Resistance to the Antitumor Agent Gallium Nitrate in Human Leukemic Cells Is Associated with Decreased Gallium/Iron Uptake, Increased Activity of Iron Regulatory Protein-1, and Decreased Ferritin Production*

The mechanism of drug resistance to gallium nitrate is not known. Since gallium can be incorporated into ferritin, an iron storage protein that protects cells from iron toxicity, we investigated whether ferritin expression was altered in gallium-resistant (R) CCRF-CEM cells. We found that the ferritin content of R cells was decreased, while heavy chain ferritin mRNA levels and iron regulatory protein-1 (IRP-1) RNA binding activity were increased. IRP-1 protein levels were similar in gal- lium-sensitive (S) and R cells, indicating that R cells contain a greater proportion of IRP-1 in a high affinity mRNA binding state. 59 Fe uptake and transferrin receptor expression were decreased in R cells. In both S and R cells, gallium inhibited cellular 59 Fe uptake, increased ferritin mRNA and protein, and decreased IRP-1 bind- ing activity. Gallium uptake by R cells was markedly diminished; however, the sensitivity of R cells to gallium could be restored by increasing their uptake of gallium with excess transferrin. Our results suggest that R cells have developed resistance to gallium by down-regulat-ing their uptake of gallium. In parallel, iron uptake by R cells is also decreased, leading to changes in iron home-ostasis. Furthermore, since gallium has divergent ef- fects on iron uptake and ferritin synthesis, its action may also include a direct effect on ferritin mRNA induc- tion and IRP-1 activity.

Gallium nitrate, a group IIIa metal salt with antineoplastic activity (1), is currently undergoing evaluation as a chemotherapeutic agent. A number of clinical studies have shown gallium to be effective in the treatment of lymphoma and bladder cancer (2)(3)(4). Although gallium is in clinical use, information regarding its action at the cellular and molecular levels is largely incomplete. It has been known for some time that gallium resembles iron in certain respects. Gallium binds avidly to the iron transport protein Tf 1 (5) and is incorporated into cells by Tf receptor-dependent and -independent transport systems (6 -8). Furthermore, we have shown that gallium inhibits the growth of several leukemic cell lines by inhibiting cellular iron uptake and by blocking the activity of ribonucleotide reductase (9 -11).
Iron taken up by cells is stored in ferritin, a high molecular weight protein composed of 24 subunits of H and L chains (12). Ferritin sequesters excess intracellular iron and protects cells from the toxicity of iron overload. An increase in the delivery of iron to cells therefore serves as a powerful stimulus for ferritin synthesis. It is now established that iron-dependent ferritin and transferrin receptor gene expression is regulated by the binding of two iron regulatory proteins (IRPs), IRP-1 and IRP-2, to sequences termed iron-responsive elements (IREs) present at the 5Ј-untranslated region of ferritin mRNA and the 3Ј-untranslated region of Tf receptor mRNA (13)(14)(15)(16)(17). In irondepleted cells, IRPs bind with high affinity to the IREs, resulting in suppression of ferritin mRNA translation and stabilization of Tf receptor mRNA (increased Tf receptor synthesis). Conversely, under conditions of intracellular iron excess, IRPs are converted to low affinity forms that no longer bind IREs. As a result, ferritin synthesis proceeds unhindered in iron-repleted cells, and Tf receptor synthesis is down-regulated. The iron-induced decrease in IRP-1 binding activity occurs through a switch between apoprotein (high affinity) and [4Fe-4S] forms (low affinity) without changes in IRP-1 levels (18), whereas the decrease in IRP-2 binding activity is due to iron-induced IRP-2 proteolysis (19,20). IREs are also present on the 5Ј-untranslated regions of erythroid 5-aminolevulinate synthase (21) and mitochondrial aconitase (22) and on mitochondrial succinate dehydrogenase subunit b in Drosophila melanogaster (23). In addition to iron, nitric oxide (24 -26), oxidative stress (27,28), and ascorbic acid (29) have also been shown to regulate IRP binding to the IREs (reviewed in Ref. 30).
Tumor cell resistance to antineoplastic agents is a major obstacle to the successful treatment of malignancy. Recently, we reported the development of a line of human leukemic CCRF-CEM cells resistant to the growth inhibitory effects of gallium nitrate (31). In the present investigation, we have attempted to advance our understanding of the basis for this drug resistance. Because gallium interacts with various steps in iron metabolism, we hypothesized that the development of gallium resistance may be associated with changes in the cellular handling of iron and iron proteins. Moreover, since gallium can be incorporated into ferritin (7,32) and ferritin protects cells from iron toxicity, we questioned whether this protein might play a role in modulating the cytotoxicity of gallium. Our studies show that the development of gallium resistance is due to a down-regulation in the transport of gallium into cells and that iron transport is affected in parallel, resulting in changes in the regulation of ferritin gene expression.

EXPERIMENTAL PROCEDURES
Materials-Gallium nitrate was obtained from Alpha Products (Danvers, MA). Heparin, bovine serum albumin, MTT, Tf, and dithiothreitol were obtained from Sigma. RNase T1 was obtained from Boehringer Mannheim. Na 125 I, 59 FeCl 3 , [ 32 P]GTP, and [ 32 P]dCTP were purchased from Amersham Corp. 59 Fe-Tf was prepared as described by Bates and Schlabach (33), and 125 I-Tf was prepared by the chloramine-T method (34).
Cells-Human T lymphoblastic leukemic CCRF-CEM cells (S cells) were obtained from American Type Tissue Collection (Rockville, MD) and were grown in RPMI 1640 medium supplemented with 10% fetal calf serum in an atmosphere of 6% CO 2 at 37°C. A clone of CCRF-CEM cells resistant to the growth inhibitory effects of gallium nitrate (R cells) was developed in our laboratory through a process of continuous exposure of these cells to stepwise increments of gallium nitrate over the course of several months. R cells were routinely cultured in medium containing 150 M gallium nitrate. For experiments, cells grown to confluency were harvested, washed with complete medium to remove gallium nitrate, and then subcultured in fresh medium in the presence or absence of gallium nitrate. At specified times, cells were harvested and analyzed as described below.
Cell Proliferation Assay-The effect of gallium on the proliferation of S and R cells was determined by MTT assay as described by Mossman (35). Cells were plated at an initial density of 2 ϫ 10 5 cells/ml in 96-well microwell plates and incubated for 72 h in the presence of 0 -1000 M gallium nitrate. At the end of the incubation, 10 l of MTT (5 mg/ml stock solution) was added to each well, and the cells were incubated at 37°C for an additional 4 h. Cells were then solubilized by the addition of 100 l of 0.04 N HCl in isopropyl alcohol to each well, and the absorbance of each well was determined spectrophotometrically at a dual wavelength of 570/630 nm using an EL 310 microplate auto-reader (Biotech Instruments, Winooski, VT). The effect of gallium nitrate on cell proliferation was determined by comparing the absorbance of the wells containing gallium nitrate with that of wells in which the drug was omitted. In some experiments, cell number was determined by directly counting cells using a hemocytometer.
cDNA Probes and Antiserum-cDNAs for rat L and H ferritin cloned into pGEM4Z and pSP65 vectors (Promega, Madison, WI), respectively, and rabbit antiserum to IRP-1 were generously provided by Richard Eisenstein and have been previously described (36). cDNAs for L and H ferritin were excised from the plasmids with PstI and EcoRI, respectively, and were 32 P-labeled using a RadPrime DNA labeling system from Life Technologies, Inc. Human ␤-actin cDNA probe was obtained from CLONTECH (Palo Alto, CA). 32 P-Labeled IRE mRNA for the RNA band-shift assay was prepared, using as a template, a 1000-base pair rat L ferritin pseudogene that contains the conserved IRE sequence. The plasmid (p66-L gene) containing this insert (generously provided by Elizabeth Leibold) (13) was linearized with SmaI (Life Technologies, Inc.) and used for in vitro transcription of IRE mRNA. Transcription was carried out with Sp6 RNA polymerase using a Riboprobe transcription system from Promega.
Measurement of Cellular Ferritin-Ferritin in S and R CCRF-CEM cells was measured before and after a 24-h incubation in fresh medium. Cells were harvested; washed with 10 mM KPO 4 , pH 7.4, 150 mM NaCl buffer (PBS); and disrupted by sonication. Cellular debris was removed by centrifugation (30,000 ϫ g for 30 min), and the cytoplasmic fraction (supernatant) was assayed for protein content using a Pierce BCA protein assay and for total ferritin using a Bio-Rad Quantimune ferritin assay. The ferritin content is expressed as ng/mg of protein.
RNA Isolation and Northern Blotting-Total cellular RNA was isolated from S and R cells by a modification of the method of Chomczynski and Sacchi (37) using RNAzol (Tel-Test Inc., Friendswood, TX) according to the manufacturer's recommendations. The integrity of the RNA was verified by agarose gel electrophoresis. Twenty micrograms of RNA from each preparation was electrophoresed on a 1% agarose gel containing 2.2 M formaldehyde and transferred to Nytran membranes (Schleicher & Schuell) using the capillary-blotting method. H and L ferritin and ␤-actin mRNA were detected by sequential hybridization of the membranes to the corresponding 32 P-labeled cDNA probes (1.2 ϫ 10 6 cpm/ml) using QuickHyb hybridization solution (Stratagene, La Jolla, CA) according to the manufacturer's recommendations. Autoradiography of the membranes was carried out by exposing the membranes to XAR-5 film (Eastman Kodak Co.) with intensifying screens at Ϫ70°C for 24 -48 h. For quantitation of band intensities, the autoradiograph was scanned on an AMBIS optical imaging system. The band intensities obtained with 32 P-labeled ␤-actin were used to monitor for equal loading of RNA on the gels and to normalize the results obtained with hybridization using 32 P-labeled L and H ferritin cDNA probes.
RNA Band-shift Assay-The binding of IRP-1 to ferritin IRE mRNA was examined by an RNA band-shift assay as described by Leibold and Munro (13). For preparation of cytoplasmic extracts, S and R cells were incubated for 24 h in fresh medium with or without gallium. Cells were then washed by centrifugation with PBS and were lysed in 20 mM HEPES, pH 7.6, containing 5% glycerol, 0.5 mM EDTA, 25 mM KCl, 1 mM dithiothreitol, 1% Nonidet P-40, and 1 mM phenylmethylsulfonyl fluoride. Cell lysates were centrifuged at 30,000 ϫ g for 30 min, and the protein content of the supernatant was measured. For the band-shift assay, 40 g of cytoplasmic extract from cells was incubated with 200,000 cpm of 32 P-labeled IRE mRNA in binding buffer containing 10 mM HEPES, pH 7.6, 3 mM MgCl 2 , 40 mM KCl, 5% glycerol, and 1 mM dithiothreitol. To specifically identify IRP-1 binding to the IRE, incubation conditions were as described above, except that cytoplasmic extracts were incubated in binding buffer with 5 l of antiserum to IRP-1 for 1 h at 4°C prior to the addition of 32 P-labeled IRE mRNA. After a 30-min incubation at room temperature, 1 unit of RNase T1 was added to the reaction mixture, and incubation was continued for an additional 10 min. Heparin (5 mg/ml) was then added to the reaction, and incubation was continued for another 10 min at room temperature. The mixture was resolved on a 5% nondenaturing polyacrylamide gel, and autoradiography of the gel was performed at Ϫ70°C. For quantitation of band intensities, the autoradiograph was scanned on an AM-BIS optical imaging system.
Western Blotting-S and R cells were incubated for 24 h in fresh medium containing 0 -250 M gallium nitrate. Approximately 5 ϫ 10 7 cells were harvested; washed twice with ice-cold 10 mM Tris, pH 7.6, 150 mM NaCl buffer; and then resuspended in 150 -500 l of the same buffer. Phenylmethylsulfonyl fluoride (final concentration of 1 mM) was added to the cell suspension, and the cells were disrupted by sonication. The cell lysate was then centrifuged at 30,000 ϫ g for 45 min at 4°C, and the supernatant was assayed for protein content and used for Western blotting to detect IRP-1. SDS-polyacrylamide gel electrophoresis of the samples was performed according to the method of Laemmli (38), and proteins were transferred from the gel onto a nitrocellulose membrane as described by Towbin et al. (39) using a Transblot system (Bio-Rad). Membranes were incubated in 50 mM Tris, pH 7.5, 150 mM NaCl buffer (TS buffer) containing 2.5% bovine serum albumin for 1 h at room temperature, followed by a 90-min incubation in the same buffer containing rabbit antiserum against IRP-1 (1:1000 dilution). Following sequential washes in TS buffer and in TS buffer containing 0.05% Triton X-100 (TST), membranes were incubated for an additional 90 min at room temperature in TS buffer/bovine serum albumin containing 125 I-protein A (200,000 cpm/ml). The membranes were finally washed in TS and TST buffers, and autoradiography of the membranes was carried out at Ϫ70°C. 59 Fe Uptake Studies-S and R cells were washed twice with medium and replated (0.5 ϫ 10 6 cells/ml) in 1-ml 24-well plates in fresh medium with or without 150 M gallium nitrate. 59 Fe-Tf (6 g of Tf, 10,000 cpm of 59 Fe) was added to each well, and incubation was continued for 24 h. Following this, cell counts were determined, and the cells were removed from the wells and washed twice with ice-cold PBS. 59 Fe cpm in the cell pellet was counted using a Wallac Compugamma ␥-counter. In addition to measuring 59 Fe uptake by R cells that had been continuously grown in gallium-containing culture medium, 59 Fe uptake by R cells was also determined after these cells had been washed to remove gallium and cultured for 72 h in culture medium without additives.
Gallium Uptake Studies-To measure the cellular uptake of gallium, cells were plated in triplicate in fresh medium (5 ϫ 10 5 cells/ml) in 1-ml multiwell plates. Gallium nitrate (100 -1000 M) containing 67 Ga as a tracer (1 Ci of 67 Ga/1 mM gallium nitrate) was added to wells at the onset of incubation. After 48 and 72 h of incubation, cell counts were determined, and the cells were harvested and washed by centrifugation with PBS. 67 Ga radioactivity in the cell pellet was counted in a Wallac Compugamma ␥-counter, and the amount 67 Ga incorporated per 10 6 cells was determined.
125 I-Tf Binding Studies-Cell-surface Tf receptor density was determined by 125 I-Tf binding to intact cells. Confluent S and R CCRF-CEM cells were washed and reincubated in fresh medium without gallium nitrate. After 24 and 48 h of incubation, cells were harvested, washed with ice-cold PBS containing 0.1% bovine serum albumin, and assayed for 125 I-Tf binding at 37°C as described previously (40). In separate experiments, R cells that had been grown for 8 weeks in the absence of gallium were also assayed for 125 I-Tf binding. Maximal Tf binding and Tf receptor affinity for Tf were determined according to the method of Scatchard (41). (31), a gallium-resistant CCRF-CEM cell line was developed by us. To confirm that these cells continued to display a gallium-resistant phenotype, their growth in the presence of gallium was compared with that of S cells (the parent cell line). Based on the concentration of gallium required to inhibit cell growth by 50%, R cells were ϳ8-fold more resistant to growth inhibition by gallium nitrate than S cells (Fig. 1).

Demonstration of Resistance to Growth Inhibition by Gallium Nitrate-As previously reported
Ferritin Is Decreased in Gallium-resistant CCRF-CEM Cells-Since prior studies have shown that gallium can be incorporated into ferritin (7, 32), we first investigated whether gallium resistance could be due to an increase in ferritin content, leading to sequestration of intracellular gallium. S and R cells were therefore analyzed for ferritin content. However, as shown in Table I, R cells were found to contain significantly less ferritin than S cells. At stationary phase (confluency, 0 h), the ferritin content of R cells was ϳ27% that of S cells. Following 24 h of growth in fresh medium without gallium nitrate, ferritin in both S and R cells decreased significantly; however, the decrease in ferritin was greater in R cells than in S cells. As a result, after 24 h, the amount of ferritin in R cells was only 9% that in S cells. When the 24-h incubation was carried out in the presence of a concentration of gallium nitrate that did not inhibit cell growth, the amount of ferritin in R cells decreased, but remained almost 2-fold higher than that in cells incubated without gallium nitrate. In contrast, in the presence of gallium nitrate, the ferritin content of S cells increased rather than decreased over this period (Table I).
To determine whether R cells were able to increase their production of ferritin in response to iron, cells were incubated with ferric ammonium citrate for 24 h. Table I shows that 50 M ferric ammonium citrate produced 2.7-and 1.4-fold increases in ferritin in S and R cells, respectively.
Ferritin mRNA Levels in Gallium-sensitive and -resistant Cells-Steady-state H and L ferritin mRNA levels in S and R cells were examined to determine whether the differences in ferritin content were due to changes in mRNA. Northern blot analysis of mRNA, however, revealed that although R cells had markedly less ferritin protein than S cells, they displayed increased levels of H ferritin mRNA and equivalent levels of L ferritin mRNA (Fig. 2, upper panel, compare the band intensities of S and R cells incubated without gallium). Unexpectedly, incubation of cells with gallium induced a marked increased in H and L ferritin mRNAs in both R and S cells (Fig. 2, upper  panel). As illustrated further in the lower panel of Fig. 2, H ferritin mRNA levels in S cells increased in a dose-dependent manner following exposure to increasing concentrations of gallium. S cells incubated with 50 and 150 M gallium nitrate displayed 1.6-and 3.7-fold increases in H ferritin mRNA, respectively, relative to cells grown without gallium.

IRP-1 Activity Is Increased in Gallium-resistant Cells-
The above experiments suggested that although H and L ferritin mRNA levels were increased in R cells, their translation was inhibited. Since ferritin synthesis is regulated at the translational level by the binding of IRPs to H and L ferritin mRNAs, IRP binding activity in S and R cells was examined by bandshift assay. As shown in Fig. 3, IRP mRNA binding was increased ϳ2.4-fold in R cells incubated for 24 h in the absence of gallium (compare lanes 3 and 4 with lanes 1 and 2). Antibody to IRP-1 produced a supershift in these bands (lanes 9 and 10), indicating that they specifically represent IRP-1 binding to ferritin mRNA. Fig. 3 also shows that in the presence of gallium, IRP activity in both S and R cells was decreased; however, it still remained greater in R cells than in S cells.
To exclude the possibility that the differences in IRP-1 activity in R cells were due to changes in the amount of IRP-1 protein, cells were analyzed by Western blotting after a 24-h incubation with increasing concentrations of gallium nitrate (0 -250 M). In three separate experiments, S and R cells were found to contain comparable amounts of IRP-1, which did not change significantly following exposure of cells to gallium (data not shown).
Iron Uptake by Gallium-sensitive and -resistant CCRF-CEM Cells-Since prior studies showed that gallium inhibits the cellular uptake of iron (9), we investigated whether the differences in ferritin content between S and R cells could be explained by an effect of gallium on iron transport into cells. Cells were therefore examined for 59 Fe uptake in the presence and absence of gallium nitrate. As shown in Fig. 4, 59 Fe uptake by S cells was decreased by ϳ31% in the presence of 150 M gallium nitrate. In contrast, 59 Fe uptake by R cells was markedly decreased even in the absence of gallium nitrate and was ϳ60% lower than that by S cells. With the addition of gallium nitrate to the incubation, 59 Fe uptake by R cells was decreased further, albeit to a lesser extent than that seen with S cells.
Prior to being used in these experiments, R cells were washed extensively to remove gallium present in the stock cultures; however, to confirm that the decrease in iron uptake by R cells was not due to residual gallium in the system, 59 Fe uptake studies were also performed using R cells that had washed, incubated for 3 days in fresh medium without gallium, and then washed again. 59 Fe uptake by these cells (R 1 cells) was identical to that seen with washed R cells taken directly from stock cultures (Fig. 4). Even after 12 days (three passages)  of culture of R cells in gallium-free medium, 59 Fe uptake by R cells remained only 40% that of S cells (data not shown). Hence, the decrease in iron incorporation into R cells was not a result of an inhibition of iron uptake by the presence of residual gallium.
Gallium Transport into Resistant Cells Is Decreased-To determine whether the mechanism of drug resistance to gallium involves changes in its transport into cells, gallium uptake by R cells was compared with that of S cells. To correlate gallium transport into cells with the cytotoxicity assays, gallium uptake by cells was examined after 48-and 72-h incubations in the presence of increasing concentrations of gallium nitrate with 67 Ga as a tracer. When incubated with 100 M gallium (a noncytotoxic concentration), both S and R cells incorporated similar amounts of gallium/cell over 48 and 72 h (Fig. 5). Incubation with higher concentrations of gallium nitrate resulted in increases in gallium uptake by both S and R cells; however, gallium incorporation into R cells was significantly lower and, in the presence of 1000 M gallium nitrate, was only 34 -36% that into S cells (Fig. 5).

125
I-Tf Binding Studies-125 I-Tf binding to S and R cells was examined to determine whether the decrease in gallium and iron uptake by R cells could be explained on the basis of a decrease in Tf receptors. As shown in Fig. 6A, 125 I-Tf binding to R cells was ϳ72 and 54% that of S cells after 24 and 48 h of incubation, respectively, in the absence of gallium. To examine whether this decrease in Tf receptors was related to a residual effect of exogenous gallium, 125 I-Tf binding to R cells was measured again after 8 weeks of growth of these cells in medium without gallium. Drug resistance to gallium was retained even after growth of these cells in the absence of gallium (data not shown). Scatchard analysis of 125 I-Tf binding to cell-surface Tf receptors (shown in Fig. 6B) revealed that maximal 125 I-Tf binding to R cells was ϳ53% that of S cells at 48 h without a significant change in Tf receptor affinity for Tf (K D ϭ 9.8 ϫ 10 Ϫ9 M for S cells and 6.71 ϫ 10 Ϫ9 M for R cells).
Transferrin Increases Gallium Uptake by R Cells and Restores Sensitivity to Gallium-The above experiments show that R cells have a decrease in gallium uptake and changes in transferrin receptor number. Since Tf is known to enhance gallium uptake by cells (6,7,9), further experiments were performed to determine whether increasing the amount of exogenous Tf would enhance gallium uptake and increase its cytotoxicity in R cells. As shown in Fig. 7, the presence of 1 mg/ml transferrin resulted in a marked increase in the uptake S and R cells were incubated for 24 h in the presence or absence of gallium nitrate. Total cellular RNA was isolated, electrophoresed on a formaldehyde-agarose (1%) gel, and transferred to a membrane. The membrane was hybridized sequentially to 32 P-labeled cDNAs for H ferritin, L ferritin, and ␤-actin. The membrane was stripped of the respective 32 P-labeled cDNA probes between each hybridization. Data shown are from a single membrane and are representative of six separate experiments. Lower panel, effect of gallium on H ferritin mRNA levels in gallium-sensitive cells. Cells were incubated for 24 h in the presence or absence of gallium nitrate. Total cellular RNA was isolated, electrophoresed on a formaldehyde-agarose (1%) gel, and transferred to a membrane. The membrane was hybridized to 32 P-labeled cDNA for H ferritin. Equal RNA loading in each well was confirmed by hybridization of the membrane to 32 P-labeled ␤-actin (not shown). Data shown are representative of three separate experiments. of gallium by R cells (upper panel) and a progressive inhibition of cell proliferation (lower panel). When added to R cells incubated with 1000 M gallium, Tf produced a 3.4-fold increase in gallium uptake and completely inhibited cell proliferation. DISCUSSION As an iron storage protein, ferritin plays an important role in protecting cells from the toxicity of excessive intracellular iron. In addition to sequestering iron, however, ferritin can also bind gallium (7,32). This latter property of ferritin prompted us to initially investigate whether the development of drug resistance to gallium was associated with changes in ferritin gene expression. Our studies revealed that whereas ferritin mRNA levels were increased in R cells, ferritin protein content was markedly diminished. Further investigation demonstrated that IRP-1 RNA binding activity was increased, thus suggesting that the decrease in ferritin production in R cells was the result of a repression of ferritin mRNA translation. IRP-1 protein levels were equivalent in R and S cells, indicating that the increase in IRP-1 binding activity in R cells was due to a greater proportion of IRP-1 existing in a high affinity mRNA binding state. Since IRP/RNA interactions are directly influenced by changes in cellular iron, iron uptake studies were carried out that showed that 59 Fe incorporation into R cells was Ͻ50% that into S cells even in the absence of gallium. Hence, it appears that the switch in IRP-1 from a low to a high affinity mRNA binding state in R cells is due to a down-regulation of iron uptake and a depletion of an intracellular iron "pool" responsible for influencing IRP binding activity.
Whereas the inhibitory effect of gallium on iron uptake by S and R CCRF-CEM cells was consistent with earlier results in other gallium-sensitive cell lines (9), an unexpected finding was that iron uptake by R cells was consistently less than that by S cells even after extended growth of R cells in the absence of gallium. Since earlier studies have shown that cells incorporate iron and gallium by similar Tf-dependent and -independent transport systems (6 -8), we questioned whether the uptake of gallium by R cells was also decreased. Gallium uptake studies were therefore performed that showed that when challenged with increasing concentrations of gallium nitrate, R cells incorporated significantly less gallium than S cells. Hence, it appears that R cells have a decrease in the activity of a metal uptake transport system that affects both gallium and iron. This decrease in gallium uptake would serve to protect cells from the cytotoxicity of gallium, while the parallel decrease in iron uptake would lead to changes in intracellular iron homeostasis.
While the gallium uptake studies strongly suggest that the development of gallium resistance is primarily due to a downregulation of gallium transport into cells, the specific transport mechanisms involved remain to be fully elucidated. One explanation for the decrease in gallium uptake is that the lower number of Tf receptors expressed on R cells during proliferation may serve to limit transferrin-mediated uptake of Tfgallium complexes (formed by the binding of gallium to Tf in the culture medium). Alternatively, the decrease in gallium uptake by R cells may be secondary to a decrease in the activity of a transferrin-independent gallium/iron transport system. It is conceivable that both transport systems work in concert to decrease the uptake of gallium by R cells. Regardless of the membrane transport system involved, our studies indicate that the decrease in gallium uptake is central to the mechanism of drug resistance to gallium. Further evidence for this mechanism is provided by the demonstration that the sensitivity of R cells to gallium could be restored by increasing their uptake of gallium with excess Tf (Fig. 7). Increasing the amount of Tf favors the formation of Tf-gallium complexes and stimulates Tf receptor-mediated uptake of gallium, thereby overriding the basal decrease in gallium transport into R cells. Studies are in progress to elucidate how Tf-dependent and -independent gallium transport pathways are regulated in R cells and to determine their relative roles in gallium resistance.
Although the decrease in ferritin in R cells appears to be secondary to the decrease in iron uptake, the effect of gallium on ferritin gene expression appears to be complex. Since gal-lium inhibited iron uptake by both S and R cells, cells incubated with gallium would be expected to contain less ferritin than cells incubated without gallium. Instead, cells incubated with gallium contained more ferritin mRNA, lower IRP-1 binding activity, and more ferritin protein than cells incubated without this metal. These results suggest that gallium may affect ferritin expression through mechanisms that are independent of its inhibitory action on cellular iron uptake. Although there is no evidence that gallium interacts directly with IRP-1 to alter its mRNA binding activity, Cowley et al. (42) have reported the synthesis of a cubic gallium-sulfur cluster with gallium in place of iron. This raises the intriguing possibility that intracellular gallium may be capable of interacting with the iron-sulfur cluster of IRP-1 to produce a change in its affinity for the IREs. However, an alternative explanation for the gallium-induced increase in ferritin may be that intracellular gallium displaces iron from its binding to different ligands/macromolecules, thereby leading to an increase in an iron pool, which, in turn, produces a decrease in IRP activity and an increase in ferritin synthesis.
While a direct interaction of gallium with IRP-1 remains speculative at this time, our studies clearly demonstrate that gallium, in a dose-dependent manner, increases H ferritin mRNA and, to a lesser degree, L ferritin mRNA. The mechanism for this increase is unknown and is under investigation. Several studies have shown that ferritin gene expression can be induced by a number of stimuli that are unrelated to iron. Cytokines such as interleukin-1␣, tumor necrosis factor-␣, and interferon-␥ induce H ferritin mRNA transcription and increase H ferritin synthesis in cells (43,44), whereas interleukin-1␤ increases the synthesis of both H and L ferritin by increasing mRNA translation (45). More recently, thyroid hormone has been shown to increase ferritin synthesis by modulating the interaction between IRP and ferritin mRNA (46). Hence, it is conceivable that gallium may influence ferritin mRNA transcription and IRP activity through mechanisms that are independent of iron metabolism.
In conclusion, these studies are the first to show that the development of drug resistance to gallium involves a diminution in the transport of gallium into cells and that this alteration in gallium transport also affects iron transport and ferritin gene expression. Continued investigation of the interaction of gallium with biological systems may enhance our understanding of the mechanisms of its transport, cytotoxicity, and drug resistance and may allow us to increase its efficacy as an antineoplastic agent.