Corresponding author’s email address: cornelia.schmutzler@charite.de

Introduction

The classical molecular genetic view of cancer states that carcinogenesis is a multistep process whereby cells accumulate mutations in essential genes that give them growth advantages and finally develop into malignantly transformed cancer cells with the ability to invade neighboring tissues and blood vessels to give rise to metastases [1]. Recently, this view is being complemented by conceding a major role for interactions between tumors and their stromal environment [2] as well as for so-called epigenetic changes, such as histone modification and DNA methylation, that influence gene activity without altering gene sequence [3]. In the course of these events, cells abandon the expression of genes encoding tissue-specific proteins and consequently lose the ability to perform their proper physiological functions, i.e., they undergo de-differentiation. Concerning thyroid cells, this includes, probably due to a deficit in thyroid-specific transcription factors [4-6], the loss of thyroperoxidase (TPO), thyrotropin receptor (TSHr), thyroglobulin (Tg) and, as a comparably early event, of the sodium iodide symporter (NIS) [7] which then also has therapeutic consequences in terms of the feasibility of radioiodide therapy. Also, this process may induce slowly growing thyrocytes, showing only minimal probability to undergo division during their lifetime, to give rise to one of the most aggressive and rapidly proliferating cancers in humans, an anaplastic thyroid carcinoma (ATC) [8].

A recent alternative to cytotoxic drug treatment of cancers is re-differentiation therapy. In case of the thyroid gland this option is tested especially for those cases that do no longer respond to established treatment protocols [9]. Here, it is the aim to reactivate at least some of the features of differentiated cells, e.g. to control rapid tumor growth by stimulating appropriate cell cycle regulation or an adequate level of apoptosis, or to reinduce typical functions, such as radioiodide uptake, being exploitable for treatment. This must again involve changes in gene expression and may be achieved by various types of drugs targeting different, although mutually interdependent, mechanisms of the control of gene activity: 1) by providing transcriptional signals, 2) by influencing the accessibility of genes for the transcriptional machinery or 3) by altering the long-term activation or silencing of genes. As it is of course impossible to revert chromosomal rearrangements, deletions or mutations of essential tumor suppressor or oncogenes, it is the epigenetic level that has become a main focus of re-differentiation therapy. Concerning thyroid carcinomas, all these principles have been applied, either experimentally in vitro and in vivo or in clinical trials, and will be discussed in the following sections.

1) Modulating gene expression using nuclear receptor ligands

Retinoic acid
A class of substances widely utilized for chemoprevention and therapy of hematological and solid tumors are the vitamin A-derived retinoic acids (RA) and some of their synthetic derivatives [10]. RA are key regulators of morphogenesis, proliferation, and differentiation during development and are involved in vision, spermatogenesis and the maintenance of healthy, functional tissues, such as the lung, in the adult. Various effects of RA in cancer cells have been described. They have been shown to affect signaling pathways regulating growth, differentiation and apoptosis via proteins such as AP-1, MAPK, PI3 kinase, Akt, cyclins, cyclin-dependent kinases and their inhibitors, Bcl proteins and caspases [11].
RA exert their effects via nuclear receptors acting as ligand dependent transcription factors. There are two subfamilies of RA receptors, retinoic acid receptors (RAR) which bind all-trans RA and 9-cis RA and retinoid X receptors (RXR) which bind only 9-cis RA. They usually function as RAR-RXR heterodimers and interact with RA response elements in the regulatory region of RA-responsive genes. In the absence of ligand, they recruit so-called co-repressor complexes. These contain histone deacetylase (HDAC) activities which remove acetyl residues from the tails of histones H3 and H4 and thereby induce a compact and closed conformation of the chromatin. In the presence of ligand, nuclear receptors recruit co-activator complexes. These include histone acetyl transferase (HAT) activities adding acetyl residues to histones. This causes an opening of the chromatin making DNA accessible for the transcriptional machinery which then initiates transcription [12]. Necessarily, this is a global process affecting many RA-responsive genes simultaneously. However, as RA-regulated genes are involved in cell cycle regulation and apoptosis, a growth-limiting effect may be the consequence. Moreover, as it is the case in the thyroid gland, therapeutically “useful” genes may be induced (see below).
From that it is evident that cells must have an adequate supply of RARs and RXRs as a prerequisite for RA signal transduction. Several publications report that RARs and RXRs are present in thyroid carcinomas. However, their expression levels may vary between tumor cell lines as well as tumors and control thyroid tissues [13-17]. There may also be a redistribution of the normal nuclear localization of the receptor to the cytoplasm as demonstrated by immunocytochemistry of thyroid carcinoma specimen [17]. Tang et al. [16] reported that loss of mRNA expression for one or more RAR or RXR correlated with increased proliferation and altered histological features of PTC tumors. There may even be an association between the responsiveness of thyroid carcinoma cells to RA and the loss of certain receptor subtypes, namely RXRγ or RARβ [18-20] as only those lines expressing these isoforms exhibit growth reduction after RA treatment. Thus, the RAR/RXR repertoire of thyroid tumors may determine, if they will respond to RA re-differentiation therapy.
Nevertheless, a couple of experimental studies demonstrated that RA treatment can indeed alter the expression of certain differentiation-related markers in thyroid carcinoma cells. These are, among others, Tg [21], ICAM-1 [22], and type I 5’-deiodinase (5’DI) [23,24] which are up-regulated and CD97 [25] and urokinase which are down-regulated [26]. Growth reduction in vitro or in a xenotransplantation model was also observed [19,25]. Most interesting, NIS mRNA expression and activity are increased by RA [27,28], which is due to a RA response element in the human NIS promoter [29]. This is therapeutically promising, as it opens the possibility to treat patients exhibiting non-accumulating tumors or metastases with radioiodide after applying RA. Interestingly, this RA responsiveness is also exhibited by NIS aberrantly overexpressed in the mammary cancer cell line MCF-7 so that NIS is discussed also for radioiodide therapy of breast cancer (reviewed in [30]). Furthermore, a combination between NIS gene therapy and RA differentiation therapy has been considered for the treatment of prostate cancer [31]. 5’DI is a differentiation marker in the thyroid, as 5’DI expression and enzyme activity are high in the healthy gland but progressively lost in thyroid carcinoma in correlation with increasing malignancy. Stimulation of 5’DI mRNA and protein expression as well as enzyme activity was observed in follicular, but not in anaplastic thyroid carcinoma cell lines [23] and, furthermore, in a fine needle aspiration biopsy obtained from a thyroid carcinoma patient treated with RA [32]. A RA responsive element is also present in the 5’DI promoter [33,34].
Several studies have been conducted to elucidate RA effects on thyroid carcinoma in patients. In a multi-center pilot study [35-37], 75 patients with advanced thyroid cancer and without the therapeutic options of operation or radioiodide therapy were treated with 13-cis RA at a dosage of 1.5 mg/kg body weight daily over 5 weeks. Of the 50 patients evaluated, 13 showed a clear increase in radioiodine uptake, and eight a mild increase. Tg levels (a tumor marker for thyroid cancer progression) were unchanged or decreased in 20 patients. Tumor size was assessable in 37 patients; regression was observed in 6, and there was no change in 22. Enhanced glucose uptake into tumors results from anaerobic metabolism and is indicative of rapid malignant growth. Glucose uptake was measured by ¹⁸F-FDG PET, and there was an increase after RA in 1, a decrease in 6 and no change in 25 of 32 patients. In total, a response, classified as a reduction of tumor size and/or Tg levels with or without concomitant increase in radioiodide uptake, was observed in 10 of 50 patients; a stabilization of these two parameters was observed in another 9 patients. Response to retinoid therapy did not always correlate with increased radioiodine uptake, so other direct antiproliferative effects had to be assumed. Several other studies with smaller numbers of patients report on similar [38-42] or lower efficiencies [43] of RA with respect to the parameters mentioned. However, therapeutically relevant doses of radioiodide were accumulated by some of the lesions with previously insufficient or absent radioiodide uptake after RA treatment [35,38]. In one study, two of 25 patients were completely free of symptoms after a follow-up of two years [42]. A summary of the various studies is presented in Table 1.

Table1 as PDF

To further elucidate the potential of this therapeutic modality, a multi-center, randomized, prospective phase II/III trial (“MSSR study”) [44] has now started to recruit patients. 120 participants will be enrolled in two arms, one receiving 13-cis RA at 1.5 mg/kg body weight daily for six weeks alone, the other arm ablative ¹³¹I -radioiodide therapy in addition. Monitoring will include whole body scan, determination of Tg serum levels, ¹⁸F-FDG PET and determination of tumor size.

RA is the only compound discussed for re-differentiation of thyroid cancer for which already therapeutic experience is available from several clinical trials. A couple of other compounds are being evaluated in vitro and in animal models, and so far, some first studies in patients have been initiated. The respective data are reviewed in the following sections.

PPARγ ligands
PPARγ is, like RAR and RXR, a member of the nuclear receptor family. This receptor which functions as a heterodimer with RXR is, above all, known for its role in lipid metabolism and adipocyte differentiation. Its ligands are, on the one hand, natural compounds such as polyunsaturated fatty acids and eicosanoids, on the other hand, synthetic drugs such as glitazones or thiazolidinediones [45]. In the context of thyroid carcinoma, PPARγ came into focus after the description of a chromosomal translocation, t(2;3)(q13;p25), in follicular thyroid carcinomas and later also in follicular thyroid adenomas [46-50]. The product, a fusion protein of thyroid transcription factor PAX8 to peroxisome proliferator-activated receptor &gamma (PPARγ), is a dominant-negative inhibitor of PPARγ signal transduction, accelerates cell proliferation, reduces apoptosis and permits anchorage independent and growth without contact inhibition [46,51], indicating that PPARγ is involved in tumor suppression as well as epithelial differentiation in the thyroid gland. Consistent with these observations, restoration of PPARγ expression in PPARγ-negative NPA cells derived from a papillary thyroid carcinoma (PTC) decreased cell growth, and PPARγ agonists induced further inhibition, both associated with increased p27(kip1) protein levels and apoptotic cell death [52]. A dose-dependent growth-inhibitory effect was also observed in the PPARγ-positive ATC cell lines OCUT-1 and ACT-1 after treatment with the PPARγ ligands troglitazone and 15-deoxy-delta 12,14-prostaglandin J2. Cells were arrested in G1 via a p53-independent, but p21- and p27-dependent cytostatic pathway [53].

1α,25-dihydroxyvitamin D₃ (VitD3)
VitD3 also activates a nuclear receptor that belongs to the same family as RAR and PPARγ and also forms heterodimers with RXR. VitD3 plays a role in Ca²⁺ and phosphate metabolism, in the differentiation of bone and skin and has shown anti-tumor effects in several preclinical and clinical studies [54-56]. VitD3 as well as its synthetic analogs, 22-oxa-1,25(OH)2D3 (OCT) and EB1089 exhibited a dose-dependent, growth-inhibitory effect on human PTC-derived NPA cells in vitro [57,58]. VitD3 and EB1089 were also shown to induce a G1-phase arrest accompanied by an increase in the expression of the cyclin-dependent kinase inhibitor, p27(kip1) due to reduced degradation. Dackiw et al. [59] used a model of 4- to 5-week-old female SCID mice xenografted with WRO cells which are derived from a human follicular thyroid carcinoma (FTC). Mice were treated i.p. three times a week with 0.75 µg/kg VitD3 or vehicle and killed after 21 d. Tumors from vehicle-treated animals demonstrated morphological features of epithelial malignancies with characteristics of insular carcinoma and multiple metastases to the lungs, whereas tumors from VitD3-treated animals demonstrated signs of differentiation with restoration of Tg staining, associated with a marked accumulation of p27(kip1) immunoreactivity in the nuclear compartment.

2) Modulating gene expression via HDAC inhibitors:

As described above, a consequence of nuclear receptor-dependent transcriptional regulation is an alteration in chromatin modification by interaction with histone modifying enzymes. Instead of this indirect mechanism, histone modifying proteins may be targeted directly, e.g., by the use of HDAC inhibitors. This is a quite heterogeneous class of substances, including, e.g., hydroxamic acids (Trichostatin A (TSA), suberoylanilide hydroxamic acid), carboxylic acids (valproic (VPA) and butyric acid), benzamides (MS-275, N-acetyldinaline), and fungal metabolites (depsipeptide (FR901228, FK228), apicidin). The rationale for their use against cancer is the empirical observation that HDAC activity is generally increased in cancer cells, resulting in altered gene transcription. Furthermore, there is evidence that genes for HDACs are disrupted in certain malignancies or that there is an aberrant recruitment of HDACs by oncogenic transcription factors. In normal cells, the pattern of histone acetylation together with that of other histone modifications define the so-called “histone-code” which is involved in discriminating domains of active from those of inactive chromatin. The effects of the application of HDAC inhibitors on gene expression are, of course, more global, as the result is not the regulation of very specific responsive promoters, but a more general relief of gene silencing. Nevertheless, a significant up- or down-regulation is displayed by no more than 4 – 10 % of genes in cultured cells. Many of these genes are involved in growth arrest, apoptosis, angiogenesis and differentiation, e.g., the one coding for cyclin-dependent kinase inhibitor p21^WAF, which is up-regulated by all known HDAC inhibitors. A couple of publications describe anti-cancer properties of HDAC inhibitors in several cancer cell lines and animal models, and several phase I/II clinical studies indicate that they are well-tolerated drugs [60,61].

Valproic acid (VPA)
VPA has been used for years as an antiepileptic drug before its properties as a HDAC inhibitor were discovered [62], and both its pharmacological properties and the spectrum of its side effects are well characterized. Pharmacologic concentrations determined in patients taking VPA for anticonvulsant therapy reach 0.7 mM. Catalano et al. [63] demonstrated effects of VPA in the poorly differentiated PTC cell lines NPA and BHT-101. At concentrations ≥ 1 mM, they observed a decrease in cell viability, an increase of apoptosis via caspase 9, and an induction of cell cycle arrest at G1. The latter was accompanied by up-regulation of p21^WAF and down-regulation of cyclin A mRNAs and proteins. Furthermore, NIS mRNA expression is increased in NPA cells at VPA concentrations ≥ 1 mM and in the ATC cell line ARO at concentrations ≥ 1.5 mM. This was accompanied by an induction of NIS protein, which was localized in the plasma membrane of NPA but intracellularly in ARO cells. Iodide uptake was only increased in NPA cells [64].

Depsipeptide (FR901228, FK228):
Furuya et al. [65] described the powerful effects of depsipeptide on BHP18–21v (derived from a PTC) and ARO cells. At concentrations from 1 to 10 ng/ml, NIS, TPO and Tg mRNA and protein expression were stimulated. Iodide uptake was increased from 3 ng/ml on in both cell lines. Furthermore, in the treated cells, accumulated iodide was organified into a protein species of more than 210 kDa molecular weight, probably by depsipeptide-induced TPO, as this effect could be blocked by the TPO inhibitor methimazol at a concentration of 300 µM. Depsipeptide also increased accumulation, retention and organification of iodide in tumors that had grown from xenotransplanted BHP18-21v cells on nude mice. In these tumors depsipeptide could also induce the expression of NIS, Tg and TPO mRNAs, whereas TSH-R mRNA was not stimulated. Depsipeptide also increased dose-dependently the expression of TTF-1, but not of PAX8, mRNA in BHP-18-21v and ARO cells.
Kitazono et al. [66] performed all their experiments using FTC-133, FTC-236, Kat-4 and SW-1736 cells at a depsipeptide concentration of 1 ng/ml, which is only minimally cytotoxic as determined by MTT assay. After 48 and/or 72 h, they observed increased histone acetylation by Western blot and immunocytochemistry, elevated iodide uptake after 72 h and enhanced activity of a Tg promoter/luciferase reporter construct in all cell lines. Depsipeptide also restored expression and function of p53 in SW-1736 cells which are “pseudo-null” for this protein [67].
It may be added that the concentrations required to see effects in thyroid carcinoma cell lines can easily be achieved in patients, as in phase I studies concentrations up to 500 ng/ml have been tolerated without significant toxicity [66].

Trichostatin A
A reduced cell viability at TSA concentrations up to 250 ng/ml was described for TPC-1, FTC-133 and XTC-1 cells, deriving from a PTC, an FTC and a Huerthle cell carcinoma (HTC), with cytotoxicity occurring from 1000 ng/ml on. NIS mRNA was induced by 50 and 100 ng/ml TSA and basal pendrin gene expression was reduced by these two concentrations [68]. In cell lines derived from primary ATC, sodium butyrate and TSA repressed proliferation independent of the p53 status through the induction of apoptosis and cell cycle arrest in G1 or G2/M. Apoptosis was associated with the appearance of the cleaved form of the caspase substrate, poly-(ADP-ribose) polymerase (PARP), while a reduced expression of cyclins A and B, an increased expression of the cyclin-dependent kinase inhibitors, p21^WAF1 and p27(kip1), a reduced phosphorylation of the retinoblastoma protein and a reduction in cdk2- and cdk1-associated kinase activities accompanied cell cycle arrest. However, in ATC cells overexpressing cyclin E, drug treatment failed to replicate these events [69].

3) Modulating gene expression by inhibition of DNA methyl transferases

DNA methylation is a covalent modification of CpG dinucleotides in genomic DNA which plays a role in the regulation of gene activity, namely transcriptional repression, and results, e.g., in long-term silencing of genes by X chromosome inactivation or genomic imprinting. Two mechanisms have been described: One is the direct inhibition of transcription factor binding to a methylated response element. The other is based on proteins such as MBD 1-3 and MECP2 that specifically bind to methylated CpG dinucleotides, interact with histone methylases or HDACs and thereby contribute to the generation of inaccessible chromatin [70]. Abnormal patterns of DNA methylation are observed in human cancers [71], among them those of the thyroid gland [72], and include general genomic hypomethylation in combination with regional hypermethylation. Hypermethylation-triggered silencing affects genes involved in almost all of the important steps of cancerogenesis: cell-cycle regulation, DNA repair, drug resistance and detoxification, apoptosis, differentiation, angiogenesis, and metastasis, e.g., the genes coding for CRABP1, CDKN2/p16INK4A, RASSF1, LKB1, MT1G, maspin, E-cadherin, TSH receptor, NIS, pendrin and 5’DI in thyroid cancer and/or carcinoma cell lines [73-85]. Recently, Hoque et al. [86] showed hypermethylation for Rassf1A, TSHR, RAR-beta2, DAPK, CDH1, TIMP3, and TGF-beta, and, furthermore, a trend toward multiple hypermethylation in thyroid cancer tissues. Hypermethylation of 2 or more markers was detectable in 25% of hyperplasias, 38% of adenomas, 48% of thyroid cancers, and 100% of cell lines, and a subset of these markers were epigenetically modified in concert, probably reflecting an organ-specific regulation process. E.g., a positive correlation was found between the BRAF mutation and RAR-beta2, and a negative correlation was found between BRAF and Rassf1A.
Interference with DNA methylation to release gene silencing and thereby activate dormant tumor suppressor genes for experimental and clinical purposes makes use of derivatives of cytidine derivatives, namely 5-aza-cytidine or 5-aza-2’-deoxycytidine (decitabine). They get incorporated into DNA during replication (or, as for 5-aza-cytidine, also into RNA) and then inhibit the action of DNA methyltransferases. Several clinical phase-I/II trials have been performed for solid tumors, however, the highest potential for these substances has so far been demonstrated in the case of hematological malignancies [87].
As for thyroid cancer, Venkataraman et al. [4] analyzed a collection of human thyroid carcinomas with different pathologies and reported variable methylation patterns of the hNIS promoter and first exon which, however, did not correlate with NIS mRNA expression or clinical iodide uptake. When seven human thyroid carcinoma cell lines lacking NIS mRNA were treated with 5-azacytidine or sodium butyrate, NIS mRNA expression was re-induced in four and iodide transport in two cell lines. This was associated with demethylation of NIS DNA in the untranslated region within the first exon and with restoration of the expression of TTF-1.
An association has been demonstrated between de-differentiation of thyroid carcinomas and E-cadherin expression in ATC, and E-cadherin expression is an independent prognostic factor for highly differentiated thyroid tumors. Hypermethylation of CpG islands in the proximal promoter region of the E-cadherin gene could be demonstrated in thyroid tumors and in E-cadherin-negative cell lines derived from thyroid carcinomas [88]. On the other hand, treatment with 5-aza-2’-deoxycytidine restored E-cadherin expression in 5 thyroid carcinoma cell lines [89].
Results from our group indicate that methylation of a GC-rich region close to a thyroid hormone responsive element in the 5’DI promoter may be responsible for silencing of the 5’DI gene in human thyroid FTC cell lines FTC-133, FTC-238 and HTh74. Treatment of these cell lines with 5-aza-2’-deoxycytidine drastically increases both 5’DI mRNA and enzyme activity [85]. Interestingly, this increase was also seen in the ATC cell line HTh74, where 5’DI was not stimulated by RA [23].

4) Miscellaneous

Phenylacetate
Phenylacetate and phenylbutyrate are reported to inhibit growth and to modulate differentiation in a variety of cancer cell lines as well as in patients at concentrations exhibiting minimal toxicity. Inhibition of HDACs and activation of PPARγ are discussed as possible mechanisms [90,91]. In five FTC cell lines, cell cycle arrest in the G0-1 phase occurred at a dose of 2.5-10 mmol/L phenylacetate. Increased radioiodide uptake (in 2 out of 5 cell lines) and decreased secretion of vascular endothelial growth factor were also observed [92]. Furthermore, combination of phenylacetate and RA had synergistic effects: In FTC-133 cells, RA (2.5 µmol/L) inhibited growth by 16% and PA (10 mmol/L) by 35% versus controls, whereas the combination inhibited growth by 60% at 5 days [93].

Resveratrol
Resveratrol, trans-3,5,4'-trihydroxystilbene, is a fruit constituent of various plants, including grapes, berries and peanuts. Besides cardioprotective effects, resveratrol exhibits anticancer properties in a wide variety of tumor cells. The growth-inhibitory effects of resveratrol are mediated through cell-cycle arrest and activation of caspases; suppression of angiogenesis has also been described. In vivo, resveratrol shows both chemopreventive and therapeutic effects against cancer and seems to be pharmacologically quite safe [94]. In two PTC and two FTC cell lines, treatment with 1-10 µM resveratrol induced apoptosis by increasing p53 expression and serine phosphorylation via a Ras-MAPK kinase-MAPK signal transduction pathway [95].

3-Hydroxymethyl-3-methylglutaryl coenzyme A (HMG-CoA) inhibitors: Statins
Statins have been approved for the treatment of lipid disorders. However, studies in cell culture, with experimental animals and first trials in patients also indicated antitumor properties such as induction of growth arrest and apoptosis as well as inhibition of metastasis and of angiogenesis [96]. In thyroid carcinoma cell lines, lovastatin induced apoptosis, and several different mechanisms have been discussed, such as inhibition of geranylgeranylation of Rho proteins, lamin B proteolysis and cytochrome C release from mitochondria [97].

Concluding remarks

From the data reviewed above it is clear that there is a broad selection of drugs inducing (partial) re-differentiation and exerting anti-tumor effects on experimental models for thyroid cancer, which now have to be tested in patients. To date, there are five clinical trials evaluating re-differentiation therapy for thyroid cancer listed by the NIH [98]: a phase I and a phase II study on depsipeptide, a phase II study on 5-aza-2’-deoxycytidine, and a phase I study on rosiglitazone; a phase I on azacytidine is no longer recruiting patients (Table 2).


Table 2 as PDF

A multi-center phase II/III trial on RA has just been started (“MSSR study”)[44). Data published on RA show that a certain number of patients seems to experience benefit from this kind of therapy, however, most patients did not respond. Whereas studies initially reported response rates of about 20 % [35,39], they may be lower after longer follow-up intervals [42]. For an extensive discussion, see references [34-42]. Other re-differentiating agents might be more efficient, especially as PPARγ ligands, HDAC inhibitors or DNA methylation inhibitors show positive effects in ATC cell lines where RA was inactive [52,64,85]. The most promising feature of these drugs, however, might be their use in combination. As they target several different aspects of the same process, i.e., control of gene activity, their combination may be synergistic. Thus, HDAC inhibitors could open up chromatin and give nuclear RA or VitD3 receptors access to gene regulatory sequences. Similar considerations hold true for a combination of HDAC and DNA methylase inhibitors due to the interaction of methylated CpG-binding proteins with HDACs. Combination of drugs may also help to reduce doses of single compounds to circumvent possible dangerous side effects that are elicited by methyl transferase inhibitors [99]. Conforming with these considerations, the HDAC inhibitor m-carboxycinnamic acid bis-hydroxamide (CBHA) at 50, 100, and 200 mg/kg/day inhibited growth of SMS-KCN-69 neuroblastoma tumor xenografts on immunodeficient mice in a dose-dependent fashion, with 200 mg/kg CBHA resulting in a complete suppression of tumor growth. The efficacy of 50 and 100 mg/kg CBHA was enhanced by the addition of 2.5 mg/kg all-trans RA, while this dose of all-trans RA was ineffective when administered alone [100]. In a mouse model for lung cancer, treatment with low doses of 5-aza-2'-deoxycytidine (0.5 mg/kg) decreased the incidence of neoplasms by 30%. When the methyl transferase inhibitor 5-aza-2'-deoxycytidine was combined with the HDAC inhibitor phenylbutyrate (300 mg/kg), lung tumor development was significantly reduced by >50%, while no effect was seen with phenylbutyrate alone [101].
Thyroid cancer in general is an easily treatable disease with a good prognosis. However, in a considerable number of cases, de-differentiation of the tumors makes classical therapy obsolete and, therefore, other approaches are required. Starting from the first attempts using RA, the field of re-differentiation therapy for thyroid cancer has developed and now offers promising new prospects in vitro and in experimental animals. Hopefully, it will be possible, after they have shown their potential in the ongoing clinical trials, to transform them into new therapeutic alternatives for otherwise untreatable thyroid cancer patients.