One Method, Multiple Imaging Modalities
How to Maximize the Data Retrived from a Single Specimen
- One Method, Multiple Imaging Modalities - How to Maximize the Data Retrived from a Single Specimen
- Fig. 1: Composite image of Daphnia using the NHM’s Nikon A1 SI Laser Scanning Microscope. Scan taken using fluorescence emission from 4 lasers (405 nm, 488 nm, 561 nm, 640 nm), 20X magnification (numerical aperture 0.7) and 1.2 au pinhole. 193 Z-steps at 2.075 μm. Pixel size 1.2 μm. Scale bar 500 μm. R– rostrum, Ce– compound eye, Cg– cerebral ganglion, P– phyllopod, C-carapace, Ac-abdominal claw, As– apical spine, H- heart, Am-antennae muscles, G-gut, D-diverticulum, An-antennae.
- Fig. 2: Rendered µ-CT scans of specimens from all four staining groups (A-D), scanned wet and dry. All renderings were performed in Drishti . All specimens were scanned on transmission target except A1+ A2, which were scanned on reflection target. Scan parameters ranged between: 100-120 kV, 65-90 μA, 500-1000ms exposure, 1100-3142 projections, and final voxel sizes were between 5-7 μm. To remove noise- C1+C2 were subsampled (X2 in x, y, z) and dilated, D1+D2 were subsampled (X2 in x, y, z), and C3 was subsample (X2 in x, y, z), filtered and dilated.
- Fig. 3: SEM of Daphnia stained with iodine and PTA and dried with HMDS. Image taken on Zeiss Leo 1455 VP SEM using 20 kV and a chamber pressure of 20 Pa. Scale bar 200 μm. Key as in figure 1.
Description-rendered micro-CT scan of a polychaete worm parasitized by a copepod parasite, this specimen was prepared using the method described here.
The application of micro-CT (µ-CT) to biological specimens allows full three-dimensional examination of internal and external morphology, but has a number of limitations. Complementing µ-CT data with other imaging techniques can mitigate these limitations, and also serves to maximize the amount of information gathered from different biological scales. This article outlines a workflow that is being used at the Natural History Museum (NHM) to unify preparations for different modalities.
µ-CT utilizes x-rays to create 3-dimensional (3D) models of specimens and artifacts. The x-rays are attenuated in response to atomic mass, analogous to density, and these attenuated x-rays are recorded as grayscale 2-dimensional (2D) projections taken at regular intervals during a 360° rotation. A Feldkamp back-projection algorithm  is then employed to reconstruct this data. The main limitations for biological materials are the resolution and the lack of contrast between tissues so that morphology can be hard to discern. In order to increase the contrast, staining and drying agents can be applied that artificially increase the x-ray attenuation between neighboring tissues [1,2]. Studies involving µ-CT can maximize the amount of morphological detail by using a suite of techniques to incorporate data from different biological scales. Scanning electron microscopy (SEM) and confocal laser scanning microscopy (CLSM) complement µ-CT by offering greater resolution of external (SEM and CLSM) and internal (CLSM) features, and can provide quantitative analysis of chemical (SEM) and physiological (CLSM) composition. Between these three techniques a wealth of complementary data can be gathered, but how does one prepare a specimen in order to use this kind of comparative approach?
Methods and workflows that can unify these techniques, whilst maintaining the longevity of specimens housed within museum collections, such as those held at the NHM, are essential.
This study aimed to test a range of staining and drying methodologies for µ-CT and compare these results with those from SEM and CLSM.
This study utilized the crustacean model Daphnia sp., as Crustacea are under-represented in µ-CT associated literature and presents new challenges for the application of µ-CT.
Method and Workflow
Daphnia were fixed in glutaraldehyde and left overnight before imaging with the Natural History Museum’s Nikon A1 SI CLSM (fig. 1). The specimens were then split into four staining groups (A-D) before dehydrating all of them to 100% ethanol. The staining groups were as follows: A - stained in phosphotungstic acid (0.3% PTA in 50% ethanol) for 48 h; B - stained in PTA for 24 h and then transferred to iodine solution (1% in 95-100% ethanol) for the next 24 h; C - stained in iodine for 48 h; and D - stained for 24 h in iodine and then 24 h in PTA. All specimens were then mounted into low density plastic tubes, using lens tissue to secure them in place and prevent movement during the scans. All µ-CT scans were performed with the NHM Nikon Metrology HMX ST 225 system.
Post-staining CLSM images were then collected for comparison with preliminary CLSM scans. Specimens were dried with hexamethyldisilazane (HMDS), placing them in the solution for 1 h before removing and leaving them to air dry in a fume cupboard. Chemical drying with HMDS was chosen as the process is reversible, which is a key concern for collection material which ideally needs to be returned undamaged.
The second round of µ-CT scans were then conducted. Finally all specimens were imaged in the Natural History Museum’s Zeiss Leo 1455 VP SEM using 20 kV, and a chamber pressure of 20 Pa. Whole specimen images and EDX maps were taken to determine the effectiveness of each of the staining protocols.
From the CLSM results we could see that the daphnia were highly autofluorescent. Organs that can be seen are clearly defined (e.g. antennule muscle fibres) although those found deeper within the body (e.g. cerebral ganglion) are obscured (fig. 1).
Results of the µ-CT scans reveal that the surface features are defined best in specimens stained in PTA or a combination of PTA and iodine (A2, B2, D2, fig. 2). Streaking artefacts can be seen in specimens that were scanned wet (D1 and C2, fig. 2). PTA and combinations of PTA and iodine produced the best internal anatomical information (A2 and B2, fig. 2). While dried specimens produced datasets with clearer boundaries between tissue types and reduced overall noise (B3 and C3, fig 2), although some specimens did incur serious damage during the application of HMDS.
Scanning electron micrographs resolved surface features clearly e.g. setae of the antennae (fig. 3).
Key internal features of Daphnia sp. can be seen in SEM, CLSM and µ-CT images. In the CLSM images features near the emission source are resolved well but features found deeper within the specimen, the cerebral ganglion and the gut, are not (fig. 1). They are obscured due to the diffraction of light that occurs in the preceding layers, resulting in less signal returning from features deeper within the sample. This limits the z depth of the data and exemplifies the reasons to use CLSM in combination with other techniques when studying thicker biological material. CLSM images taken after specimens were treated with staining agents revealed the same problems.
µ-CT results from stained specimens showed a marked improvement over non-stained specimens to resolve both internal and external features. Staining with PTA, whether alone or in combination, gave the best results in terms of both external and internal features (fig. 2); since PTA is more dense than iodine .
In comparison with the CLSM results the µ-CT data allowed tissues and organs found within the specimen to be imaged regardless of position within the organism, and allowed for virtual histology. µ-CT results do, however, fall short of CSLM in terms of resolution and clarity when specimens are scanned with too much ethanol. Ethanol diffracts the x-rays, which causes streaking artefacts and reduces visibility (C2, fig. 2). This can be mediated by removing excess alcohol before scanning but this may cause desiccation, a problem that is amplified when working with small specimens (<10mm).
Another way to mitigate this is to chemically dry specimens.
Drying using HMDS produced variable results, in some cases the application of HMDS improved definition of tissues and removed streaking artefacts. In other cases the technique caused extreme shrinkage of the specimen. The µ-CT scans, SEM micrographs and EDX analysis of dried specimens proved that both iodine and PTA are unaffected by the HMDS drying process. This result is significant as dried and stained specimens are less likely to move during a µ-CT scan, have increased contrast and can be imaged using SEM and SEM-hosted nano-CT.
The workflow presented here is successful in allowing a single specimen to be imaged with multiple techniques and has the potential to unify the information obtained at different scales. The inconsistencies in the results of HMDS application require attention and further study.
 Feldkamp LA. et al.: J. Opt. Soc. Am. 1(6), 612–619 (1984)
 Metscher B.: BMC Physiology, 9 (11), doi:10.1186/1472-6793–9-11
 Pauwels E. et al.: J. Microscopy 250 (1) 21–31 (2013)
 Limaye A.: Vis, Baltimore (2006)
Rebecca Anne Summerfield (corresponding author via e-mail request)
The Natural History Museum