Hepatocellular carcinoma (HCC), originating from epithelium of hepatocytes and accounting for 80% of primary liver cancers, ranks as 4th place in causing tumor-related deaths globally.1 HCC affects more than half a million people annually and the comparable incidence to its mortality rate demonstrates its dismal prognosis.1 About 80% of HCC is found in patients with cirrhotic liver2 with hepatitis B and C being the main causes of liver cirrhosis. The incidence of HCC in hepatitis B patients is 200 times as high as that of non-infected people and patients with hepatitis C have fivefold more chance to develop HCC than patients with hepatitis B.3 Other cases of non-viral related liver cirrhosis have also been found to be positively associated with HCC, such as nonalcoholic steatohepatitis, hemochromatosis, alcoholic liver disease, alpha-1 antitrypsin deficiency, and autoimmune hepatitis. Moreover, some environmental toxins, such as aflatoxin B1, are also reported to incite the development of HCC.4 Generally, men are more vulnerable to HCC than women; especially in some areas, such as Africa and Southeast Asia, the ratio of male-to-female could reach 3.7.5
Current treatments of HCC
Presently, partial hepatectomy remains the standard treatment for patients with resectable HCC and without obvious liver cirrhosis. However, growing evidence has suggested that liver transplantation and radiofrequency ablation of the tumor could provide comparable benefit on survival as well, compared to partial hepatectomy, especially when the tumors are smaller than 3 cm.6 For example, in Child–Pugh class A patients with a single tumor, the 5-year-survial rate could be improved to 70% after these radical therapies as compared to 65% 3-year-survial without any treatments.2 On the other hand, the advanced HCC patients, who are unfit for receiving radical therapies and are poor respondents to traditional chemotherapy and radiotherapy, usually have a survival time of less than 6 months.7 Finally, most HCC patients (70–80%)7 are diagnosed at intermediate-advanced stage and there is no effective treatment available at the present time.2,8 Under these bleak conditions, developing a new therapeutic regimen against HCC has been a priority. Several new chemotherapeutic drugs, including octreotide, interferon, interleukin-2, and estrogen receptor antagonist, tamoxifen, have been evaluated for their therapeutic efficacy on HCC patients. Unfortunately, no single agent or combination of agents administered systemically provided beneficial effects on survival rate.9,10 Recently, with the advances in understanding the molecular biology of HCC, new therapeutic strategies to treat HCC have emerged.11 For instance, the clinical efficiency of sorafenib, an oral multi-kinase inhibitor of several kinases (vascular endothelial growth factor [VEGF], platelet-derived growth factor [PDGF], c-kit receptor, Raf) with antiproliferative and anti-angiogenesis prosperities in HCC, has been studied in clinical trials. Although sorafenib could benefit advanced-HCC patients, the effect is rather limited.12 Vitamin D, identified as a hormone to maintain blood calcium homeostasis and promote skeletal mineralization, has been demonstrated to exert additional functions, including antiproliferative, pro-differentiation, pro-apoptotic, anti-angiogenesis and anti-invasive characteristics in many cancer cell systems during the past two decades. These antitumor activities have led to several clinical trials. Regarding HCC, vitamin D has also been investigated from in vitro, pre-clinical animal studies to clinical trials. In this review, we discussed the source and metabolism of vitamin D, a new emerging aspect of the genomic and non-genomic actions of vitamin D, vitamin D receptor (VDR) polymorphism, clinical trials and the latest development of vitamin D analogs for potential HCC therapy.
Source and metabolism of vitamin D
There are two common forms of vitamin D: vitamin D2 and vitamin D3. They originate from two distinct sources. Vitamin D2 (ergocalciferol) is synthesized from ergosterol of yeast and vitamin D3 (cholecalciferol) is produced from 7-dehydrocholesterol (7-DHC) of lanolin. (Fig. 1). Vitamin D (representing D2 and D3) is rare in foods, only few foods contain sufficient vitamin D naturally.13 Therefore, fortification of vitamin D in foods or taking a supplement is gaining popularity in some countries. To most humans, exposure to sunlight accounts for about 90% of vitamin D requirements.13 Vitamin D is inert biologically. Vitamin D is first hydroxylated in the liver by vitamin D-25-hydroxylase (25-OHase)14 to form 25-hydroxyvitamin D [25(OH)D], the circulating form of vitamin D. 25(OH)D is considered as the most reliable index of human vitamin D status. 25(OH)D circulates as a vitamin-D-binding-protein (DBP)-bound form in the blood stream. Upon entering the kidneys, 25(OH)D is further hydroxylated mainly in the proximal tubules at the C-1 position to form 1α,25-dihydroxyvitamin D [1α,25(OH)2D], catalyzed by 25(OH)D-1α-hydroxylase (1α-OHase or CYP27B1). Among all of the vitamin D metabolites, 1α,25(OH)2D is the most active form. Another renal enzyme that also plays a crucial role is 25(OH)D-24-hydroxylase (24-OHase or CYP24A1). 24-OHase is responsible for the degradation of 1α,25(OH)2D to terminate the actions of 1α,25(OH)2D to form 1α,24,25(OH)3D. In addition, when there is an excess of 25(OH)D, 24-OHase in the kidneys can convert it to 24,25(OH)2D14 to prevent over-production of 1α,25(OH)2D. Of note, originally it was believed that 1α-OHase and 24-OHase exist exclusively in the kidneys, however, the two enzymes have been demonstrated in many extra-renal tissues.15–17 Given that anephric individuals had no detectable 1α,25(OH)2D in their circulation, it is believed that 1α,25(OH)2D generated in the extra-renal tissues acts and degrades only locally in an autocrine/paracrine fashion. This autocrine pathway seems to be regulated in a tissue-specific manner, which is not associated with systemic calcium homeostasis. Once 25(OH)D is internalized into the cells, the fate of 25(OH)D may depend on the relative expression levels of 24-OHase to 1α-OHase. In the cells with dominant expression of 1α-OHase, 25(OH)D will be converted to 1α,25(OH)2D to exert its non-calcemic functions. At the same time, the locally generated 1α,25(OH)2D will upregulate the expression of 24-OHase within the cells to hydroxylate 1α,25(OH)2D and excess 25(OH)D to form 24-hydroxylated metabolites leading to their catabolism. On the other hand, in cells dominated with the expression of 24-OHase, the generated 1α,25(OH)2D will be degraded very quickly with little or no biological actions.
Functions of vitamin D
The genomic action of 1α,25(OH)2D is mediated through its binding to vitamin D receptor (VDR) to modulate the expression of genes in a cell- and tissue-specific manner18 (Fig. 2). VDR is an endocrine member of the nuclear receptor superfamily.19 So far, there are 2776 VDR binding sites being identified by a chip-sequencing method located within 229 vitamin-D-regulated genes.20
Since the initial identification of VDR in tissues not associated with the regulation of calcium and bone metabolism by Stumpf et al.21 in 1979, many non-calcemic actions of vitamin D have been described. At the present time, the 1α,25(OH)2D-induced antiproliferation, anti-inflammatory response, pro-differentiation, pro-apoptosis and immune regulation are well established and found to be tissue- and cell-specific.17,22 For example, at least 23 human cancer cell lines have been found to express VDR23–26 and 1α,25(OH)2D has been shown to exhibit growth inhibitory effect on those cells, including prostate, breast, lung, liver, and pancreatic cancer cells.
Similar to other members of the nuclear receptor family, the liganded VDR requires further dimerization with retinoid X receptor (RXR) to form a heterodimer to bind to vitamin D response element (VDRE)27 located in the promoter region of vitamin D responsive genes to exert its genomic functions, including the inhibition of cancer cell growth and the prevention of cells from malignant transformation. The VDR-mediated gene expression is further modulated by a multiple of co-factors.28 As 1α,25(OH)2D binds to VDR, phosphorylation occurs and leads to subsequent conformational change of VDR, which, in turn, results in the release of co-repressors and recruitment of co-activators,29,30 such as HDAC3,31 a corepressor, and NCoA62/SKIP,32 a coactivator. Besides the genomic pathway, 1α,25(OH)2D exhibits the ability of changing some transmembrane signals rapidly, which results in instant biologic reaction at the plasma membrane or in the cytoplasm.18 Although this kind of action may not affect gene expression directly, it can still modulate transcription through cross-talk with various signaling pathways.33 At present, the exact mechanism for this rapid non-genomic action of 1α,25(OH)2D is not well understood, it is believed that this rapid action is associated with the non-classical membrane VDR34 to activate protein kinase C and protein phosphatase PP1c, leading to ion channel regulation etc.35,36 The non-trascriptional rapid effects of 1α,25(OH)2D may play some critical roles in controlling cancer cell proliferation.35,36
Vitamin D and HCC: biological evidence, in vivo and in vitro studies
As mentioned previously, 1α,25(OH)2D exerts antiproliferative, pro-differentiation, pro-aptotosis effects on many cancer cells which express VDR.17,25,26 In terms of HCC, Pourgholami et al. reported that 1α,25(OH)2D demonstrated growth inhibition on HCC cell lines, including four human and one rat HCC cell line, with greatest effect found on two human HCC cell lines, HepG2 and Hep3B37 (Fig. 3). The antiproliferative effect of 1α,25(OH)2D on HCC is mainly attributable to cell cycle arrest at G0/G1, leading to increased fraction of cells at G0/G1 phase and decreased fraction of cells at S phase.38 Previously, it has been shown that the observed cell cycle arrest at G0/G1, which is characteristic of 1α,25(OH)2D3 action, is through the induction of p21 and p27, leading to suppression of cyclins (D1, E and A) and cyclin-dependent kinases 2 and 4 in many cancer cell lines.39–41 Since systemically administered 1α,25(OH)2D3 can cause calcemic side-effects, 1α,25(OH)2D3 is not suitable for treating cancers. To prevent the lethal side-effect of 1α,25(OH)2D3 and to obtain a more potent antiproliferative effect, thousands of vitamin D analogs have been synthesized and studied in anticancer research. For HCC, two analogs of vitamin D, EB 1089 and CB 1093, have been shown to possess a prominent growth inhibitory effect in vitro.42 Of note, induction of apoptosis has also been found in HCC cells when they were exposed to EB 1089,43 indicating a new mechanism whereby vitamin D analogs inhibit HCC cell growth. Previously, we have reported a new vitamin D analog, 19-nor-2α-(3-hydroxypropyl)-1α,25(OH)2D3 (or MART-10), which was shown having about 1000-fold greater activity than 1α,25(OH)2D3 in inhibiting the proliferation of prostate cells derived from normal or cancerous cells in vitro.25,44 Recently, we have studied this analog in HepG2 cells. Our results demonstrate that MART-10 is about 100-fold more potent than 1α,25(OH)2D3 in inhibiting the proliferation of HepG2 cells. We further analyzed the mechanism of the antiproliferative effect on HCC and conclude that the effect is attributable to the cell cycle arrest at G0/G1 phase by upregulating p21 and p27 tumor suppressor genes (unpublished observation). Therefore, our data suggest that MART-10 is a promising candidate to be further studied as a new therapeutic regimen against HCC. In addition to using vitamin D analogs, it was found that when 1α,25(OH)2D3 was combined with fish oil, the antiproliferative effect on HCC was greatly enhanced.26 Besides an in vitro study of liver cancer cells, two animal studies using either 1α,25(OH)2D3 or EB1089, a less calcemic analog of 1α,25(OH)2D3, have been reported by Morris and colleagues.37,45 Using a xenograft animal model, they demonstrated that 1α,25(OH)2D3, at a dose of 0.5 ug/kg/ per day for 21 days, successfully inhibited the growth of SKHEP-1 cells without causing hypercalcemia in animals.37 The same group also reported that systemically administered EB 1089 was effective in repressing HCC growth in a xenograft animal model without inducing hypercalcemia.45 In addition, Sahpazidou et al. employed C3H/Sy virgin female mice, a strain capable of developing spontaneous HCC, to study the chemopreventive effect of EB1089 on HCC. They reported that the animals receiving 0.5 ug/kg of EB 1089 every other day for 2 months had 3.9% incidence of HCC as compared to the controls with 36.4% incidence, indicating the chemopreventive role of EB 1089 in HCC.46
Vitamin D and HCC: clinical trial
It is well known that hypercalcemia and hypercalciuria are the major side-effects of 1α,25(OH)2D when it is administered systemically. To overcome these drawbacks, efforts have been made to synthesize vitamin D analogs that retain most of the non-classical actions of 1α,25(OH)2D, but have much lower calcemic activity in vivo. Phase I and phase II clinical trials using 1α-hydroxyvitamin D2 (Hectorol), a pro-drug of 1α,25,(OH)2D2, have been conducted in hormone-refractory prostate cancer patients.47,48 It was reported that Hectorol was well tolerated,47 and 30% of the patients had disease stability greater than 6 months and a median survival of 21 months, which is higher than the 17.7 months predicted by the survival nomogram for that patient group.48 Although the results are less than conclusive, the encouraging findings do warrant further studies with vitamin D analogs. Other vitamin D analogs or structural VDR activators, such as Maxacalcitol (OCT), 16-ene analogs, 19-nor analogs, 1α-hydroxyvitamin D5, and LG190119, C-20 cyclopropylcalcitriol, elocalcitol, Gemini vitamin D analogs have been developed and tested in pre-clinical studies.49 These compounds may have promise as therapeutic agents for cancer and other diseases, with fewer side-effects than 1α,25(OH)2D and 1α-hydroxyvitamin D2.
An unblinded clinical trial on 10 postmenopausal women at risk for colon cancer also reported a potential role of vitamin D on the chemoprevention of colorectal neoplasia by estrogen (0.5–1 mg).50 Primary end-points in this study were the expression of VDR, CYP24A1, CYP27B1 and E-cadherin in rectal mucosa. Eight of the 10 subjects showed significant upregulation of VDR and E-cadherin, a downstream target of vitamin D action, suggesting that the chemopreventive action of hormone replacement therapy on colon cancer may result partially from changes in vitamin D activity.
As no effective regimens are available for advanced HCC at the present time, new strategies are urgently needed. In this regards, 1α,25(OH)2D3 and its analogs have been shown to possess an antiproliferative effect on HCC in vivo and in vitro, 1α,25(OH)2D3 will be a promising therapeutic regimen for advanced HCC. Knowing that a pharmacological dose of 1α,25(OH)2D3 is usually required to be therapeutically effective in treating cancers, and the serious hypercalcemic side-effect accompanying the massive dose of 1α,25(OH)2D3, Morris DL et al.51 conducted a phase I clinical trial, in which 1α,25(OH)2D3 was dissolved in 5 mL lipiodol and was injected through the hepatic artery. They reasoned that lipiodol would be preferentially retained by HCC, and by injecting 1α,25(OH)2D3 into the hepatic artery they could avoid the 24-OHase-mediated degradation of 1α,25(OH)2D3 in the liver before reaching the tumor, and therefore could obtain higher concentrations of 1α,25(OH)2D3 in HCC.52–54 Eight cases of refractory HCC were included in this study. The subjects were administered with either 50, 75, or 100 µg 1α,25(OH)2D3. Although three out of eight patients developed hypercalcemia, none of them was over grade III hypercalcemia, indicating this was a safe way to deliver 1α,25(OH)2D3. However, no obvious benefit on survival was observed in spite of transient stabilization of tumor marker, alpha-fetoprotein. EB 1089 has also been investigated in a clinical trial.55 In this trial, 56 patients with inoperable HCC were treated with EB1089 orally for up to one year with doses of EB 1089 titrated according to their serum calcium concentrations. Most of the patients could tolerate 10 µg/day of EB1089 orally. Although the survival benefit could not be obtained because no controls were included in this study, however, two patients did have the size of tumor decreased and 12 patients had stable disease.55 Further control studies are warranted to determine the survival benefit of EB 1089 on HCC.
VDR polymorphism and HCC
Human VDR cDNA was cloned in 1988 by Baker et al.,56 and the major parts of the genomic structures of the human VDR gene was described 10 years later by Miyamoto et al.57 The location of the VDR gene was later determined at the chromosome 12q13.1 region.58 The gene itself is quite large (just over 100 kb). The VDR gene has an extensive promoter region with capability of generating multiple tissue-specific transcripts.59 Recent studies have provided the existence of many subtle sequence variations (polymorphisms) in the VDR gene. Because VDR is essential for vitamin D actions, and, in addition, numerous epidemiologic studies have indicated a positive association between vitamin D deficiency and the incidence of some forms of malignancy,60 it is not surprising to find that VDR polymorphisms have been related to some malignancies and even influenced the prognosis of patients with some specific malignancies, including prostate cancer, breast cancer, and renal cell carcinoma.61,62 There are at least 25 polymorphisms of VDR gene being identified most frequently in the Caucasian population. The polymorphisms of the VDR gene could affect its binding affinity with 1α,25(OH)2D and VDRE, which, in turn, influence the transcriptional activities and proteins synthesized.63 For example, Fok1, one kind of VDR gene polymorphism, has a polymorphic site (a single nucleotide transition from T to C) in exon 2 at the 5′-end of the VDR gene. This “f” site transition allows protein translation to begin at the first initiation codon rather than from the second “F” codon, resulting in a protein with three extra amino acids. This longer VDR is a less active transcriptional activator and was found to be positively associated with breast cancer.64 In terms of HCC, Falleti et al. conducted a VDR gene polymorphism study which included 240 patients who underwent liver transplantation.65 A group of 236 healthy volunteers served as controls. HCC was detected in the explanted livers from 80 patients. Genomic DNA was isolated from the whole blood of these 80 patients and the matched controls. Then, VDR polymorphisms were investigated for FokI, BsmI, ApaI and TaqI. The study concluded that carriage of the b/b genotype of BsmI, and the T/T genotype of TaqI was significantly associated with the higher incidence of HCC. For alcoholic liver disease, existence of the BAT A-T-C and G-T-T haplotypes was a risk factor for HCC.65 On the contrary, a study conducted by Huang et al. in which VDR polymorphisms were studied in patients with hepatitis B virus (HBV) carriers, reported that VDR polymorphisms could influence the distinct clinical phenotypes in Taiwanese HBV carriers, but not associated with the development of HCC.66 Therefore, more studies are needed to further clarify the precise correlation between VDR polymorphisms and HCC.
Other vitamin-D-related signaling pathways as possible molecularly targeted therapies against HCC
Several oncogenic pathways implicated with the development of HCC have been shown to be responsive to 1α,25(OH)2D3. The pathways include VEGF67,68 and epidermal growth factor receptor.69 It has been shown that 1α,25(OH)2D3 is capable of downregulating epidermal growth factor receptor expression, that leads to the inhibition of mitogen-activated protein kinase (MAPK), and the subsequent induction of cellular differentiation, apoptosis, and growth inhibition.70 Similarly, 1α,25(OH)2D3 can inhibit endothelial cell proliferation and therefore inhibits angiogenesis,71,72 suggesting that 1α,25(OH)2D3 may have the ability to prevent the VEGF-mediated hepatocarcinogenesis through the inhibition of blood vessel formation. Furthermore, VEGF overexpression and high vessel density in the tumors of HCC have been found to influence the prognosis negatively.67,68,73 Additional HCC oncogenic pathways with a less defined vitamin D role include transforming growth factor-β1 (TGF-β1)74 and insulin-like growth factor-I and II (IGF-I and II)-mediated signaling pathway.75 In light of the recent progress relating to the non-classical actions of vitamin D, its effects on liver cells proliferation and differentiation further support the use of vitamin D compounds for the treatment and prevention of HCC.