**4. Discussion**

The stemness program plays an important role in maintaining the properties of CSCs due to self-renewal, which is a hallmark of cancer-initiating cells. Recent studies have shown that SOX2 is aberrantly expressed and involved in the maintenance of properties of colorectal CSCs, including spheroid-like growth and metastatic potential [14,15]. Here, we extended these studies to demonstrate that SOX2 is regulated by the PI3K/AKT pathway and contributes to the induction of colorectal CSCs in response to radiation. By comparative analysis of radiation-induced population of CSCs in both radioresistant and radiosensitive colorectal cancer cells, we found that radioresistant cells such as HCT116 and DLD1 specifically increased the CD44+ population after irradiation, which is one of the properties of CSCs. Interestingly, we also found that the radiation-induced activation of PI3K/AKT pathway functions as an upstream regulator of SOX2-dependent induction of CSCs in colorectal cancer.

In this study, we report that radiation-enriched CD44+ cells exhibited colorectal CSC properties including resistance to radiation, enhanced in vitro metastatic potential, and a spheroid growth pattern. CD44 is a receptor of hyaluronan and is a transmembrane glycoprotein that participates in many cellular processes, including growth, survival, differentiation, and motility [31–34]. CD44 is considered a more selective marker of colorectal CSCs than CD133, because the properties of colorectal CSCs are not regulated by CD133 modulation [19,20,35,36]. Consistent with this, our comparative study between radioresistant and radiosensitive colorectal cancer cells showed that CD44 expression, but not that of CD133, was selectively increased in radioresistant colorectal cancer along with acquiring the properties of colorectal CSCs after irradiation. A recent study reported that SOX2 expression primarily coincided with CD44+ and ALDH1+ population in pancreatic CSCs [37] and CD44+ and CD24+ in colorectal cancer [14]. Indeed, we also observed that FACS-sorted CD44+ cells showed an upregulation of SOX2 expression and demonstrated its important role in modulating the CD44+ population growth and the properties of CSCs in colorectal cancer using both knockdown and overexpression of SOX2, which is consistent with previous reports. Notably, in our study, this functional relationship occurred in response to radiation, indicating that radiation affects SOX2-dependent induction of CD44+ population.

Factors that are important for self-renewal in stem cells are found to be dysregulated in human malignancies. SOX2 expression has been implicated in the control of colorectal CSC properties; however, the related signaling pathways are less understood. In a previous study, SOX2-induced CSCs in cervical and pancreatic cancer have been linked to epithelial-mesenchymal transition (EMT)-related factors [37,38]. However, these studies were considered controversial. Han et al. reported a role for SOX2 in EMT and increased in vitro metastatic potential, such as in migration and invasion in colorectal cancer [24], while Lundberg et al. reported that SOX2 mediated induction of CSC characteristics in an independent manner [14]. In our system, the ability of migration and invasion was dramatically regulated by SOX2-modulated Snail expression, known as a master regulator of EMT. However, we observed that Snail did not affect the induction of colorectal CSC properties, including CD44+ population growth, resistance to radiation, and the ability of tumoursphere formation. This suggested that the Snail-mediated EMT process might not be involved in SOX2-dependent induction of colorectal CSCs, although we could not exclude the possibility of involvement of other regulators of EMT process or factors related to tumor microenvironment [39] affected by irradiation. Therefore, it is likely that SOX2 modulates either EMT process or CSC induction through alternative pathways, at least in response to radiation. Further studies are required to clarify the relationship between EMT and CSCs induction.

An elucidation of the signaling pathways that govern the SOX2-dependent induction of CSCs is also required for devising an optimal targeted therapy. Considering that MAPK and PI3K/AKT pathways, in addition to being activated by radiation [40], were associated with resistance to therapy and tumorigenicity in cancer cells [41,42], we investigated

the effect of the inhibitors of MAPK and PI3K/AKT for induction potential of CSCs. It was found that radiation activated the genes of both MAPK and PI3K/AKT pathways, consistent with previous reports, except ERK. Interestingly, the induction of SOX2 and CSC characteristics, including CD44+ cells in colorectal cancer, were only affected by the inactivation and downregulation of PI3K/AKT following irradiation. This finding is contradictory to the report by Wang et al., [43] who showed that activation of both AKT and MAPK pathways was involved in the induction of properties of colorectal CSCs, such as the colony formation ability in primary colon cancer cells. These differences can be explained using target cells with differential markers of CSCs and the response to stresses. To isolate colorectal CSCs, Wang et al. used CD133, a colorectal CSC marker, and characterized cells with or without CD133 expression under non-stress conditions for CSC properties, whereas we used CD44, which was specifically induced by radiation stress. Furthermore, the involvement of genes in the PI3K/AKT pathway in SOX2 regulation in breast and nasopharyngeal carcinoma has recently been reported [28,29]. Therefore, this is an interesting finding that radiation-activated PI3K/AKT pathway genes were essential for the SOX2-dependent induction of colorectal CSCs, and it is potentially an effective therapeutic target for CSCs in colorectal cancer activated by radiation.

**Author Contributions:** Designed the experiments and drafted and edited the manuscript: M.-J.K. and S.B.L. Performed the experiments, interpreted the data, and drafted the manuscript: J.-H.P. and Y.-H.K. Performed the screening of experimental conditions and provided assistance for all the experiments: S.S. (Sehwan Shim), A.K. and H.J. Provided valuable suggestions and feedback: S.-J.L. and S.P. Performed the application of statistical techniques to analyze data: S.S. (Songwon Seo) Acquisition of the financial support for publication: W.I.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by a grant from the Korea Institute of Radiological and Medical Sciences (KIRAMS), funded by the Ministry of Science and ICT (MSIT), Republic of Korea (grant number 50535-2020).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**

