6.1. Trimming to the biological question

Sample trimming is an often ignored step during sample preparation. We recommend users trim their sample to their biological question (Figure 6.1). While there is often a temptation to image bigger samples (such as whole brains), this is often unnecessary (for example, if labeled neurons are restricted to a small part of the brain) and can add significant costs. Smaller samples clear faster, can be immunostained more quickly and with less antibody, are easier to image (no matter how transparent samples look, there will always be distortions when light propagates through them) and can be imaged more quickly. In addition, trimming can make samples with unconventional shapes easier to mount in the microscope by turning them into a block. In short, trimming a sample is a simple way for users to save time and money. Note that if a sample is trimmed down to be very small (on the order of 1-2 mm in diameter), this can make it challenging to mount in a sample holder. I address this challenge in the sample mounting tips section (7).

Figure 6.1.

Examples of trimming samples to match the expected distribution of neurons under study. Note that all relevant regions are included, but samples are scaled properly to those regions of interest. This saves time and reagents for clearing and labeling (if (more...)

6.2. Clearing and labeling

There are many effective labeling and clearing methods in the literature, as well as several good reviews that summarize them (see Ariel 2017 for references to the methods I have found most useful). One of the main advantages of the LaVision BioTec UltraMicroscope II is its compatibility with organic solvents, which allows users access to some of the fastest and most effective clearing techniques. In the UNC Core, we have had great success with iDISCO+ (full disclosure: I am an author on the paper describing the first iteration of iDISCO), and BABB-based methods (which we image in DBE). I have also seen acceptable results with uDISCO (if the endogenous fluorophores are not RFPs and are expressed at a very high level), CUBIC (in the superficial 2-3 millimeters of cleared samples) and PACT (though the samples are very soft and harder to mount). These observations should not be taken as rigorous comparisons between any of the methods, and it is important to note that the success of any clearing method is highly sample dependent. When I consult with people on which clearing/labeling method to use, I advise them to copy one that has already worked for a similar sample type and biological question, that has not only been used by the lab that created it, that requires minimal equipment and has a low implementation cost. In the UNC Core, that frequently means iDISCO+. However, in your own lab or core, you may find other techniques more suitable for your own research. Several clearing and labeling methods (particularly in the CLARITY family) require specialized equipment but this may not be a limiting factor if there is a sample clearing/labeling service at your institution (that is not the case at UNC).

With respect to labeling samples, using fluorophores with longer wavelengths is a big advantage, as there is less autofluorescence, scattering and absorption at longer wavelengths (Figure 3.3). In fact, the combination of these factors can make it very difficult to image successfully in the green, cyan or blue channels, unless the fluorophore concentration is very high. Thus, even if fluorescent proteins are in the sample, it can be advantageous to immunolabel them to amplify signals (the typical multiplicative effect of primary and secondary staining) and spectrally shift them to longer wavelengths. In our core, users have had particular success combining AlexaFluor dyes 568, 647 and 790 with the iDISCO+ technique. While longer wavelengths lead to lower resolution, the effect is negligible on this microscope, given that its maximal resolution is not high enough for this wavelength dependence to make an appreciable impact on image quality.