X-ray fluorescence microscopy methods for biological tissues

Abstract

Synchrotron-based X-ray fluorescence microscopy is a flexible tool for identifying the distribution of trace elements in biological specimens across a broad range of sample sizes. The technique is not particularly limited by sample type and can be performed on ancient fossils, fixed or fresh tissue specimens, and in some cases even live tissue and live cells can be studied. The technique can also be expanded to provide chemical specificity to elemental maps, either at individual points of interest in a map or across a large field of view. While virtually any sample type can be characterized with X-ray fluorescence microscopy, common biological sample preparation methods (often borrowed from other fields, such as histology) can lead to unforeseen pitfalls, resulting in altered element distributions and concentrations. A general overview of sample preparation and data-acquisition methods for X-ray fluorescence microscopy is presented, along with outlining the general approach for applying this technique to a new field of investigation for prospective new users. Considerations for improving data acquisition and quality are reviewed as well as the effects of sample preparation, with a particular focus on soft tissues. The effects of common sample pretreatment steps as well as the underlying factors that govern which, and to what extent, specific elements are likely to be altered are reviewed along with common artifacts observed in X-ray fluorescence microscopy data.

Keywords: X-ray fluorescence microscopy; XFI; XFM; biological samples; chemical speciation; imaging; mapping.

Graphical Abstract

Studying elemental distributions in biological specimens with synchrotron-based X-ray fluorescence microscopy requires knowledge of instrument techniques and capabilities, an understanding of how X-rays interact with matter, adaptation of standard sample preparation methods, and an awareness of the myriad technical challenges and data artifacts that can arise. These challenges must often be met while also operating within the confines of limited periods of competitively allocated access to facilities. We provide an overview to X-ray fluorescence microscopy, examples of sample preparation methods and pitfalls, present commonly encountered technical challenges, and show examples of X-ray fluorescence microscopy in mouse research.

Fig. 1

Example of elemental mapping in a C57BL/6 mouse, sectioned along the sagittal plane, off the midline. The whole body was frozen within a block of 2.5% methylcellulose and cryosectioned to 100 μm thickness, collected onto adhesive Kapton film and allowed to air-dry for 48 h at −20°C before thawing. Data acquired at SSRL BL7-2.

Fig. 2

General considerations for synchrotron-based XFM characterization of biological specimens, from initial experiment planning through to data collection and processing.

Fig. 3

Examples of XFM characterization of macroscale plant and animal tissue as well as fragile microscale structures. (A) Young Arabidopsis thaliana with attached roots (courtesy Dr K.L. Haas), (B) Balb/C mouse brain with an ischemic stroke lesion in the cortex, (C) lateral ventricle of a C57BL/6 mouse and associated Cl-rich choroid plexus, (D) same choroid plexus structure shown in (C), highlighting elemental distributions. (A), (B) and (C) were acquired at SSRL BL 10-2 at 30 μm pixel resolution, while (D) was acquired at SSRL BL 2-3 with a nominal pixel resolution of 2 μm.

Fig. 4

Effects of conventional tissue preparation methods on elemental levels. Sample preparation conditions are as follows: full cryo = decapitation directly into liquid N₂ (no fixation), the brain was chiseled from the frozen skull and cryosectioned; 2 h PFA = cervical dislocation, brain removed at room temperature under 2 min postmortem and soaked in 4% paraformaldehyde (PFA) for 2 h before being flash-frozen in liquid N₂-cooled isopentane and cryosectioned; 56 h Prot. = cervical dislocation, brain removed at room temperature under 2 min postmortem, soaked in 4% PFA for 24 h, followed by 12 h in a 10% sucrose solution and an additional 20 h in 20% sucrose before flash-freezing in liquid N₂-cooled isopentane and cryosectioned; 24 h PFA = cervical dislocation, brain removed at room temperature under 2 min postmortem and soaked in 4% PFA for 24 h before being flash-frozen in liquid N₂-cooled isopentane and cryosectioned; paraffin = cervical dislocation, the brain removed at room temperature under 2 min postmortem and fixed in 4% PFA for 24 h, followed by 3× 24 h washes in 70% ethanol and finally embedded in paraffin wax for sectioning. All PFA solutions were prepared as 4% PFA in 1× phosphate-buffered saline (PBS). Note that the PFA solution for sample “24 h PFA^*” was spiked with additional NaCl (total Cl = ∼400 mM).

Fig. 5

Most tissue sectioning procedures, particularly with fixed or frozen tissues, result in damaged cells in the cut plane. (A) Shows a schematic representation of a tissue specimen comprising layers of cells. In this example the cut plane is parallel to a plane of cells, but could just as likely be a less ordered arrangement. Preparation of the tissue block and defining the cut-plane required sectioning through the top surface of the specimen (depicted by the rough, irregular edges) which is likely to have cut through cell membranes within the initial cut plane. The subsequent cut may also break apart cell membranes on the other side of the tissue section. If the hypothetical cells were ∼10 μm in diameter (such as shown in B) then a 20 μm tissue may only contain one layer of intact cells. Damaged cells (B) will readily lose much of their cytosolic contents, particularly if exposed to buffers or other solutions.

Fig. 6

Examples of the types of chemical environments and bonding that govern the retention of elements in biological specimens. Note that while a protein is shown as an example, the types of bond interaction examples are universal. The “coordinated metal ion” depicts donation of electron density from the ligands toward the central metal. The “chemically bonded element” shows sulfur in the amino acid cysteine as an example, but this is equally applicable to elements like phosphorus and sulfur in similar environments. “Labile elements” may include any solvated ions (such as K⁺ or Cl⁻), as well as elements in low molecular weight complexes, which may readily diffuse (taurine, phosphate, hydrated, or metabolite complexes of transition metals, among others)

Fig. 7

Relationship between the fluorescence yield and emission energy and the detectability of elements in biological specimens (concentrations are low, whereas geological or engineered materials may have weight percent elemental composition). Here the concentrations are typically low (trace level) and therefore any additional factors that decrease signal can impact the likelihood of detection.

Fig. 8

Basic X-ray fluorescence mapping setup and experimental data. (A) Computer-aided drawing of the essential components at beamline 7-2 at the Stanford Synchrotron Radiation Lightsource, and (B) a schematic representation of the components in A. (C) Ion chamber reading from I₀ (measuring incident beam intensity before the sample), (D) the total incoming photon count recorded by the detector, and (E) the I₁ ion chamber of the X-ray beam transmitted through the sample. The X-ray fluorescence spectrum from a single pixel in an area of high Zn K_α fluorescence (F) shows the relative peak heights of the emitted fluorescence from the sample compared to the scatter, with the Zn K_α and K_β peaks indicated. The Zn Ka fluorescence map (G) is shown in grayscale, along with a false color lookup table (LUT) applied (H), which converts intensity values at each pixel to a red–green–blue color value.

Fig. 9

Emission energies for commonly encountered elements, as well as representative element tags found in some contrast agents (e.g. I, Gd, Au) and other contaminants (e.g. Br and Pb). Emission peaks were convoluted with a Gaussian function, using a full width half-maximum (FWHM) of 100 eV, which matches the broadening typically encountered with silicon drift Vortex detectors (with Xspress3 digital signal processing electronics) employed for much of the data shown in this article. Sets of emission lines are normalized and sum to 100% intensity, without any assumption as to the excitation energy (affecting absorption cross section) or effects of line broadening with increasing Z, due to shorter corehole lifetimes. More detailed information is available.^,

Fig. 10

Effect of high background signal (red line; primarily due to Fe and elevated Ni, Cu, and Pb) on trace element detectability in a 20-μm-thick section of brain tissue. Note that excitation energy, mapping resolution (beam spot and step size), dwell time, and sample/detector geometry were unchanged between the two scans. Mitigation of background signal improved detector performance (black line), with demonstrable increase in trace element signals.

Fig. 11

(A) Shows the X-ray fluorescence spectrum from a 30-μm-thick section of brain (blue line) compared with the spectrum from a purified GaP quantification standard—note the presence of high-energy peaks at roughly double the energy of the Ga fluorescence peaks, indicating the detector is saturating with the standard in this case. (B) The procedure of defining energy windows, or bins, for photon counting versus the more quantitatively accurate peak-fitting procedure. (C) Maps of P, K, and Ca for an ischemic stroke bran, comparing the intensity differences between binned and peak-fitted data.

Fig. 12

Acquisition of X-ray absorption spectroscopic data at a point (pixel) (μXAS) data from an identified region of interest. The schematic diagram shows a coronal section of mouse brain with the highlighted region in the secondary motor layer of the cortex (hypothesized to contain a microglial cell), and the corresponding Fe K_α map is shown below with the same region of interest highlighted. Higher-resolution data, acquired at 13 450 eV shows a localized Fe signal much higher in concentration that the surrounding tissue (note: the trailing tail shape of the high Fe signal is an artifact, discussed in the whole body and organ mapping subsection of Examples of biologival XFM in mice, and shown in Fig. 13). Mapping with an incident energy just above the Fe K-edge ensures the precise position of the high Fe signal is identified when acquiring spectroscopic data. The acquired Fe K-edge spectrum matches a model compound spectrum for the ferritin form of Fe(III), which is a mineralized form of iron contained within the ferritin protein, commonly associated with microglial cells.

Fig. 13

Some of the common types of artifacts encountered in XFM experiments with biological specimens. (A) Artificially distorted shapes due to high count rates; (B) adhesion of tissue to the metal cryostat stage before transfer to the mounting medium creates surface irregularities; (C) bone fragments created during whole body sectioning drag through soft tissue in the direction of sectioning; (D) trapped air pockets under the tissue creates bubbles that distort the surface; and (E) wrinkles and folds in tissue are common and is one of the reasons floating fixed tissues on a water bath is employed for other techniques.

Fig. 14

High-resolution imaging (30 μm pixel size) of the C57BL/6 mouse data presented in Fig. 1, with select regions shown as tricolor plots. (A) Optical image of the mouse sagittal cross section prepared from a frozen whole body nitrocellulose block and collected on Kapton tape; (B) head with skull and brain: Cl (red), Zn (green), and Cu (blue); (C) gut region with partially digested food pellets: Fe (red), Zn (green), and Cu (blue); (D) liver and gallbladder: Fe (red), Zn (green), and S (blue); (E) nose and mouth/tongue: Fe (red), Zn (green), and S (blue); and (F) heart and adjacent spine: Fe (red), Zn (green), and P (blue).

Fig. 15

RML-infected FVB mouse brains (provided by Drs. J. Stöehr and H. Wille). All specimens demonstrated significantly elevated Fe throughout the corpus callosum in XFM maps, and this was confirmed with Perls’ stain and Fe K-edge X-ray absorption spectroscopic data at a point (pixel) (μXAS). Fe K-edge spectra acquired from multiple regions of interest in the specimens matched a ferritin standard purified from horse spleen. The Fe was deposited in the tissues as insoluble hemosiderin, which were resolvable with light microscopy. The tissue was unfixed and cryosectioned.

Fig. 16

Using elemental mapping to characterize metabolic changes after a stroke injury. The elements P, S, Cl, K, and Zn are also used for clustering the map (Gaussian-based clustering method is employed, seeded from an initial K-means clustering of the data to improve reproducibility). This example shows signs of consolidation of the lesion and repair.