|Home | About | Journals | Submit | Contact Us | Français|
We describe an instrument to record x-ray diffraction patterns from diseased regions of human brain tissue by combining an in-line visible light fluorescence microscope with an x-ray diffraction microprobe. We use thiazine red fluorescence to specifically label and detect the filamentous tau protein pathology associated with Pick’s disease, as several labs have done previously. We demonstrate that thiazine red-enhanced regions within the tissue show periodic structure in x-ray diffraction that is not observed in healthy tissue. One observed periodicity (4.2 Å) is characteristic of cross-beta sheet structure, consistent with previous results from powder diffraction studies performed on purified, dried tau protein.
Tau, a soluble microtubule-binding protein, transitions into filamentous aggregates during the progression of several neurodegenerative diseases, including Alzheimer’s disease (AD) , cortical basal degeneration (CBD) , progressive supranuclear palsy (PSP) , and Pick’s disease . The pathological triggers of these diseases change tau from a relatively unstructured protein into an assembly competent macromolecule that then forms a highly stable, well-ordered neurofibrillary tangle, or NFT . This self-assembly phenomenon is of interest in its own right as a biophysical process, but this interest is made more acute by the potential relevance of its mechanism to human neurodegenerative disorders wherein filamentous protein aggregates can cause dysfunction and disease. Understanding how and why tau polymerizes could open avenues for the development not only of therapeutic agents directed at AD, but at compounds to treat the broader base of neurodegenerative diseases.
Powder diffraction patterns of tau filaments extracted, purified, and dried down from AD brains  and similar structures formed from dried truncated peptide fragments of tau in vitro [6,7] indicate that these tau aggregates form cross-beta sheet structures having approximately 4.7 Å periodicity. This is striking, considering that native tau has very little secondary structure in solution . Tau aggregation inside cells is a complex process, involving differential proteolysis and processing of toxic tau proteins throughout disease progression by phosphorylation, nitration, and other mechanisms, and leading to different tau ultrastructure in NFTs from different pathologies. This is supported by immunohistochemical evidence indicating that the epitopes displayed by the tau pathology in PSP and CBD are distinct from those displayed in AD  and by the apparent lack of structural organization in Pick’s disease NFTs relative to Alzheimer’s NFTs . Thus, while experiments on isolated, dried tau proteins provide encouraging evidence that tau is highly ordered in its pathological form, it cannot be taken for granted that tau aggregates formed in vitro or extracted from diseased tissue behave the same way as aggregated tau in vivo.
The complexity and diversity of tau aggregation inside cells motivates the development of instrumentation to determine the structure of aggregated tau directly from diseased brain tissue. Our approach is to take advantage of the fact that there are several well-established fluorescent probes for specifically detecting filamentous tau pathology, by coupling an in-line fluorescence microscope into an x-ray microdiffraction beamline setup (Fig. 1). Uncertainty regarding the validity of prepared or extracted samples is replaced with an experimental difficulty: ensuring that the regions of tissue examined correspond to the location of high concentrations of tau aggregates.
To demonstrate this approach, we have chosen to study the dense isolated aggregates of tau found in Pick’s disease. Pick’s disease is characterized by a very high density of filamentous tau structures (Pick bodies) inside the cells of the dentate gyrus of the hippocampus and in the frontal and temporal regions of the cerebral cortex . Because of the high density of cells containing tau aggregates, Pick’s disease was an ideal proof-of-principle for our method. Our apparatus allows individual Pick bodies to be localized using an in-line fluorescence microscope with excellent spatial registration to an x-ray diffraction microprobe.
Experiments were conducted using the x-ray microdiffraction instrument on the Bio-CAT beamline 18 ID at the Advanced Photon Source . The apparatus is shown schematically in Fig. 1. The x-rays were focused to roughly match the typical size of a dentate gyrus cell filled with Pick bodies (~ 10 × 10 µm2) at the sample position using a pair of Kirpatrick-Baez mirrors (not shown). The tissue sample could be scanned with sub-micron precision around the x-ray beam in three dimensions. A separate linear stage allowed a helium-filled 1.25 m flight tube and x-ray CCD detector (Mar 165, Mar USA Evanston, Illinois) to be moved out of the beampath in order to insert a visible light camera for recording fluorescent images. The tissue sample was illuminated with a 550 nm light-emitting diode located approximately 30 cm from the sample (10 mW maximum power, wavelength spread <10 nm FWHM), and the images were recorded through an emission filter (peak transmission wavelength = 593 nm, 40 nm width) onto an RGB camera. The position of the light-emitting diode was adjusted to obtain the brightest fluorescence image possible. The fluorescence sensitivity was improved by using only the red and green channels. A zoom lens assembly was used to magnify the sample image. Registration of the visible light camera with the x-ray beam was done by imaging a sheet of x-ray burn paper placed at the sample location. By scanning the sample stage holding the burn paper, we could calibrate our pixel size (0.81 microns/pixel.). Each time an x-ray exposure was made, that region was also imaged using the in-line fluorescence microscope. This was critical to ensure accurate registration of x-ray and optical data to within the 10 × 10 µm2 size of the x-ray microprobe beam and Pick body-containing cells.
Four hippocampal sections were obtained from one advanced case of Pick’s disease. Three were fixed and labeled with thiazine red, and one was fixed, but not labeled. Thiazine red fluorescence (excitation peak = 510 nm, emission peak = 580 nm) has been widely used as a marker for the presence of polymerized tau pathology in a number of different diseases and different tissue preparations, as the dye intercalates efficiently into cross-beta structures but does not bind to soluble tau [13,14]. Each section was cut to 40 µm thickness. The sections were rinsed 3 times in phosphate buffered saline (PBS), then incubated in a solution of 0.002% thiazine red for 20 minutes, rinsed again 3 times in PBS, and fixed in formalin (2% for 8 hours). Colocalization of thiazine red staining with Pick bodies was verified during sample preparation using fluorescence and differential interference contrast microscopy, combined with Microbrightfield’s StereoInvestigator software (Fig. 2A). Sections were mounted on 0.0025” thick mylar which was in turn fastened to teflon washers using epoxy resin, resulting in a contained compartment consistent with level 2 biohazard regulations for Argonne National Laboratory. Preparation of all samples and assembly of compartments were accomplished at Northwestern University Feinberg School of Medicine.
Initially, all four tissue sections were examined by x-ray diffraction at the APS as described above without the in-line fluorescence microscope. All sections showed the periodicities reported in Fig. 2B within some regions. Pick body dense regions were positively identified using the in-line fluorescence microscope as described above on one of the three thiazine red-labeled tissue samples. The others sustained visible-light damage from setting up the in-line fluorescence system. On that sample, a series of eighteen continuous regions were mapped out prior to x-ray exposure. A fluorescence image was acquired while each region was at the center of the x-ray beam before beginning x-ray exposures. The size of individual cells within the dentate gyrus is approximately that of the x-ray microprobe (10 × 10 µm2), therefore, different dentate gyrus cells were imaged within different regions. A total of ten, one-second x-ray exposures at 12 keV were taken using the full beamline flux in each region. Following the x-ray exposures, the samples were removed to Northwestern University where they were imaged using the same microscope used for sample preparation to verify that the intended regions had indeed been exposed to x-rays, as indicated by some slight modification of the tissue appearance resulting from x-ray damage.
To determine fluorescence intensity as a measurement of tau aggregate concentration, an area the size of the focused x-ray beam dimensions was set in each visible light image corresponding the darkest exposed region of the burn paper. The camera response was summed over this area for each of the eighteen regions. Quantified in this manner, the visible light fluorescent intensity changed by nearly a factor of three across the tissue sample areas studied, indicating a wide range in the concentration of tau filaments.
X-ray diffraction patterns typical of high and low areas of fluorescence intensity are shown in Fig. 2, along with periodicities calculated from the experimental geometry. Each pattern was obtained by averaging the ten exposures, which did not show any systematic reduction in diffracted intensity over the course of the measurement. An air scattering background was also recorded and subtracted from these images. The total diffracted intensity for a particular spacing was calculated by azimuthally integrating the diffraction patterns using the Fit2d software package  at a particular diffraction angle. All three periodicities observed were seen to some limited extent in all regions as well as in the unlabeled tissue. While the same periodicities were consistently observed, it may or may not be appropriate to attribute them to particular structural features of the tau aggregates in our samples. It is reasonably likely that some periodic features observed here are fixation artifacts due to the stringent fixation conditions used to ensure that our samples were Biohazard Level II compliant. Nonetheless, periodicities near 5 Å have been observed in previous diffraction studies of purified tau from AD brains and may correspond to the 4.7-Å periodicity found in β-sheets and found previously in dried tau samples [5,6,16]. The 38 Å repeat, which has not been previously reported, may be indicative of long-range order in the tau aggregates or may result from applying specific fixation and staining protocols to tau aggregates.
Correlations between x-ray and fluorescent intensities are shown in Fig. 3. The long-range 38 Å repeat in particular shows a very strong correlation with fluorescence intensity, while the 4.2 Å repeat shows only a moderate correlation. For comparison, regions in between the diffraction rings (exemplified by the scattering pattern at 15 Å) shows no correlation between x-ray diffraction and fluorescence intensity. While preliminary, this experiment has demonstrated that periodic tau filament structures can be colocalized and visualized by X-ray diffraction.
Although the dense and isolated distribution of Pick bodies in brain tissue made them ideal for this initial demonstration, the best structural information on tau NFTs would likely be obtained from the more highly oriented and ordered structures in AD NFTs. To date, we have made several measurements of diffraction patterns from AD brain tissues in the same manner as described in this paper. In particular, we have used x-ray contrast agents such as uranyl acetate that selectively bind to tau filaments and found that they reveal several new periodicities. Demonstrating that these periodicities correspond to tau pathology in AD was challenging using our preliminary setup described here because the tau aggregates occur at far lower density in AD making them more both more difficult to locate with fluorescence and yielding diffraction data with poorer signal-to-noise ratios. Improvements to the fluorescence camera with better optics and closer positioning to the sample would be helpful, as would the possibility of conducting these measurements on fully hydrated tissues without the need for mounting them between mylar film. Performing the experiment in a Level III or IV Biohazard facility would eliminate the need for stringent fixation protocols, which may have affected the periodicities observed in the preliminary experiments reported here.
The in-line fluorescence microscope/x-ray microprobe experimental setup described here could prove to be tremendously valuable. In addition to thiazine red, a wealth of fluorescent probes and contrast reagents exist for detecting highly specific forms of tau pathology as well as other filamentous protein pathology. By combining these sophisticated and highly optimized detection reagents with an x-ray diffraction microprobe, this method can in principle be used to unambiguously determine periodic structures found in diseases having specific forms of filamentous pathology.
We would like to thank M. Vukonich and D. Rodi of Argonne for help with safety procedures, and E. Bigio of the CNADC at Northwestern University for providing the brain tissue samples. E.L. was funded in part by the DePaul University College of Liberal Arts and Sciences. J.O. and O.A. were funded by the National Science Foundation (Grant #MCB-0644015 CAREER). S. Rice was funded in part by the Northwestern Alzheimer’s Disease Center, pilot project grant # PHS AG 13854. Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Basic Energy Sciences, Office of Science, under contract No. W-31-109-ENG-38. BioCAT is a National Institutes of Health-supported Research Center RR-08630. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Center for Research Resources or the National Institutes of Health.