Small chemical tweaks during tissue preparation make a big difference for imaging thick, complex biological samples.
Understanding biological complexity often requires that cellular populations be analyzed within tissues. But imaging in thick tissue samples is notoriously difficult. Two recent papers show how seemingly minor modifications in sample preparation can have big effects.
The work of Kwanghun Chung and colleagues at MIT shows that tissue samples up to a millimeter thick that have been fixed with paraformaldehyde (PFA) can be treated with glutaraldehyde (GA) to form a stable 'tissue gel' that can be subjected to multiple rounds of antibody stripping and probing without undue loss of structural integrity or antigenicity. The researchers successfully labeled nine antigens in the same human brain sample and demonstrated that 12 rounds (at minimum) of labeling produce no image distortion. The constructs can also withstand high temperatures (80 °C), so that tissue can be optically cleared rapidly, within days.
A general problem in working with thick samples is that any agent used to modify the tissue—whether a fixative or a labeling antibody—needs to diffuse into its interior. In addition, superficial layers can act as a 'sink' for reagents that interact with their targets; typically, superficial layers are modified much more readily in a given time than deeper ones. Chung and colleagues demonstrate an approach to tackle this problem.
Making use of the fact that GA does not cross-link protein under acidic conditions, they essentially render the fixation conditional. After GA has been allowed to penetrate the tissue at pH 3, the buffer is returned to neutral, enabling uniform fixation throughout the sample. Remarkably, this principle can also be applied to other agents, such as molecular labels; all one needs is to identify the right chemical on-off switch. For instance, antibody binding to antigen and dye labeling of lipid can be inhibited by SDS and therefore rendered conditional by this additive (of course, samples must already be fixed for this to work).
Powerful though the GA-based fixation approach may be, it does still require prior fixation with PFA as well as long incubations under harsh conditions. Although Chung and colleagues found that antigenicity and sample structure were well preserved, this could vary depending on the parameters being studied. It also remains a fact that fixation of any kind is likely to cause some perturbation to the sample. Indeed, a second paper from Kevin Briggman and colleagues at the US National Institutes of Health revisits a known artifact of aldehyde-based fixation: extracellular space is lost in aldehyde-fixed tissue, the researchers remind us, in contrast to rapidly frozen tissue. Briggman and colleagues sought to determine whether they could manipulate this artifact, to wit, increase the extracellular space and thus improve automated cell segmentation in electron micrographs for the reconstruction of neuronal circuits.
In a spirit similar to that of the MIT group, Briggman and colleagues modified the buffer in which they fixed mouse neural tissue, increasing the osmolarity of the solution to prevent cell swelling and loss of extracellular space. They observed that this increases the accuracy of cell segmentation in 2D and 3D electron micrographs, does not cause macroscopic changes to the sample, can preserve ultrastructure, and makes it easier to identify gap junctions than in conventionally prepared samples. In addition, perhaps not surprisingly, antibodies penetrate better into samples in which extracellular space has been preserved.
It is an agreeable irony that our increasingly sophisticated ability to image complex, multidimensional biological samples seems to benefit in no small way from an understanding of fundamental, low-tech methods of tissue preparation.