Cancer-associated fibroblasts (CAFs) secreted high level of IL-6 in colorectal cancer.
As mentioned above, CAFs act as major stromal cells to promote development of colorectal cancer, we thus hypothesized that CAFs is likely to release several key cytokines that links to the development of CRC. Based on this hypothesis, we began to isolate CAFs and normal fibroblasts (NFs) from 10 colorectal tumor tissues and adjacent healthy tissues. To validate that we have successfully isolated CAFs and NFs, we subsequently measured expression of several CAF-specific genes, including myofibroblast marker α-SMA, CD90,fibroblast activation protein (FAP) in CAFs, NFs and epithelial cell controls CaCO-2. As shown in Fig. 1A, CAFs and NFs expressed higher level of α-SMA, CD90,FAPand FSP1, compared to epithelial cell controls(P < 0.01). Furthermore, the expression of these markers is more significantly up-regulated in CAFs than those in NFs (P < 0.01), suggesting that we have successfully isolated CAFs and NFs from colorectal tumor and adjacent normal tissues.
We therefore collected culture medium from CAFs and NFs for cytokine profiling by Bio-Plex assay. Interestingly, we observed several differential cytokines between CAFs and NFs (Figure B). Obviously, the level of IL-6 in CAF culture medium is up to 6 fold change (CAFs vs NFs, P < 0.01), while other differential cytokines is only up to 1–2 fold change, such as TNF-α, CCL25, CCL17, and IL-1β, implicating that IL-6 that released by CAFs may play an important role in CRC.
Cancer-associated fibroblasts (CAFs) released increased IL-6 toresponse to hypoxia.
Previous studies have shown that hypoxia has significant impact on CRC carcinogenesis(12, 13), we therefore evaluate whether hypoxia is able to affect CAFs to secrete IL-6. To mimic tumor microenvironment (TME), CAFs were co-cultured with HCT116 cells under either normoxia (20% O2) or hypoxic (1% O2)condition (Fig. 1C). We collected culture medium from CAFs from Day1 to Day5, and measured IL-6 level by ELISA assay. In spite of the fact that IL-6 level from co-culture of CAFs and HCT116 is significantly higher than the level from CAFs cultured aloneunder normoxia, hypoxia, however, led to a pronounced increased IL-6 level from CAFs cultured alone or together with HCT116 (Fig. 1D). In consistent, we also collected CAFs cells and found the expression of IL-6 is dramatically induced by hypoxia (Fig. 1E), suggesting that hypoxia substantially contributes to the release of IL-6 by CAFs.
CAF-derived IL-6promotes cell proliferation of colorectal cancer cell lines HCT116 and SW480.
We next investigated whether CAF-derived IL-6 can affect CRC cell growth in vitro.
Therefore, we collected conditional medium from CAF (CM-CAF) or NF (CM-NF), and treated colorectal cancer cell line HCT116 and SW480 with CM-CAF or NF for 72hrs. The cell proliferation was determined by using CCK-8 assay. As shown in Fig. 2A, HCT116 and SW480 cells with CM-CAF showed an obvious rapid growth. To investigate whether IL-6 exhibits similar biological effects of CM-CAF on colorectal cancer cells, we treated HCT116 cells with recombinant IL-6 (10ng/ml or 20ng/ml). As shown in Fig. 2A, IL-6 has the capacity to promote proliferation of HCT116 and SW480 in a dose-dependent manner, suggesting that CAF-derived IL-6 contributes to the cell growth of CRC.
Since hypoxia promotes CAFs to release IL-6, and IL-6 has tumor-promoting activity as shown above, we next interrogated the correlation between hypoxia and proliferation of colorectal cancer cells. Therefore, HCT116 and SW480 cells were co-cultured with CAFs under hypoxia for 72hrs, and collected for CCK-8 assay. As expected, cells under normoxia showed lower cell growth than those under hypoxia (Fig. 2A), highlighting role of hypoxia in colorectal carcinogenesis.
Hypoxia enhanced IL-6/STAT3 signaling in CRC cells.
We next investigated the mechanisms for crosstalk of IL-6-mediated-carcinogenesis with hypoxia. Accumulating evidences have demonstrated that IL-6/STAT3 pathway is aberrantly hyper activated in a variety of cancer type, and its hyper activation is often associated with poor prognosis. Biologically, IL-6/STAT3 signaling acts to drive the proliferation, invasiveness and metastasis of cancer cells(14, 15). Consistently, HCT116 or SW480 treated with STAT3 inhibitor showed lower cell proliferation than control cells showed (Fig. 2B).
HIF-1α is well recognized as a major transcription factor, that induced by the presence of hypoxia. We then measured expression of HIF-1α in HCT116 or SW480, which co-cultured alone or with CAFs. As shown in Fig. 2C, the HIF-1α expression in HCT116 or SW480 is significantly induced in the presence of hypoxia, confirming thatHIF-1α can be sufficiently induced by hypoxia. Moreover, we noticed that CAFs can significantly induceSTAT3 phosphorylation in HCT116 or SW480 under normoxia, and strikingly, STAT3 phosphorylation can be further up-regulated in the presence of hypoxia (Fig. 2C), indicating that hypoxia is able to enhance IL-6/STAT3 signaling in CRC cells.
To confirm hypoxia can regulate IL-6/STAT3 signaling, HCT116 or SW480 cells were pre-treated with HIF-1α inhibitor chloride (CdCl2) for 12hrs, and subsequently cultured with CAFs under hypoxia. As Fig. 2D showed, the expression of HIF-1α is significantly reduced by CdCl2 in HCT116 and SW480 cells. Moreover, STAT3 phosphorylation is also suppressed byCdCl2, highlighting the role of hypoxia in enhancingIL-6/STAT3 signaling in CRC.
HIF-1α target pyruvate kinase M2 is involved in activation of IL-6/STAT3
Numerous studies have shown that HIF-1α target pyruvate kinase M2 (PKM2) promotes the activation of STAT3. Therefore, we hypothesized that the enhanced activity of IL-6/STAT3 by hypoxia maybe, in part, correlated with PKM2. HIF-1α and PKM2 was observed up-regulated in HCT116 cell sunder hypoxia (Fig. 3A). Moreover, inhibition of HIF-1α by CdCl2 can suppress hypoxia-induced PKM2, confirming that PKM2 is direct target of HIF-1α.
To examine whether PKM2 increase STAT3 phosphorylation, HCT116 were treated with PKM2 siRNA under hypoxia. The results clearly showed that knockdown of PKM2 inhibited STAT3 phosphorylation under hypoxia (Fig. 3B). We then investigated whether knockdown of PKM2 can suppress the biological effect of CAFs or IL-6 on CRC cells. As shown in Fig. 3C, suppression of PKM2 obviously inhibited STAT3 phosphorylation in HCT116 cells co-cultured with CAFs or treated with IL-6 under hypoxia.