Chapter 1
Perspectives on Receptor Autoradiography in Human Brain
Michael J. Kuhar, Neuroscience Branch, Addiction Research Center, National Institute on Drug Abuse, Baltimore, Maryland 21224
Publisher Summary
This chapter presents perspectives on receptor autoradiography in human brain. Autoradiography combines biochemistry and microscopic anatomy. Earlier, brains were dissected, tissues were homogenized, and receptors were measured in biochemical binding assays. The anatomical resolution was limited by the size of the piece of tissue that could be dissected and practically assayed. Receptor maps are beneficial as the receptors can be localized to small brain regions and, therefore, to specific neuronal pathways with a greater accuracy and precision that helps to explain the action of drugs in the brain. A key feature of technical approach is that the labeling of receptors in tissue sections is sectioned and thaw-mounted onto microscope slides. The in vitro labeling approach is advantageous because it is significantly cheaper as it is possible to label individual sections on slides that make more efficient use of tissues and radioactive compounds. Positron emission tomography (PET) scanning provides receptor maps without excision of tissue.
Autoradiography has been a powerful tool for many years. It combines biochemistry and microscopic anatomy and has been involved in many major discoveries in biology. It has been an important tool, particularly where radiolabeled compounds have been available, and part of ongoing experimental strategies. Thus, it is not surprising that when biochemical receptor binding burst into the neurosciences in the early 1970s (Snyder and Bennett, 1976), there was an immediate effort to use autoradiography to answer important questions. This chapter is an effort to describe some of the experiences and developments that helped create the field of drug and neurotransmitter receptor autoradiography as it exists today and how it produced the material in this book. For practical reasons it is not a comprehensive review but hopefully describes the foundations of the field as well as some unique contributions made by many workers during the past years.
Some of the earliest biochemical receptor binding studies, those for the nicotinic acetylcholine receptor, the opiate receptor, and the muscarinic cholinergic receptor, created a great deal of excitement. The opiate receptor was especially alluring because it was generally acknowledged that opiates were not found in the body, and therefore, everyone wondered what the natural, endogenous substrate of the receptor was. Because of this, regional localization and light microscopic localization of receptors to specific neuronal tracts was a prime interest. It was not until a couple of years later that enkephalins and endorphins were identified.
The earliest studies on the anatomical distribution of receptors involved the biochemical âgrind and bindâ approach. Brains were dissected, tissues were homogenized, and receptors were measured in biochemical binding assays (Yamamura et al., 1985). The main difficulty was that anatomical resolution was limited by the size of the piece of tissue that could be dissected and practically assayed. Certainly smaller and smaller pieces of tissue could be dissected and assayed under scaled-down conditions. However, this approach reached a practical limit, and increasingly finer dissection became impractical and eventually impossible. The application of light microscopic autoradiographic techniques became imperative at this point.
How do we localize receptors by autoradiography? Receptor binding was a new technique, so how was autoradiography applied? Neuroscientists and pharmacologists working on receptors in the early and mid 1970s were fortunate in that previous workers had developed techniques for looking at steroid hormone receptors (Stumpf and Roth, 1966; Roth and Stumpf, 1969). In addition, there had been a strong effort to produce generally applicable techniques for localizing diffusible substances by autoradiography (Roth and Stumpf, 1969). When radioactive steroid hormones were injected into animals and localized by autoradiography, it was clear that they (and their receptors) were localized to the nuclei of specific cells in brain (Stumpf and Grant, 1975). Similarly, it seemed possible that radiolabeled ligands for brain neurotransmitter receptors could be injected into animals and then localized by autoradiography. This should in fact localize receptors, because we could identify conditions where the bulk of the ligand in the brain appeared to be associated with specific receptors. The evidence for this was that the regional distribution of radioactive ligand about an hour after intravenous administration had the same regional distribution as receptors as was found by regional dissection and in vitro biochemical binding assays. Also, this selective regional distribution was prevented by pretreating the animals with high doses of unlabeled drugs that blocked the receptors. Thus, there was an in vivo pharmacological and regional specificity for ligands for receptors that were among the first studied [i.e., the muscarinic cholinergic receptor (Yamamura et al., 1974) and the opiate receptor (Pert and Snyder, 1975; Pert et al., 1975)]. Given that it was possible to label receptors in vivo, it was, therefore, feasible to apply techniques originally developed for steroid hormones to this problem. This was successfully done and autoradiograms of receptor distributions in brain were produced. For example, in our laboratory, receptor autoradiographic maps for muscarinic cholinergic receptors in rat brain were produced in February 1974 and published soon after (Kuhar and Yamamura, 1974, 1975).
These and other initial successes were very exciting. It was possible to identify receptor distributions and localize receptors at the light microscopic level in intact brain sections. In general, the observations were reasonably compatible with current concepts of neurotransmission. For example, most receptors were found in the neuropil where the synapses were found and where neurotransmission occurred. But some observations were more difficult to understand. For example, in the stratum oriens and stratum radiatum of the hippocampus, muscarinic cholinergic receptor distributions as determined by autoradiography were fairly uniform over those dendritic fields; by contrast, acetylcholinesterase staining, thought to be a useful marker for presynaptic cholinergic axons in nerve terminals, appeared to be more restricted in those regions (Lewis and Shute, 1967). This apparent mismatch between neurotransmitter marker and receptor became known much later as the âmismatch problemâ in receptor mapping (Kuhar, 1985b; Herkenham and McLean, 1986).
A major benefit of receptor maps is that receptors could be localized to small brain regions and, therefore, to specific neuronal pathways with a greater accuracy and precision than ever before. This provides the opportunity to help explain how drugs exert their action in the brain. Although it is important to say that a receptor exists in the brain, it is only by knowing which neuronal circuits contain the receptors can we fully explain how drugs exert their actions. Given the promise of this approach, it is not surprising that many investigators expressed a keen interest in receptor autoradiography.
However, investigators were faced with a serious limitation at this point in time. Although we had been successful in mapping receptors in brain after labeling of receptors by intravenous administration, it was not possible to carry out this procedure with humans. Obviously, it was becoming important to measure receptor distributions in human brain with increased anatomical resolution. Receptor distributions had been measured by dissection and binding, but it was clear that we would ultimately need to be able to do it by techniques with greater resolution (i.e., light microscopic autoradiography). Thus, the existing methodology for neurotransmitter and drug receptors, which was derived mainly from procedures that had been developed for steroid receptors (i.e., in vivo injection and then dissection), was impossible to use with human tissues. Some studies had appeared localizing nicotinic acetylcholine receptors using radiolabeled alpha-bungarotoxin in muscle biopsies (Fambrough et al., 1973) after labeling of the receptors in vitro. It was clear that being able to carry out autoradiography with brain sections that were labeled in vitro after postmortem sectioning would be an essential advance for this field.
Receptors obviously could be labeled in tissues in vitro. It has been mentioned that nicotinic acetylcholine receptors were labeled and localized by autoradiography after in vitro incubation of muscle biopsies with radiolabeled bungarotoxin. Similarly, it seemed feasible to take blocks of brain tissue, incubate them with ligands, and then section these blocks and localize the receptors by autoradiography, using techniques specifically developed for localizing diffusible substances. But there were many practical difficulties with this. After incubating a tissue block, we had to be careful with the plane of sectioning because not all the receptors throughout the tissue block had equal opportunity to bind radioactive ligand, because the ligand did not diffuse throughout the block. Thus, in practical terms, we could not be sure whether the regional variation of receptors in a section was caused by true differences in receptor presence or whether it was caused by differences in the availability of ligand to the receptor. This was a real issue because it was known that receptor ligands did not fully penetrate into tissue slices of manageable thickness. Hence, a different approach would be required.
The approach that eventually worked was a combination of labeling receptors in slide-mounted tissue sections and then doing autoradiography with techniques for localizing diffusible substances (Young and Kuhar, 1979a). It was advantageous to label receptors in tissue sections already mounted to slides so that autoradiography by dry-emulsion apposition could be carried out. In this case, highly reliable results were obtained because all receptors facing the emulsion had an equal opportunity for binding ligand, and regional variations surely reflected regional variations in receptor density rather than problems with diffusion barriers. Rotter et al. (1979) and Polz-Tejera et al. (1975) successfully used slide-mounted tissue sections with irreversible binding ligands to localize muscarinic cholinergic and presumed nicotinic receptors. However, the vast bulk of ligands available for binding to receptors did not bind irreversibly, rather they were reversible binding ligands. Our earliest experiments showed that after labeling receptors in slide-mounted tissue sections with reversible ligands, dipping the sections into molten emulsion resulted in an almost total loss of and rapid diffusion of radioligand from receptor sites. Thus, although slide-mounted tissue sections were capable of binding ligands relatively specifically and selectively, a special approach would be needed for those ligands that bound in a reversible mannerâwhich was the bulk of the available ligands at the time. This part of the problem had been addressed by Roth and co-workers (Roth et al., 1974). They used a dry, emulsion-coated coverslip, which was apposed to a slide-mounted tissue section containing radioactivity. Because the emulsion was dry, there would be minimal opportunity for the radioactive drug to diffuse away from its site. After exposure, the emulsion could be developed and the tissue could be treated, because at that point, the autoradiogram was already completed and loss of ligand had no practical consequence. The combination of labeling neurotransmitter receptors with reversible ligands in slide-mounted tissue sections and ...