Effect of microbial siderophores on mammalian non-malignant and malignant cell lines

Background Iron is a vital nutrient for all cells, and malignant cells have a higher requirement for the metal due to their rapid multiplication. Bacterial siderophores can be used to reduce free ferric ion concentration from the cellular environment. Methods In the present study, we have evaluated effect of three siderophores – exochelin-MS, mycobactin S and deferoxamine B on the proliferation of mammalian cell lines using MTT assay. Results These siderophores caused a significant decrease in the viability of malignant cells, without significantly affecting non-malignant cells. Conclusions Based on these results, we suggest that iron-chelation therapy could be explored as an adjunctive therapeutic option against cancer along with other therapies.


Background
Iron is a vital nutrient for all cells. In eukaryotes, the highly iron-dependent activity of ribonucleotide reductase is essential for cell division [1]. Under iron-deficient conditions, cells cannot transit from G1-phase to S-phase in the cell cycle. Neoplastic cells have a higher requirement for iron than normal cells as they proliferate at a faster rate. Therefore, iron depletion by iron-chelating agents may deprive the rapidly dividing malignant cells of the DNA precursors required for replication, resulting in the inhibition of their proliferation [2,3].
Bacterial iron chelators termed siderophores can be used for reducing free iron concentration in the cellular environment. Growing microbial cells in iron-deficient laboratory growth media allows isolation and characterization of the siderophores.
Exochelin-MS (Exo-MS) is a water-soluble siderophore and mycobactin S (MBS) is a lipid-soluble siderophore produced by Mycobacterium smegmatis. Water-soluble deferoxamine mesylate B (DFO-B), originally from Streptomyces pilosus, is now commercially available.
Effect of Exo-MS, MBS and DFO-B were evaluated individually on malignant and non-malignant cell lines of murine and human origin. The concept tested here was whether iron deprivation by microbial siderophores affects multiplication of mammalian cells. This was assessed in vitro by MTT assay.

Methods
For the production of siderophores from M. smegmatis mc 2 155, iron-deficient minimal medium was used [4,5]. The culture of M. smegmatis mc 2 155 was inoculated and incubated at 37°C for 10 days on the shaker at 200 rpm for Exo-MS production; whereas, for the production of MBS, the culture flasks were incubated at 37°C for 10 days under static conditions.

Extraction, detection and purification of Exo-MS (a) Extraction of Exo-MS:
The10-day old culture broth of M. smegmatis was centrifuged at 10,000 rpm for 10 min to pellet the cells. The supernatant, in batches of 100 mL was freeze-dried, subsequently redissolved in 10 mL of distilled water and the pH was adjusted to 3.5. Then, it was saturated with ammonium sulfate (76 g/100 mL), and kept at 4°C overnight. Next day, the solution was centrifuged and the supernatant was transferred to a 50 mL tube and treated with 2.0 mL of benzyl alcohol. The aqueous and the solvent phases were mixed vigorously, centrifuged and the upper phase of benzyl alcohol fraction was collected in a clean test tube. This extraction step was repeated three times and all the benzyl alcohol fractions were pooled [6]. The pooled extract was mixed with 8.0 mL of diethyl ether for partitioning the water soluble Exo-MS. Exo-MS was extracted with 1.0 mL of distilled water. The solvent-water mixture was centrifuged, and the lower aqueous phase was transferred to a clean tube. This step too was repeated thrice with 1.0 mL distilled water each time. All the aqueous fractions were pooled and washed with diethylether to remove traces of benzyl alcohol. This aqueous Exo-MS preparation was concentrated by freeze-drying which also removed any residual diethyl ether.

(b)Purification and Detection of Exo-MS:
The Exo-MS extract was purified on a column of neutral grade alumina. When the column was eluted with petroleum spirit (bp = 60 to 80°C), the Exo-MS was firmly bound to the alumina, whereas the fastermigrating compounds were washed away with no retardation. Hydrophobic impurities were removed from the alumina-bound Exo-MS by sequential washes with cyclohexane, toluene, diethyl ether, ethyl acetate, and chloroform; all the washes were discarded. The adsorbed Exo-MS was then eluted from the column with the methanol-formic acid (4:1, v/v) mixture [7]. The eluate was concentrated by freeze-drying and neutralized to pH 6.5. The concentration of this Exo-MS preparation was determined by UV spectrophotometer.
The Exo-MS preparation was also subjected to highresolution liquid chromatography/mass spectrometry (HR-LC/MS) using Agilent Technologies Q-TOF Mass Spectrometer system at Sophisticated Analytical Instrument Facility, Indian Institute of Technology Bombay, India. The solvent system was 95 mL of 0.1% of Trifluoroacetic acid (TFA) in water with 5 mL of 0.1% TFA in 90% Acetonitrile (ACN) for 22 min. Then, 5 mL of 0.1% of Trifluoroacetic acid (TFA) in water with 95 mL of 0.1% TFA in 90% Acetonitrile (ACN) for 9 min. This was followed by the first solvent mixture for 5 min. The flow-rate was 0.4 mL/min for 35 min with maximum pressure of 1200 bar.
Two dimensional nuclear magnetic resonance -total correlation spectroscopy (2D-NMR-TOCSY) was carried out at National Facility for High-Field NMR, Tata Institute of Fundamental Research, Mumbai, India. The solvent system was 10% D 2 O and 90% water, which was run for 2 h on 500 MHz instrument. The 10-day old culture broth of M. smegmatis mc 2 155 was centrifuged and the cell pellet was transferred to a flask containing 100 mL of ethanol and kept overnight in the refrigerator. Next day, this extract was mixed with equal volume of chloroform. Sufficient water was added to this mixture to form two layers. The upper aqueous layer was discarded, and the lower chloroform layer containing MBS was washed with water, dehydrated over anhydrous Na 2 SO 4 and evaporated to dryness. The dried residue was extracted with methanol at room temperature. The methanol-soluble MBS was thus separated from the other insoluble impurities [8].

(b)Purification and Detection of MBS:
Methanol-soluble MBS was evaporated to dryness and then dissolved in a minimum amount of tertiary butanol : cyclohexane (1:1, v/v) ratio. After dissolving the residue, cyclohexane was added further to make up the volume to 100 mL. This solution was then adsorbed onto neutral grade alumina in a beaker and washed with 50 mL of cyclohexane. The wash was discarded. The adsorbed material was then sequentially washed with 50 mL each of petroleum ether, toluene and diethyl ether, which were also discarded. Finally, MBS was eluted using chloroform: acetone mixture (1:1, v/v) [9]. This eluate was evaporated to dryness again, dissolved in absolute ethanol and subjected to Sephadex LH 20 column chromatography. MBS was eluted using absolute ethanol. UV spectrophotometer was used to determine the concentration of this MBS preparation.
Reverse-phase HPLC was used to confirm the purity of MBS after adding 10 μL of 10% ethanolic ferric chloride to 100 μL of the MBS preparation to get ferri-MBS, which was carried out on Waters Xterra MS MS system at Bombay College of Pharmacy, Mumbai, India. The gradient solvent system used for HPLC was methanol: water (80:20, v/v) which was passed through a C 18 column for 30 min and then with 100% methanol for the next 15 min. The column used was, C 18 , 5 μm, 4.6 × 100 mm at the flow-rate of 1.5 mL per min. The eluted peaks were determined by monitoring the absorbance of the sample at A 220nm and A 450nm .

Purity analysis of DFO-B
DFO-B manufactured by Novartis, Switzerland, and obtained as 500 mg dry powder was dissolved in 2 mL sterile water at a concentration of 250 mg /mL. For HPLC, ferrioxamine-B was prepared by mixing 10 μL of 10% aqueous ferric chloride to 100 μL of 1.0 mg/mL DFO-B. Gradient solvent system used was 0.1% Trifluoroacetic acid (TFA) in water for 10 min, then equal volumes of 0.1% aqueous TFA and acetonitrile (ACN) (1:1, v/v) for 30 min and 0.1% TFA in water for 10 min. The column used was C18, 5 μm, 4.6 × 100 mm, at a flow-rate of 1.5 mL per min. The peaks were monitored by measuring the absorbance at 420 nm. This analysis was carried out on Waters Xterra MS MS system at Bombay College of Pharmacy, Mumbai, India.

Determination of the effect of Exo-MS, DFO-B, and MBS on mammalian cell lines
The effect of Exo-MS, DFO-B, and MBS was evaluated using malignant and non-malignant mammalian cell lines. A working stock solution of DFO-B at 100 mg/mL was prepared from 250 mg/mL stock solution before use. Exo-MS was dissolved in water and MBS in methanol for use.
The cell lines were procured from NCCS, Pune, India. They were maintained in T-25 flasks in Dulbecco's Modified Eagle's Medium (DMEM) with 10% Fetal Bovine serum (FBS) at 37°C in a CO 2 incubator with 5% CO 2 in humidified air.
Cell culture growth medium: DMEM, FBS and Trypsin-EDTA were obtained from Cell Clone (Genetix Biotech Asia, India).
Cell viability assay using MTT reagent: MTT {3-(4, 5dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide} assay was used to determine the viability of cells. This assay is a useful quantification method which measures metabolic activity of cells [10,11]. Metabolically active cells reduce yellow-colored tetrazolium and form purple crystals of formazan. These crystals upon solubilization with DMSO (dimethyl sulfoxide) were quantified by measuring the absorbance at 570 nm and 655 nm using a microplate reader.
All the cells used in the study were seeded in triplicates at a concentration of 6 × 10 4 cells/mL of complete DMEM in 12-well plates (total volume was 1.0 mL) and incubated for 24 h. Next day, different concentrations of each of the siderophores -Exo-MS, DFO-B, and MBS were added and incubated for another 24 h. Since MBS was prepared in methanol, methanol control was also maintained, where the cells were incubated with only methanol. The assay contained: Blank wells with only medium; Control wells with untreated cells; Control wells with cells treated with sodium azide; Test wells with cells treated with the siderophores.
Next day, from adherent cells, spent medium was discarded, and 1.0 mL of MTT reagent in DMEM was added to the wells and incubated at 37°C for 4 h. For the non-adherent cell line, the entire suspension was centrifuged in a 1.5 mL microfuge tube to pellet the cells. The supernatant was carefully poured out, and the cell pellet was then resuspended in 1.0 mL of MTT reagent, reloaded in the same wells and incubated for 4 h at 37°C. At the end of the incubation period, 1.0 mL of 100% DMSO was added to solubilize the visible formazan crystals. The plates were placed on a rocker for 20 min, and the absorbance of each well was measured at 570 nm and 655 nm using Epoch 2 (Biotek Instruments, USA) instrument.
MTT assays for every cell line were performed as three independent experiments, each in triplicate. Statistical analytical software -GraphPad Prism 7.01 and SigmaStat 4.0were used to compare the effect of Exo-MS, DFO-B and MBS on each cell line. The readings of control wells containing untreated cells were considered as 100% survival. The readings of control wells containing cells treated with 0.1% sodium azide were considered as no survival.

Purity analysis of siderophores
i) The HR-LC/MS of ferri-Exo-MS showed many peaks at 420 nm from 0 to 30 min ( Fig. 1a and Table 1). The major peak was between 24 and 25 min. The area under this curve was considered to be 100% and the relative concentrations of the other peaks were measured. The MS-MS results showed a significant peak corresponding to a molecular weight of 663 indicating presence of Exo-MS (Fig. 1b). The second major peak had a molecular weight of 326 indicating the presence of doubly charged Exo-MS. For 2D NMR-TOCSY of Exo-MS, the concentration used was 800 μg/mL. The analysis was done on a 500 MHz instrument with 10% D2O in water for 2 h. The chromatogram obtained showed the spin systems of threonine, β-alanine and γ1 β-ornithine (Fig. 2). ii) HPLC of ferri-MBS: HPLC analysis of ferri-MBS showed peaks in order of increasing acyl-chain length, which were visualized as seven peaks at 450 nm (Fig. 3a), when 10 μL of 4 mg/mL concentration of MBS was used. Three peaks were seen between 5 and 10 min, one major peak between 10 and 15 min along with a minor peak. The sixth and seventh peaks were seen between 20 and 25 min and between 35 and 40 min, respectively. The absorbance of ferri-MBS at 220 nm (Fig. 3b) was measured to know the extent of purity of the MBS preparation. The peaks at 220 nm were at the same positions as that seen at 450 nm. iii) The HPLC chromatogram of ferrioxamine B showed a single sharp peak between 15 and 20 min (Fig. 4). Gradient Trifluoroacetic acid-Acetonitrile (TFA-ACN) solvent system was used for analysis of 10 μL of 1.0 μg/mL concentration of ferrioxamine B.
To determine the concentration of purified siderophores, A 420nm of Exo-MS and A 450nm of MBS were measured using UV-VIS spectrophotometer. As per Ratledge and Ewing [12], the extinction coefficients are: Exo-MS: E 1cm 1% = 15.8 in water; MBS: E 1cm 1% = 43 in water; Using the above correlations, the concentrations of siderophores purified for use in these assays were Exo-MS = 19 mg/mL and MBS = 4 mg/mL.

Effect of Exo-MS, MBS, and DFO-B on mammalian cell lines
The MTT reagent exhibited negligible background absorbance values in blank wells without the cells. The absorbance values of untreated cells were considered as 100%. The absorbance reading from each well, (A 570nm -  . 2 The two-dimensional nuclear magnetic resonance spectroscopy 2D NMR-TOCSY of ferri-Exo-MS. Exo-MS at a concentration of 800 μg/mL was analyzed on a 500 MHz instrument with 10% D2O in water for 2 h. The spin systems of threonine, β-alanine and γ1 β-ornithine are seen A 655nm ) was calculated, and then the following formula was used to determine the percent survival; The intensity of purplish pink colour produced is directly proportional to the number of viable cells. The readings of control wells containing untreated cells which were considered as 100% survival showed a dark purple colour. The readings of control wells containing cells treated with 0.1% sodium azide showed either very pale purple colour or were colourless (data not shown).
Murine cell lines -RAW 264.7 and NIH/3 T3: A statistically significant decrease in the percent survival of RAW 264.7 was seen at Exo-MS concentrations of 0.2 mg/mL, where percentage of viable cells was 34% (Fig. 5a). No significant effect was seen on normal mouse fibroblast cell line NIH/3 T3 at Exo-MS concentrations up to 0.5 mg/mL. Percent cell survival was 100, 83 and 67% at 0.1 mg/mL, 0.2 mg/ mL and 0.5 mg/mL, respectively (Fig. 5b). Figure 5c shows the IC 50 of Exo-MS for RAW 264.7 which was 0.128 mg/mL.
A statistically significant decrease in the percent survival of RAW 264.7 was seen at DFO-B at Fig. 3 HPLC of pure ferri-mycobactin S. MBS at 4 mg/mL concentration was analyzed by injecting 10 μL volume in Waters Xterra MS MS HPLC system. The column used was C 18 , 5 μm, 4.6 × 100 mm at the flow-rate of 1.5 mL per min. Solvent system was methanol: water (80:20, v/v) for 30 min followed by 100% methanol for 15 min. (a) at 450 nm; (b) 220 nm. Seven peaks of different intensities were observed depending upon the acyl chain length that varies from 9 to 19 carbons; the shortest chain being eluted out first and the longest last concentrations up to 1.0 mg /mL, where percent cell survival was 38, 30 and 11% at 0.1 mg/mL, 0.2 mg/ mL and 1 mg/mL, respectively (Fig. 6a). No significant effect was seen on normal mouse fibroblast cell line NIH/3 T3 at DFO-B concentrations up to 1.0 mg/mL. Percent cell survival was 100, 94 and 87% at 0.1 mg/mL, 0.2 mg/mL and 1 mg/mL, respectively (Fig. 6b). Figure 6c shows that the IC 50 of DFO-B for RAW 264.7 which was 0.366 mg/mL.

Human Cell Lines -K 562, MCF-7, HEK 293
No statistically significant decrease in the percent survival of K 562 and MCF-7 at Exo-MS concentrations at 0.5 mg/mL and 1.0 mg/mL, respectively, was observed. For K 562, percent survival was 94 and 82%, respectively (Fig. 7a) and for MCF-7, it was 100 and 88.5%, respectively (Fig. 7b). Similarly, no significant decrease in the percent survival of HEK 293 at Exo-MS concentrations up to 1.0 mg/mL was observed. Percent cell survival at Exo-MS concentrations 0.2 mg/mL and 0.5 mg/mL for HEK 293 was 92 and 90%, respectively (Fig. 7c).
The percent viability of K 562, MCF-7, HEK 293 in the presence of DFO-B. For K 562, a statistically significant reduction in cell survival at DFO-B concentration of 0.5 mg/mL was observed, (P <0.01). Percent cell survival at DFO-B concentrations -0.5 mg/mL and 1 mg/ mL for K 562 was 55 and 54%, respectively (Fig. 8a).
There was a statistically significant reduction in the percent cell survival of breast cancer cell line MCF-7 at DFO-B concentration of 1.0 mg/mL, (P <0.01). Percent cell survival at DFO-B concentration of 1 mg/mL for MCF-7 was 62% (Fig. 8b).
There was no statistically significant decrease in the percent survival of HEK 293. Percent cell survival at DFO-B concentrations -0.5 mg/mL and 1 mg/mL for HEK 293 was 92 and 88%, respectively (Fig. 8c).
There was a statistically significant decrease in viability of HEPG2 cells at MBS concentrations from 10 μg/mL onwards (P <0.001). Percent cell survival at MBS concentrations of 5 μg/mL, 10 μg/mL and 20 μg/mL were 92.4, 48 and 23%, respectively (Fig. 9a). Methanol did not have any statistically significant effect on the cells since 95% survival of cells was observed (data not shown). Figure 9b and c show that there was no statistically significant decrease in percent survival of HEPG2 at the Exo-MS and DFO-B concentrations as high as 1 mg/mL and 10 mg/mL, respectively.

Discussion
HPLC and HR-LC/MS analysis confirmed the presence of Exo-MS with some impurities. NMR technique not being sensitive, low concentrations of Exo-MS could not allow a clear correlation in TOCSY, though threonine, β-alanine and γ 1 β-ornithine could be deciphered. The column used was C18, 5 µm, 4.6 x 100 mm, at a flow-rate of 1.5 mL per min. Solvent system was 0.1% trifluoroacetic acid (TFA) in water (v/v) for first 10 min, followed by equal volumes of 0.1% aqueous TFA and acetonitrile (1:1 v/v) for the next 30 min. And then 0.1% TFA in water (v/v) for the next 10 min, peaks monitored at 420nm. The chromatogram showed a single peak between 15-20 min HPLC analysis of ferri-MBS indicated about 98% purity. When the siderophore is complexed with iron, the red colored complex shows maximum absorbance at 450 nm, whereas the peptide bonds of MBS have an absorbance maximum at 220 nm. The peaks seen at A 220nm corresponded with the peaks observed at A 450 nm indicating pure MBS preparation. These were similar to the peaks obtained for MBS by Ratledge and Ewing [13].
Only one peak was obtained in the HPLC analysis of ferrioxamine-B indicating it was pure.
To be used for iron-depletion from microbial environments, siderophores need to be in a desferri form and not saturated with iron. Therefore, desferri form of siderophores were obtained using the isolation protocols for Exo-MS and MBS, but without the addition of iron during the entire isolation process. Iron was added to small aliquots only for determining purity and concentration of the siderophores.

Effect of Exo-MS, MBS, and DFO-B on mammalian cell lines
Non-malignant cells and malignant cells lines were used to study the effect of the iron-chelators on mammalian cells. The exposure with the siderophores was done for 18-24 h because the average time of division for the mammalian cells in vitro is 18-28 h. Besides, siderophores are not respiratory inhibitors, and are expected to affect proliferation of cells. When some pathogenic bacteria were treated with these siderophores for 18-24 h, they were inhibited (manuscript under preparation). This was also taken into consideration whilst subjecting mammalian cells to siderophores for 18-24 h.
The experiments with mammalian cells showed that concentrations of Exo-MS and DFO-B that significantly inhibited the murine cancer cell line RAW 264.7 were not inhibitory to the normal cell line NIH/3 T3. It should be noted that both Exo-MS and DFO-B had no significant effect on the non-malignant human cell line HEK 293 up to 0.5 mg/mL and 1 mg/mL, respectively.
DFO-B also inhibited the proliferation of human breast cancer (MCF-7) and human leukemia (K 562) cell lines. However, it had no effect on human liver cancer cell line (HEPG2) even at 10 mg/mL. Exo-MS did not inhibit any of the human cancer cell lines tested. DFO-B has already been shown to have anti-proliferative effect on cancer cells [14], and hence was used here as a control. When ferri-siderophores were evaluated for activity using the same MTT assay, no inhibitory effect on the cell proliferation was observed (data not shown). In this form, the siderophores were already saturated with iron, and therefore, could not deprive the cells of iron.
Earlier, some investigators have reported that a lipidsoluble siderophore from Mycobacterium tuberculosis (Mtb), mislabelled by the authors as (desferri-)exochelin 772SM, induces death by apoptosis in human breast cancer cells without harming normal breast epithelial cells [15]. However, the correct term for the siderophore referred to by the authors should have been 'Carboxymycobactin' , since water-soluble exochelin is not produced by Mtb.
Hoke et al. reported that DFO-B could be used as an adjunct to chemotherapy along with doxorubicin drug to inhibit breast tumor growth without any cardiotoxicity [16]. DFO-B was also found to be cytotoxic to malignant cells of neural origin [17,18]. When DFO-B was given during the initial stages of tumor formation, it resulted in either regression or slower tumor growth due to decrease in intracellular Fe 3+ concentration [19]. DFO-B was found to be inhibitory to leukemia cells in vivo as evident from the reduction in the cell numbers [20].
Since the water-soluble siderophores Exo-MS and DFO-B had no effect on HEPG2 even at high concentrations, effect of the lipid-soluble siderophore MBS was evaluated on HEPG2. MBS significantly decreased the  percent survival of HEPG2 cells at very low concentrations (10-20 μg/mL). This concentration is almost 50-1000 times lower than Exo-MS and DFO-B concentrations that inhibited RAW 264.7 cell line. Due to its lipophilic nature, MBS may traverse eukaryotic cell membrane. This lipid-soluble characteristic makes MBS much more efficient at lower concentrations in trapping intracellular iron than the hydrophilic siderophores. Since very low concentrations of MBS showed antiproliferative activity in vitro, it would be interesting to investigate its effects in vivo.
In the light of our results with the lipid-soluble MBS, conjugating water-soluble siderophores with lipid molecules may make them more effective against tumor cells. These siderophore conjugates could be evaluated as anticancer agents by themselves, or along with other therapies.

Conclusions
These investigations provide a "proof of concept" that iron chelation therapy may be useful against malignant cells without any significant cytotoxicity on nonmalignant cells. The anti-proliferative effect of siderophores on malignant cells needs further experimental and clinical studies to determine their role and utility in cancer treatment. The in-vivo toxicity of water-soluble siderophores may not be of concern, since DFO-B has already been approved by US-FDA for treating secondary iron-overload in patients suffering from thalassemia.