X-ray vision for molecules
New imaging techniques are being developed to show changes in biochemistry associated with diseases and treatments. Although x-rays have been widely used to image anatomical structure since their discovery in 1895, it is difficult to distinguish between different types of soft tissue in x-ray images. Three-dimensional x-ray computed tomography is more sensitive, but it does not provide molecular information. Two new methods—x-ray fluorescence and x-ray excited optical luminescence—overcome these obstacles by tagging specific molecules and imaging them through tissue. These x-ray techniques for molecular analysis have the potential to detect cancer biomarkers located on the surface of cells and to measure changes in acidity associated with infection on implanted medical devices.
X-rays are widely used in structural imaging of bone, teeth, microcalcinations, lungs, and orthopedic devices. In conventional x-ray imaging, different types of tissue are distinguished by how strongly they attenuate x-rays, a characteristic that depends on the tissue’s elemental composition. For example, soft tissue is mostly composed of low-atomic-number elements (e.g., carbon, hydrogen, oxygen, and nitrogen) that are relatively transparent to medical x-rays, whereas bone is more radiopaque because it contains more high-atomic-number elements, namely calcium and phosphorous. When a patient in Britain swallowed a felt-tip pen, it could not be seen on regular abdominal x-ray images because it was made of organic material with an x-ray attenuation similar to tissue. Twenty-five years later, the pen was detected using 3D x-ray computed tomography (CT) and removed (still in working order).
Although CT is commonly used for studying anatomy, the technique does not provide molecular sensitivity. For this reason, tissue samples collected from biopsies are frequently stained with fluorescent labels and imaged with fluorescent microscopes. Unfortunately, fluorescent microscopy cannot be used to image through deep tissue because tissue often has a large fluorescent background that obscures the label’s signal and because tissue scatters light, which dramatically reduces the image’s resolution. Recently, two x-ray imaging approaches were developed to detect specific molecules noninvasively through tissue: x-ray fluorescence (XRF) and x-ray excited optical luminescence (XEOL). These techniques enable functional imaging, which can be used to measure changes in biological function, including metabolism, blood flow, regional chemical composition, and biochemical processes.
X-rays can interact with matter in several distinct ways. They can be attenuated through absorption and scattering, and the remaining x-rays transmitted through the tissue can be detected by x-ray projection imaging or CT. In CT, images are collected at different angles and reconstructed to generate a 3D image. Alternatively, absorbed x-rays can produce a secondary x-ray emission with an element-specific energy, which can be detected by XRF (see Figure 1). XRF provides low-background detection for molecular labels made of elements not typically found in the body. X-rays can also generate visible light in materials known as scintillators, which is the basis of XEOL. Scintillators are commonly used to observe x-rays in conventional x-ray projection and CT imaging. (Röntgen discovered x-rays because a scintillator in his lab lit up when he turned on a cathode ray tube.) Recently, researchers began using XEOL to generate visible light within tissue for low-background molecular detection. XRF and XEOL can provide high-resolution images because the external x-ray beam maintains focus through tissue.
XRF has been widely used to analyze native elemental concentrations and map elemental distributions. For example, XRF has been used to map iron, calcium, and phosphorus distributions in healthy and cirrhotic human liver tissue sections; to map iron, copper, and zinc in healthy and cancerous breast tissue samples; and to non-invasively detect lead concentrations in children’s bones. XRF detects the presence and concentration of element-specific secondary x-ray emissions. The technique has a much lower background and can be more sensitive than CT to low elemental concentrations. In XRF, secondary x-rays are generated when a primary x-ray photon is absorbed by an atom, resulting in the ejection of an inner-shell electron. This process leaves behind a vacant hole within that shell, which is filled by an electron from a higher energy level, resulting in the emission of secondary x-rays. The energy of a secondary x-ray is unique to each element and is related to the element’s atomic number (i.e., the number of protons in the atom’s nucleus). XRF is well-suited for the detection of elements with a high atomic number, but is limited in detecting light elements because they have low absorption cross-sections and low fluorescent yields, and they generate low-energy secondary x-rays that are readily absorbed by a sample.
In addition to identifying native elements in biological samples, XRF has recently been extended to detecting molecules by labeling them with nanoparticles made of elements not normally found in tissue. The use of nanoparticles in research and medicine as imaging contrast agents and drug delivery vehicles is increasing because their chemical, magnetic, and optical properties can be controlled during synthesis. In XRF applications, nanoparticle labels can be chemically functionalized to adhere to molecules found on a specific type of cell surface (e.g., to aid in cancer detection). Because each element has a unique x-ray fluorescence spectrum, many different target molecules can be imaged simultaneously by using a different nanoparticle label for each target. This simultaneous detection, termed ‘multiplexing,’ is advantageous when single-analyte detection is insufficient for conclusive identification. The use of multiple labels increases specificity and allows researchers to investigate the correlation between different species of interest.
Ming Su’s research group at the University of Central Florida employed this multiplexed x-ray fluorescence approach to detect several different biomarkers using bismuth, tin, indium, and lead-tin alloy nanoparticle labels. Each species of nanoparticle was modified with a different molecular targeting moiety, namely a thiolated probe single-strand DNA (ssDNA) that binds to a complementary ssDNA-modified substrate. The researchers simultaneously detected four different DNA strand sequences using the four different nanoparticle labels, although up to 50 different elements could have been distinguished. They also demonstrated the feasibility of using the approach for in vivo measurements by detecting the XRF signal from bismuth nanoparticles labeling a surface coated with prostate-specific antigens and acquiring the measurement through polymethylmethacrylate (i.e., tissue-mimicking) substrates. Similarly, the Lei Xing group at Stanford University used multiplexed XRF to simultaneously detect and distinguish gold, gadolinium, and barium in a saline solution containing a mixture of the components within a water phantom, and superimposed the XRF image on a conventional CT image.
An alternative molecular imaging approach employs XEOL in scintillators (e.g., rare-earth-doped gadolinium oxysulfide nanoparticles). Scintillator particles can be injected into tissue and serve as an internal light source that is turned on only where and when it is irradiated with an x-ray beam. Red and near-infrared light generated by scintillators can propagate and be detected through the skin. High-resolution images of the location of scintillators can be obtained by scanning a narrow x-ray beam through the tissue and measuring the luminescence spectrum at each x-ray position. Using this technique, image resolution is limited by the beam width. The principles of XEOL tomography have been demonstrated by the Xing group at Stanford, which achieved sensitive low-background imaging through tissue phantoms. XEOL generates a stronger signal than XRF because each absorbed x-ray photon produces thousands of visible photons, compared to at most one x-ray photon for XRF. XEOL also has the advantage that visible detectors are less expensive and can be applied to larger areas than XRF detectors. However, XRF may be more appropriate for deep tissue imaging because tissue absorbs visible photons more strongly than it absorbs high-energy x-rays.
Recently, our group demonstrated that scintillator luminescence can serve as a light source for spectroscopic chemical detection using indicator dyes. Indicator dyes change their absorption spectrum (i.e., their color) when they interact with specific chemicals. When they are placed near scintillator light sources, the luminescence spectrum changes according to the chemical concentration. Using this methodology, we are developing a sensor for in vivo pH detection on the surface of implanted medical devices. If bacteria colonize on an implant surface, they can form highly antibiotic-resistant biofilms, partly because the bacteria generate acidic and hypoxic regions. Therefore, we expect that noninvasive methods to detect and image pH on an implant surface will be useful for diagnosing and monitoring bacterial growth during antibiotic and photodynamic treatments. To show proof-of-principle for pH sensing with XEOL, we demonstrated that pH could be determined from the luminescence spectrum of a scintillator film coated with methyl-red-dyed pH indicator paper. We also demonstrated sub-millimeter imaging of scintillator spectra through 10mm-thick tissue with a resolution that was limited by the x-ray beam width. We are currently developing more robust biocompatible sensor layers and applying them to implanted medical device surfaces to image changes in pH during bacterial growth.
Combining XEOL with indicator dyes provides a general approach for detecting a wide variety of chemical analytes for which there are indicator dyes. For example, we have demonstrated the ability to observe dissolution of 5nm-thick layers of silver deposited on a scintillator surface. We chose silver because it is a known antimicrobial agent and may be useful for preventing implant infections. The light signal was modulated by the optical absorption of the silver films, and silver dissolution by hydrogen peroxide was observed through 10mm-thick pig tissue. We also detected release of a colored chemotherapy drug, doxorubicin, from nanoparticle drug delivery vehicles. Overall, we believe combining XEOL with indicator dyes will be a versatile approach to chemical analysis on implanted medical devices.
The development of x-ray techniques for molecular analysis is in its nascent stage. Proof-of-principle has been demonstrated for high-resolution imaging through tissue with low background and high sensitivity using both XRF and XEOL. When compared with other medical imaging techniques, which provide either poor molecular sensitivity or limited spatial resolution, these x-ray approaches show tremendous promise for researching, diagnosing, and monitoring the molecular aspects of disease. Additional work is needed to improve the photon collection efficiency, and to control the composition, surface chemistry, and size of nanoparticle contrast agents to provide molecule-specific signals from the tissue of interest with minimal long-term toxicity. The applications of the technique will be constrained by inherent tradeoffs between the acquired signal intensity and the x-ray exposure dose, image resolution, concentration of contrast agents, and image acquisition rate. For example, relatively low x-ray doses would be required for 2D imaging of an implanted medical device surface coated with a scintillator and pH sensing film at approximately 100µm resolution, but extremely high doses would be needed for nanometer-resolution imaging of intracellular molecules in excised samples using focused x-ray beams. Overall, these novel x-ray techniques offer new methods for high-resolution, low-background molecular imaging to detect and study diseases.
- O. R. Waters, T. Daneshmend, and T. Shirazi, An incidental finding of a gastric foreign body 25 years after ingestion, BMJ Case Reports, 2011.
- F. Helmchen and W. Denk, Deep tissue two-photon microscopy, Nat. Methods 2, pp. 932–940, 2005.
- H. Chen, M. M. Rogalski, and J. N. Anker, Advances in functional x-ray imaging techniques and contrast agents, Phys. Chem. Chem. Phys. 14, pp. 13469–13486, 2012.
- F. Le Naour, C. Sandt, C. Peng, N. Trcera, F. Chiappini, A.-M. Flank, C. Guettier, and P. Dumas, In situ chemical composition analysis of cirrhosis by combining synchrotron Fourier transform infrared and synchrotron x-ray fluorescence microspectroscopies on the same tissue section, Anal. Chem. 84, pp. 10260–10266, 2012.
- G. R. Pereira, H. S. Rocha, C. Calza, M. J. Anjos, I. Lima, C. A. Pérez, R. T. Lopes, 3D elemental distribution images in biological samples by XRFµCT, X‐Ray Spectrom. 40, pp. 260–264, 2011.
- J. A. Hoppin, A. C. A. Aro, P. L. Williams, H. Hu, and P. B. Ryan, Validation of K-XRF bone lead measurement in young adults, Environmental Health Perspectives 103, pp. 78–83, 1995.
- A. C. Todd and D. R. Chettle, In vivo x-ray fluorescence of lead in bone: review and current issues, Environmental Health Perspectives 102, pp. 172–177, 1994.
- C. J. Fahrni, Biological applications of x-ray fluorescence microscopy: exploring the subcellular topography and speciation of transition metals, Current Opinion Chem. Biol. 11, pp. 121–127, 2007.
- W. Bambynek, B. Crasemann, R. W. Fink, H.-U. Freund, H. Mark, C. D. Swift, R. E. Price, and P. V. Rao, X-ray fluorescence yields, Auger, and Coster-Kronig transition probabilities, Rev. Mod. Phys. 44, pp. 716–813, 1972.
- N. Ahmed, H. Fessi, and A. Elaissari, Theranostic applications of nanoparticles in cancer, Drug Discovery Today 17, pp. 928–934, 2012.
- M. Hossain, C. Wang, and M. Su, Multiplexed biomarker detection using x-ray fluorescence of composition-encoded nanoparticles, Appl. Phys. Lett. 97, p. 263704, 2010.
- Y. Kuang, G. Pratx, M. Bazalova, B. Meng, J. Qian, and L. Xing, First demonstration of multiplexed x-ray fluorescence computed tomography (XFCT) imaging, IEEE Trans. Med. Imaging 32, pp. 262–267, 2013.
- H. Chen, A. L. Patrick, Z. Yang, D. G. VanDerveer, and J. N. Anker, High-resolution chemical imaging through tissue with an x-ray scintillator sensor, Anal. Chem. 83, pp. 5045–5049, 2011.
- H. Chen, D. E. Longfield, V. S. Varahagiri, K. V. T. Nguyen, A. L. Patrick, H. Qian, D. G. VanDerveer, and J. N. Anker, Optical imaging in tissue with x-ray excited luminescent sensors, Analyst 136, pp. 3438–3445, 2011.
- G. Pratx, C. M. Carpenter, C. Sun, and L. Xing, X-ray luminescence computed tomography via selective excitation: a feasibility study, IEEE Trans. Med. Imaging 29, pp. 1992–1999, 2010.
- A. Vertes, V. Hitchins, and K. S. Phillips, Analytical challenges of microbial biofilms on medical devices, Anal. Chem. 84, pp. 3858–3866, 2012.
- P. S. Stewart and J. W. Costerton, Antibiotic resistance of bacteria in biofilms, The Lancet 358, pp. 135–138, 2001.
- F. Wang, Y. Raval, H. Chen, T.-R. J. Tzeng, J. D. DesJardins, and J. N. Anker, Development of luminescent pH sensor films for monitoring bacterial growth through tissue, Adv. Healthcare Mater., 2013.
- G. Gosheger, J. Hardes, H. Ahrens, A. Streitburger, H. Buerger, M. Erren, A. Gunsel, F. H. Kemper, W. Winkelmann, and C. von Eiff, Silver-coated megaendoprostheses in a rabbit model—an analysis of the infection rate and toxicological side effects, Biomaterials 25, pp. 5547–5556, 2004.
- H. Chen, T. Moore, B. Qi, D. C. Colvin, E. K. Jelen, D. A. Hitchcock, J. He, et al., Monitoring pH-triggered drug release from radioluminescent nanocapsules with x-ray excited optical luminescence, ACS Nano 7, pp. 1178–1187, 2013.