Differentiation therapy with transcription factors might present as an ideal strategy for the treatment of cancer†

Reply: We were very interested to read the letter from Hisashi Moriguchi.1 In his letter, an interesting viewpoint was raised that the four reprogramming factors, i.e., Oct3/4 (octamer3/octamer4), Sox2 (sex-determining region Y box 2), Klf4 (Kruppel-like factor 4), and c-Myc, may present as novel targets for human hepatocellular carcinoma (HCC) therapy and serve as markers for the therapeutic evaluation of human HCC as well. Substantial evidence has demonstrated that cancer stem cells (CSCs), which escape the normal limits of self-renewal and undergo aberrant differentiation, have the capability to sustain tumor growth and metastasis. The CSCs share similar properties with the normal stem cells including self-renewal, prolonged survival, and the ability to give rise to more differentiated cells. It has been proven that these four factors (Oct3/4, Sox2, Klf4, and c-Myc) exert the ability to reprogram adult human fibroblasts to pluripotent stem cells that exhibit the essential characteristics of embryonic stem cells.2 These genes were suppressed with the differentiation of hepatoma cells into hepatocytes by hepatocyte nuclear factor 4α (HNF4α).3 Therefore, it may be reasonable to hypothesize that down-regulation of these four genes might improve the efficacy of human HCC therapy as suggested by Dr Moriguchi.1 Nevertheless, a study by Yu et al. has revealed that another combination of four factors (Oct3/4, Sox2, NANOG [Nanog homeobox], and LIN28 [lin-28 homolog]) is also sufficient to induce pluripotent stem cells from human somatic cells,4 indicating that more genes are involved in the establishment or maintenance of pluripotency. In this case, the therapy of HCC directing to one or several of these genes may be insufficient to achieve satisfactory efficacy. On the other hand, a recent report by Gupta et al. showed that salinomycin might selectively eliminate breast CSCs and inhibit metastasis by inducing the differentiation of CSCs. They also suggest that it is preferable to treat cancer using agents that target both the CSCs and non-CSCs, because non-CSCs might transform into CSCs and thus eradication of CSCs alone may not obtain complete regression of an established tumor.5 Interestingly, our study indicated that HNF4α could induce the differentiation of both hepatoma cells and its CSCs. The suppression of CSCs was accompanied by the inhibition of a cluster of genes which contribute to the pluripotency of human stem cells, including β-catenin, Oct3/4, SMO(smoothened homolog), Bmi, Sox2, NANOG, c-Myc, Klf4, LIN28, and ESG1 (enhancer of split groucho 1).3 More recently, we also demonstrated that up-regulation of HNF4α remarkably ameliorated hepatic fibrosis6 and prevented the development of HCC in rats accompanied by the revision of epithelial-mesenchymal transition (unpublished data). Thus, we believe that differentiation therapy with HNF4α, a central regulator for hepatocyte differentiation, might be an ideal strategy for the treatment of human HCC. Moreover, this strategy may be extended to other cancer types through the induction of differentiation using their corresponding key transcription factors. Chuan Yin*, Wei-Fen Xie*, * Department of Gastroenterology, Changzheng Hospital, Second Military Medical University, Shanghai, China.