Role of Transferrin Receptor and the ABC Transporters ABCB6 and ABCB7 for Resistance and Differentiation of Tumor Cells towards Artesunate

The anti-malarial artesunate also exerts profound anti-cancer activity. The susceptibility of tumor cells to artesunate can be enhanced by ferrous iron. The transferrin receptor (TfR) is involved in iron uptake by internalization of transferrin and is over-expressed in rapidly growing tumors. The ATP-binding cassette (ABC) transporters ABCB6 and ABCB7 are also involved in iron homeostasis. To investigate whether these proteins play a role for sensitivity towards artesunate, Oncotest's 36 cell line panel was treated with artesunate or artesunate plus iron(II) glycine sulfate (Ferrosanol®). The majority of cell lines showed increased inhibition rates, for the combination of artesunate plus iron(II) glycine sulfate compared to artesunate alone. However, in 11 out of the 36 cell lines the combination treatment was not superior. Cell lines with high TfR expression significantly correlated with high degrees of modulation indicating that high TfR expressing tumor cells would be more efficiently inhibited by this combination treatment than low TfR expressing ones. Furthermore, we found a significant relationship between cellular response to artesunate and TfR expression in 55 cell lines of the National Cancer Institute (NCI), USA. A significant correlation was also found for ABCB6, but not for ABCB7 in the NCI panel. Artesunate treatment of human CCRF-CEM leukemia and MCF7 breast cancer cells induced ABCB6 expression but repressed ABCB7 expression. Finally, artesunate inhibited proliferation and differentiation of mouse erythroleukemia (MEL) cells. Down-regulation of ABCB6 by antisense oligonucleotides inhibited differentiation of MEL cells indicating that artesunate and ABCB6 may cooperate. In conclusion, our results indicate that ferrous iron improves the activity of artesunate in some but not all tumor cell lines. Several factors involved in iron homeostasis such as TfR and ABCB6 may contribute to this effect.


INTRODUCTION
Artemisinin is a sesquiterpene isolated from Artemisia annua L., which is used in traditional Chinese medicine for the treatment of fever and chills [1]. Artemisinin reveals profound activity against Plasmodium falciparum and Plasmodium vivax [2,3]. Artesunate and artemether are semi-synthetic derivatives of artemisinin with improved pharmacological features [4]. In addition to their antimalarial activity, artemisinin and its derivatives are also active against cancer cells [5][6][7][8][9]. In both cases, the activity of the drugs is associated with the presence of iron. Iron is present in large excess bound to hemoglobin in erythrocytes, where the Plasmodia parasites are located. The active moiety of artemisinin-like drugs is an endoperoxide bridge, whose reductive homolysis is promoted by iron(II)-heme leading to C4-centered alkylating radicals [10]. These radical molecules cause macromolecular damage by alkylating essential malarial proteins inducing cell death of parasites [11][12][13]. On the other hand, iron content is higher in tumor cells than in normal cells [14] making them more susceptible to artemisinins. We and others have shown that the susceptibility of tumor cells to artemisinins can further be enhanced by the addition of transferrin or ferrous iron [15,16].
The role of artemisinins and iron for malaria treatment has been intensively investigated during the past years [17], whereas the role of iron for tumor treatment with artemisinins is far less understood. The iron-binding protein, transferrin, is internalized into cancer cells after binding to the transferrin receptor (TfR, CD71). This is a transmembrane glycoprotein involved in iron uptake by internalization of transferrin. TfR exerts growth regulatory functions and is over-expressed in rapidly growing tumors [18]. The expression of TfR is of prognostic significance for several tumor types [19][20][21]. Ferrous iron can either be bound to transferrin or to other proteins before uptake.
The ATP-binding cassette (ABC) transporters ABCB6 and ABCB7 are involved in iron homeostasis. They are located in the mitochondria and transport heme and protoporphyrins into these organelles [22]. Most ABC transporters are involved in the active transport of phospholipids, ions, peptides, steroids, polysaccharides, amino acids, bile acids, pharmaceutical drugs and other xenobiotic compounds [23].
In humans, 49 different ABC transporters have been identified, which are classified into seven sub-families (A-G) [24]. In healthy organs several ABC transporters protect against the harmful effects of xenobiotics taken up with food. A high expression can, therefore, be found in the gastrointestinal tract, liver, and kidney. A protective function is also given as components of the blood brain barrier and the blood placenta barrier.
ABC transporters have been intensively investigated in the past years. ABCB1 (P-glycoprotein, MDR1), ABCC1-C6 (MRP1-6) and ABCG2 (BCRP) confer resistance to cytostatic drugs of tumors and contribute to the failure of tumor chemotherapy [25]. The mutated gene product of ABCC7 (CFTR) is involved in the pathogenesis cystic fibrosis, and some other mutated ABC transporter genes contribute to a number of hereditary diseases and disorders [24]. The pathogenic function of the majority of ABC transporters still awaits elucidation. To gain insight into functional aspects, we have developed a low density microarray for ABC transporters [26,27]. Applying this microarray, we identified ABCA2 and ABCA3 as further genes involved in tumor drug resistance [28,29].
In the present investigation, we focused on TfR, ABCB6, and ABCB7. The aim of the present study was to address the question, whether these proteins are involved in the response of tumor cells to artesunate.

Cell Lines
CCRF-CEM were cultured in RPMI medium and MCF-7 and MEL cells in DMEM medium each supplemented with 10% fetal calf serum under standard conditions (37uC, 5% CO 2 ). Cells were passaged twice weekly.
Twenty-four cell lines out of Oncotest's 36 cell line panel were established at Oncotest from patient-derived tumor xenografts passaged subcutaneously in nude mice [30]. The origin of the donor xenografts have been previously described [31]. The cell lines T24, Panc1 and 22RV1 were obtained from ATCC (Rockville, MD, USA), the cell line LnCAP from DSMZ (Braunschweig, Germany) and the other 8 cell lines were kindly provided by the National Cancer Institute (Bethesda, MA, USA).
The cultivation of 55 cell lines (leukemia, melanoma, non-small cell lung cancer, colon cancer, renal cancer, ovarian cancer, brain tumors, prostate cancer, and breast cancer) of the Developmental Therapeutics Program of the NCI has been previously described [32].

Effect on cell growth
MEL cells were cultured with and without 1.0 mg/ml artesunate (2.6 mM) for 24 and 48 h. Cell numbers were determined by trypan blue exclusion.

Effect on differentiation
MEL cells (1610 6 cells/ml) were cultures in the presence of 2% DMSO with or without 1.0 mg/ml artesunate (2.6 mM) for 28h. The content of heme was estimated by the conversion of heme to protoporphyrin with oxalic acid [33 Taketani et al., 2003].

Sulforhodamine B assay
Drug sensitivity of the NCI cell lines has been determined by the sulforhodamine B assay [34 Rubinstein et al., 1990]. The 50% inhibition concentration (IC 50 ) values for artesunate have been reported [7] and those of other anti-cancer drugs have been deposited in the database of the NCI database (http://dtp.nci.nih.gov).

Propidium iodide (PI) assay
A modified propidium iodide assay [35] was used to assess the compound's activity in Oncotest's 36 cell line panel. Briefly, cells were harvested from exponential phase cultures by trypsinization, counted and plated in 96 well flat-bottomed micro-titer plates at a cell density depending on the cell line (4.000-10.000 cells/well). After a 24 h recovery period to allow the cells to adhere and resume exponential growth, 10 ml of culture medium (six control wells/plate) or of culture medium containing the test compounds were added to the cells. The compounds were applied in triplicates at five concentrations. Following four days of continuous drug exposure, medium or medium with test compound, including all dead cells suspended in the culture medium, was aspirated and replaced by 200 ml of an aqueous propidium iodide (PI) solution (7 mg/ml). To measure the amount of living cells, cells were permeabilized by freezing the plates, resulting in the death of all cells that had remained attached to the bottom of the well after the incubation period. After thawing of the plates, fluorescence was measured using the Cytofluor 4000 micro-plate reader (excitation 530 nm, emission 620 nm), providing a direct relationship to the total viable cell number.

Cell line array
Arrays of cell lines from the Oncotest Human Cell Line Collection were assembled using a tissue arrayer (Beecher Instruments, Sun Prairie, Wisc., USA) as recently described [36]. Pellet biopsies (0.6 mm in diameter) were taken from embedded cells and arrayed in duplicate in a new recipient paraffin block. Four-micrometer sections of the resulting microarray block were cut and transferred onto glass slides using the paraffin-sectioning aid system (Instrumedics, Hackensack, NJ, USA) and further processed for immunohistochemical staining. The arrays were stained with the primary antibody 71C03 against TfR (Neomarkers, Fremont, CA, USA). Staining was analyzed by light microscopy and semiquantitatively evaluated as described [36].

Quantitative real-time PCR
Quantitative PCR was done as described [37]. The ABI Prism 7700 Sequence Detector and Pre-Developed Assay Reagents (Applied Biosystems, Weiterstadt, Germany) were used for the quantification of all genes. The expression of the ABC transporters was standardized for the expression of two housekeeping genes, b-2-microglobulin (B2M) and Abelson gene 1 (ABL1). The ABI Prism 7700 Sequence Detector and Pre-Developed Assay Reagents (Applied Biosystems, Weiterstadt, Germany) were used for the quantification of all genes. The assay IDs were: B2M: Hs00187842_m1; ABL1: Hs00245445_m1; ABCB6: Hs00180568_m1 and ABCB7: Hs00188776_m1.

Western blotting
MEL cells (1610 6 cells/ml) were cultured in the presence of 2% DMSO. After 24, 48, 72, or 96 h the cells were collected and lysed. The cellular proteins were analyzed by SDS-PAGE and transferred onto PVDF membranes. Immunoblotting was performed with antibodies for ABCB6, ABCB7, and actin as the primary antibodies [33].

Transfection of oligonucleotides
Phosphorothioate sense and antisense oligonucleotides corresponding to the mouse ABCB6 gene were transfected using DOTAP transfection reagent (Roche Molecular Biochemicals, Mannheim, Germany) as described [33] Statistics Fisher's exact test was used to calculate significance values as a measure for the dependency of two variables. This test was implemented into the WinSTAT Program (Kalmia, Cambridge, MA, U.S.A.). The TfR expression in the Oncotest panel of tumor cell lines has been determined by a tissue micro-array technique adapted for the examination of cell line and immunohistochemistry [36]. The results have been validated by conventional immunohistochemistry and Western blotting [36]. The mRNA expression values for TFR (clone M11507, GC102405) determined by microarray analysis [38,39] were selected from the NCI database (http://dtp.nci.nih.gov). The microarray-based expression levels of ABCB6 and ABCB7 have been validated by RT-PCR [40].

Transferrin Receptor (TfR)
To investigate the role of TfR and iron for cellular sensitivity to artesunate, Oncotest's 36 cell line panel was used. The 50% inhibition concentrations (IC 50 ) for artesunate are shown in Table 1. The IC 50 values varied between 0.197 and 47.779 mg/ml (0.512 and 124.295 mM respectively). The mean IC 50 value was 2.791 mg/ml (7.261 mM). Furthermore, we tested a combination of artesunate and iron(II) glycine sulfate (10 mg/ml; FerrosanolH) ( Table 1). Here, the mean IC 50 value was 0.978 mg/ ml (2.544 mM) with a range from 0.013 and 22.96 mg/ml (0.026 and 59.748 mM, respectively). The data in Table 1 are arranged according to the IC 50 values for artesunate. In the majority of cell lines, the combination of artesunate plus iron(II) glycine sulfate resulted increased inhibition rates. However, in 11 out of the 36 cell lines (31%; OVXF 1619L, CNXF, 498NL, MCF7, RXF 393NL, MEXF 514L, MEXF 394NL, LXFL 529L, H460, 22RV1, RXF1781L, OVXL 899L) the combination treatment was not superior to treatment with artesunate alone (degree of modulation ,1.2). Furthermore, degrees of modulation (IC 50 for artesunate divided by the IC 50 for artesunate plus iron(II) glycine sulfate) did not correlate with IC 50 values for artesunate, e.g. cell lines with high IC 50 values for artesunate were not more modulated by the addition of iron(II) glycine sulfate than cell lines with low IC 50 values for artesunate (Table 1).
Then, we correlated the expression of TfR with the degrees of modulation. We found that cell lines with high TfR expression significantly correlated with high degrees of modulation ( Table 2) indicating that high TfR expressing tumor cells could be more efficiently inhibited by a combination of artesunate and iron(II) glycine sulfate than low TfR expressing ones. TfR expression did not correlate with IC 50 values for artesunate alone or artesunate plus iron(II) glycine sulfate.
Furthermore, we investigated the association of TFR for sensitivity or resistance to artesunate in 55 cell lines of the NCI drug screening panel. We found a significant relationship between cellular response to artesunate and TFR expression ( Table 3) indicating that TFR might be a determinant of artesunate sensitivity in tumor cells.

ABCB6 and ABCB7
Since ABCB6 and ABCB7 are involved in iron homeostasis of cells, we investigated a role of these two proteins for cellular response to artesunate. Again, we used the IC 50 values for artesunate of 55 NCI cell lines. These data were correlated with the microarray-based expression of ABCB6 (clone AF070598, GC56272) and ABCB7 (clone AA056272, GC10226). As shown in Table 3, a significant correlation was found for ABCB6, but not for ABCB7.
Then, MCF-7 breast cancer cells and CCRF-CEM leukemia cells were incubated with artesunate in a range of 0.05 to 50 mg/ ml (0.13 to 130 mM respectively) for 24 h. Afterwards, mRNA was isolated and the expression of ABCB6 and ABCB7 determined by real time PCR. As can be seen in Figure 1 the expression of ABCB6 increased with increasing doses of artesunate, while the expression of ABCB7 decreased accordingly. The induction of gene expression by artesunate might indicate a role of these two genes for cellular response to artesunate treatment. This can, however, not be taken as evidence, that the two genes are causatively related to artesunate response, and an epiphenomenal association can not be excluded.
Finally, we treated MEL cells with artesunate (1 mg/ml, 2.6 mM). Mouse erythroleukemia (MEL) cells were used as model to study the role of ABCB6 and ABCB7 during cellular differentiation. While cell growth was almost stopped, only 20% of cells died (Figure 2A). Treatment of MEL cells with 2% DMSO leads to erythroid differentiation and heme production (33). DMSO also induced the expression of ABCB6 and ABCB7 proteins ( Figure 2B). Furthermore, the addition of artesunate completely blocked DMSO-induced erythroid differentiation in MEL cells in a dose-dependent manner ( Figure 2B). Under these conditions most cells were alive. This indicates that artesunate is a potent inhibitor of proliferation and differentiation in MEL cells. Artesunate induced protein expression of both ABCB6 and ABCB7 ( Figure 2C).
The IC 50 value regarding heme production for artesunate in MEL cells was 0.18 mg/ml (0.468 mM). We previously reported that an antisense oligonucleotide directed against ABCB6 depressed the erythroid differentiation of MEL cells to 75% [33]. When MEL cells were transfected with this antisense oligonucleotide and treated with artesunate (0.18 mg/ml, 0.468 mM) in combination, the heme content decreased to about 30% control ( Figure 2D). This indicates an additive effect.

DISCUSSION
The aim of the present study was to explore the role of proteins related to iron homeostasis for response of tumor cells to artesunate. For this purpose, we focused on TfR and the ABC transporters, ABCB6 and ABCB7.
We observed a wide heterogeneity of tumor cell lines towards artesunate. Whereas iron(II) glycine sulfate was able to increase sensitivity of tumor cells to artesunate in most cell lines, there was no correlation between the IC 50 values and high degrees of modulation by iron(II) glycine sulfate. This indicates that iron(II) glycine sulfate is not a typical resistance modulator comparable to modulators of multidrug resistance, where modulating agents completely or partially overcome resistance of a drug-resistant subline to a cytostatic drug towards the level of a parental drugsensitive cell line [25].
It is important to point out that several cell lines did not show increased inhibition by the combination of artesunate and iron(II) glycine sulfate compared to artesunate alone. Some of these cell lines did even reveal decreased efficacy to artesunate under cotreatment with iron(II) glycine sulfate. Iron is an important regulator of cell growth and tumor cells may take advantage of external iron supply to grow faster. Iron(II) glycine sulfate is an approved drug, FerrosanolH applied for many years in the clinic to treat iron deficiency. Therefore, iron(II) glycine sulfate seems to be a safe and suitable adjunct for combination treatments with artesunate at first sight. Our results do, however, indicate that there may be patients, who would not profit from addition of iron(II) glycine sulfate during artesunate treatment. In this context, iron(II) glycine sulfate is a double edged sword improving treatment response in the majority of tumor cell lines but worsening it in others. Other iron preparations have also been reported to improve the response of tumor cells to artemisinins, e.g. ferrous sulfate or iron-loaded transferring (holotransferrin) [9,15]. It remains to be analyzed, whether the same is true for these iron preparations.
The fact that in some cases iron(II) gylcine sulfate could also lead to reduced response of tumor cells towards artesunate opens an option for the testing of tumors before treatment in a clinical setting. If predictive tests, e.g., determination of TfR expression or sensitivity testing by MTT assay, indicate enhanced sensitivity,  [41].
On the other hand, three out of the five melanoma cell lines were clearly above-average sensitive to artesunate (individual IC 50 ,4.2 mM), and both more resistant melanomas (IC 50 .12 mM) were strongly sensitized (degree of modulation .15) by the addition of iron(II) glycine sulfate indicating that this entity is especially suited for this treatment regimen. Interestingly, these results are supported by recent clinical data. Artesunate has been applied to two patients with otherwise drug-resistant and refractory uveal mealoma [42]. The prognosis of refractory uveal melanoma patients is generally worse (3-5 months). One artesunate treated patient with uveal melanoma had a stable disease following artesunate treated, but finally the tumor progressed and the patient died after 24 months. The second patients with uveal melanoma was treated with artesunate plus iron(II) glycine sulfate and is still living after more than 50 months [42]. This may be taken as a clue that the combination of artesunate plus iron(II)glycine sulfate is indeed especially suited for this tumor type. Further analyses have to prove this hypothesis.
Interestingly, the expression of TFR correlated with IC 50 values for artesunate alone in the panel of 55 NCI cell lines, but not in the panel of 36 Oncotest cell lines. This is a clue that TFR expression is only one determinant of response of tumor cells towards artesunate. While endogenous iron might contribute to artesunate sensitivity, exogenous iron application during artesunate treatment is favorable for those tumors with high TFR expression levels.
In addition to TfR, iron is transported in cells by ABCB6 and ABCB7. Therefore, we also investigated the role of these two proteins for treatment of tumor cells with artesunate. ABCB6 is involved in the biosynthesis of heme via interaction with ferrochelatase, which is regulated by iron [33]. We found that the microarray-based mRNA expression of ABCB6 but not of ABCB7 correlated with IC 50 values for artesunate in the NCI cell line panel. While ABCB6 was induced upon artesunate treatment in MCF7 and CCRF-CEM cells, ABCB7 expression was downregulated. These results speak at least in the case of ABCB6 for a role in determining sensitivity to artesunate.
In addition to the anti-proliferative activity of artesunate on tumor cells, we also analyzed its effect on differentiation. For this reason, we used MEL cells, which are an established model to study ABCB6 and ABCB7 during cellular differentiation [43]. DMSO induced both differentiation and up-regulation of ABCB6 and ABCB7 protein expression. Artesunate was able to reverse DMSO-induced differentiation was measured by cellular heme production. Down-regulation of ABCB6 by transfection with antisense oligonucleotides inhibited DMSO-induced heme biosynthesis indicating that this protein might play a role in differentiation.
The role of ABCB6 and ABCB7 for drug resistance has been sparsely investigated as of yet. Boonstra et al. [44] did not find that ABCB6 expression in small lung cancer cells was associated with resistance to mitoxantrone. Thirteen different ABC transporters including ABCB6 and different anti-apoptotic Bcl2 gene family members were reported to be amplified in various drug-resistant cell lines [45]. While the other factors may contribute to clinical response towards neo-adjuvant chemotherapy, the specific role of ABCB6 remains unanswered in this study. Park et al. [46] analyzed the ABC transporter gene expression profiles in breast cancer patients who underwent neo-adjuvant chemotherapy.