Sodium iodide symporter (NIS) in extrathyroidal malignancies: focus on breast and urological cancer

The demonstration of NIS presence in lactating breast [23] has suggested that this protein could be expressed also in breast cancer (BC). Accordingly, in the seminal study in which NIS expression in lactating breast was discovered, it was shown that this protein is expressed in more than 80% of both invasive and in situ BCs [23]. However, both plasma membrane and intracellular immunohistochemical signal was detected (Table 2), which is in contrast to the only basolateral membrane signal detected in lactating breast. The notion that the NIS protein is expressed in a large number of breast carcinomas was confirmed by the same group by investigating a larger cohort of samples [13]. In this study it was found that NIS is also expressed in about 80% of fibroadenomas. Again, in breast carcinomas, the NIS protein was predominantly located in the cytoplasm, suggesting that in BC a deficiency of NIS trafficking from cytoplasm to plasma membrane occurs. High levels of NIS positivity in BC by immunostaining has also been described in other studies [31, 32]. It should be pointed out, however, that such a large positivity, when obtained by immunohistochemistry in the cytoplasmic compartment, could be due to non-specific staining [33]. In order to understand the molecular basis of NIS inability to target the plasma membrane in a large fraction of BC, genes whose expression is associated to NIS plasma membrane localization have been recently identified by microarray analysis [34]. Interestingly, the cysteinyl-tRNA synthetase gene is highly associated with cell surface NIS protein levels only in the estrogen receptor (ER)-positive BC subtype, suggesting that molecular mechanisms responsible for reduced plasma membrane localization of NIS may be different in a distinct subtype of BC [34].

Triple-negative BCs (TNBCs) are defined by the absence of ER, the progesterone receptor (PR), and the human epidermal growth factor receptor 2 (HER2) expression [38]. Because of absence of ER, PR, and HER2, TNBC cannot be treated by hormonal therapy or HER2-targeting compounds, leaving chemotherapy as the only therapeutic tool. Patients with this disease have a worse outcome than patients with other BC subtypes [38, 39]. It has been shown that NIS is expressed in about 65% of TNBCs and that in a fraction of them a strong plasma membrane localization is present [40]. Accordingly, in the same study, efficient iodine uptake was detected by ¹²³I scintigraphy in a patient. The notion that the NIS protein expressed in BC is able to allow radioiodine uptake has been reported in other studies as well. In fact, by studying women with infiltrating duct carcinoma, high NIS expression at both transcriptional and translational level and its ability to transport iodine in cancer tissue has been demonstrated [35]. Recently, Damle and coworkers reported that the radioiodine uptake in breast cancer specimens was significantly higher as compared to that observed in the normal tissue from the same patients [41]. In this study, 50% of breast cancer samples were positive for radioiodine uptake as well as NIS gene expression [41].

Expression and function of NIS has been investigated also in metastatic BC. Wapnir and coworkers investigated 23 patients with metastasis predominantly at the level of lung, liver, bone, and lymph node/soft tissues [30]. Eight of these subjects showed protein NIS expression, and iodide uptake was noted in two of eight NIS-expressing tumors. The same group has more recently investigated NIS expression in brain metastasis by immunohistochemistry [40]. In 75% of cases a predominant cytoplasmic signal was detected; however, plasma membrane immunoreactivity was detected only in 23.8% of NIS-positive samples. Altogether these data would indicate that NIS protein is correctly located and is able to accumulate iodine only in a small fraction of BC metastasis.

Besides immunohistochemical studies, high expression of NIS mRNA has been shown by quantitative reverse transcriptase polymerase chain reaction (RT-PCR) evaluation. Oh and coworkers have shown that NIS gene expression was present in approximately one-third of BC tissues, and no relationship was found between NIS mRNA levels and hormonal receptors expression [42]. More recently, Ryan et al. confirmed that NIS expression levels are significantly higher in BC and fibroadenoma than in normal tissue, with the highest levels of NIS mRNA observed in fibroadenoma tissues [37]. At present, detection of NIS expression levels has no prognostic value: in fact no significant relationship has been detected between NIS mRNA levels and clinical characteristics of the tumors [37]. In addition, immunohistochemistry of a subset of tumor tissues in the same cohort confirmed the presence of NIS protein both in selected malignant carcinomas and benign fibroadenomas [37].

A strategy attempted to achieve significant radioiodine uptake by the BC cells is using gene therapy to introduce an "active" exogenous NIS gene. Montiel-Equihua and coworkers have generated a replication-incompetent adenovirus, AdSERE, in which the expression of NIS is directed by an estrogen-responsive promoter [43]. Therefore, this vector would be active only in ER-positive BC (about 60% of all BC). In vitro, AdSERE mediated human NIS expression and iodide uptake in ER+ cell lines (MCF7 and ZR75-1). Moreover, the authors show that ZR75-1 AdSERE-positive xenografts in nude mice can be imaged after ^99mTc injection and their growth suppressed with therapeutic doses of ¹³¹I [43]. The use of a non-replicative adenovirus has been recently reported by the Santisteban group [44]. In this virus, NIS transcription is driven by promoters of human telomerase subunits RNA (hTR) and human telomerase reverse transcriptase (hTERT). Telomerase is a ribonucleoprotein that is essential in most human cancers but is not expressed in most normal tissues [45‐47]. Thus, hTR and hTERT promoters would be active only in cancer cells. When the BC cell line MDA-MB-231 was infected by this virus, expression of NIS protein, iodine uptake, as well as reduced cell survival after radioiodine administration was observed. A conditionally replicating adenovirus (CRAd) in which the E1a gene is driven by the tumor-specific promoter Mucin 1 (MUC-1) has also been generated [48]. This virus can efficiently replicate only in MUC-1 overexpressing cells, including BC cells [49]. In addition, this virus contains the transcriptional cassette RSV promoter-hNIScDNAbGH polyA in the E3 region, which permits NIS to express at high levels. After infection of the MUC-1-positive BC cell line T47D, virus replication, cytolysis, and release of infective viral particles, as well as iodide uptake, were observed [48].

The increase of the exogenous, virus-mediated expression of the NIS gene by pharmacological treatment has been also investigated. Treatment with retinoic acid (RA) has been shown to increase NIS expression in MCF7 cells infected by a non-replicating adenovirus in which NIS expression is controlled by the potent cytomegalovirus (CMV) promoter [50]. Indeed, the CMV promoter contains an RA-responsive element [51]. A large increase of iodine uptake has been also described in virus-infected, RA-treated MCF7 cells.

Though NIS expression has been demonstrated in most BCs, only in very few patients would spontaneous NIS expression allow efficient radioiodine uptake. For this reason, a large body of investigation has been undertaken to identify compounds that are able to increase NIS expression, its localization in plasma membrane, and iodine uptake. The major inductor of NIS expression in breast cancer cells is certainly RA. Several compounds of the RA family stimulate NIS expression, including all-trans RA, 13-cis RA, and AGN190168, all of which are already used for medical purposes [9]. Among them, the one used most to activate NIS expression in BC cells is all-trans RA. NIS expression has been induced in several BC cell lines including MCF7, T47D, and BT474 [52]. Several data indicate that RA induces NIS expression primarily by activating RARβ/RXRα heterodimer receptors. Hormone-bound receptor may act through two mechanisms. The first is binding to an element located in cis to the NIS gene [9]. It has been demonstrated that in MCF7 cells, treatment by RA induce retinoic acid receptor-alpha (RARα) binding to a retinoic acid response element located in intron 2 of the NIS gene [53]. It must be mentioned, however, that the NIS regulation by RARα was not confirmed in a different study performed on MCF7 cells [9]. The second mechanism is activation of the phosphoinositide 3-kinase (PI3K) pathway and the p38MAPK pathway. In MCF7 cells, Ohashi and coworkers have shown that either treating cells with the PI3K inhibitor LY294002 or inducing knockdown of p85alpha (a regulatory subunit of PI3K) decreases RA-induced NIS expression. Moreover, the AKT inhibitor VIII decreases iodine uptake in MCF7 cells in a dose-dependent manner [54]. Kogai and coworkers, by using both gain and loss of function experiments, have shown that p38β plays a role in the RA-induced NIS expression increase in MCF7 cells [55]. Interestingly, in the same study it was shown that in FRTL5 thyroid cells not the β but the p38α isoform has a role in NIS control of expression. Moreover, NIS induction was also observed in mouse MCF7 xenograft [56, 57], although this finding was not confirmed by another group [58]. These different results using MCF7 cells might be due to heterogeneity of this cell line [59].

In addition to gene expression, the PI3K pathway regulates NIS localization. Glycosylation of NIS protein is necessary to plasma membrane localization [60]. In MCF7 cells, overexpression of PI3K increases the non-glycosylated NIS protein [61]. In the same study, it was shown that the presence of NIS in the plasma membrane as well as iodine uptake was reduced by an active mutant of PI3K. It appears, therefore, that activation of the PI3K signaling pathways exerts opposite effects on NIS: expression is activated while NIS localization in the plasma membrane is inhibited.

Several compounds cooperate with RA in inducing NIS expression in BC cell lines. RA-induced enhancement of NIS is increased by hydrocortisone, dexamethasone, troglitazone (a peroxisome proliferator–activated receptor γ, PPARγ, agonist), histone deacetylase (HDAC) inhibitors (tricostatin A and sodium butyrate), and carbamazepine [58, 62‐64]. Hydrocortisone, dexamethasone, troglitazone, and carbamazepine cooperate with RA also in inducing iodine uptake. Interestingly, by using MCF7 xenografts in nude mice, it has been shown that RA alone is not able to increase iodine uptake; however, significant increase in ¹²³I accumulation occurs when RA is used in combination with dexamethasone [65]. Other stimulators, such as prolactin, insulin, and insulin growth factor (IGF)-I and II, are able to increase NIS mRNA levels in MCF7 cells also in the absence of RA [66]. Fortunati and coworkers reported that the HDAC inhibitor LBH589 significantly induced NIS mRNA and protein levels as well as iodine uptake in several BC cell lines [67]. Table 3 summarizes the data regarding the stimulation of iodide uptake in breast cancer cells.

In 2010 Navarra et al. analyzed the expression of NIS in tissue specimens from a large series of patients with prostate adenocarcinoma [70]. They demonstrate that approximately half of prostate cancers express the NIS at both mRNA and protein levels (Table 4). In addition, NIS expression correlates with aggressive features of the tumors such as Gleason score and pathologic stage, thus suggesting the hypothesis that these changes are a result of the dedifferentiation process occurring during a late stage of malignant transformation. A quantitative evaluation of NIS protein levels, using more sensitive methods than immunohistochemistry, will provide more details on the role of NIS as biomarker for prostate cancer aggressiveness, as reported for beta-catenin using fluorescence microscopy [71]. In prostate tumor cells expressing NIS, it appears primarily in the cytosolic fraction of the acini as a result of an incomplete maturation or too low levels of expression, as hypothesized in some thyroid and breast cancers [72, 73]. In any case, the observed strong staining of the cytoplasm makes it difficult to discern plasma membrane immunoreactivity, so that the presence of a functional NIS in the tumor cell plasma membranes could not be proved.

A recent report, proposing the function of cytoplasmic NIS as an element of a pathway involved in tumor cell invasive capacity [76], suggests a role of cytoplasmic NIS in tumor aggressiveness, strengthening the hypothesis of using NIS expression as biomarker for defining individuals with biologically active prostate cancer.

In 2003 Wapnir et al., by analyzing a few specimens of testicular tumors by immunohistochemistry, first evidenced the expression of NIS in some cores of these tumors [13]. In a larger study including a series of 107 testicular tumors, we have recently demonstrated that NIS is expressed in the plasma membrane of the large majority of seminomas and embryonal carcinomas of human testis, while it is absent in Leydig cell cancers [75]. Our data also demonstrated a significant association of the expression of NIS protein with lymphovascular invasion, a well-known marker of aggressiveness. We believe that the association between NIS expression in the tumor cells and lymphovascular invasion may reflect the different biological aggressiveness of testis tumors, suggesting the presence of NIS as an unfavorable prognostic factor. Also, its presence in the plasma membrane compartment of the tumor cells suggests that it may serve as potential carrier of radioiodine for an ablative treatment of cancer tissue.

A successful prostate cancer xenograft model has been first described that accumulates 25–30% ID/g in the tumors [77]. For comparison, poorly differentiated thyroid cancer xenografts accumulated only 4.9-9.3% ID/g and were not effectively treatable with radioiodine [78]. A NIS gene delivered with an adenovirus vector and a tissue-specific gene promoter, the prostate-specific antigen gene (PSA) promoter, conferred efficient functional NIS expression in prostate cancer xenografts [79, 80]. In a recent report, Trujillo and coworkers, by using a prostate tumor–specific CRAd in a xenograft model of prostate cancer, demonstrated that the efficacy of radioiodide therapy depends mainly on an efficient viral tumor spread and a decrease in the rate of the efflux of radioisotope [81]. To achieve synergistic or additive cytotoxic effects, combined treatments with NIS gene therapy and a tumor targeting strategy, such as utilization of an oncolytic vector [82], are also under experimentation.

Induction of NIS expression has been obtained in vitro in two testicular cancer cell lines. Findings from our laboratories revealed that NIS expression may be enhanced in vitro in a human embryonal testicular carcinoma cell line by the histone deacetylase inhibitor (HDACi) [75]. Histone acetylation is a known epigenetic mechanism of regulation of gene expression, and its alteration has been reported in many human cancers [83]. In many cell lines of thyroid and non-thyroid cancer, HDACi have been successfully tested to induce radioiodine uptake due to increased NIS expression [84‐86]. The same result was obtained in the NTERA cells in our study, showing that, at least in vitro, embryonal testicular tumor cell susceptibility to radioiodine administration may occur, and suggesting the possibility of using radioiodine after pharmacological induction of NIS expression even in this rare tumor histotype. It is noteworthy that these drugs are being tested in clinical trials at doses compatible with those effective in vitro.

Recently, Maggisano et al. analyzed the effects of the HDACi suberoylanilide hydroxamic acid (SAHA) and valproic acid (VPA) on NIS expression and function in rat Leydig testicular carcinoma cells (LC540) [87]. LC540 cells were exposed to SAHA 3 μM and VPA 3 mM (alone and in combination), and NIS mRNA and protein levels were evaluated by using, respectively, real-time RT-PCR and western blotting. Also, NIS function was analyzed by iodide uptake assay. They found that both HDACi, used alone, were able to stimulate the transcription of NIS gene, but not its protein expression, while the association of SAHA and VPA increased both NIS transcript and protein levels, resulting in a significant enhancement of radioiodine uptake capacity of LC540 cells. These data demonstrate the presence of an epigenetic control of NIS expression in Leydig tumor cells, suggesting the possibility of using the combination of these two HDACi for a radioiodine-based treatment of these malignancies.

Considering altogether data obtained in breast, prostate and testicular cancer, an important difference seems to emerge. In fact, in testicular and prostate cancer NIS expression, evaluated by immunohistochemistry, appears to be related to the degree of dedifferentiation and aggressiveness [70, 75, 76]. However, such a relationship is not present in breast cancer [37]. Such a difference seems not due to different methodologies in detecting NIS expression. In fact, the lack of correlation between NIS expression and dedifferentiation was detected in breast tumors both by using RT-PCR and immunohistochemistry [37]. Thus, discrepancy observed between testicular/prostate tumors and breast tumors would depend on the difference of originating tissues. It would be relevant to test such a possibility in a single study, in which testicular, prostate and breast cancer are investigated by the same methodology.