Dual Expression of Magnetosome-Associated Genes mms6 and magA to Enhance Magnetic Resonance Imaging Contrast in Mammalian Cells

Abstract

Reporter genes are employed to improve visualization of specific tissues by creating a higher contrast in MR imaging. The genes on the magnetosome gene island (MAI) in magnetotactic bacteria have been of interest as contrast agents due to their ability to facilitate the formation of magnetic nanoparticles. In this work, we evaluated the influence of co-expressing mms6 and magA- genes within MAI- in mammalian cells and their effect on transverse relaxation rate when cultured with and without iron supplementation. A transgenic colorectal adenocarcinoma (HT-29) cell line co-expressing mms6 and magA was produced by transgene delivery using Sleeping Beauty. The iron uptake was measured using an iron assay kit and viability of cells was determined by trypan blue and MTT assays. The cells were supplemented with ferric citrate and imaged using a 3T Siemens Prisma. Data from a multiecho sequence was used to calculate R₂ (R₂ = 1/T₂) values by monoexponential fitting. We evaluated the R₂ of the cells co-expressing the two genes with cell lines expressing either gene individually and non-expressing control cells. The result showed dual expression of mms6 and magA lead to a significant increase in R₂ and iron uptake compared to cells expressing either gene individually.

Introduction

Reporter genes that can be introduced into a cell and create a detectable enzyme or protein are an important tool in molecular imaging.^1,2 Reporter genes can be used to monitor cellular interactions, protein localization, and expression levels in various organisms. ³ In MRI, reporter genes generate contrast by modifying transverse or longitudinal relaxation rates^4,5 and have been employed to track structural or functional changes in tissues on a cellular level in vivo. ⁶ The initial MRI reporter genes were enzymatic reporter genes, which encode for an enzyme that produces a signal. For instance, Tyrosinase is a crucial enzyme in the melanin synthesis pathway, and overexpression of this gene will result in higher oxidation of Dioxyphenylaline (DOPA). DOPA converts to melanin^7–9 and can be imaged with MRI. ¹⁰ However, by binding to iron, melanin produces highly reactive oxygen species (ROS), which can be toxic to cells. ¹¹ Ferritin is an iron-based MR reporter gene that causes the cells to sequester endogenous iron without compromising cell viability. ¹² Multiple studies have shown that overexpression of ferritin can alter relaxivity,^13–15 though the effect is limited by iron availability and accumulation, yielding modest sensitivity.^16,17 The demand for more sensitive MR reporter genes has led to investigation of dual-modality reporter systems. ¹⁸ For instance, significant T₂ weighted contrast in MRI images was observed by coupling, LacZ gene which encodes for β-galactosidase enzyme, with monocrystalline iron oxide nanoparticles. ¹⁹ In Patrick et al., when the expression of the Oatp1a1 reporter gene is accompanied by hepatobillary contrast agent, a rapid and intense signal enhancement in T₁-weighted MR images of cells in vivo was reported. ²⁰ However, the drawback to these approaches is the need for injection of an exogenous contrast agent or substrate. In this work, we sought to investigate a dual gene approach that generates MR contrast endogenously.

Genes derived from magnetotactic bacteria (MTB) have been increasingly used as MRI reporter genes.^21–23 MTB produce magnetic iron oxide nanoparticles within a substructure known as magnetosomes.^24,25 Magnetosome formation in MTB is coordinated by a series of genes within the magnetosome gene island.^26–28 The magnetosome membrane specific 6 (mms6) gene is of interest due to its encoding a protein capable of binding to iron and playing an important role in regulating the shape of nano-sized magnetite crystal in biomineralization within MTB.^29–31 Another gene that contributes to the magnetic properties of MTB is magA, which is not directly associated with magnetosome formation ³² but encodes as an iron transporter across the magnetosome membrane.^33,34

Multiple studies have evaluated the capacity of mms6 and magA individually to act as MRI reporter genes. Zurkiya et al. engineered mammalian cells that can express the magA gene and found cells expressing the magA gene exhibited a significantly higher T₂ contrast. ³⁵ Another study investigated the ability of magA to produce endogenous contrast and observed significantly higher intracellular accumulation of iron and improved contrast with the presence of magA. ³⁶ Expression of mms6 has also been shown to enable non-invasive tracking of live cells and prior work has observed a 2.1-fold increase in R₂ of transgenic mammalian cells expressing the mms6 gene as compared to wild-type cells. ³⁷ Prior work also demonstrated the capability of mms6 to produce contrast in vivo without the need for exogenous supplementation of iron. ³⁷

In this study, we engineered human colorectal cancer cells to stably express both magA and mms6. We assessed the iron accumulation and the MR contrast in vitro. We found that the co-expression construct with magA and mms6 genes resulted in a higher iron uptake than constructs encoding either gene individually. Further, co-expression of the genes resulted in a higher MR contrast when compared to individual expression of either gene. These results indicate that co-expression of magA and mms6 is a more robust approach for MRI reporter compared to either gene individually.

Materials and Methods

Cell Culture

Human colorectal adenocarcinoma (HT-29) cells are an adherent epithelial cell line isolated from a primary tumor derived from a 44-year-old female patient (ATCC HTB-38, Manassas, VA). ³⁸ HT-29 cells were cultured in McCoy's 5A (Modified) Medium (ThermoFisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (ThermoFisher Scientific, Waltham, MA), penicillin-streptomycin (100 U/mL-100 µg/mL) (Sigma-Aldrich, St. Louis, MO). Cells were routinely cultured in 90 mm cell culture plates containing 8 mL of cell culture medium and incubated at 37 °C and 5% CO₂.

mms6 and magA Expression Vector and Viral Transduction

The expression vector for mms6 was modified to fit the mammalian expression system (NC_007626.1). ³⁹ The pTwist_mms6 vector was created by inserting the mms6 gene into pTwist-Lenti-SFFV-Puro plasmid and synthesized by Twist Bioscience (South San Francisco, CA). The pTwist_mms6 vector was co-transfected using the TransIT-Lenti Transfection Reagent (Mirus Bio, Madison, WI) with the lentiviral packaging plasmid (psPAX2, Addgene) and the envelope plasmid (pMD2.G, Addgene) in HEK 293TN cells (System Biosciences, LV900A-1). The viral media was harvested at 72 h post-transfection and used to transduce HT-29 parental cells (HT-29_WT). pTwist_mms6 transduced cells (HT-29_mms6) were incubated at 37 °C and 5% CO₂ for 48 h. To establish a pure and stable population, HT-29_mms6 cells were routinely passaged in fresh culture medium supplemented with 1.5 µg/ml puromycin (InvivoGen, San Diego, CA) for three weeks. The same protocol was utilized for the magA gene to establish the HT-29_MagA cell line.

Design of the mms6/magA Dual Expression Vector

To achieve dual expression of our target genes of interest we utilized two minimal CMV promoters in the reverse and forward orientations flanking a CMV enhancer, matching the sequence at 3069-574 of pBI-CMV1. ⁴⁰ The bidirectional CMV promoter constitutively expresses the mms6 gene ⁴¹ in the forward direction (terminated by rbGlob poly(A) signal) and magA gene^42,43 in the reverse direction (terminated by SV40 late poly(A) signal). For visual confirmation and selection of transgene expression, we included an mCherry reporter fused with a puromycin resistance gene via 2A peptide (mCherry-T2A-puroR) driven by the EF1α core promoter with WPRE sequence directly downstream, terminated by the BGH poly(A) signal. This custom gene sequence was inserted within a Sleeping Beauty transposon backbone plasmid and synthesized (pSB; VectorBuilder Inc., Chicago, IL). The resulting plasmid (pSB-magA-BiCMV-mms6-EFS-mCherry-T2A-PuroR) is 9233 bp with a 6323 bp transposable segment (Figure S1).

Multitransgenic Cell Line Generation

HT-29 colorectal adenocarcinoma (ATCC, Gaithersburg, MD) cells were cultured in McCoy's 5A (Modified) Medium (ThermoFisher Scientific, Waltham, MA) supplemented with fetal bovine serum (10%), penicillin-streptomycin (100 U/mL-100 µg/mL), and 1 µg/mL amphotericin B (ThermoFisher Scientific, Waltham, MA) incubated at 37 °C and 5% CO₂. Transgene delivery was accomplished using Sleeping Beauty as previously described.^44,45 Briefly, cells were cultured to a 90% confluent population in a 6-well plate and placed in fresh culture medium prior to transfection. Cells were double transfected with the pSB-magA-BiCMV-mms6-EFS-mCherry-T2A-Puro plasmid DNA and SB100 transposase mRNA (Genscript, Piscataway, NJ) in a 5:1 DNA/RNA mass ratio using Endofectin Max transfection reagent (GeneCopoeia, Rockville, MD). Transfected cells were incubated for 48 h at 37 °C and 5% CO₂ before passage to a 90 mm plate and selection with 2 µg/mL puromycin. Cells were again passaged 5 days post-transfection and the puromycin concentration was reduced to 1 µg/mL. Cells were selected for two more weeks to generate a stable transgenic cell population.

Viral and non-viral integrating vectors such as lentivirus and hyperactive transposase variants of the Sleeping Beauty transposon system share several similar features. These include the ability to integrate transgenes into genomic DNA of a broad range of cell types at high efficiency achieving stable long-term transgene expression.^44–49 Both methods have advantages and disadvantages. Lentivirus has demonstrated more efficient gene transfer but is limited by the size of its genetic cargo (∼10 kb),^49–51 whereas Sleeping Beauty typically has lower efficiency but can deliver genetic material greater than 100 kb (although efficiency at this size is significantly reduced). ⁴⁸ Further, lentiviral integration of transgenes is biased towards transcribed regions, increasing the potential for endogenous gene disruption, although the impact is still being studied.^52–54 In comparison, Sleeping Beauty has a near random integration profile targeting TA dinucleotides resulting in increased insertion at genomic safe harbor sites, leading to reduced potential for insertional mutagenesis and oncogene activation.^55,56 Both methods provide a reasonable approach for generating stable transgenic cell lines and have been utilized in this research. Specifically, a second generation lentiviral system was used to generate the HT-29 cells independently expressing mms6 or magA and a Sleeping Beauty transposon system with hyperactive (SB100) transposase mRNA was utilized to generate the HT-29 cells co-expressing mms6 and magA.

Fluorescent Microscopy

Cells were imaged using a DMi8 inverted microscope and DFC9000 GT sCMOS camera (Leica Microsystems, Chicago, IL). Phase contrast was used for cell morphology and mCherry expression was imaged using a Texas Red filter set.

Quantitative Real-Time PCR (qPCR)

A total of 1 × 10⁷ HT-29_mms6, HT-29_MagA and HT-29_Bi-CMV cells were trypsinized and collected. The RNA was isolated from the cell pellets separately using a lysis buffer (ThermoScientific). The RNA was then purified using the GeneJET RNA Purification Kit (ThermoScientific, Waltham, MA) following manufacturer's instructions. Amplification and quantification of cDNA were performed in triplicates using the SensiFast SYBR Lo-ROX kit (Bioline, London, England) on the mic qPCR cycler machine (Bio Molecular Systems, Australia). The polymerase was activated at 95 °C for 2 min followed by 40 cycles of 5 s at 95 °C for denaturation, 10 s at 60 °C for annealing and 5 s at 72 °C for extension. The primers designed for the mms6 gene were as follows: Forward, 5’-GTGGTTGGCGGCACTATTTG-3’ and Reverse, 5’-CTCCGACCACTCCCAGGATA-3’. The primers designed for magA gene were as follows: Forward, 5’-AGATCGGCGAGTTCTCGTTC-3’ and Reverse, 5’-AAATTCCAGAGCCAGGTCCG-3’. Ribosomal Protein Lateral Stalk Subunit P0 (RPLP0) primers were utilized for normalization control when quantifying the relative gene expression. The primers for Ribosomal Protein Lateral Stalk Subunit P0 (RPLP0) are: Forward, 5’-TGGTCATCCAGCAGGTGTTCGA-3’ and Reverse, 5’-ACACACACTGGCAACATTGCGG-3’. The gene expression levels were then compared quantitatively using an improvement method of the 2^−ΔΔCT method. ⁵⁷ The PCR products were analyzed by running a 1% gel electrophoresis. The agarose gel was imaged using the ChemiDoc imaging system (Bio-Rad, Hercules, CA).

Cytotoxicity Assay

Cell growth was determined by counting the number of viable cells using trypan blue (Invitrogen, Waltham, MA). Following 1–4 days of incubation, HT-29_Bi-CMV and HT-29_WT cells were trypsinized and resuspended in 1:1 dilution of trypan blue and culture media. The cells were then incubated for 3 min at room temperature and loaded into a hemocytometer. The cell growth was measured by counting the number of unstained cells at time point.

Metabolic activity was determined by MTT assay. A total of 1 × 10⁶ cells of HT-29_Bi-CMV and HT-29_WT (non-transfected parental cell line that serve as the control group) were seeded and cultured under two conditions: medium supplemented with 200 µM ferric citrate (Sigma Aldrich, St. Louis, MO) and medium without iron supplementation, each in triplicates. Following 1–4 days of incubation, the experimental media was replaced with 500 µL of fresh media that contains 10% (v/v) MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution. The cells were then incubated at 37 °C for 5 h. Each well is then carefully aspirated and 500 µL of DMSO is added to wells and mixed with pipetting, well plate is covered in foil and mixed in an orbital shaker at 125 RPM and 37 °C for 15 min to allow for formazan to fully dissolve. Four technical replicates of each sample are then measured at 570 nm in the SpectraMax plate reader. Absorbance values from iron-treated samples were normalized to the corresponding non-treated samples within each cell type at each time point. The data are presented as mean ± standard deviation (SD).

Iron Uptake Measurement

HT-29_mms6, HT-29_MagA, HT-29_Bi-CMV and the wild-type cells (HT-29_WT) were seeded in 60 mm plates containing 5 ml of fresh media supplemented with 0, 25, 50, 100, 200, or 400 µM of iron (ferric citrate) and incubated for 72 h. All conditions were tested in triplicates. Following the protocol for iron assay kit MAK025 (Sigma Aldrich, St. Louis, MO), a total of 2×10⁶ cells from each plate were harvested and counted using a hemocytometer. Cells were then rapidly homogenized using a Dounce tissue grinder and mixed in an iron assay buffer. The homogenate was then centrifuged at 16,000×g for 10 min at 4°C to remove insoluble material. The supernatant was mixed with 100 µL of iron Assay Buffer and transferred to a 96-well plate. A series of iron standards were prepared by adding 0, 2, 4, 6, 8, and 10 µL of the 1 mM standard solution into a 96 well plate. Iron Assay Buffer was added to each well to bring the total volume of standards to a 100 µL. The plate was incubated in a horizontal shaker for 30 min at 25 °C. 100 µL of Iron probe was added to each well containing standard and test samples and the plate was incubated at room temperature for 60 min and the 593 nm optical density was read by a SpectraMax M2 plate reader.

Absorbance values are corrected for background by subtracting the 0 (blank) value from each reading. A standard curve was established using the iron standard absorbance readings and their known concentrations. The slope of the standard curve was determined by applying linear regression fitting. The iron concentration of each sample was calculated using the formula and converted to pg/cell:

$$\mathrm{Iron}\mathrm{Concentration}={\displaystyle \frac{{\mathrm{OD}}_{\mathrm{Sample}}-{\mathrm{OD}}_{\mathrm{Blank}}}{\mathrm{Slope}}}$$

MRI of Cell Cultures

HT-29_mms6, HT-29_MagA, HT-29_Bi-CMV and HT-29_WT cells were seeded at the same density in 90 mm with 8 mL of media supplemented with 200 µM of ferric citrate and incubated for 72 h. Post incubation, cells were trypsinized and a total of 10⁷ cells were collected and resuspended in 1 mL of PBS without Calcium and Magnesium in 1.5 mL centrifuge tubes. The pellets of all the samples were allowed to settle by gravity at 4.4 °C for 3 h. Centrifugation can also be used to achieve the same results; however, a standard fixed angle rotor will result in a tilted pellet which we found not to be ideal for MR imaging. Imaging data were acquired using a 3T Siemens Prisma (Siemens Healthineers, Malvern, PA). Data from a multi-echo 2D spin echo sequence (echo time (TE) 1/ΔTE/TR =20/20/1500 ms, field of view = 120 mm×56 mm, in plane resolution = 0.47 × 0.47, slice thickness = 0.6 mm, 20 echoes) was used to calculate R₂ (R₂ = 1/T₂) values by fitting the signal curve to a monoexponential decay. Imaging data were processed in MATLAB (MathWorks, Natick, MA). Sample R₂ values are presented as mean ± SD.

Statistical Analysis

All statistical tests were performed using GraphPad Prism version 9.0.0. The threshold for statistical significance was set to p < 0.05. The expression level of each gene was assessed by measuring the number of cycles needed for the fluorescence to pass the set threshold level (C_q value) for each gene of interest subtracted by the Cq of the house keeping gene (denoted ΔCq). Two-tailed unpaired t-tests were used to compare ΔCq of mms6 and magA gene in their respective cell line to the expression in Bi-CMV dual gene cell line.

A two-way analysis of variance (ANOVA) was performed to determine if there is a difference between viability of HT-29_Bi-CMV cell line and HT-29_WT cell line over 4 days when supplemented with 200 µM of ferric citrate. We treated the viability of cell lines in each day as an independent comparison. Therefore, a post-hoc Sidak test was used to assess the viability of the cell lines with multiple comparison corrections for each day.

The difference between intracellular iron uptake for all four cell lines pre- and post-iron supplementation was determined using a two-way ANOVA. Pairwise comparison was performed using a Tukey post-hoc test. The difference between relaxation rate for all 4 cell lines pre- and post-iron supplementation was also determined using a two-way ANOVA. Pairwise comparison was performed using a Tukey post-hoc test.

Results

Consistent Expression Levels of Genes in Mammalian Cells

The presence of mms6 and magA genes in HT-29 cells was confirmed by performing quantitative PCR and running gel electrophoresis (Figure 1A). The expression levels of mms6 and magA were stable, as verified by examining the number of cycles needed for the fluorescence to pass the set threshold level (C_q value), which were consistent within each cell line. To address the potential for different levels of gene expression in each cell line, we utilized the improved 2^−ΔΔCT method ⁵⁷ and compared the relative fold change in expression levels of each gene in HT-29_Bi-CMV cells versus HT-29_MagA and HT-29_mms6 which was 0.9 and 0.97, respectively (Figure 1B), and no statistically significant difference was observed between gene expression levels across different cell lines (mms6: p > 0.99; magA: p > 0.99). To verify protein expression in cells co-expressing mms6 and magA, fluorescence microscopy was used to detect the red fluorescent protein mCherry, as seen in Figure 2. We can also observe that co-expression of mms6 and magA does not alter the morphology of the cells compared to wild-type cells.

Figure 1.

Analysis of magA and mms6 expression in HT-29 cells. (A) Gel electrophoresis (1% Agarose) demonstrates the presence of mms6 (97 bp) and magA (306 bp) in HT-29_Bi-CMV cells, RPLP0 used as housekeeping gene; the presence of mms6 and magA are also confirmed in their respective cell lines. (B) q-PCR analysis demonstrates similar expression of mms6 gene in HT-29_mms6 and HT-29_Bi-CMV cells and similar expression of magA gene HT-29_MagA and HT-29_Bi-CMV cells. The ΔC_q values are obtained by deducting the control gene's (RPLP0) C_q values from the C_q value of the target. It is expected for parental cells to have no expression of mms6 and magA, non-detection is indicated as 0 in this plot. Error bars indicate ± SD (n = 6).

Figure 2.

Fluorescence microscopy images show expression of mCherry reporter protein from pSB-magA-BiCMV-mms6-EFS-mCherry-T2A-Puro dual expression vector in the HT29_Bi-CMV cells under Phase (Left), mCherry (Middle) and Merge of both channels (Right). HT-29_WT cells were used as control to demonstrate expression.

Cell Viability

The effect of dual expression of mms6 and magA on cell growth and viability was assessed. Figure 3A shows the growth of each cell line when cultured under normal conditions. Two-way analysis of variance (ANOVA) revealed no significant difference in cell growth across cell types (p = 0.18; F = 1.8) demonstrating that dual expression of mms6 and magA does not affect cell growth. In Figure 3B, the metabolic activity of each cell line cultured in media + iron was normalized to that of respective cell line cultured in media only at the same time point. There is a slight decrease in normalized metabolic activity of both cell line when cultured in 200 µM ferric citrate which can be attributed to a decreased growth rate of cells when cultured in high concentrations of iron over multiple days. Sídak multiple comparison test revealed that the HT-29_Bi-CMV cells had a significantly higher metabolic activity than control cells on day 4 (p = 0.008), suggesting that the HT-29_Bi-CMV cell may be able to adapt better to high iron concentration environments. The incubation period for subsequent experiments was set at 3 days due to the observation that by day 4, metabolic activity of HT-29_WT cells began to decline significantly when cultured in 200 µM of ferric citrate (p < 0.001, from day 3 to day 4).

Figure 3.

Effect of concurrent expression of mms6 and magA on cell viability. (A) HT-29_Bi-CMV and HT-29_WT cells were cultured under normal conditions (37 °C and 5% CO₂) in McCoy's 5A (Modified) Medium with 10% fetal bovine serum and penicillin-streptomycin. The number of viable cells were counted with a hemocytometer using trypan blue stain. (B) Cells were cultured in culture medium containing 200 µM of ferric citrate for 1–4 days and an MTT assay was performed at each day. Absorbance values from iron-treated samples were normalized to the corresponding non-treated samples within each cell type at each time point (n = 12).

Increased Iron Uptake in Transgenic Cell Lines

Iron uptake was measured to investigate the effect of iron supplementation on each cell line. HT-29_Bi-CMV and HT-29_WT cells were cultured in medium with iron concentrations between 0 and 400 µM for 3 days. Figure 4A shows that iron uptake in both groups plateaued after 200 µM, which was therefore used in the subsequent experiments. Higher iron concentrations may lead to increased oxidative stress within the cells.^58,59 Iron uptake in HT-29_Bi-CMV, HT-29_MagA, HT-29_mms6 and HT-29_WT cells was assessed and iron uptake in each cell line is summarized in Table 1. Figure 4B shows the iron uptake of each cell line before and after iron supplementation. A significant main effect of iron supplementation was (p < 0.001; F = 291.4) was observed on iron uptake. Tukey's post-hoc comparison revealed that the presence of iron supplementation led to a significant increase in iron uptake across all cell lines (ps < 0.001). The three transgenic cell lines had a significantly higher iron increase compared to the wild-type cells (ps < 0.001). The iron uptake increased by 4.2-fold in HT-29_Bi-CMV cells when cultured in media with iron supplementation and these dual gene expression cells had higher uptake than HT-29_mms6 (p < 0.001) and HT-29_MagA (p < 0.001). Hence, HT-29_Bi-CMV cells exhibit a significantly greater ability to accumulate intracellular iron compared to the other three cell lines.

Figure 4.

Iron uptake analysis. All the cells were incubated with or without iron supplementation for 72 h and iron uptake of cells were determined by iron assay kit (A) Iron uptake within both HT-29_WT and HT-29_Bi-CMV cells plateaus after 200 µM of ferric citrate supplementation. (B) HT-29_Bi-CMV cells show a significant increase in iron uptake. Error bars indicate ± SD (n = 6, ****p < 0.001).

Table 1.

Iron Uptake in Each Cell Line.*

Cell Line (n = 6)	Without Supplementation [pg/cell]#	With Supplementation [pg/cell]
HT-29 WT	0.068 ± 0.003	0.184 ± 0.016
HT-29_mms6	0.124 ± 0.025	0.354 ± 0.029
HT-29_MagA	0.136 ± 0.021	0.446 ± 0.03
HT-29_Bi-CMV	0.146 ± 0.022	0.612 ± 0.032

* A total of 2×10⁶ cell were counted using a Hemocytometer.

^#Iron uptake is for each sample is measured in triplicates and reported as mean ± SD.

Transverse Relaxation is Increased Significantly in Dual Gene Cells

After incubating all cell lines in 200 µM ferric citrate for 3 days, a total of 10⁷ cells were trypsinized and resuspended in 1 ml of PBS, and the pellets were allowed to settle. Once the pellets settled by gravity, MR image data was collected. Figure 5A presents R₂ maps of the four cell lines, cultured with and without iron supplementation, respectively. HT-29_Bi-CMV cells produced visually detectable elevation in R₂ compared to the other three cell lines when supplemented with iron. Two-way ANOVA was used to evaluate the effect of iron supplementations on cell lines with different genes on mean R₂. A significant interaction was observed between iron supplementation and cell lines with different genes (p < 0.001, F = 403.6). Figure 5B shows a comparison of R₂ values before and after iron supplementation. R₂ values before and after iron supplementation increased by 157%, 60%, and 65% in the HT-29_Bi-CMV, HT-29_mms6, and HT-29_MagA, respectively. Tukey's multiple comparison test (Figure 5C) confirmed that these increases were statistically significant (ps < 0.001). However, the HT-29_WT cells did not experience a significant increase in R₂ when supplemented with ferric citrate (p = 0.36). The mean R₂s in Table 2 indicate that concurrent expression of mms6 and magA gene leads to a significant enhancement in relaxation rates in the presence of iron supplementation compared to the other 3 cell lines. The mean R₂ of HT-29_Bi-CMV cells exhibited a 1.73-fold increase (p < 0.001), 1.56-fold increase (p < 0.001), and 3-fold increase (p < 0.001) in comparison to the HT-29_mms6 cells, HT-29_MagA cells, and HT-29_WT cells, respectively. Interestingly, comparing HT-29_mms6 and HT-29_MagA, we did not observe a significant difference in R₂ with (p = 0.08) or without iron supplementation (p = 0.71).

Figure 5.

Magnetic resonance imaging of cell pellets with or without ferric citrate supplementation after 72 h of incubation. (A) T₂ weighted images taken from a CPMG sequence. The MR image shows that the dual gene expression leads to a visible MR contrast. (B) Result indicates a significant increase in R₂ value of HT-29_Bi-CMV cells with 200 µM of iron supplementation in comparison to any other group with iron supplementation. (C) Post-hoc Tukey pairwise multiple comparison of all four cell lines pre- and post- iron supplementation. Error bars indicate ± SD (n = 12, ****p < 0.001).

Table 2.

R₂ Values in Each Cell Line.*

Cell Line	No Iron Supplement [S−1]	With 200 µM of Iron Supplement [S−1]
HT-29 WT	2.476 ± 0.158	2.655 ± 0.289
HT-29_mms6	2.840 ± 0.263	4.563 ± 0.270
HT-29_MagA	3.046 ± 0.461	5.043 ± 0.426
HT-29_Bi-CMV	3.052 ± 0.402	7.875 ± 0.293

* Data from a multiecho 2D spin echo sequence was used to calculate R₂ (R₂ = 1/T₂) values by fitting the signal curve to a monoexponential decay.

Discussion

This study examined metabolic activity, iron uptake and R₂ of dual expression of two MR reporter genes derived from MTB, mms6 and magA. In addition, we compared the performance of the dually expressed MR reporter genes with individually expressed mms6 and magA genes in colorectal cancer cells (HT-29). There was no difference in cell growth between HT-29_Bi-CMV and HT-29_WT cells suggesting that the dual expression of mms6 and magA is well tolerated by the cells and does not induce detectable toxicity. Further, co-expression of mms6 and magA in cells resulted in more iron accumulation and higher R₂ values as compared to cells expressing a single MTB gene or to wild-type cells.

Iron-based MR reporters have been incorporated to improve visualization of tumors. Many studies have had success in elevating MR relaxation rates through overexpression of the ferritin heavy and light chains.^12–14,17 The potential of ferritin as a reporter gene was demonstrated by constructing an adenovirus vector encoding for the heavy and light chains of ferritin and expressing them in human lung adenocarcinoma (A549) cells which resulted in hypointensity in T₂ weighted MR images. ¹² One study was able to achieve a 1.2-fold increase in R₂ in vitro by overexpressing the heavy chain of ferritin in C6 rat glioma cells. This increase may be attributed to ferritin's ability to redistribute the existing intracellular iron that leads to an increase in net iron uptake. ¹⁴ While ferritin-based reporter genes hold great promise, the sensitivity of this system is dependent on the iron availability in the tissue which varies, and the time needed to accumulate enough iron for a detectable change in contrast. ¹⁶ The genes used in this work suffer from the same limitations that may affect the sensitivity.

An alternative approach has utilized the whole magnetoactic bacteria engineered into Magneto-endosymbionts (ME) capable of cell tracking in vivo. Human breast adenocarcinoma (231BR) cells labeled with MEs exhibited strong MR contrast both in vitro and in vivo, and the system was demonstrated to be safe for clinical applications. MR relaxation characterization of the cells revealed a very strong contrast both in vitro and in vivo. ⁶⁰ Unlike iron-based reporter genes, this approach is independent of the iron levels in the tissue. However, the long-term stability of this system remains an issue, as MEs are subject to digestion through the autophagy pathway. ⁶¹

Prior work demonstrated that R₂ is dependent on the gene expression level as well as the availability of free iron in transgenic kidney cells expressing magA. ³⁵ Therefore, we ensured that mms6 and magA expression levels were similar across all cell lines before evaluating the iron uptake and the R₂ values. An earlier study reported a significant increase in iron uptake in rat glioma cells expressing mms6 gene (9LS) compared to rat glioma wild-type (9L) cells under identical culture conditions, ³⁷ consistent with our findings. A 2.8-fold increase in iron uptake was observed in 9LS cells versus the 9L cells ³⁷ whereas we observed a 1.7-fold increase in iron uptake in HT-29_mms6 cells in comparison to wild-type HT-29 cells. A similar pattern was observed in cells expressing the magA gene where a significant increase in iron uptake was detected in previous studies. Earlier work expressing magA alongside a tetracycline response element in human kidney cells found a 4-fold increase in R₂ over cells not expressing magA. ³⁵ In another paper by Pereira et al. human embryonic kidney (HEK) cells expressing magA were supplemented with iron and a significant decrease in T₂ was seen compared to wild-type cells (p = 0.046). ³⁴ Here, we found that HT-29 cells expressing magA exhibited a 1.9-fold increase in R₂ over wild-type HT-29 cells. This discrepancy in iron uptake between the prior work and our study may be dependent on the host cell type, number of cells imaged, or duration of culture. Taken together, these findings confirm that mms6 and magA genes can be successfully expressed in mammalian cells and contribute to enhanced iron uptake.

The highest intracellular iron content was found in the HT-29_Bi-CMV cells. Prior work suggests that magA is involved in iron transport ³² but not involved in magnetosome formation in Magnetospirillum magneticum bacteria.^37,62 The formation of magnetite in mammalian cells expressing the magA gene and the increase in MR contrast, as reported in earlier studies, may be attributed to elevated intracellular iron concentrations.^37,62 Whereas mms6 could facilitate nucleation of iron into nanoparticles, alleviating the cells from build-up of excessive free iron in the cytoplasm that can ultimately lead to oxidative stress. ⁶³ Thus, we posit that the improvement in iron accumulation is a result of more efficient iron transport and crystallization when both MTB genes are expressed.

The intracellular iron content in the transgenic cell lines with no iron supplementation was similar to that of the wild-type cells when supplemented with 200 µM ferric citrate. However, no significant increase in R₂ was observed in wild-type cells with and without iron supplementation, which may be due to the iron concentrations being lower than the detection threshold. As compared to baseline iron uptake measurements, the HT-29_mms6 cells had a 2.85-fold increase in intracellular iron content whereas the HT-29_magA cells had a 3.28-fold increase. Despite the higher increase in intracellular iron content in HT-29_magA cells, no statistically significant difference in R₂ was observed between the two cell types (p = 0.3605). Since magA is an iron transporter gene (22), some of the iron transferred into the HT-29_magA may not be MR visible. In contrast, iron is crystalized by the mms6 gene, and this may generate higher MR contrast in cells containing mms6. In multi-transgenic cells, the higher relaxation rates observed may be attributed to the combination of magA's transporter abilities and mms6's iron crystallization thus creating more MR visible iron.

Future work can explore mms6 and magA dual gene expression in vivo, assessing their potential to accumulate endogenous iron naturally available in body tissue. This work validates the expression of mms6 and magA on a gene level. There is a possibility of translational or post-translational regulations that may affect the protein's interaction and function in the mammalian host cells. The protein expression and interaction of the genes can be investigated to strengthen the molecular characterization of the system. A recent study demonstrated that individual expression of other essential magnetosome-associated genes, including mamI, mamL, mamB, and mamE, in human melanoma cells (MDA-MB-435) led to increased relaxation rates. ⁶⁴ These findings highlight the potential of exploring expression of other combinations of magnetotactic genes in mammalian cells to further enhance MR contrast.

Conclusion

In summary, we successfully engineered HT-29 cells to co-express mms6 and magA transgenes. We evaluated the dual gene expression as an MR reporter and compared its performance to cells expressing each gene individually. The dual gene expression was found to work synergistically, resulting in higher iron uptake and increased transverse relaxivity in the cells expressing both genes compared to cells expressing either gene independently. Furthermore, the concurrent expression of both genes did not demonstrate any increase in cellular toxicity. The dual expression of mms6 and magA has the potential to be used as an MR reporter for imaging labeled cells.

Supplemental Material

sj-tiff-1-mix-10.1177_15353508261431711 - Supplemental material for Dual Expression of Magnetosome-Associated Genes mms6 and magA to Enhance Magnetic Resonance Imaging Contrast in Mammalian Cells

Supplemental material, sj-tiff-1-mix-10.1177_15353508261431711 for Dual Expression of Magnetosome-Associated Genes mms6 and magA to Enhance Magnetic Resonance Imaging Contrast in Mammalian Cells by Vala Masjedizadeh, MS, William Darch, MS, Jason Langley, PhD, Joshua T. Morgan, PhD and Xiaoping P. Hu, PhD in Molecular Imaging

Supplemental Material

sj-tiff-2-mix-10.1177_15353508261431711 - Supplemental material for Dual Expression of Magnetosome-Associated Genes mms6 and magA to Enhance Magnetic Resonance Imaging Contrast in Mammalian Cells

Supplemental material, sj-tiff-2-mix-10.1177_15353508261431711 for Dual Expression of Magnetosome-Associated Genes mms6 and magA to Enhance Magnetic Resonance Imaging Contrast in Mammalian Cells by Vala Masjedizadeh, MS, William Darch, MS, Jason Langley, PhD, Joshua T. Morgan, PhD and Xiaoping P. Hu, PhD in Molecular Imaging