The Center for Materials Science and Engineering at MIT, a Materials Research Science and Engineering Center sponsored by the National Science Foundation, maintains and supports, amongst others, an Electron Microscopy Shared Experimental Facility. The purpose of this paper is to highlight selected recent research results for high-resolution investigations performed in that facility.

The facility owns the first VG HB603 intermediate-voltage FEG-STEM, which operates at 250KeV and is equipped with a high-solid-angle x-ray detector and a Gatan Digi-Peels. It was intended to be, and has been, used for high sensitivity, high spatial resolution microanalysis. It is well-known that the “resolution” of an x-ray analysis is intimately (and inversely) related to its sensitivity; one extreme situation occurs when analyzing, for example, a diffusion profile, when the need is to determine the composition to the highest precision. An example of such an analysis is given in fig. 1. In this case, the sample is a 1.4Cr-0.8C pearlitic steel, and the chromium analysis is carried out across a cementite plate. During the growth of the pearlite, the chromium, which is not thermodynamically required to redistribute, nevertheless diffuses along the growth interface towards the cementite, resulting in a comparatively wide depletion profile in the ferrite, and a very narrow enrichment in the cementite.

The pressure of shrinking budgets and downsizing has made it increasingly difficult to meet requests for comprehensive microstructural characterization. In many laboratories, especially those in industry, the expense of maintaining and staffing electron, X-ray and ion beam facilities has resulted in fewer analytical capabilities. At the same time, modern materials and products continue to diversify and increase in complexity, requiring increasingly sophisticated techniques. This growing paradox has increased the need for multiclient, shared-access ‘clusters’ of microanalytical instruments.

The above federal laboratory is mandated to serve both the Canadian public (in a broad sense) and Canadian industry (in the specific sense of collaborative and contract research projects). As a result, the Characterization Group has evolved into an integrated lab competency which compliments other competencies in materials processing (full-scale foundry, industrial rolling facility, Gleeble processing simulator) and mechanical behavior (fracture mechanics, stress and deformation modeling). Building on established expertise in metallography, microprobe, X-ray and TEM, the group has doubled in size (to 20 professionals and 10 major capabilities) and now includes SIMS and surface analysis (Table 1).

The Biological Electron Microscope Facility (BEMF) at the University of Hawai‘i at Mānoa (UHM) is located 2400 miles over water from the next nearest research university. BEMF is a multi-user core facility, administered by the Pacific Biomedical Research Center (PBRC), an organized research unit at the UHM. The mission of the BEMF is to provide state-of-the-art instrumentation, services and training for electron microscopy to the biomedical and biological researchers in Hawai‘i and the Pacific region. The BEMF was established in 1984 under the direction of Dr. Richard D. Allen, and has since grown steadily in its instrumentation, expertise, and use. In the past 5 years it has served researchers from over 50 laboratories in PBRC and the colleges of Natural Sciences, Tropical Agriculture and Human Resources, Engineering, Medicine, and Ocean and Earth Sciences and Technology, as well as visiting investigators from other Hawai‘i, mainland and foreign institutions.

The BEMF has a full line of instrumentation for conventional transmission and field emission scanning electron microscopy as well as a complete line of instruments for cryoelectron microscopy.

The application of nano- and microfabrication to biology is an emerging area linking engineering and biology. We are using microfabrication techniques to make patterns of specific bioactive molecules on glass and silicon surfaces to study central nervous system neurons and glia. Cell adhesion and tissue reaction to implanted silicon based electronic prosthetic devices are being studied, and the methods expanded to interface neurons with integrated circuits to study their electrical properties. The number, type and distribution of cells adhering to particular portions of a substrate can be influenced by patterning amine groups, proteins or polypeptides. We are using microcontact printing as an alternative to lithography for chemical patterning.

Primary cultures of rat cortical astrocytes and hippocampal neurons, and continuous cultures of transformed astrocytes were prepared on microcontact printed patterns produced by an elastomer stamp made from a silicon master. Fig. 1 shows two such patterns prepared by depositing self-assembled monolayers of N1[3-(Trimethoxylsily)propyl] diethylenetriamine (DETA) and octadecyltrichlorosilane (OTS).

The application of nano- and microfabricated devices based on silicon electronics technology is an emerging interdisciplinary area combining engineering and biology. The placement of electrically active probes in damaged or diseased tissues of the central nervous system could have enormous impact on the health and quality of life of large numbers of individuals by restoring lost function, or by treating or controlling disease states. Such probes have been fabricated at a high level of engineering sophistication. Unfortunately, when inserted into the brain a tissue reaction is initiated forming a scar that surrounds and electrically isolates the probe within a few weeks. This reaction is thought to primarily involve glial cells, and is undoubtedly dominated by the bulk surface of the probes which have a silicon oxide layer on top of single crystal silicon.

Model probes (Fig. 1) were microfabricated by photolithography with a 1×1mm tab used for gripping and inserting probes.

High pressure freezing enables the rapid arrest of developmental events without prefixation. Standard chemical fixation is a time dependent event and may cause artifacts in sensitive cytoskeletal components. We are studying two developmental events in embryonic Caenorhabditis elegans: that involve changes in the cytoskeleton: spindle alignment and membrane fusion. The mitotic spindle undergoes rapid rotational alignment prior to certain differentiative divisions. We are trying to capture these events by anticipation their timing and rapid freezing. Precursor hypodermal cells of embryonic C. elegans undergo a transition from individual cells to a syncytium at the onset of morphogenesis. In an effort to visualize the fusion events, embryos were stained with the vital probe FM4-64 to highlight cell membranes. Development was monitored by fluorescent microscopy using multiple-photon excitation imaging to minimize photobleaching while providing clear images of deep sections. Small cellulose capillary tubes, as described by Hohenberg for isolation and high pressure freezing of individual cells, were not of sufficient optical quality for monitoring by a laser-scanning microscope.

The Okazaki National Research Institutes, consisting of three cooperative research Institutes, the Inst. of Molecular Science (NIMS), the Inst. of Basic Biology (NIBB) and the Inst. of Physiological Sciences (NIPS), employs about 700 researchers and 200 full time technical staff. To support cooperative research projects, ONRI has several special shared research facilities in addition to those conventionally used for bioscience laboratories: a 1000 kV high voltage electron microscope designed for three dimensional morphometry and computer tomography of biological structures, a 300 kV ultra cryoelectron microscope equipped with an energy filtering system, which is now being set up, and a two photon excitation laser confocal microscope. In addition, there is an MRI, a rapid MRI, SQUID with 37 + 37 channels, and micro SQUID for non-invasive examination of brain function in human and experimental animals. A UVSOR (UV synchrotron orbital radiation) lab., a laser lab., and an ultra-low temperature lab. are available for neuroscience collaboration with Japanese and international scientists.

The central nervous system of a vertebrate organism exhibits a very complex spatial organization structure and function. These relationships are the subject of intense study for over a century, and recent developments in imaging have attracted ever increasing effort devoted to the understanding of brain function. One can produce any number of quantitative images that provide maps of the anatomy and function of nerve tissues. For example, autoradiography can yield functional images (2-deoxy glucose), maps of neurotransmitters receptors (over 100 know types), and gene expression labeled with complementary mRNA (in-situ hybridization). Immunohistochemistry produces maps of a large variety of neuroactive components, such as transmitters.

To illustrate a typical procedure, we describe the mapping of brain function with 2-deoxy glucose (2DG). A rat performing some task is injected with a solution of 2DG radiolabeled with 14C¨ Subsequently, the animal is sacrificed, the brain is cryosectioned (ca. 20 μm), and contact autoradio-grams of these sections are made on X-ray film.

For several decades the US scientific community has recognized the value of pooling specialized skills and equipment in centralized user facilities. Significant efficiencies are realized within the user community by avoiding unnecessary duplication and by moving projects more quickly to completion. The user facility concept served the nanofabrication community well for over two decades at the National Nanofabrication Facility (NNF) at Cornell. In 1994, the National Science Foundation expanded the concept by integrating NNF-Cornell (Now CNF) with several other nanofabrication programs throughout the country to form the National Nanofabrication Users Network (NNUN). Within this framework of a “users” network, scientists and engineers have access to state of the art equipment and expertise.

NNUN consists of two “full service” hub facilities at Cornell (the Cornell Nanofabrication Facility) and Stanford (the Stanford Nanofabrication Facility), in association with specialized facilities at Howard, Penn State and UC-Santa Barbara. The network is managed by the 5 site directors, who are responsible to NSF program management and to a Network Advisory Board.

Multiphoton molecular excitation by the strongly focused femtosecond pulses of infrared light generated as an 80 MHZ pulse train by a mode locked laser provides intrinsic submicron three dimensional spatial resolution of fluorescence excitation and photochemistry for laser scanning fluorescence microscopy. Because two-photon excitation requires simultaneous (∼10-16 seconds), absorption of two-photons focused laser intensities of about 1022 photons/cm2s are required. Since the rate of absorption is proportional to the square of the intensity, excitation is limited to the focal volume and is negligible elsewhere along the double cone of the focused illumination. Therefore, out of focus photodamage and fluorescence are generally negligible and laser scanning fluorescence microscopy with multiphoton excitation is intrinsically three dimensionally resolved with no out of focus background. Since the appropriate wave lengths are infrared for multiphoton excitation of ultraviolet or visible absorbing molecules, out of focus photodamage is eliminated. This allows imaging of useful ultraviolet absorbing indicators, vital DNA stains and autofluorescence in living cells with minimal, but not necessarily negligible, photodamage.

Imaging small structures substantially below the tissue surface in living specimens poses special challenges mainly because light is scattered by ever present refractive index inhomogeneities. Confocal microscoy removes the blurring caused by scattered and out-of-focus light but does so only at the expense of photodynamic damage that is often unacceptable when observing live specimens.

Multi-photon absorption microscopy[l] solves these problems because excitation is virtually limited to the focal plane. Out-of-focus photobleaching and photodamage are therefore eliminated. In scattering samples substantial improvements accrue even for the focal plane because, different from confocal microscopy, where only ballistic fluorescenc photons can be used, in the multi-photon microscope scattered photons can be utilized in addition [2-4], provided whole-field detection is used[5].

Many questions in the study of the nervous system require the investigation of intact portions of neural tissue in order to preserve the multiply branched processes of neurons, often extending over hundreds of microns, together with the local nervous circuitry.

The intensity of two-photon induced fluorescent shows quadratic dependency to the excitation intensity, as a result, one can achieve optical sectioning in fluorescent microscopy without the use of a confocal aperture. When the excitation power is properly adjusted, the fluorescent and photobleached volume is restricted to the vicinity of the focal points; therefore, it is possible to use this mechanism to achieve 3D data storage and microstructure fabrication (stereolithography).

The recording medium used in this experiment was poly-HEMA block doped with a high efficient two-photon fluorophore, APSS, (4-[N-(2-hydroxyethyl)-N-methyl) amino phenyl]-4’-(6-hydroxyhexyl sulfonyl) stilbene). The block was made by dissolving APSS in HEMA with an initiator, 2,2’-azobisisobutyronitril (AiBN) (1% mole ratio) and polymerized at 60° C for 48 hours. The polymer was then trimmed into a small rectangular block (3×5×1 mm3) and surfaced by using a ultramicrotome with a glass knife. A modified Biorad MC500 LSM equipped with a home-made computer controlled stage scanner was used for both writing (stage scanning) and read-back (beam scanning).

Two-photon excitation microscopy (TPEM) provides attractive advantages over confocal microscopy for three-dimensionally resolved fluorescence imaging and photochemistry. It provides three-dimensional resolution and eliminates background equivalent to an ideal confocal microscope without requiring a confocal spatial filter, whose absence enhances fluorescence collection efficiency. This results in inherent submicron optical sectioning by excitation alone. In practice, TPEM is made possible by the very high local instantaneous intensity provided by a combination of diffraction-limited focusing of a single laser beam in the microscope and the temporal concentration of 100 femtosecond pulses generated by a mode-locked laser. Resultant peak excitation intensities are 106 times greater than the CW intensities used in confocal microscopy, but the pulse duty cycle of 10−5 limits the average input power to less than 10 mW, only slightly greater than the power normally used in confocal microscopy. Because of the intensity-squared dependence of the two-photon absorption, the excitation is limited to the focal volume.

Multiphoton fluorescence excitation imaging is an optical sectioning technique for fluorescence microscopy. At very high photon densities, two or more photons may coordinately excite an energy transition in a fluorophore that corresponds to the sum of the energies of the individual photons. by this means, a fluorophore may be excited by a wavelength that is considerably longer than its single photon excitation wavelength. Ultra-fast pulsed (femtosecond) lasers can produce the peak power densities in the focal volume of an objective lens needed to provide sufficient 2- or 3- photon excitation events for imaging. The use of short-pulse lasers provides the high peak powers necessary for imaging yet with modest mean power levels that do not thermally damage biological specimens. Production of multiphoton events depends on the square of photon density for 2-photon excitation and the cube of photon density for 3-photon excitation. The power density therefore rapidly falls off away from the focal volume of an objective lens, thereby confining fluorescence excitation to the focal volume.

Atherosclerotic plaque at the carotid bifurcation is strongly correlated with the incidence of clinically significant events, such as transient ischemie attacks or stroke. Large multi-center trials have demonstrated that surgical removal of the atheroma is more effective in reducing these clinical events than is medical treatment alone. Patients are selected for surgery from an assessment of the severity of disease at the carotid bifurcation. The measure of disease severity is conventionally taken to be the degree of narrowing of the diseased vessel compared to a normal segment of vessel. However, using this criterion alone, many patients receive endarterectomy surgery who would probably not have progressed to other neurological events, while others are excluded who do progress to neurological events. For this reason, there is substantial interest in methods that could evaluate the composition of the atherosclerotic plaque in vivo, with the hope that this information would improve the predictive power of pre-surgical imaging of the diseased vessel.

Traditional methods for studying human coronary artery disease have significant limitations. Angiography allows evaluation only of the geometry of the remaining lumen; it cannot provide information on the structural or cellular composition of the arterial wall, which is essential to understand the processes involved in the progression of atherosclerosis. Intravascular ultrasound (IVUS) imaging is a new technology that permits tomographical visualization of a cross section through the vessel wall (Fig 1 A.) The development and refinement of IVUS has provided a powerful in vivo method to assess plaque morphology. Recent clinical studies have documented its sensitivity in detecting atherosclerosis, and it is increasingly employed to assist in selecting an appropriate therapeutic intervention. Perhaps more importantly, the potential of IVUS to quantify the structure and geometry of normal and atherosclerotic coronary arteries will allow one to characterize specific lesions and to differentiate the plaques that lead to various clinical syndromes. This study presents a three-dimensional (3D) reconstruction technique suitable for clinical use that accurately preserves 3D geometric information throughout the cardiac cycle without requiring mechanical IVUS catheter withdrawal aids.

For several decades, contrast angiography has served as the principal imaging modality used to assess the anatomic severity of coronary artery disease. Recent technical advances in acoustics and microelectonics have permitted development and refinement of miniaturized ultrasound devices capable of real-time tomographic coronary imaging. This approach, intravascular ultrasound, represents an emerging alternative to angiography for direct visualization of coronary anatomy during diagnostic and interventional catheterization. Clinical studies comparing angiography to intravascular ultrasound have demonstrated frequent differences in quantitative or qualitative findings. Accordingly, intraluminal ultrasound imaging is increasingly utilized to confirm, refute, or supplement angiographie data in patients with coronary disease.

The discrepancies between angiography and intravascular ultrasound can be explained by the characteristics of these two different imaging techniques. Angiography, which portrays the vessel as a silhouette of the lumen, whereas ultrasound depicts the coronary from a tomographic perspective, providing cross-sectional images that portray not only the lumen, but also the deeper intramural structures.

Coronary artery balloon catheter angioplasty for the treatment of obstructive atherosclerosis has an unacceptably high restenosis rate of about 40% within 6 months. Local delivery of pharmaceuticals and nucleotide-based compounds is under investigation as one means to decrease recurrence. Labeled molecules, when delivered to arterial angioplasty sites in aqueous solution using porous balloon catheters, are quickly lost from the tissue. Alternative strategies include drug encapsulation in microparticles, and implantation of drug-containing polymers. We tested the feasibility of a catheter-based system to deposit a 70-μm thick layer of polyethylene glycol-lactate based hydrogel (polymerized by cross-linked end-terminal acrylate groups) onto the luminal surface of coronary arteries in pigs after balloon angioplasty, and examined the histologic consequences at one hour and seven days using light, scanning electron, and transmission electron microscopy.

At one hour the gel layer was readily visualized by all three microscopic techniques and appeared well adhered to and in intimate contact with the blood vessel surface (Figs. 1-2).

Build-up of visible atherosclerotic plaque in the arteries is readily quantifiable. The mouse and the rabbit provide useful models for understanding the pathogenesis of atherosclerosis by investigating the effects of genetic and dietary perturbations.

Although the wild type mouse does not develop atherosclerosis, atherosclerosis susceptibility genes have been identified in some laboratory mouse strains which do. Furthermore, transgenic technology and gene targeting have produced genetically modified mice that express various apolipoproteins, enzymes and cofactors involved in human lipoprotein metabolism. Gene “knockout” technology allows transgene expression without interference from homologous genes. One notable “knockout” mouse, deficient in apolipoprotein E, develops spontaneous atherosclerosis on a normal chow diet. Transgenic modulations of the atherosclerotic responses of these highly susceptible mice are more pronounced and easily measured. Small, cheap and fast breeding, mice are convenient animal models. But to make mice susceptible to atherosclerosis, their genetic background has to be so drastically altered that the resulting lipoprotein metabolism may not model the human metabolism accurately enough.

Recent data suggests that reactive oxygen species (ROS) play a role in the progression of atherosclerosis. Superoxide (O2˙−) and its free radical metabolites can participate in the oxidation of LDLs, activation of proto-oncogenes and promotion of cellular growth. Endothelial and vascular smooth muscle cells are capable of producing O2˙− and other ROS. A major source of O2˙- and other reactive ROS in vascular tissues is a membrane associated NADH oxidase. This paper presents a study of NADH oxidase by cytochemical localization in developing atheromas in a rabbit model of diet-induced atherosclerosis.

NADH oxidase generates O2˙− which dismutates to hydrogen peroxide (H2O2). Cerium perhydroxide, an electron dense precipitate of cerium and H2O2, is the reaction product of NADH oxidase. Cytochemical localization of NADH oxidase was done by the cerium method in developing atheromas in abdominal aorta and iliac arteries of New Zealand rabbits made atherosclerotic by feeding a diet of 0.2% cholesterol/5% peanut oil over four months.