In the present study, we successfully established various canine PAC cell lines from a dog. Cellular morphology varies among cell lines, including epithelial, mesenchymal, and intermediate types. Moreover, certain cell lines were suspected to have developed EMT with increased vimentin expression and high motility, and they showed CSC-like phenotypes characterized by high sphere-forming ability and high levels of SP cells. Sensitivity to vinorelbine in vitro varied among the cell lines. These findings are summarized in Table 1. CPACd-AB4 and CPACd-CG11 exhibited both EMT- and CSC-like phenotypes and showed resistance to vinorelbine. Therefore, the two clones may have high malignancy rates among the newly developed cell lines in the present study.
Table 1
Characteristics of the canine PAC cell lines.
|
CPACd
|
CPACr
|
-p
|
-AB4
|
-AC12
|
-AD4
|
-AE3
|
-BB4
|
-CG11
|
-DC10
|
-p
|
-AD4
|
-AH4
|
-BC9
|
-BH12
|
-CB2
|
-DC3
|
-DD10
|
Cellular morphology a
|
E
|
E
|
E
|
E
|
I
|
E
|
E
|
M
|
E, I
|
E
|
M
|
I
|
I
|
E
|
I
|
E
|
Proliferative capasity b
|
++
|
+
|
++
|
++
|
++
|
+++
|
++
|
++
|
++
|
+++
|
++
|
+++
|
++
|
++
|
+++
|
++
|
EMT
|
Expression of vimentin
|
+
|
++
|
+
|
-
|
++
|
-
|
+
|
+
|
+
|
+
|
+
|
++
|
+
|
-
|
++
|
+
|
Migration ability
|
(WHA) c
|
+
|
++
|
+
|
±
|
++
|
±
|
++
|
+
|
+
|
+
|
±
|
++
|
+
|
++
|
++
|
+
|
(TMA) d
|
+
|
++
|
+
|
±
|
±
|
±
|
++
|
+
|
+
|
+
|
+
|
++
|
±
|
±
|
±
|
±
|
Invasion ability d
|
±
|
++
|
±
|
±
|
±
|
±
|
+
|
±
|
+
|
+
|
±
|
++
|
±
|
±
|
±
|
±
|
CSC
|
Sphere forming ability e
|
±
|
+
|
±
|
+
|
++
|
+
|
+
|
±
|
+
|
+
|
+
|
±
|
++
|
+
|
+
|
±
|
Population of SP f
|
±
|
+
|
±
|
±
|
±
|
±
|
++
|
++
|
±
|
+
|
±
|
±
|
±
|
±
|
±
|
++
|
Sensitivity to vinorelbine g
|
+
|
±
|
±
|
+
|
++
|
++
|
±
|
+
|
++
|
++
|
++
|
+
|
++
|
+
|
++
|
+
|
EMT: Epithelial Mesenchymal Transition, CSC: Cancer Stem Cell, WHA, Wound-Healing Assay; TMA, Traswell Migration Assay; SP, side population
|
a E: Epithelium, I: Intermediate between epithelium and mesenchmal, M: Mesenchyma
|
b +: >30 hr of doubling time (DT), ++: 20–30 hr of DT, +++: <20 hr of DT
|
c ±: <40% of wound closure, 40–70% of wound closure, ++: >70% of wound closure
|
d ±: <60 migration or invasion cells, +: 60–120 maigration or invasion cells, ++; >120 migration or invasion cells
|
e ±: <15 spheres, +: 15–30 spheres, ++; >30 spheres
|
f ±: <5% SP, +: 5–10% SP, ++; >10 SP
|
g ±: >1000 nM of IC50, +: 10-1000 nM of IC50, ++: <10 nM of IC50
|
Traditionally, a tumor was considered a group of cells with similar characteristics. However, currently, a tumor tissue is recognized as an assembly of heterogeneous tumor cells with distinct cellular properties, including morphology, proliferative potential, motility, and function. In the present study, we successfully extracted and established various tumor clones with infinite proliferative capacity from canine PAC tissues. Although it is unknown why various cell lines are obtained from one tissue, this variety may reflect the heterogeneity and differentiation potential of tumor cells.
EMT is associated with malignancy and is considered a crucial therapeutic target [11–15, 33, 34]. Certain cell lines in the present study showed a mesenchymal-like shape, which is a morphological characteristic of EMT. In the cloned cell lines, CPACd-AE3, CPACr-AH4, CPACr-BC9, and CPACr-DC3 showed mesenchymal-like morphologies and relatively high vimentin expression. In contrast, CPACd-DC10, which had a spindle shape, did not show high vimentin expression, whereas CPACd-AB4, which had an epithelial shape but a relatively wide gap between cells, showed relatively high vimentin expression. According to the migration and invasion analyses, CPACd-AB4, CPACd-CG11, and CPACr-BC9 showed high motility. CPACr-BC9 showed mesenchymal morphology, whereas CPACd-AB4 and CPACd-CG11 showed epithelial morphologies but exhibited relatively high vimentin expression. Therefore, EMT may have occurred in these three cell lines. In addition, these findings indicated that EMT cannot be determined using cellular morphological analysis alone.
CSCs are a subset of tumor cells characterized by self-renewal, differentiation capacity, and chemoresistance. These properties make CSCs crucial targets for tumor therapy [14, 17, 18]. In the present study, we assessed CSC properties using two different methods: sphere-formation and SP assays with DCV. CPACd-AB4, CPACd-CG11, and CPACr-AD4 showed relatively high sphere-forming ability and SP population; however, in CPACd-AE3, CPACd-DC10, CPACr-AH4, and CPACr-BH12, the results of the two assays were not concordant. Assessing CSC-associated antigens is widely used for detecting CSCs. In human oncology, several antigens are used as CSC markers, including CD44, CD133, CD24, and EpCAM; however, the expression patterns appear to vary with respect to tumor origin [16, 18]. Although CD44 and CD133 expression have been previously assessed in a xenograft model of canine PAC, their utility as CSC markers has not been indicated [35]. In human PAC, these two antigens have been reported to have poor diagnostic value in CSC detection [36]. Reliable markers for the detection of CSCs may be determined by further assessment using developed canine PAC cell lines.
EMT and CSCs are closely linked [14, 16]. In our study, CPACd-AB4 and CPACd-CG11 showed both EMT and CSC phenotypes. Several transcription factors are associated with the development of EMT and maintenance of CSCs. Transforming growth factor-β (TGF-β) has pleiotropic functions in tumor progression [37]. In epithelial cells, TGF-β acts as a tumor suppressor by inhibiting cell growth and inducing apoptosis. In contrast, in the advanced stages of tumorigenesis, TGF-β induces EMT and stimulates cell proliferation and survival in tumor cells [38]. TGF-β overexpression has been reported to be associated with malignancy in human tumors [39]. In human liver tumor cells, autocrine stimulation of TGF-β was reported to induce EMT [40]. Importantly, EMT stimulated by TGF-β induces the selection and expansion of CSCs [41, 42]. Moreover, previous reports have indicated that MAPK and PI3K signaling play important roles in EMT or the maintenance of CSCs by EMT [43–46]. In our study, relatively high phosphorylation of Erk1/2 and expression of vimentin were observed in CPACr-AH4 and CPACr-DC3. However, Erk1/2 phosphorylation did not appear to correlate with cellular motility. CPACd-AB4 induced phosphorylation of Akt at a relatively high level; however, the other cell lines with CSC phenotypes did not show high phosphorylation of Akt. Further assessment of other signals, including the TGF-β signal, is needed to reveal the detailed mechanism of EMT and CSCs in the canine PAC.
Canine PAC resembles human lung tumors in never-smokers. Adjuvant chemotherapy, including vinorelbine, is indicated for patients with lung cancer. However, in some lung tumors that are intrinsically resistant to chemotherapy, acquired resistance has been shown to develop rapidly even in initial responders [47]. Chemotherapy-resistant cells contribute to relapse. Although few reports are available for canine cases, chemotherapies, including vindesine, mitoxantrone, vinorelbine, and toceranib, are not fully effective [8, 9, 48–50]. Additionally, the mechanisms of resistance have not been elucidated. In the present study, the sensitivity to vinorelbine differed among the cell lines. CPACr-p and the clones, appeared to be relatively sensitive to vinorelbine, and Ras prenylation levels were lower than that in CPACd-p and cloned cell lines. This finding indicates that Ras signaling may be related to the mechanism that determines susceptibility to vinorelbine. However, Erk1/2 and Akt signaling may not be involved in the mechanisms common to these cell lines. Importantly, cell lines with EMT- and CSC-like phenotypes, such as CPACd-AB4 and CPACd-CG11, were resistant to vinorelbine. This finding indicates that EMT and stemness may contribute to chemoresistance in canine PAC. In contrast, although CPACr-AC12 showed resistance to vinorelbine, it showed poor expression of EMT and CSCs phenotypes. Further studies are needed to elucidate the mechanism of resistance to vinorelbine in canine PAC.
The present study had several limitations. First, we did not perform transplantation in mice. Second, we did not assess the transcription factors associated with EMT, including Zeb, SNAIL, and SLUG [34]. These experiments can possibly further clarify the EMT and CSC phenotypes and the detailed mechanisms underlying these phenotypes. Finally, cell lines with various properties may not reflect the heterogeneity in tumor tissue but an in vitro selective pressure during the culture process. However, this possibility is difficult to be excluded.
In conclusion, we successfully generated novel canine PAC cell lines with various characteristics. Moreover, we assessed the cellular properties in vitro, focusing on the EMT and CSC phenotypes. Our results indicate that EMT and stemness may play a role in resistance to vinorelbine. Therefore, these cell lines will facilitate the development of novel therapeutic strategies targeting EMT and CSCs and elucidation of the contribution of tumor heterogeneity to tumor biology, including chemoresistance, in human and canine PAC.