The discovery of X-ray radiography in 1896 showed that a position-sensitive detector could detect the locations of photon removal occurring in an object with which it is in close contact. After this, radiographs of small objects were produced and examined under the light microscope to obtain a useful early form of X-ray microscopy. However, the early form of X-ray microscopy was not sufficient to provide satisfactory images to biomedical scientists due to insufficient image resolution as well as many technological limitations. Despite many technological difficulties, consistent efforts at technical refinement eventually enabled development of more “user-friendly” and powerful X-ray microscopy, which now can be applicable to analysis of biological specimens (Kirz et al., 1995).
The current X-ray microscopic technique is well suited to provide high resolution and sensitivity in microscopic analysis of biological specimens without the use of special sample processing such as fixation and staining required for electron microscopy (EM). The great advantages of the X-ray microscopy technique are (1) the high contrast provided by the density of the elements in the specimens, thereby avoiding processing of a specimen; (2) X-ray microscopy permits the observation of thicker specimens than that in EM, without sample preparation or alteration; (3) the ability to record X-ray absorption spectra of localized regions by means of image sequences at a range of energies; (4) the actual location of a specific element within a cell can be visualized when the proper wavelengths of X-rays are used; and (5) three-dimensional observations are potentially possible with a single exposure with divided X-ray beams in multiple directions.
There are several technical issues that are difficult to resolve by using ordinary light and electron microscopy in the observation of living and hydrated biological specimens. For example, a hydrated specimen, such as a bacterial organism, results in motion blurring. Therefore, a pulse exposure time, which must be sufficient enough to overcome the issue of kinetic movement, is required in addition to other technical aspects when viewing a hydrated biological specimen. However, current advanced technology enables us to overcome these technical issues. In this article, we review the principal as well as technical aspects of two types of X-ray microscopy that are currently developed and extensively used in analysis of biological specimens with our current data and discuss possible future directions in application to biomedical science.
The types of structures that become visible under microscopy depend on the interaction process of the radiation with the sample. For instance, with visible and ultraviolet light, this process is dominated by the interaction with the valence electrons of the atoms. Their electronic states are strongly influenced by chemical bonds in the molecules. In contrast, X-rays mainly interact with inner-shell electrons of atoms. Therefore, X-ray microscopy can be highly sensitive to elements and bonds between atoms (Schneider, 1998).
The X-ray microscopy techniques reviewed in this article use soft X-rays with an energy of approximately 100 to 1,000 eV, which is well-matched to K shell absorption edges of low Z atoms such as carbon and oxygen or L shell edges of atoms such as calcium. The wavelength of these X-rays is in the 1- to 10-nm range (as opposed to 350–700 nm for visible light). At the diffraction limit, shorter wavelengths give much higher spatial resolution than longer wavelengths. X-ray microscopy uses the intrinsic absorption properties, described above, of the sample and, thus, does not require staining or fluorescent probes, which are commonly used in light microscopy. On the other hand, hard X-rays are an excellent tool for studying matter on an element-by-element basis, but the spatial resolution routinely achieved is only ∼1 μm in the hard X-ray vs. 50 nm or better in the soft X-ray region (Hitchcock, 2001). Good contrast in a natural aqueous environment can be obtained by using the wavelengths between the inner-shell absorption edges of oxygen and carbon, 2.3 and 4.4 nm, known as the water window (Schneider, 1998). In this region, the absorption of carbon is around 10 times that of oxygen and, hence, water as well. Thus the X-rays of such wavelengths enable examination of thick (up to 10 μm) and hydrated cells. Even thicker samples of approximately 100 μm can be investigated at wavelengths around 0.3 nm (Schneider, 1998).
The contrast of image with high resolution is an important issue for analysis of biological specimens. In this regard, the total radiation dose is critical in the contrast of an X-ray image obtained. The absorption of X-rays used for imaging leads to ionization of matter, which causes radiation damage by breaking chemical bonds in the sample. Therefore, higher does of radiation to obtain a clear contrast may cause sample damage (Shinohara and Ito, 1991). One of the strategies to avoid such technical limitations is to use flash X-rays, which must be fast enough to capture the image before morphological damage can occur (Ito and Shinohara, 1992). The flash X-rays can also avoid motion blurring of the object. Using frozen, hydrated specimens is another way to avoid such radiation damage as well as motion blurring. The use of vitrified specimens can make it possible to increase the radiation dose 1,000-fold without bringing about structural changes as can be observed in wet biological specimens at the resolutions obtainable with the X-ray microscope (Methe et al., 1997).
Despite many technological difficulties, consistent efforts at technical refinement enabled development of more “user-friendly” and powerful X-ray microscopy applicable to analysis of biological specimens.
Visual ultrastructural studies are possible with variation of EM, although specimen preparation steps such as fixation, dehydration, resin embedding, ultra-thin sectioning, coating, and staining are very technical, expensive, and may introduce artifacts in the original samples. Most ultrastructures of eukaryotic cells have been visualized and analyzed by EM, even though such biological specimens are hydrated in natural status but visualized in fixed and dehydrated form. Cell surface structures covered by hydrophilic substances, such as mucopolysaccharides, may have unique morphology that may affect the function of the cells. For example, the surface structure of cells may be involved in attachment, cell–cell communications, and phagocytosis. However, high-resolution analysis in a hydrated and viable state are difficult by ordinary microscopy techniques. In this regard, X-ray microscopy may offer high-resolution transmission images of thick biological samples maintained in aqueous solution by the reasons described above. Several X-ray microscopes can image biological specimens at approximately 30-nm resolution (Jacobsen, 1999). With a unique set of capabilities, X-ray microscopy, therefore, is in-between those of visible light and electron microscopy (Kirz et al., 1995; Jacobsen, 1999).
Laser Plasma X-ray Contact Microscopy
A laser-produced plasma can emit strong X-rays with high conversion efficiency from a target such as yttrium (Kondo and Tomie, 1994). However, early work with laser plasma X-ray sources required the use of major facilities (dedicated to laser fusion) due to the necessity of a high-power laser. Current advances in laser technology have provided a compact table-top Nd:glass laser system, which generates a high-power laser with 10J–20J of energy at a 1,064-nm wavelength in a 5-ns pulse and 5 × 1013 W/cm2 of intensity on the target (Richardson et al., 1993). The pulsed laser beam is sufficient to produce laser plasma X-rays from a target. The pulse width of the X-ray emission in the water window region is approximately the same as the laser pulse width, 5 ns. Therefore, current laser plasma X-ray microscopy does not necessarily require specific facilities. Several compact types of laser plasma X-ray microscopes are now available institute with conventional laboratories, such as a laser plasma X-ray contact microscope (LXM) at the Center for Research and Education in Optics and Lasers, University of Central Florida, Orlando, FL. In this article, an outline of the LXM is described as an example of the this type of microscope.
Because soft X-rays are absorbed in air, the sample holder and other elements of an X-ray microscope, except for the laser itself, are installed in a vacuum chamber. Therefore, the sample holder is designed to keep wet specimens protected while in vacuum. The specimen is placed on an X-ray photo resist (PMMA; polymethyl methacrylate) supported by a silicon wafer, which can be obtained from Silson, Ltd., Northants, UK, and mechanically shielded with a thin silicon nitride window, 100 nm in thickness. The resolution of PMMA is as high as 5 nm and is adequate for analysis of cell structures, because the resolution of a photo resist is critical for the final spatial resolution of X-ray microscopy (Sayre and Chapman, 1995; Kado et al., 1999). The specimen holder is placed 1 cm away from the target at an angle of 45 degrees from the target. A single X-ray exposure is made on each resist. The outline of the entire instrument is shown in Figure 1. The photo resists after exposure to X-rays are rinsed with sodium hypochlorite solution to remove any remaining specimen from the resist and then developed by dissolving the X-ray–damaged material in a mixture of methyl isobutyl ketone and isopropyl alcohol (1:1, v/v) for 3–5 min. The development time is varied according to the laser energy and the thickness of the specimens to obtain the best images.
The three-dimensional topological absorption image that results from development is then observed with a differential interference microscope. An atomic force microscope (AFM) is used to reproduce the images from the developed photo resists. The AFM is a recent innovation that relies on a mechanical probe for generation of magnified images without coating or making a replica of the target resist. AFM is a good technique for examining low-contrast topography. Moreover, the relief height can be measured with high precision, which allows us to discuss the density of specimens quantitatively (Tomie et al., 1991).
Transmission X-ray Microscopy
Transmission X-ray microscopes (TXM), which use zone plates as objective lenses for high-resolution imaging, were first developed by Niemann et al. (Niemann et al., 1976). There are several transmission X-ray microscopes available at synchrotron radiation centers in Berlin (Germany), Århus (Denmark), and Berkeley (California) (Jacobsen, 1999). For example, the system XM-1 at the Advanced Light Source (ALS), Lawrence Berkeley National Laboratory, Berkeley, CA, has been extensively used for analysis of variety of biological specimens (Meyer-Ilse et al., 2001). The great advance of the XM-1 microscopy system is that it combines visible light microscopy with X-ray microscopy. Therefore, the visible light image positions and X-ray image positions can be mutually indexed to each other, allowing easy adjustment and focusing of specific cells.
The sample is mounted on a sample carrier, which can be precisely moved from the visible light microscope to the X-ray microscope, while maintaining registration of sample position and focus. The instrument allows multiple-wavelength imaging with high spatial resolution and modest spectral resolution. The specimen holder has been designed to accommodate multiple angular orientations for stereo-imaging or tomography (Kirtz et al., 1995). The sample, 5–10 μm thick, is imaged at atmospheric pressure and is in a liquid environment. The wet sample is mounted between two silicon nitride films, each approximately 100 nm thick. The sample holder is placed between a condenser zone plate and an objective zone plate as shown in Figure 2.
The X-ray wavelength of the microscope can be adjusted, allowing spectromicroscopy techniques without compromising spatial resolution. An enlarged image of the transmitted radiation is formed by an objective zone plate. Exposure time of specimens to X-rays for obtaining images is usually relatively long, such as 1 s. The images are recorded directly on an X-ray sensitive CCD camera, by using a magnification of typically 2,400×, and stored digitally with other parameters of the microscope. The field of view is approximately 10–15 μm. A detail of the XM-1 can be found at the laboratory Web site (http://www-als.lbl.gov/als/microscopes).
ANALYSIS OF CELL SURFACE STRUCTURES BY X-RAY MICROSCOPY
Murine elicited peritoneal macrophages are chosen as a target cell for X-ray microscopy study due to their well-known morphological features. In general, macrophages are characterized as a professional phagocytic cell with many immunological functions, such as production of cytokines as well as antigen presentation. The high phagocytic property of the macrophages first appears in an irregular cell shape and a large number of ex-tensions of the cell membrane (Ryter, 1985). In stimulated macrophages, the cell surface forms many ruffles, which constitute a large plasma membrane area available for rapid internalization of foreign material and for cell spreading (Cohn, 1978). The initial and obligate event involved in the phagocytosis is the contact and adhesion of molecules or particles to the cell surface. Most of such events have been observed and analyzed by EM. For example, the phagocytosis of bacteria by macrophages has well been studied morphologically by EM. It is known that some bacteria are phagocytized by coiling phagocytosis, characterized by a long phagocyte pseudopod coiled around the bacterium as the organism is internalized (Wyrick and Brownridge, 1978; Chang, 1979; Horwitz, 1984). Such morphologically characteristic phagocytosis may contribute to their intracellular process. Thus, the importance of morphological study of macrophages has been recognized, particularly in the study of cell–cell interaction.
For analysis of cell structures by X-ray microscopy, macrophages are placed on a photo resist supported by a silicon base or a window with silicon films and exposed to soft X-rays generated by a high-powered pulse laser or a synchrotron radiation. Mouse peritoneal macrophages are prepared from 8-week-old female BALB/c mice 4 days after intraperitoneal injection of thioglycollate broth (Yamamoto et al., 1988). The cells are suspended in RPMI 1640 medium supplemented with 10% fetal bovine serum and antibiotics (penicillin and streptomycin) and cultured on either a photo resist for LXM or a window with silicon nitride film for TXM for 2 h at 37°C in 5% CO2. Nonadherent cells are removed by washing the photo resist or window with the medium. The resulting macrophage monolayers, hydrated with the medium on the photo resist or window, are then analyzed by either LXM or TXM. The X-ray shadow of a specimen recorded on the photo resist is developed and observed by an AFM in the case of LXM. An AFM is valuable for examining developed photo resists, because the relief on a nonconducting resist surface can be examined directly without any additional process. In the case of TXM, the X-ray images are directly captured by X-ray sensitive CCD camera and digitally stored in a computer.
Figures 3 and 4 show representative X-ray images of a hydrated macrophage in medium without any fixation and sample processing observed by LXM. In comparison with conventional transmission electron microscopy (TEM) (Figure 5), the X-ray images showed some interior structural components of the macrophages, but these components seemed not rigid, which means not clear compared with TEM images. The dissimilarities between X-ray images and TEM are apparent in the figures; namely, there are no obvious organelles (e.g., nucleus, endoplasmic reticulum) in the X-ray images (Figures 3 and 4), compared with the TEM images (Figure 5). The reason for the absence of contrast inside the cell at X-ray images may be due to the large thickness and, thus, overlap in X-ray direction of many different structures.
Unlike TEM specimens, the specimens for X-ray microscopy are not sliced or stained with heavy metals, such as lead citrate and uranyl acetate, which permit high contrast of specimens, but instead are whole cells. Furthermore, X-ray images are based on elemental density. That is, the elements of the specimen are the principal reaction sites. They act as X-ray photon removal sites and as particle emission sites emitting several classes of photons and electrons. In addition, each element (carbon, nitrogen, and others) has its own individual set of cross-sections for the removal and for the several emission behaviors, and these element cross-sections are further individualized by the way in which they vary with the energy of the photons present and with the chemical state of the element (Da Silva et al., 1992). Therefore, images of a cell observed by X-ray microscopy may provide physiological rather than merely morphological information.
Nevertheless, the most impressive structures observed in X-ray images of macrophages are outer surface structures. Some fibrillar surface structures of macrophages are obvious (Figures 3 and 4). These surface structures are observed even in glutaraldehyde-fixed macrophages in the medium by LXM (data not shown). The fibrillar structures seemed to be hydrophilic, because the edge of the structures is not clear. Similar surface structures are also observed by TXM (Figure 6). That is, there are many fluffy but diffused structures on the surface of macrophages. The outline of the fibrilla is also observed by ordinary light microscopy (Figure 7) but is less clear due to low resolution. Scanning electron microscopy (SEM) is widely used for analysis of surface structures of cells because the resolution is extremely high. However, the ordinary protocol for preparation of a biological specimen for SEM requires fixing, dehydrating, and coating, similar to that for TEM. Figure 8 shows a typical SEM image of a macrophage, which is cultured on a photo resist similar to the preparation for LXM. The image shows the fine surface structures of the macrophage, but the surface edge appears rigid and shrunken, which may have been caused by the sample processing. The function of the fibrillar surface structures observed by X-ray microscopy is not known. Macrophages ingest microbes, communicate with other immune cells, and secrete many cytokines. Furthermore, macrophages adhere to tissue culture plates in vitro; therefore, it is reasonable to speculate that the fibrilla may be involved in some or all of these important functions of macrophages.
Both types of X-ray microscopy, LXM and TXM, provide similar X-ray images of macrophages in medium, without any essential difference. That is, well-maintained fragile outer structures of macrophages are observed by X-ray microscopy, providing acceptable resolution for the study of cell–cell interactions. The application of X-ray microscopy to biological specimens, such as chromosome fibers, protozoa, bacteria, sperm, and cultured cells, has also been studied with some success (Tomie et al., 1991; Kinjo et al., 1994; Kirz et al., 1995; Methe et al., 1997; Rajyaguru et al., 1997a, b; Kado et al., 1999; Scharf and Schneider, 1999). In particular, current equipment supported by advanced technologies provides practical X-ray microscopes, such as those used in this study and for similar cell biology studies by biological scientists. The study reported here describes a finding indicating the presence of fragile structures sensitive to sample processing required for other methods, including electron microscopy. The nature of the structures composed of hydrophilic substances may be detected only when hydrated. Therefore, analysis of such structures necessities hydrated conditions. In this regard, X-ray microscopy may provide a powerful tool for studying such hydrophilic structures of cells in a native form.
X-ray microscopy may provide a powerful tool for studying hydrophilic structures of cells in a native form.
X-ray microscopy is a relatively new technique in biomedical science for imaging the intact structure of a biological specimen at high resolution without the introduction of possible artifacts, because X-ray microscopy does not require sample processing, such as staining or fixing. The principal of X-ray microscopy is based on elemental density as described previously. Therefore, X-ray microscopy has a potential to provide a variety of information concerning biological specimens, such as location of specific elements in a cell in a natural form, rather than merely morphological information, if the system is properly set and chosen. For instance, iron is essential for several normal brain functions and accumulates in high concentrations in specific regions of the brain. Although the function of a regionally high brain iron content is unknown, the homeostasis of brain iron is thought to be necessary for normal brain function, especially for learning and memory (Yehuda and Youdim, 1989; Youdim et al., 1989; Connor et al., 2001). Furthermore, knowledge concerning the roles of iron and iron-binding proteins in immune cells has been developed recently (Kemp, 1993). For instance, intracellular iron plays a critical role for the antimicrobial activity of macrophages (Byrd and Horwitz, 1989; Gebran et al., 1995). Iron is also an essential participant in many metabolic processes, including DNA, RNA, and protein synthesis, and as a cofactor of many heme and nonheme enzymes (Gerlach et al., 1994). A deficiency of iron metabolism, therefore, would be expected to alter some or all of these processes. Thus, the distribution, concentration, and chemical states of iron between regions of a cell are critical in homeostasis of the human body.
There are several techniques for determining cellular iron distribution. For example, EM with X-ray spectrometry represents one of the most widely available instruments for identifying and localizing trace elements present in biological specimens (Perl and Good, 1992) and is designed to collect and analyze the emitted X-rays for either their energy levels or their wavelength. However, due to the limitations of the secondary electron imaging capabilities and the relatively poor detection limits for this technique, it is unlikely that this technique will contribute substantially to characterizing cellular and subcellular iron distribution in a cell (Perl and Good, 1992). Through X-ray microscopy, not only can one visualize iron distribution but also possibly determine chemical states, such as oxidation. Such a capacity of X-ray microscopy would be highly valuable in developing a detailed understanding of the function of iron in cells and very useful in biomedical research.
There are many other variants of X-ray microscopy currently available, such as spectromicroscopy and full-field transmission X-ray microscopy, which provide multiple wavelength imaging. These variants might provide more detailed analytical capabilities, such as chemical mapping by means of X-ray absorption spectroscopy (Hitchcock, 2001). Thus, there are potential possibilities to provide a new direction for biomedical study by X-ray microscopy. However, only limited trials have been performed to date, due at least in part to the limited access to X-ray microscopes, which are found only in a few institutes worldwide. Nevertheless, recognition of the usefulness of X-ray microscopy in biomedical science may eventually contribute to the advancement of many areas of research, from metabolism to functional anatomy.
The authors thank Drs. M. Kado (Japan Atomic Energy Research Institute, Ibaraki, Japan), M. Richardson (University of Central Florida, Orlando, FL), G. Denbeaux (Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA), T. Tomie (Electrotechnical Laboratory, Tsukuba, Japan), S. Shioda (Showa University, Tokyo, Japan), H. Yoshimura (Meiji University, Kawasaki, Japan), and Y. Kinjo (Tokyo Metropolitan Industrial Technology, Tokyo, Japan) for their cooperation in this study. The authors also thank Dr. H. Friedman (University of South Florida, Tampa, FL) for his critical review.
Dr. Yoshimasa Yamamoto is an Associate Professor in the Department of Medical Microbiology and Immunology at University of South Florida College of Medicine. His research team applies X-ray microscopy to the study of macrophage function. Dr. Kunio Shinohara is a Professor and Director of the Radiation Research Institute in Graduate School of Medicine at the University of Tokyo. His research focuses on the development of X-ray microscopy techniques and the application of them to analyze biomedical specimens.