The main objective of this work was to study the effect of wound healing factors on cancer cell aggressiveness. We chose to do this by collecting wound fluid from patients operated for benign head and neck disease (to avoid effects of concurrent cancer treatment or tumor-host responses), and adding this to HNSCC cell cultures. This enabled us to quantitatively measure several facets of cancer cell aggressiveness and also to explore the cellular signaling events involved in the processes. One difficulty associated with this methodology lay in defining the conditions with which to compare the HWF treatment, as ordinary cell culture conditions do not necessarily mirror ordinary physiological growth conditions for tumor cells. Apart from the fact that the in vitro culture consists of a single cell type growing on a plastic surface, the soluble factors are different. Fetal bovine serum is normally added to the culture medium, providing among other things the necessary growth factors to sustain a high growth rate. These conditions diverge from the physiological state, as the soluble factors are bovine and fetal rather than human and adult, and in addition there is the fact that serum is a product of blood coagulation – an early wound healing process under which several soluble factors, not normally present in the tissue, are released. This means that the characteristics of cells grown under ordinary in vitro conditions to some degree might be similar to those of cancer cells remaining in a surgical wound. Although well aware of this, we nevertheless used the “gold” variant of FBS as an artificial zero level for comparing the effects of HWF, mainly because this is a comparatively well-defined product with low batch-to-batch variations enabling comparisons over time. However, we also used HS to avoid species-dependent issues. Most importantly, this enabled us to compare the effects of HWF between different cell lines.
For all four cell lines, proliferation, migration, and invasion were supported as well or better by HWF compared with FBS or HS. When comparing the cell lines, HN-7 differed markedly in the response to HWF compared with HS, being highly stimulated in all the measured parameters. The only other cell line with a significant difference between the HWF and HS effects was HN-4, which had a higher proliferation rate with HWF. For proliferation, migration, and scattering (invasion was not assayed in this respect), these effects on HN-7 decreased for HWF collected at later time points after surgery, indicating a relation to the wound healing response.
Taken together, this shows that HWF has the ability to increase the aggressive behavior of HNSCC in a cell-line-dependent manner, thus indicating that soluble factors produced in the early wound healing response can affect the propensity for recurrent growth. As deduced from the earlier discussion, we cannot exclude the possibility that the other cell lines were affected to some extent by wound healing factors present in the sera, but it is evident that the HN-7 cell line reacted beyond any such effects, resulting in a higher migration rate and more invasive growth than any of the other cell lines (proliferation was not recorded in absolute measurements comparable between the cell lines).
Previous work has indicated differences in the response to wound healing factors between cell lines. For example, Roh et al. studied three different murine cancer cell lines and found a large effect of wounding on a squamous cell carcinoma cell line, a smaller effect on a melanoma cell line, and no effect on a colon cancer cell line , while Bogden et al. found an up to seven-fold difference in tumor growth stimulation of wounding in a xenograft model using 16 different tumors . The basis for these differences was, however, not further explored. The present experimental setup allowed us to make a molecular characterization of the wound healing effects on the HN-7 cell line. The question of whether the cells undergo EMT is an important one, because such processes could facilitate the formation of distant metastases . The HN-7 cell line did at least partially change into an EMT-like phenotype characterized by a spindle-shaped morphology, decreased cell-cell adhesion, and increased migratory capacity. These changes could facilitate the re-colonization of tissues close to the surgical wound. However, several molecular markers suggested as major criteria for EMT did not change appropriately under the tested conditions, and we did not see persistent phenotypic changes when the HWF was withdrawn from the cells. This suggests that the effect of wound healing for the metastatic capacity of these cells might be less important. However, we cannot exclude the possibility that wound healing factors might permanently or temporarily change the responsiveness to factors present in tissue that could affect the metastatic propensity of the cells or that long term exposure of tumor cells to wound healing processes per se could result in permanent cellular changes.
Great efforts have been made over several years to design pharmaceutical agents targeting different intracellular signaling molecules of importance in cancer development and malignity. Several such substances have recently been approved for the treatment of cancer , and many more are presently in clinical trials. It is therefore of interest to know which intracellular signaling pathways are involved in the stimulatory effect of HWF. We examined the activation of three important signaling molecules (ERK1/2, STAT3, and Akt), representing different intracellular pathways, in response to HWF. All three were to some extent activated by HWF in HN-7, but the most pronounced activation was seen for STAT3 (Figure 4b). For this protein, there was also a small increase in activation for the HN-4 and HN-5 cell lines (approximately 1.5 times), which also displayed a low HWF-dependent increase in migration over the effect of FBS. By using a STAT3 inhibitor, we found that it was possible to radically, but not completely, decrease the effects of HWF on the HN-7 cell line (Figure 4). It thus seems that STAT3 has a major influence on the described effects of HWF. It is probable, however, that other signaling pathways also contribute to some lower degree in this respect.
We further investigated the receptor signaling responsible for the STAT3 activation. HGF is a growth factor that stimulates cell scattering and migration in a way similar to the HWF effect on the HN-7 cells, and is known to be produced during wound healing [30, 31] and to activate STAT3 We further investigated the receptor signaling responsible for the STAT3 activation. HGF is a growth factor that stimulates cell scattering and migration in a way similar to the HWF effect on the HN-7 cells, and is known to be produced during wound healing [30, 31] and to activate STAT3 [25, 26]. In this work, however, we could not find any evidence that HGF was involved in the HWF-stimulated processes. Likewise, we found no evidence for the involvement of EGFR family receptors in the HWF response though previous work has shown that EGFR family receptors may be involved in the response of cancer cells to wound healing [12, 13]. We can not exclude, however, that the increased expression of c-Met could be important for the long-term activation of tumor cells in vivo. The concentration profile of different molecular factors changes over time in the wound healing process, and activation of c-Met by HGF might have an effect in later stages. It is also possible that EGFR could be of importance in a basal wound healing response, although not being involved in the extra sensitivity displayed by the HN-7 cell line.
In this work, however, we could not find any evidence that HGF was involved in the HWF-stimulated processes. Likewise, we found no evidence for the involvement of EGFR family receptors in the HWF response though previous work has shown that EGFR family receptors may be involved in the response of cancer cells to wound healing [12, 13]. We can not exclude, however, that the increased expression of c-Met could be important for the long-term activation of tumor cells in vivo. The concentration profile of different molecular factors changes over time in the wound healing process, and activation of c-Met by HGF might have an effect in later stages. It is also possible that EGFR could be of importance in a basal wound healing response, although not being involved in the extra sensitivity displayed by the HN-7 cell line.
As a third alternative stimulator we investigated IL-6, which is a well-known STAT3 inducer  that has been shown to affect HNSCC cell lines [33, 34] and has also been suggested as a predictive marker for recurrence of HNSCC . We showed that part of the STAT3 activation as well as the migration increase in response to HWF could be attributed to IL6R stimulation. Proliferation, on the other hand, was slightly faster with tocilizumab. The fact that IL6R inhibition, in contrast to STAT3 inhibition, did not decrease proliferation is somewhat puzzling. The incomplete inhibition of STAT3 phosphorylation by tocilizumab indicates that STAT3 is activated by extracellular stimuli other than IL-6, and it is possible that these signals could have other cellular outcomes. Different routes for STAT3 signal transduction in a single cell type have been described , and it has been proposed that differently activated STAT3s might result in different transcriptional outcomes . Of course, such mechanistic hypotheses need further experimental verifications.
The fact that HN-7 became radically more aggressive when challenged with HWF while the other cell lines were largely unaffected indicates that some HNSCC tumors might have an inborn predisposition for wound healing stimulated recurrence after surgery. A biomarker distinguishing sensitive from non-sensitive tumors could in that case be a valuable tool for treatment decisions, both for presently available treatments and for possible future target specific therapies. We found that HN-7 was the only cell line in the study expressing the active IL6R at detectable levels. As indirectly evident from our results, IL-6 is not the only wound healing derived factor influencing the aggressive behavior of HNSCC cells, but it seems to play an important role. Therefore, IL6R expression might be evaluated as part of a panel of biomarkers predicting wound healing reactivity of HNSCC.