Metallome of cerebrovascular endothelial cells infected with Toxoplasma gondii using µ-XRF imaging and inductively coupled plasma mass spectrometry

In this study, we measured the levels of elements in human brain microvascular endothelial cells (ECs) infected with T. gondii . ECs were infected with tachyzoites of the RH strain, and at 6, 24, and 48 hours post infection (hpi), the intracellular concentrations of elements were determined using a synchrotron-microfocus X-ray fluorescence microscopy (μ-XRF) system. This method enabled the quantification of the concentrations of Zn and Ca in infected and uninfected (control) ECs at the sub-micron spatial resolution. T. gondii -hosting ECs contained less Zn than uninfected cells only at 48 hpi (p < 0.01). The level of Ca was not significantly different between infected and control cells ( p > 0.05). Inductively Coupled Plasma Mass Spectrometry (ICP-MS) analysis revealed infection-specific metallome profiles characterized by significant increases in the intracellular levels of Zn, Fe, Mn and Cu at 48 hpi ( p < 0.01), and significant reductions in the extracellular concentrations of Co, Cu, Mo, V, and Ag at 24 hpi ( p < 0.05) compared with control cells. Zn constituted the largest part (74%) of the total metal composition (metallome) of the parasite. Gene expression analysis showed infection-specific upregulation in the expression of five genes, MT1JP, MT1M, MT1E, MT1F, and MT1X, belonging to the metallothionein gene family. These results point to a possible correlation between T. gondii infection and increased expression of MT1 isoforms and altered intracellular levels of elements, especially Zn and Fe. Taken together, a combined μ-XRF and ICP-MS approach is promising for studies of the role of elements in mediating host-parasite interaction. response to T. gondii infection, especially in tissues with high metabolic demand such as the CNS. This study establishes a novel use of synchrotron-microfocus X-ray fluorescence microscopy (μ-XRF) to determine changes in the cellular levels of Fe, Cu, Zn, and other elements in ECs challenged with T. gondii . μ-XRF is a chemical, non-invasive, element imaging technique, which can be used to generate x-ray fluorescent 2D elemental maps of biological samples, 11-12 with detection sensitivity and spatial resolution well-suited to characterize host-parasite interaction. μ-XRF imaging can enable in situ interrogation of the spatial distribution of one or more elements to submicron spot sizes and offers enough sensitivity and precision to detect metal distribution even at the single cell and subcellular 13 However, using method We therefore compared these results with those obtained from coupled plasma mass spectrometry currently gold-standard for determination absolute trace element and widely for elemental analysis of various tissues. 14


Introduction
Toxoplasma gondii is an obligate intracellular apicomplexan pathogen causing diseases by reiterating its lytic cycle, comprising host cell invasion, parasite replication and parasite egress. The ability of this parasite to invade host cells, disseminate through tissues and cause disease depends critically on its ability to locate and engage with the target host central nervous system (CNS). The blood-brain barrier (BBB) plays an important role as the biological interface that separates the host neural tissues from circulating blood. Human brain microvascular endothelial cells (ECs) constitute the fundamental component of the BBB, and together with basal lamina, astrocytic foot processes, pericytes, and tight junctions strictly regulate substance entry to neuronal tissue. [1][2][3] Overcoming this protective barrier is a fundamental step in the establishment of brain infection. 4 T. gondii invades and survives in host cells by modulating host cell processes and evading innate defenses, but the mechanisms are not fully defined. Metal dyshomeostasis during T. gondii infection implicates a significant role for metals in mediating parasite-host interaction. Fe, Cu and Zn are essential elements required for a multitude of cellular functions, such as enzymatic reactions, DNA synthesis, metabolic processes, and gene expression. 5 These metals play key roles in the host response to infection, including the development of the immune response and influencing the virulence of microorganisms. [6][7][8] The intracellular homeostasis of these elements is tightly regulated, and alteration of their levels can have adverse impacts on the host cell and its ability to respond to microbial infection. T. gondii infection has been shown to alter the levels of Zn, Fe, Mg and Cu in the blood of seropositive sheep 9 and humans, 10 compared to their seronegative counterparts.
Therefore, it is reasonable to hypothesize that host cell elemental content will change in response to T. gondii infection, especially in tissues with high metabolic demand such as the CNS.
This study establishes a novel use of synchrotron-microfocus X-ray fluorescence microscopy (μ-XRF) to determine changes in the cellular levels of Fe, Cu, Zn, and other elements in ECs challenged with T. gondii. μ-XRF is a chemical, non-invasive, element imaging technique, which can be used to generate x-ray fluorescent 2D elemental maps of biological samples, [11][12] with detection sensitivity and spatial resolution well-suited to characterize host-parasite interaction. μ-XRF imaging can enable in situ interrogation of the spatial distribution of one or more elements to submicron spot sizes and offers enough sensitivity and precision to detect metal distribution even at the single cell and subcellular levels. 13 However, absolute quantification of elements is not so straightforward using μ-XRF imaging, and in this respect the method is usually considered semi-quantitative. We therefore compared these results with those obtained from inductively coupled plasma mass spectrometry (ICP-MS), which is currently a gold-standard technique for the determination of absolute trace element concentrations and has been widely used for the elemental analysis of various tissues. 14

Parasite strain
Toxoplasma gondii genotype I (RH strain) tachyzoites were maintained by passage in Madin-Darby Canine Kidney (MDCK) cell cultures grown in complete Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM glutamine, and 1% antibiotic-antimycotic solution at 37°C under a humidified atmosphere of 5% CO2. The tachyzoites were purified from their feeder MDCK cell cultures by passage through PD-10 desalting columns filled with Sephadex, as described previously. 15 The purified parasites were centrifuged at 800×g, re-suspended in fresh medium, and quantified using a hemocytometer. The final volume of pure tachyzoite suspension was adjusted with fresh Roswell Park Memorial Institute (RPMI) 1640 medium.

Cell culture
Human brain microvascular endothelial cells (ECs) were used at passage 9 and were maintained in vitro, as described previously. 16,17 Briefly, ECs were propagated in tissueculture medium composed of RPMI 1640 medium supplemented with 20% (v/v) heat inactivated FBS, 2mM L-glutamine, 1mM sodium pyruvate, 1% non-essential amino acids, 1% MEM vitamins, and 2% penicillin/streptomycin (Invitrogen, GIBCO, UK). Cells were maintained in an incubator in a humidified atmosphere at 37°C and 5% CO2. ECs were trypsinized twice a week using trypsin-EDTA (Invitrogen, GIBCO, UK). Cells were examined daily under light microscopy and were considered confluent when their expansion had reached a point where cells touched each other on all sides, leaving no intercellular gaps.
To exclude the possibility that cell viability has influenced elemental concentration of the ECs or parasite interactions with the ECs and any subsequent measurement, the viability of the cells was assessed for a minimum of 100 cells using 0.15% trypan blue exclusion assay prior to use in any experiment.

Quantification of elements using ICP-MS
ECs were seeded in T-175 cm 2 tissue culture flasks at 10 10 cells/flask and grown in RPMI medium, as described above. Once a confluent cell monolayer was formed (~24 h), tachyzoites were added at a multiplicity of infection (MOI) of 2 (i.e., a host:parasite ratio of 1:2). The culture medium (500 μl) was collected 6, 24 and 48 hours post-infection (hpi) from the infected and non-infected (control) cultures and diluted to 10 ml (1:20 dilution) with 1% nitric acid, before running on the inductively coupled plasma mass spectrometry (ICP-MS; Varian Ultramass, Melbourne, VIC, Australia) system fitted with a direct injection nebulizer (CETAC, Omaha, NE, USA) using Rh (10 ng/l) as an internal standard in order to quantify the concentrations of the extracellular trace elements. At identical time points after infection (Supplementary file; Figure S1), infected and uninfected cells were harvested using a sterile cell scraper, followed by washing three times with deionized water. The cells (infected or uninfected) and purified tachyzoites were centrifuged for 5 min at 800xg. The cell or tachyzoite pellets were lyophilized by a Modulyo freeze-dryer (Thermo Savant, USA). These lyophilized pellets were digested for 1 h at ambient temperature using a solution containing 3 ml HNO3, 3 ml H2O, and 2 ml H2O2. The samples were heated in a microwave for 90 min.

Cell infection and sample preparation for μ-XRF imaging
Endothelial cells were seeded on quartz slides (UQG Ltd., Cambridge, England) until they formed a monolayer (~ 24 h). T. gondii tachyzoites were then added to the cell monolayer at a MOI of 2. Control samples included quartz slides that were also seeded with the same number of cells, but without addition of tachyzoites. At 6, 24 and 48 hpi, three slides from each of the infected and non-infected ECs were washed 3X in sterile phosphate buffered saline (PBS; 137 mM sodium chloride, 2.7 mM potassium chloride, 10 mM disodium phosphate, and 1.8 mM monopotassium phosphate; pH 7.4) and fixed in 4% paraformaldehyde in PBS for 30 minutes at ambient temperature. The slides were washed again 3X in PBS and left at 4ºC in deionized water (17.8 MΩ-cm) until μ-XRF imaging.

Microfocus X-ray fluorescence imaging of metals
The synchrotron microprobe of beamline I18 18 at Diamond Light Source, Oxfordshire, UK was utilized to map, and thereby quantify and compare, the individual distributions of Fe, Ca, and Zn within the ECs. Control samples were also analyzed to establish a baseline for the levels of these metals. Incubation times of 6, 24 and 48 hpi were measured for control and infected cells in order to evaluate the relative metal levels within the host cells at different stages after infection. The elemental distribution map of Fe could not be determined because the signals did not exceed the background level. It is likely that the ECs might have taken up Fe from the cell culture serum because ECs were cultured in media supplemented with animal serum. In contrast, the elemental distribution maps of Ca and Zn were successfully produced at a resolution of approximately 3 microns, over sample areas of 400 μm 2 . As the levels measured were known to be in the low parts per million (ppm) region, 4 sec/point dwell time was used for improved counting statistics. The elemental concentration was determined by measuring a reference material for μ-XRF measurements (AXO, Dresden GmbH) in the same conditions as the samples in order to deduce the photon flux on the sample. The reference material was composed of nm-thick layers of metals with known numbers of atoms per unit area. The resulting concentration maps were produced using Pymca, where the reference material and our sample were modelled in terms of main composition, density and thickness. 19 This analysis enabled fitting of the acquired spectra and translation of the fluorescence intensity maps for individual elements into estimated concentration distributions with an approximation to the composition of the cellular matrix.
The net peak areas from the sample spectra were translated into concentrations resulting in elemental maps in units of ppm (mg/g dry weight), which were then imported as tiff images in ImageJ for further statistical analysis. From the concentration maps regions of interest (ROIs) within cells were identified based on contrast to surrounding using the auto-threshold tool in ImageJ. From each ROI an average elemental value was extracted and plotted in scatter plots.
The dehydration of the samples did not change the quantity of each material element sampled by the beam, as hydrogen and oxygen were not considered here, and the X-ray beam penetrated the full thickness of the sample. However, the absolute concentrations were necessarily an estimate due to the limitations of matrix matching with the reference standards, and to the inevitable variation in the thickness profile of individual cells. Also, it was important to note that the examined cell matrix was not a homogeneous bulk sample. ECs were not uniform in shape, or volume, their diameter ranged from 35-50 μm, and the cell monolayer thickness might also vary. Therefore, considerable heterogeneity effects might have occurred, which can reduce the accuracy of estimated concentrations. The obtained values were tested for statistically significant differences between control and infected cells using Student's t test (two tailed). A p-value of < 0.05 was considered the cut-off for a significant difference.

Elemental analysis at a single cell level
We have developed a data-processing pipeline to estimate the elemental concentration in single cells (Fig. 1). To avoid issues of pile-up cells, we excluded spots where there was microscopic evidence of crowding and overlapping of cells. First, we aligned the bright field image ( Fig. 1a) to the μ-XRF fluorescent image (Fig. 1b) to ensure that the offset, scale, orientation and resolution of both images were completely matched. From the bright field image, ROIs (i.e. individual cells) were selected, and the contour of each individual cell was manually drawn (Fig. 1c). These outlined cells were used as masks to crop and extract the information only from the ROIs in the μ-XRF fluorescent images. The image annotated with the ROIs (Fig. 1c)

Gene expression analysis
We explored whether T. gondii infection induced metal dyshomeostasis through modulation of metal transporters or metallothionein. EC monolayers were harvested using trypsin and lysed with QIAshredder columns (Qiagen). Total RNA was isolated using an RNeasy Mini Kit (Qiagen), and eluted with nuclease-free water. RNA was stored at −80 °C for 1 week prior to microarray analysis. All subsequent sample handling, labelling and microarray Total RNA was first converted to cDNA, followed by in vivo transcription to make cRNA.
Single stranded cDNA was synthesized, end labeled and hybridized for 16 h at 45℃ to GeneChip TM Human Gene 2.1 ST arrays. All steps were performed by a GeneAtlas TM Personal Microarray system (Affymetrix). Gene expression data were analyzed using Partek Genomics Suite 7.0 software (Partek Incorporated). The raw CEL files were normalized using the RMA background correction with quantile normalization, log base 2 transformation and mean probe-set summarization with adjustment for GC content. Differentially expressed genes were considered significant if p-value with false discovery rate (FDR) was ≤ 0.05 and fold change of >2 or <-2.

Statistical analysis of the ICP-MS data
Differences in the Mass Spectrometry-based elemental concentrations between control and infected cells at different time points were assessed by two-way ANOVA and Tukey's multiple comparison test (*, p < 0.05; **, p < 0.01; ***, p < 0.001 ) using Graphpad Prism

Elemental levels in T. gondii -infected ECs using ICP-MS
The levels of elements with in culture medium and within ECs infected with T. gondii compared with uninfected cells were determined using ICP-MS. Notably, the majority of elements did not show significant differences between control and infected cells. However, our analysis identified patterns associated with T. gondii infection and detected significant differences (p < 0.001) between control and infected cell cultures (Fig. 2). We also noted that the intra-and extra-cellular measurements often differed by orders of magnitude. The The difference in elemental levels between control and infected cells was plotted against the corresponding -log (base 10) p-value for testing significant differences. As shown in the supplementary Fig.S2, the Cu inside and outside the cells was the only element that was located above the horizontal line that represents the threshold of significance. This finding, together with the result showed in Fig. 2, suggests that the level of Cu was significantly different between infected and control samples.
We also performed unsupervised hierarchical clustering of the element concentrations. In the constructed dendrograms, the samples were separated based on the mean concentration of elements in the control and infected samples at 6, 24 and 48 hpi. The heatmap provided a graphical display of the temporal changes in the concentration of each element, with one row representing the mean control (Con) and one row representing the mean infected (Inf) cells of each sample (Fig. 3). Results shown in the heatmap further point out the significant increase in the concentration of Zn (p < 0.01) and Fe (p < 0.05) at 48 hpi. Next, the dendrogram (tree) in the heatmap was divided to produce six different groups. The mean (± 1 standard deviation) of the individual replicates of the elements that constituted each group were plotted at 6, 24, and 48 hpi. Elements within each of the six cluster groups shared an overall unique pattern (Fig. 4).

Elemental content of purified parasite preparation
The elemental composition of T. gondii tachyzoites was also determined. Zn and Cu were the most abundant elements, representing 74% and 13% of the total parasite elemental content, respectively. Fe, Mn and P constituted 5%, 3% and 2%, respectively, whereas each of Na, S, and Rb accounted for ~1 % (Fig. 5).

Trace element levels in T. gondii-infected ECs using synchrotron μ-XRF
The micrometer resolution of synchrotron-based X-ray fluorescence microscopy (μ-XRF) enabled mapping and quantification of the level of elements in ECs infected with T. gondii.

Maps of Ca and Zn at 6, 24, and 48 hpi of infected and control cells are shown in
supplementary Figs. S3-4). Elemental imaging showed that Zn was significantly reduced only at 48 hpi compared with uninfected cells (p < 0.05), whereas Ca was increased at 48 hpi in infected cells, but the difference was not statistically significant (Fig. 6). Average concentration of the metal content inside single cells was calculated using a connectedcomponents labelling algorithm. The levels of Ca and Zn obtained using this image processing approach showed a similar trend to the results obtained based on the analysis of hot spots in the cell monolayer (Fig. 7).  (Fig. 9). These results were also confirmed by the heatmap and hierarchical clustering dendrogram (Fig. 10). Differential expression of MT1M and MT1JP genes was detected at 6, 24 and 48 hpi. Interestingly, MT1E, MT1F, MT1X genes showed lower gene expression at 6 hpi compared to MT1M and MT1JP. Gene expression increased gradually at 24 hpi for these three genes. The five genes showed similar gene expression after 48 hpi.

Gene expression microarray data
These results show that the expression of MT1 isoforms JP, E, F, and X increased significantly in cells infected with T. gondii compared to uninfected cells.

Discussion
The availability and restriction of elements are important aspects of the host-pathogen interaction. 7 In the present study, we determined the changes in the concentrations of 24 We also used synchrotron radiation-based μ-XRF with < 5 micron spatial resolution to assess the alterations in levels of Zn, Ca, Fe, and Cu distribution, as well as other elements in For example, 40% and 77% reduction in endogenous Fe and Zn, respectively has been observed in brain tissues fixed in formalin. 22 Substantial leaching of elements from the brain tissue into the formalin has been also detected, and the leaching varied considerably between different elements and was time-dependent. 23 Therefore, fixation of the infected ECs, for health and safety reasons, with 4% paraformaldehyde prior to μ-XRF imaging may have inevitably introduced an artificial effect and caused leaching of some elements from the cells into the fixative. It is therefore appropriate to cautiously interpret μ-XRF imaging-based metal data derived from fixed cells in the present study. Future studies could compare the elemental differences between fixed samples and fresh or frozen samples, and assess new ways of preparing the samples with less potentially-disruptive protocols for μ-XRF imaging analysis.

Microarray analysis of gene expression performed on infected and control cells using
Affymetrix GeneChip revealed infection-specific upregulation of five genes; MT1JP, MT1M, MT1E, MT1F, and MT1X. These genes belong to the metallothionein (MT) gene family, encoding small, cysteine-rich, heavy metal-binding proteins, which play an important role in the homeostasis of transition metals (e.g. Fe, Zn and Cu) and detoxification of non-essential trace elements, such as cadmium (Cd) and mercury (Hg) 24 , and cell proliferation. 25 MTs are also stress response proteins that are induced in response to triggers, such as oxidative stress, infection, inflammation, and heavy metals, 26 to protect against reactive oxygen species (ROS), 27 via their free radical scavenging ability. 28,29 Because agents that produce reactive Although the main focus of our work was a search for differential elemental levels in infected versus uninfected cells, it is noteworthy to highlight the clinical relevance of our findings. T. gondii infection stimulates immune cells to produce ROS and Th1-derived cytokines, such as interferon-gamma (IFN-γ) to limit parasite growth. 4 The increased level of Zn in infected cells might be related to its anti-inflammatory and antioxidant activities [30][31][32] to protect host cells from oxidative stress and DNA damage, 33-34 known to be associated with T. gondii infection. 35 Zn can reduce the production of tumor necrosis factor-alpha (TNF-α) and prevents the formation of free radicals. 32 This anti-oxidative stress (OS) effect of Zn has been attributed to down-regulation of the expression of ROS-producing inflammatory cytokines, such as TNF-α and interleukin-1β. 36 Zn can also limit nitric oxide (NO) production in endothelial cells by inhibiting NF-κB-dependent expression of inducible NO synthase (iNOS). [37][38] Further, Zn plays a role in maintaining the integrity of vascular endothelial cells, possibly by regulating signaling events to inhibit host cell death. [39][40][41] Zn deficiency can induce apoptosis 42 and disrupts cell membrane barrier integrity and increases the secretion of IL-8 and neutrophil transmigration. 43 Our results also showed elevated levels of Fe in infected cells, indicating that Fe is essential for parasite growth. Fe is a cofactor of many enzymes involved in diverse cellular processes including respiration and DNA replication. 44

Conclusion
Elemental profiling of ECs infected with T. gondii was accomplished using a combination of

Conflicts of interest
There are no conflicts of interest to declare.    4 Overall trends of the elements within the sub-clusters. The hierarchical clustering was divided into six groups of elements. The number of elements, including infected (Inf) and control (Con), and intracellular (Intra) and extracellular (Extra), in each cluster are listed in the box next to the corresponding group, with the exception of group 1, which contained the remaining elements. The mean (solid line) was calculated across the individual replicates, which constituted the elements in each group. The kinetic shape templates (mean ± 1 SD; dashed line) represent the span of all the elemental profiles of similar dynamic patterns in a single cluster. The x-axis denotes time in hours post infection and the y-axis represents the elemental concentration.

Supplementary Figures
Suppl. Figure 1: Schematic illustration outlining the groups of samples used to examine the levels of intracellular elements. Following collection of the culture medium the cultured cells were harvested, extracted and the concentrations of the intracellular elements were determined.
Suppl. Figure 2: Volcano plot showing the differentially abundant elements. For each element, a linear model was fitted with the measurement as the response variable. The parameter estimate for the interaction parameters between time and control/infected condition provided an estimate of the difference between the infected and control samples at each time point. The most significant differences (the time point with the smallest p-value) for intra and extra cellular elements are shown in the plot. Elements are represented as individual circles and plotted along the x-axis by the effect size (the maximum elemental concentration difference between control and infected samples across the three time points after infection), and on the y-axis by the significance level, -log10 (p value). The horizontal line indicates the significance threshold (p < 0.00088) after correcting for multiple tests, with elements above the line being significant and those below the line on-significant. Non-labelled circles represent those elements that had non-significant differences between infected and control samples. Figure 3: High-resolution elemental map of Calcium (Ca) in endothelial cells infected with Toxoplasma gondii acquired by μ-XRF at ~3 microns resolution. (A-C) Ca μ-XRF maps showing changes in the distribution and quantity of Ca at 6 hr, 24 hr, and 48 h post infection compared to the corresponding controls (D-F). Note the specific focal accumulation of the element (arrows). The rainbow-colored scale bars reflect the signal intensity and correspond to elemental concentrations in the range presented for each map, with darker pixels representing areas of lower concentration and brighter pixels representing areas of higher concentration.

Suppl. Figure 4: High-resolution elemental map of Zinc (Zn) in endothelial cells infected with
Toxoplasma gondii acquired at the μ-XRF at ~3 microns resolution. (A-C) Zn μ-XRF maps showing changes in the distribution and quantity of Zn at 6, 24 and 48 hr after infection compared to the corresponding controls (D-F). Note the specific focal accumulation of the element (arrows). The rainbow-colored scale bars reflect the signal intensity and correspond to elemental concentrations in the range presented for each map, with darker pixels representing areas of lower concentration and brighter pixels representing areas of higher concentration.