 30 sec, 95°C for 15 sec). ΔCt values were normalized to the geometric mean of the three reference genes, and relative transcription levels were calculated using the 2^−ΔCt method. Each reaction was performed in duplicate, and mean values were used for analysis.

Cell proliferation inhibition assay

To examine whether blockade of CCR4 downstream signaling suppresses cell proliferation, EO-1 cells were treated with specific inhibitors targeting molecules in the CCR4 signaling pathway, and their proliferative activity was evaluated. In brief, EO-1 cells (2 × 10⁴ cells/100 µL) were seeded into 96-well plates and pretreated with or without MBQ-167 (0.5 µM; Cayman Chemical, Ann Arbor, MI, USA), a dual Rac/Cdc42 inhibitor, or Y27632 (10 µM; Cayman Chemical), a selective ROCK1/2 inhibitor, for 12 hr under standard culture conditions (37°C, 5% CO₂). After pretreatment, recombinant murine CCL17 (100 ng/mL; PEPRO TECH) or vehicle was added and the cells were incubated for 0, 24, 48, and 72 hr. At each time point, cell proliferation was evaluated using Cell Counting Kit-8 according to the manufacturer’s instructions. After 3 hr of incubation with CCK-8 reagent, optical density was measured at 450 nm using a microplate reader (Bio-Rad). The proliferation rate of EO-1 cells pretreated with inhibitors was shown as the ratio of optical density relative to inhibitor-untreated wells.

Statistical analysis

Dunnett’s test was conducted to compare the distribution of EO-1 cells, the body weight of mice, and the size of nodules and LNs between the WT and KO groups. Cell proliferation was also compared between control cells and cells treated with chemokines by Dunnett’s test. The baseline doubling times of EO-1 cells under unstimulated conditions were compared among the three groups using the Kruskal-Wallis test. Differences in chemokine transcription levels among the groups were first evaluated using the Kruskal–Wallis test. When a significant difference was detected, pairwise comparisons between WT and each knockout group were performed using the Mann-Whitney U test. Correlations between the histopathological extent of EO-1 cell infiltration in each organ and the transcription levels of chemokine ligands were assessed using Spearman’s rank correlation coefficient. The transcription levels of CCR4 downstream signaling molecules were compared among groups by Dunn’s multiple-comparison test. For the proliferation inhibition assay, the mean proliferation ratio from three independent experiments was calculated, and differences between the CCL17-stimulated and vehicle control groups were analyzed using the Mann–Whitney U test. A value of P<0.05 was considered significant. Statistical analyses were performed using a statistical software package (JMP ver. 13.2.0, SAS Institute; Cary, NC, USA).

RESULTS

Xenograft mice

All mice gained weight until approximately day 24 after xenograft implantation. From day 25 onward, a gradual decrease in body weight was observed in the WT group, while the CCR4KO and CCR7KO groups maintained their body weight. On day 27, the body weight of the WT group was significantly lower than that of the CCR4KO and CCR7KO groups (P<0.01, Fig. 1). In addition to body weight loss, mice in the WT group exhibited decreased activity and reduced food intake beginning around day 25. These clinical signs progressed until day 28, at which point the mice met humane endpoint criteria and were euthanized. No comparable behavioral changes were observed in the CCR4KO or CCR7KO groups.

Fig. 1.

Changes in the body weight of mice based on day 0 (the date of xenograft implantation). Data are shown as the means ± SD of four mice. **P<0.01 by Dunnett’s test. WT: wild-type, CCR4KO: CC chemokine receptor 4 knockout, CCR7KO: CC chemokine receptor 7 knockout.

Distribution of xenografted EO-1 cells

In all tissues and blood, the proportion of EO-1 cells was significantly lower in the CCR4KO and CCR7KO groups than in the WT group (Fig. 2). In tissues, EO-1 cells frequently localized around blood vessels. Normal lung structures disappeared due to an increase of EO-1 cells in the WT group (Fig. 3). Iliac lymphadenopathy was observed in 4/4, 3/4, and 3/4 mice in the WT, CCR4KO, and CCR7KO groups respectively, whereas mesenteric, inguinal, and bronchial lymphadenopathies were only detected in the WT group (Table 1). Iliac LNs in the WT group (diameter: 7.9 ± 0.22 mm) were significantly larger (P<0.001) than those in the CCR4KO (5.0 ± 0.71 mm) and CCR7KO (5.0 ± 0.41 mm) groups. Lymphadenopathy was attributed to the massive infiltration of EO-1 cells.

Fig. 2.

The proportion of xenografted EO-1 cells in tissues and blood. (a) Tissue samples were analyzed by immunohistochemistry and quantified as the DAB-positive area (CD3-positive cells) relative to the hematoxylin-stained area. (b) Blood samples were analyzed by flow cytometry and quantified as the percentage of CD3-positive EO-1 cells among total white blood cells. Each bar represents the mean ± SD of four mice in the wild-type (WT), CC chemokine receptor 4 knockout (CCR4KO), and CCR7KO groups. *P<0.05, **P<0.01, ***P<0.001 by Dunnett’s test.

Fig. 3.

Representative immunohistochemical images of the lung, liver, and kidney. EO-1 cells are shown as DAB-stained CD3-positive cells. Scale bars represent 50 µm. WT: wild-type, CCR4KO: CC chemokine receptor 4 knockout, CCR7KO: CC chemokine receptor 7 knockout.

Table 1. The number of mice with lymphadenopathy in each lymph node.

	WT	CCR4KO	CCR7KO
Iliac	4	3	3
Mesenteric	3	0	0
Inguinal	2	0	0
Bronchial	1	0	0

Lymph nodes were detected only when enlarged with the infiltration of EO-1 cells. Iliac lymphadenopathy was recognized in the wild-type (WT), CC chemokine receptor 4 knockout (CCR4KO) and CCR7KO groups, whereas other lymphadenopathies were observed only in the WT group.

Chemokine transcription profiles

Real-time PCR analysis revealed no significant differences in chemokine transcription among the three groups (Fig. 4a). Spearman’s rank correlation analysis revealed a strong positive correlation between CCL22 transcription and the extent of EO-1 cell infiltration in the WT group (ρ=1.00, P<0.0001). CCL22 transcription was also strongly correlated with infiltration in the CCR4KO group (ρ=0.93, P=0.0025). In addition, CCL17 transcription showed significant positive correlations with infiltration in the CCR4KO (ρ=0.82, P=0.023) and CCR7KO groups (ρ=0.82, P=0.023) (Fig. 4b).

Fig. 4.

Tissue transcription profiles of chemokine ligands and their correlations with EO-1 cell infiltration. (a) The relative transcription levels of CC chemokine ligand (CCL) 17, CCL19, CCL21, and CCL22 in lung, spleen, skin, liver, kidney, heart, iliac lymph node (LN), and blood were quantified by real-time PCR. Transcription levels are shown as 2^-ΔCt. Higher values indicate higher relative transcript abundance. Each bar represents the mean ± SD of four mice in the wild-type (WT), CC chemokine receptor 4 knockout (CCR4KO), and CCR7KO groups (iliac LN: three mice in the CCR4KO and CCR7KO groups). (b) Correlations between EO-1 cell infiltration and chemokine transcription levels in each group. *P<0.05, **P<0.01, ***P<0.001 by Spearman’s rank correlation coefficient.

Subcutaneous nodule formation

All mice developed nodules in the subcutis at which EO-1 cells had been injected (Fig. 5). First detection date of nodules was day 18 ± 0.0 in the WT, day 21 ± 2.1 in CCR4KO and day 23 ± 1.3 in CCR7KO groups. At day 28, the size of the nodules was evaluated based on the longest diameter after skin incision. As the minor diameter was not recorded and the shapes of the nodules were frequently irregular rather than ellipsoidal, volume calculations were not performed. Nodules were significantly smaller in the CCR4KO and CCR7KO groups than in the WT group (Fig. 6).

Fig. 5.

Representative images of subcutaneous nodules. WT: wild-type, CCR4KO: CC chemokine receptor 4 knockout, CCR7KO: CC chemokine receptor 7 knockout.

Fig. 6.

The length of subcutaneous nodules. Nodule length was measured as the longest diameter. **P<0.01 by Dunnett’s test. WT: wild-type, CCR4KO: CC chemokine receptor 4 knockout, CCR7KO: CC chemokine receptor 7 knockout.

Cell proliferation assay

In the WT group, treatment with CCL17, CCL19, CCL21, and CCL22 for 48 hr resulted in increased proliferation of EO-1 cells to 282%, 302%, 312%, and 308%, respectively, compared to 246% in the control condition. These increases were statistically significant (P<0.05, Fig. 7a). At 72 hr, treatment with CCL19 and CCL21 continued to show significantly increased proliferation, whereas CCL17 and CCL22 did not elicit significant changes compared with controls. In CCR4KO and CCR7KO cells, no significant differences in cell proliferation were observed between chemokine-treated and untreated conditions at any time point (Fig. 7b, 7c). The baseline doubling times under unstimulated conditions were 43.10 ± 1.95 hr for the WT cells, 39.59 ± 2.04 hr for CCR4KO cells, and 38.93 ± 2.77 hr for CCR7KO cells. The Kruskal-Wallis test showed no statistically significant difference among the three groups (H=4.36, P=0.113).

Fig. 7.

Effects of chemokine ligands on the proliferation of EO-1 cells. (a) wild-type (WT) EO-1 cells treated with vehicle (control) or CC chemokine ligand (CCL) 17, CCL19, CCL21, or CCL22. (b) CC chemokine receptor 4 knockout (CCR4KO) EO-1 cells treated with vehicle or CCL17 or CCL22. (c) CCR7KO EO-1 cells treated with vehicle or CCL19 or CCL21. Each bar represents the mean ± SD of three independent experiments. *P<0.05, ***P<0.001 by Dunnett’s test.

Transcription analysis of CCR4 downstream signaling

Transcription levels of PREX1 and ARHGEF5 were significantly higher in EO-1 cells than in both UL-1 and CLK cells, whereas ARHGEF18 was significantly higher in EO-1 cells compared with UL-1 cells. There were no significant differences in the transcription levels of PREX2 and PLEKHG2 among the cell lines (Fig. 8).

Fig. 8.

Transcription of Rho family guanine nucleotide exchange factors (GEFs) in three canine lymphoma cell lines. PREX1, PREX2, PLEKHG2, ARHGEF5, and ARHGEF18 mRNA levels were quantified by real-time PCR in EO-1 (CC chemokine receptor 4: CCR4⁺), UL-1 (CCR4⁻), and CLK (CCR4⁻) cells. Data represent mean ± SD. Transcription levels were normalized to the geometric mean of RPL13A, RPL32, and B2M and calculated using the 2^−ΔCt method. *P<0.05 by Dunn’s multiple comparisons.

Cell proliferation inhibition assay

MBQ-167 treatment significantly reduced the proliferation ratio of EO-1 cells stimulated with CCL17 at 72 hr compared with the vehicle control (Fig. 9a). In contrast, Y27632 treatment had no significant effect on EO-1 cell proliferation (Fig. 9b).

Fig. 9.

Proliferation ratio of EO-1 cells with or without CC chemokine ligand (CCL) 17 stimulation at each time point, normalized to the inhibitor-untreated wells. (a) treated with MBQ-167, (b) treated with Y27632. Data are shown as mean ± SD from three independent experiments. *P<0.05 by the Mann-Whitney U test.

DISCUSSION

In this study, we used an in vivo NOD-SCID xenograft model along with in vitro assays to examine how CCR4 and CCR7 contribute to the metastatic behavior and proliferative capacity of EO-1 canine ECTCL cells. By comparing the WT, CCR4KO, and CCR7KO groups, we aimed to clarify the intrinsic roles of these receptors. The following discussion interprets the receptor-specific findings and considers their implications for understanding the pathogenesis and potential treatment of canine ECTCL.

Immunohistochemical and flow cytometric analyses revealed that the proportion of EO-1 cells in tissues and blood was significantly lower in the CCR4KO and CCR7KO groups than in the WT group. Large amounts of infiltrating EO-1 cells may induce multiple organ failure contributing to the decreased body weight in the WT group. Histopathological observations of the WT group also showed that EO-1 cells localized around blood vessels in metastatic tissues. Previous studies revealed that tumor cells increased in perivascular sites after extravasation when injected intravascularly [5, 29]. Collectively, these findings and the present results indicate that EO-1 cells infiltrate metastatic organs from blood vessels; therefore, the proportion of EO-1 cells in tissues has been attributed to the number of EO-1 cells in blood. According to a recently published chemokine atlas [64], CCL17 and CCL22, the ligands for CCR4, are mainly expressed in lung, whereas CCL19 and CCL21, the ligands for CCR7, are predominantly expressed in spleen in mice, excluding primary and secondary lymphoid organs such as thymus, bone marrow, and LNs. Hence, we assumed that the interruption of chemotaxis would prevent the metastasis to lung and spleen in the CCR4KO and CCR7KO groups, respectively. However, no significant differences were observed in the types of metastatic organs among groups. Importantly, the expression of chemokines in tissues is not static but can be dynamically altered by tumor infiltration and local inflammation [14, 24, 46]. We therefore conducted real-time PCR to examine the chemokine transcription profiles of the tissues from tumor-bearing NOD-SCID mice in this study. First, we compared the transcription levels of each chemokine among the WT, CCR4KO, and CCR7KO groups for each organ. However, no significant group-dependent differences were detected in any chemokines. In addition, except for the iliac LNs, CCL17 and CCL22 were expressed predominantly in the lung, whereas CCL19 and CCL21 were expressed in the spleen, in accordance with the published chemokine atlas. These findings suggest that EO-1 cell infiltration and the associated inflammatory response did not markedly alter the intrinsic tissue-specific transcription patterns of these chemokines in this study. We then analyzed whether the extent of EO-1 cell infiltration in each organ correlated with the transcription levels of individual chemokines. Some chemokines, such as CCL17 in the CCR4KO and CCR7KO groups and CCL22 in the WT and CCR4KO groups, exhibited organ-specific correlations with the extent of EO-1 cell infiltration. However, given that the CCR4KO group lacks the functional interaction between CCR4 and its ligands and that most other chemokine-receptor combinations showed no correlation, these correlations are likely to reflect incidental statistical associations rather than biologically relevant mechanisms. Therefore, these findings indicate that once EO-1 cells have entered the blood circulation, their organ tropism is largely independent of local chemokines, and is unlikely to be dictated by chemokine gradients within individual tissues.

LNs are a crucial immune site for the antigen presentation of dendritic cells and the activation of naïve lymphocytes. Naïve lymphocytes enter LNs via high endothelial venules; however, recent studies showed that activated lymphocytes returned to LNs from peripheral tissues through draining lymph vessels [4, 54]. Cancer cells have been shown to exploit draining lymph vessels for their exit from primary locations and infiltration into LNs [27]. Therefore, we focused on the migration of EO-1 cells into LNs. In the present study, LNs were only visible when enlarged with the infiltration of EO-1 cells. In the WT group, iliac lymphadenopathy was detected in all four mice, whereas the enlargement of mesenteric, inguinal, and bronchial LNs was recognized in 3/4, 2/4, and 1/4 mice, respectively. In the CCR4KO and CCR7KO groups, lymphadenopathy was only observed in iliac LNs. Since iliac LNs are sentinel LNs for the thigh, into which EO-1 cells were injected, it is plausible that EO-1 cells left the subcutis and entered iliac LNs, followed by the circulation. The size of iliac LNs was significantly smaller in the CCR4KO and CCR7KO groups, which may have reduced the number of EO-1 cells in blood. Therefore, the reduction in chemotaxis to sentinel LNs appeared to contribute to the suppression of subsequent systemic metastasis in the CCR4KO and CCR7KO groups. The most well-described chemokines for cell migration to LNs are CCL19 and CCL21 [4, 11]; however, CCL22 was also shown to be produced in large amounts by dendritic cells in LNs [47, 53], facilitating the recruitment of CCR4-positive T-cells to LNs. Consistent with these reports, CCL19, CCL21, and CCL22 expression was also detected in iliac LNs in our experiment. In the CCR4KO and CCR7KO groups, an interruption between CCR4 or CCR7 and their ligands appeared to inhibit the migration of EO-1 cells from the subcutis to sentinel iliac LNs, thereby reducing subsequent lymphatic dissemination.

Subcutaneous nodule formation was significantly less in the CCR4KO and CCR7KO groups than in the WT group. To establish whether chemokines facilitated tumor growth, we conducted a cell proliferation assay, the results of which indicated that chemokine ligands accelerated the proliferation of EO-1 cells. Notably, the results also showed that only CCL19 and CCL21 sustained proliferation beyond 48 hr, whereas CCL17 and CCL22 failed to do so. This discrepancy is most plausibly explained by distinct receptor binding kinetics. Previous studies demonstrated that CCR4 undergoes rapid β-arrestin-mediated desensitization and internalization, particularly after CCL22 engagement [23, 39], whereas CCR7 exhibits only modest ligand-induced down-regulation; CCL21 scarcely recruits β-arrestin and even CCL19 induces only partial, readily reversible internalization [2, 9]. These kinetic differences may account for the sustained proliferative effect observed with CCL19 and CCL21 at 72 hr in our study. Taken together with the reduced tumor burden observed in CCR4KO and CCR7KO groups, these findings suggest that CCR4 and CCR7 were also involved in tumor growth in canine ECTCL murine models.

CCR4 can signal through β-arrestin to activate Rho-associated coiled-coil-containing protein kinase (ROCK) [35, 36]. Because Rho activation needs Rho guanine nucleotide exchange factors (RhoGEFs) as essential upstream regulators, we hypothesized that some RhoGEFs might connect CCR4 activation with cell proliferation. We therefore checked major RhoA-oriented GEFs, such as ARHGEF family members, and Rac-specific GEFs, such as PREX1/2, to find candidates that could link CCR4 signaling to Rho GTPase activation. Transcription analysis revealed that PREX1, ARHGEF5, and ARHGEF18 were upregulated in CCR4-positive EO-1 cells compared with CCR4-negative lymphoma cell lines. We next tested pharmacological inhibitors targeting these GEFs and their downstream GTPases to assess their role in EO-1 cell proliferation. Inhibition of Rac, but not RhoA, significantly suppressed EO-1 cell proliferation, suggesting that Rac signaling contributes to CCR4-dependent cell growth. In line with our finding that Rac contributes to CCR4-mediated proliferation, previous studies have reported diverse downstream pathways involved in chemokine receptor-driven cell growth. In human cervical cancer cells, blockade of a panel of signaling molecules (Akt, JNK, STAT3, STAT5, MAPK, ERK, NFκB) under CCL17 stimulation revealed that only JNK or STAT5 inhibition suppressed proliferation [37]. In lung cancer cells, CCL21-CCR7 signaling promoted proliferation through ERK phosphorylation [62]. In metastatic head and neck squamous-cell carcinoma, autocrine CCR7 signaling maintained cell survival and proliferation via the PI3K-mTOR-Akt pathway, as shown by inhibitor and siRNA experiments [60]. Similarly, in breast cancer cells, CCL21-induced CCR7 signaling enhanced proliferation through Akt phosphorylation, which was abolished by Akt blockade [55]. These findings collectively indicate that multiple intracellular cascades including JNK, STAT5, ERK, Akt, and PI3K-mTOR may mediate CCR4- or CCR7-dependent proliferation in tumor cells. Taken together, our findings identify Rac activation as a key downstream event of CCR4 signaling that promotes EO-1 cell proliferation, yet multiple other pathways are likely involved. Therefore, comprehensive pathway analyses, such as phosphor-proteomic profiling or RNA-seq with pathway enrichment, will be required to fully delineate the mechanisms underlying EO-1 cell proliferation.

In conclusion, the present study demonstrated that CCR4 and CCR7 facilitated the migration of EO-1 cells to sentinel LNs, which contributed to subsequent systemic metastasis. CCR4 and CCR7 also induced cell proliferation in vitro and in vivo. Since systemic metastasis and local tumor growth are both serious issues in cancer, not just in ECTCL, targeting chemokine receptors may represent a promising therapeutic strategy.

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest related to the subject materials discussed in this article.