• 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br In the present study di erent ratios of PSCs


    In the present study, different ratios of PSCs could mimic the dif-ferent stages of PDAC spheroids. Some literature has shown that the prominent stromal reaction of increased alpha-smooth muscle NS638 (α-SMA) expression can lead to the worse outcomes on pancreatic cancer patients [48,49]. The literature has also shown that the ratio of the pancreatic cancer cells and PSCs are different in primary pancreatic cancers on different patients (interpatient heterogeneity) and even among different areas of primary pancreatic cancer (intra-tumoral heterogeneity) [10]. Therefore, we designated the amount of PSCs in pancreatic cancer as α-SMA index to correlates with patients’ outcome. 
    We used different ratios of pancreatic cancer cells and PSCs in the study because there was no fixed ratio of pancreatic cancer cells and PSCs in clinical patients. In addition, a higher proportion of PSCs was correlated with the higher malignancy and migration ability of cancer cells in PDAC spheroids (Fig. S11, SI). Such cell organization resembled that observed in clinical pathology of PDAC. As the initial seeding ratio of MIA cells and PSCs increased, the spheroids were even tighter.
    In literature, the population of PSCs plays a critical role in PDAC oncology and is associated with worse patient outcome [48]. The total number of co-cultured cells (MIA cells and PSCs) at 24 h significantly decreased probably because at the initial stage the PSCs started to wrap MIA cells. These results suggested that the crosstalk between PSCs and MIA cells in the beginning may inhibit the proliferation rate. The MIA-PSC 1:9 co-culture ratio gives the best morphology of core-shell struc-ture of pancreatic tumor spheroids, with PSCs on the outside and cancer cells on the inside, as shown in Fig. S2. These spheroids (after being cultured for 48 h) have a cell type population ratio as that of pancreatic cancer (Fig. S5A, SI) [42,43]. When PSCs constitute the major portion of cell population (0.66–0.83), their effect on enhancing cancer cell proliferation and invasion is the maximal [50]. In our study, the initial cell ratio (1:9) actually became roughly 1:4 after 48 h (formation of tight spheroids). This cell ratio agreed with that frequently observed in clinical tumor samples (80% PSCs) [42,43].
    Most interestingly, our co-spheroid model and computation showed that pancreatic cancer cells in spheroids moved much faster than those cultured alone and actually in similar moving rates and patterns as the co-cultured stromal cells (Fig. 3). This phenomenon suggests that, in PDAC, cancer cells and PSCs may metastasize as a whole (seeds and soil). Literature demonstrated that PSCs migrated from the primary cancer seed to distant metastatic sites [51]. The circulating micro-emboli in pancreatic cancer patients contained heterogeneous Vimentin expressing stromal cells (e.g., PSCs) [52]. Because 1:9 co-spheroids in our study revealed greater mobility, they may be more aggressive and play a critical role in metastases. In literature, cancer stem cells did tend to migrate. The breast cancer stem cells (CSCs) have a highly migratory behavior, which is correlated with the aggressiveness of the disease [53]. The migration distance is proportion to the migrate rate. In Fig. 3, we analyzed the cell migration trajectories and the average values of cell migration rate. We further analyzed the displacement (i.e. the straight distance from the end to the start point), as shown in Fig. S12 and Table S4. The displacement of MIA cells in the co-culture group was significantly higher than that of MIA cells alone, while there was no significant difference between the 1:5 and 1:9 co-culture groups.
    EMT plays an important role in the progression of cancer metastasis. During EMT, cancer cells lost their cell-cell adhesion molecules (e.g., E-cad), and up-regulated the expression of a series of classical mesench-ymal markers (e.g., TWIST, SNAIL, N-cad and Vimentin) [54]. Our ob-servation on the reduction of E-cad gene expression level in co-cultured MIA-PSC spheroids is consistent with literature findings on pancreatic cancer invasiveness [50,54]. The up-regulated expression of classical mesenchymal markers in the MIA-PSC co-spheroids may contribute to the highly invasive and metastatic properties of the current model. Increases in TWIST and SNAIL expression are consistent with the in-vasive pancreatic cancer where cancer cells are exposed to activated PSCs from stroma [55]. Moreover, the up-regulation of ECM genes (COL1, LUMICAN, and SNEDI) was observed in the current MIA-PSC co-spheroid model. In particular, the expression of MMP2 was up-regu-lated in 1:9 co-spheroids but was almost unchanged in 1:1 co-spheroids (Fig. 4A). The activation of MMP2 was associated with the invasion of pancreatic tumor [56], and metastases of melanoma in brain [57]. This finding suggest that 1:9 tumor-like spheroids may be more aggressive in nature.