The microscope is one of the most iconic symbols of science, at least equal to the white coat or bubbling neon liquids in its ability to endow scientific credibility in the eyes of TV viewers of shows like The Big Bang Theory or CSI: Miami. Yet the kinds of microscopes that most people imagine, despite being a mainstay of many labs, represent the old-school of microscopy in terms of both their technology and their impact in modern research. The new-school of microscopy involves liquid cooled lasers, glowing cells and the seemingly paradoxical ability to resolve objects smaller than the wavelength of light itself! How did all of this happen without you knowing? That I don’t know, but I can tell you how we got here, and how it’s going to change neuroscience….
The fluorescent revolution
Over the last 30 years, light microscopy has been the subject of continual innovation that has dramatically changed microscope design and capability, making them much more prominent in biological research. This is particularly true in the field of neuroscience where visualizing the intricate and diverse morphologies of neurons (brain cells) demands both specific labelling and high-resolution imaging. In fact, it was the desire to view neuronal “wiring” in three dimensions that motivated Marvin Minsky to theorize and build the first confocal microscope in the sixties. Already a paradigm shift away from the old-school, confocal microscopes of the time used narrow light beams instead of full-field illumination to repeatedly scan over biological samples. This focused illumination, combined with a pinhole aperture placed in front of the photon detector enabled early confocal microscopy to exclude background light originating outside the plane of interest, thus granting a large increase in contrast which allows for better visualization of neurons. It was not until the eighties that confocal microscopy saw widespread adoption by the scientific community.
The benefit of all types of fluorescence microscopy is that, unlike conventional light microscopy, the fluorescent molecules which compose the image can be targeted to specific features of a cell to highlight it, such as the plasma membrane or nucleus. Since these are the only things in the sample that fluoresce, we can see them, even when surrounded by non-fluorescent tissue. The figure below shows inhibitory interneurons in mouse brain tissue. These cells are genetically engineered to produce the fluorescent molecule GFP intracellularly, causing whole cells to glow green. We can trace the fine morphology of individual axons even though they are deep in the tissue.
The combination of confocal microscopy with digital cameras for image acquisition and the discovery of fluorescent proteins like GFP (left image) turned microscopy from a qualitative to a highly quantitative tool. The discovery of bio-compatible fluorescent molecules, like GFP, paved the way for engineered sensors capable of fluorescing only during neuronal communication or ‘firing’. This makes it easy to see which neurons are active and when, both temporal and spatial resolution. All you need is genetically engineered living brain tissue under the microscope! Which is actually not so easy….
Translating neuronal conversations
You see, communication between neurons consists of two components. (1) An electrical charge that accumulates at one end of the neuron can rapidly traverse down the fine dendrites and axon. (2) This triggers the release of chemical messengers into the synaptic cleft, the tiny junction between communicating neurons. This chemical component requires a local and temporally restricted influx of calcium ions at the point of contact between the two neurons, which is a smoking gun for synaptic transmission, aka neuronal communication.
With a great deal of knowledge about the mechanics of both action potential generation and synaptic transmission scientists have been able to (as is so often the case) modify proteins from nature that are already involved in neuronal communication. With a little genetic tweaking, fluorescent molecules like GFP can be attached to these proteins in such a way that light is only emitted from neurons during either a local increase in calcium concentration or a change in their electric charge which results from synaptic transmission or action potential firing respectively.
Being able to simultaneously image the electrical and chemical activity of multiple connected neurons is a dream of neuroscientists as it would allow them to link morphology with function when studying neuronal networks. When done well, direct visualization of neuronal communication can grant huge insights into how networks of neurons respond to stimuli, process information and ultimately give rise to aspects of behaviour and cognition.
So-called genetically encoded calcium indicators (GECIs) have seen widespread use in neuroscience, particularly in studying sensory systems. In a typical experiment, recording the intensity of a fluorescent pulse over time and from a single synapse is indicative of how that synapse communicates, allowing neuroscientists to watch neuronal communication, as it happens, through the lenses of a microscope. In this way, GECIs report the strength and frequency of a neurons output from a synapse, but tells us nothing about the source that drives the neuron to communicate with its neighbours in the first place. To obtain a complete picture of how neurons process incoming synaptic inputs and respond (or don’t respond), a sensor of the electrical component is needed, a genetically encoded voltage indicator or GEVI, and they’ve been a long time coming.
Since the 1940’s the gold standard procedure for recording the electrical activity of neurons has been to jam electrodes into brains or attach them to single neurons. These two approaches measure either the combined electrical activity of groups of neurons that are close to the electrode, or the extremely precise electrical changes of single neurons. Neither case is optimal as you either generalize clusters of neurons to single “units” or you attempt to understand the entire brain one neuron at a time, and there’s 86 billion of those suckers. The potential of GEVIs therefore lies in their ability to report the electric activity of neurons visually allowing researchers to discern the responses of individual neurons within a network of active neurons and thus visualize their electrical conversations, if you will. GEVIs are great indicators of the inputs fed into a neuron due to their ability to fluoresce even during small changes in membrane potential and thus report small synaptic voltage changes.
So that’s two halves of one very cool possibility, you see where this is going?
That’s right, if a neuron could genetically express both a GECI and a GEVI simultaneously researchers would be able to see which synapses stimulating a neuron caused it to fire, or, how many synaptic inputs are integrated by that neuron before firing an action potential. The potential for uncovering the complexities of neuronal computation are huge, opening up the opportunity to test many existing theories on dendritic integration. The cerebellar Purkinje cells would be a good place to start as they have planar dendritic trees making it easy to image multiple synaptic connections at once.
GEVIs and GECIs could turn Marvin Minsky’s dream of tracing the functional connectivity of neurons into a reality. Simply stimulate a single neuron with a good old-fashioned electrode and look to see which synapses nearby light up with fluorescence from the GEVI in a one-to-many experimental setup. The added benefit here is that the success of synaptic communication can be assessed by looking for GECI signals in the axon terminals of those neurons that responded to the initial electrode pulse.
Light on the horizon
Using light instead of electrodes is non-invasive, high-throughput and, I believe, represents the successor to traditional electrophysiology. Small wonder then that many research groups are exploring the development of new GEVIs. This competition is needed because current GEVIs suffer from several drawbacks including low fluorescence emission in response to small voltage changes and poor sub-cellular localization. It will also be challenging, once GEVIs are perfected, to image multiple neurons with sufficient frame rates to capture the propagation of electrical activity through them. For researchers though, overcoming these challenges will be well worth the work. In addition to GEVIs and GECIs, light (optogenetics) has already been used to stimulate neurons and interfere with intracellular trafficking; so it seems that the future biologist toolbox will have to contain a combination of genetic and laser based toys.
For a more detailed look at the applications of GEVIs and GECIs in modern research:
For a paper using both GEVIs and GECIs to study visual processing:
For an overview of recent innovations in microscopy:
Image sources were open source or belong to the author unless otherwise noted.
Joe Sheppard graduated in 2016 from the GTC Master’s program.