Retrograde Fluorescent Labeling Allows for Targeted Extracellular Single-unit Recording from Identified Neurons In vivo

1. Prepare Dye-coated Needles

1. Electrolytically sharpen a 160 μm diameter tungsten wire¹⁵. Final needle tip diameters should range from 5-50 μm. The number of needles needed depends on the size of the region being labeled. We prepared 5 needles for 3-5 injections into the posterior exterolateral nucleus (ELp).
2. The night before the experiment, place a drop (<0.25 μl) of 2 mM dextran-conjugated Alexa Fluor 10,000 MW dye onto the distal 100 μm of each needle.
3. Allow the needles to air dry at room temperature, leaving concentrated dye at the tip. Store the needles at 4 °C in a dark container to protect them from light.

2. Prepare Animal for Surgery

1. Induce general anesthesia by placing the fish in a solution of 300 mg/L MS-222 in tank water.
2. Weigh the fish and measure fork length (tip of snout to fork of caudal fin) and body depth (maximal dorso-ventral distance in the transverse plane). These measurements should fall within the ranges indicated in Table 1 so that the fish fits inside a recording chamber small enough to place under a water-immersion microscope objective (Figure 1).
3. Immobilize and electrically silence the fish by injecting 100 μl of 3 mg/ml flaxedil into the dorsal body musculature.
4. Fill the recording chamber (Figure 1A) with tank water. Place the fish ventral-side down on the platform in the center of the chamber (Figure 1). Deliver an aerated solution of 100 mg/L MS-222 using a pipette tip placed in the fish’s mouth (1-2 ml/min). Stabilize the fish with rods fixed in wax placed on both sides of the body (Figure 1C). Monitor the fish’s health by checking for continuous blood flow in the ocular vessels and a normal body color.
5. Rotate the platform along its long axis and lower the back end of the platform so that one side of the dorsal surface of the fish’s head is exposed above the water while the rest of the fish’s body remains submerged. A small piece of Kimwipe should be placed on any non-submerged portion of the skin to prevent drying.

3. Surgery (Figure 2)

The basic surgical procedure described here is well established and reliably used for blind in vivo recordings in mormyrids¹⁶. For other applications, expose the desired regions for labeling and recording. The region containing axon terminals of the cells of interest must be reachable by a dye-coated needle. The region containing more proximal segments of those same axons must have sufficient space above the tissue to accommodate the working distance of the water-immersion lens (2 mm in our case).

1. Apply a 0.4% solution of lidocaine to the exposed surface of the head using a Q-tip.
2. Using a scalpel blade, cut the perimeter of a rectangular piece of skin. Remove the rectangle using a pair of forceps. The size of the rectangle will scale with fish size, but should be approximately 3 mm X 5 mm for a 6.2 cm fish (Figure 2A). The lateral edge of the rectangle should align with the center of the eye, the anterior edge of the rectangle should be just posterior to the eye, and the medial edge of the rectangle should be just lateral of the fish’s midline.
3. Expand the exposed skull region anteromedially to expose an additional 2.5 mm square non-overlapping area (Figure 2B).
4. Completely clear and dry the exposed surface of the skull using the scalpel blade to scrape away any excess tissue and Kimwipes and forced air to dry the surface (Figure 2C).
5. Glue a metal post to the anteromedial exposed skull region using Super Glue. Wait until the glue is completely dry (Figure 2D).
6. Remove a rectangle of skull, approximately 2 mm X 4 mm for a 6.2 cm fish. Use a dental drill with a ~0.5 mm diameter ball mill carbide tip to thin the perimeter of the rectangle. Then, using a scalpel and forceps, cut the perimeter of the rectangle and peel it away to expose the underlying brain. Additional drilling or cutting with small scissors may be necessary to fully expose EL (Figure 2E). If muscle bleeding occurs, an electrocautery unit may be used.
7. Cut away both the dura mater (pigmented) and the pia mater (clear) using spring scissors or a needle and remove the cut portions with a pair of forceps. The anterior and posterior portions of the exterolateral nucleus (EL) are now visible, ELa and ELp, respectively (Figure 3A).

4. Retrograde Labeling of Axons of Interest

1. Position a manipulator with a dye-coated needle (made in Step 1) above the target region containing axons of interest, in our case ELp.
2. Swiftly insert the needle approximately 25 μm into the tissue. Wait 15-30 seconds, until all the dye has come off, and then retract the needle.
3. Repeat with additional fresh needles as needed, placing each one in a different location so that the dye is distributed throughout the target region. We used 3-5 needles per preparation.
4. Rinse excess dye from the cavity with Hickman’s ringer solution.
5. Wait at least 2 hours for dye uptake and transport.

5. Visualization of Axons of Interest

1. Place the recording chamber, along with the fish, underneath the objective of an upright, fixed-stage epifluorescent microscope. As the body of the fish occludes light penetration, both white and fluorescent light sources must come from above. Careful placement of a fiber optic light source above the skull cavity allows satisfactory brightfield images. For epifluorescence viewing, fluorescence filter specifications should match the absorption/emission spectrum of the dye.
2. Switch respiration to fresh tank water and maintain the same flow rate. Place a ground wire in the exposed brain cavity and connect to the ground of the recording headstage (see 6.3).
3. Place a pair of recording electrodes next to the base of the tail and connect to a differential amplifier and recording device (e.g. audio monitor, oscilloscope, or computer) to monitor the electric organ discharge command (EODC). After the fish recovers from anesthesia, the EODC can be used as an indicator of the fish’s condition.
4. Prepare a scaled sketch of the brain region viewed at low magnification. Include major blood vessels as landmarks (may vary from fish to fish) to identify the exact location of labeled axons visible only under high magnification (Figure 3D).
5. Confirm dye placement. First view entire tissue with brightfield illumination for orientation (Figure 3A). Then view with fluorescent illumination (Figure 3B). ELp will have diffuse labeling (Figures 3B and 3C). Minimize fluorescence excitation to limit the photodynamic and phototoxic effects of the dye.
6. Use the vessels as landmarks to locate ELa under high magnification. Illuminate with fluorescent light while searching for a labeled axon near the surface. (Figure 4).

6. Record Extracellular Activity

1. Pull suction recording electrodes using 1 mm OD, 0.58 mm ID borosilicate capillary glass with filament. Ideal tip size will depend on the diameter of the target axons, which in our case is 0.1-0.2 μm¹⁷. For our application, electrode tip diameters were 1.5 ± 0.4 μm (range: 1.0-2.4 μm) with a 5 mm long, narrow shank in order to approach labeled axons without moving the surrounding densely packed tissue.
2. Fill electrodes with filtered Hickman’s ringer solution. Final tip resistance is 45.2 ± 38.0 MΩ (range: 16 to 155 MΩ).
3. Place the electrode in an electrode holder with a pressure port and connect it to an amplifier headstage mounted on a manipulator. Run a pressure line from the pressure port to a T-junction ending in a manometer and a syringe for monitoring and controlling pressure, respectively.
4. Connect the headstage to an amplifier and an analog-to-digital acquisition device.
5. With 30 mbar outward pressure in the electrode line, place an electrode next to a labeled axon. A low-light level camera interfaced with imaging software is used to visualize pipette placement. Start near the tissue surface and advance the electrode towards the axon. As you approach the axon, the outward pressure should cause a slight, but noticeable movement of the axon.
6. While the electrode is next to the axon, (Figure 4A, top) record the potential at the electrode while presenting test stimuli (in our case, we used 100 msec monophasic positive and negative transverse pulses at an intensity of 20 mV/cm; Figure 1A). Action potentials should not be observed, although an electrical artifact confirms proper recording/stimulation (Figure 4A, bottom).
7. Release the outward pressure in the electrode and repeat stimulation/recording. Action potentials should still not be observed (Figure 4B).
8. Apply slight (125 ± 25 mbar) suction to the electrode and repeat stimulation/recording. Action potentials should now be observed in response to stimulation (Figure 4C). Spontaneous activity may also occur. If action potentials are not observed, release the suction, clear the electrode with slight pressure, move the electrode slightly, and attempt suction again. Once action potentials are visible, close the pressure line.
9. Stimulate and record as desired.

7. Termination and Disposal

1. Once all desired recordings are complete, switch to respiration with 100 mg/L MS-222 until the EODC has stopped. No EODC should be detected for at least 10 minutes.
2. Dispose of the fish according to institutional guidelines and approved animal care protocols.

8. Representative Results

For our particular application, we are interested in studying stimulus coding by central sensory neurons. Successful recordings from labeled axons allow analysis of single-unit responses to sensory stimulation¹⁸. Figure 5A shows representative action potentials evoked by transverse electrosensory stimulation using bipolar electrodes located on the insides of the left and right walls of the recording chamber. Spike times can be presented as a spike raster plot (Figure 5B). A 25 msec pre-stimulus recording window shows the low level of spontaneous activity. This particular ELa “small cell” is long-pass tuned to stimulus duration at a stimulus intensity of 6 mV/cm, increasing the number of spikes per repetition as stimulus duration increases (Figure 5C). The mean first spike latency is 4.28 ± 0.16 msec, consistent with the expected latency for small cells in ELa¹¹.

Figure 1. Specifications for a recording chamber that can fit underneath the objective of a fixed-stage epiflourescent microscope. (A) To-scale square recording chamber made from plexiglass showing top, side and back views. Paired sets of stimulating electrodes (asterisks) at the periphery allow for either transverse (red-black) or longitudinal (blue-yellow) stimulation. An additional piece of plexiglass in the corner, with a rubber-lined hole in the center (orange), holds a suction pipette that maintains a constant water level. Two vertical stainless steel posts screwed into the bottom of the chamber (solid green) connect to stainless steel posts (green outline) attached to the platform supporting the fish (light grey, detailed in C) via adjustable disc clamps. A photograph of the chamber is shown below the to-scale drawing. (B) Individual and assembled views of the circular plastic disc clamps used to secure the platform to the vertical posts. Each disc clamp has a groove (green) for a post and a center hole for the tightening screw. Disc clamps are rotated so the grooves are perpendicular to each other. Tightening the screw (red) clamps the posts in place to prevent further vertical and rotational movement of the platform. (C) Front and side to-scale views of the plexiglass platform used to hold the fish in place. The platform is coated with a layer of paraffin wax (blue) that holds wooden dowels (black bars) in place to support the fish. Tubing for respirating the fish passes through a hole in the ‘headboard’ of the platform and ends in a pipette tip placed in the fish’s mouth. A stainless steel head post (gray bar) connects to the platform via a ball joint allowing 360 degree rotation. Stainless steel horizontal posts (green) are screwed into either end of the platform. Click here to view larger figure.

Figure 2. Schematic overview of surgery looking down at the dorsal surface of the head. (A) Make four cuts, in the order indicated, to remove a rectangular piece of skin (red). (B) Extend the opening anteromedially to remove an additional rectangular piece of skin (green). (C) Scrape off any remaining fat or ligaments by moving the scalpel blade as indicated by the arrow and dry the surface completely with Kimwipes and forced air. (D) Glue a stainless steel post to the skull using Super Glue. Scale bar in A applies to A-D. (E) Use a dental drill to make four cuts, in the order indicated, to remove a rectangular piece of bone (blue), exposing the anterior and posterior exterolateral nuclei (ELa and ELp, respectively). Click here to view larger figure.

Figure 3. Fluorescent labeling in the posterior exterolateral nucleus 3 hours after injection of dextran-conjugated Alexa Fluor 568. (A) The anterior and posterior exterolateral nuclei (ELa and ELp, respectively) visualized with bright field illumination from above. Note that the extensive myelination within ELa gives it a relatively bright appearance that distinguishes it from ELp. (B) The same area visualized using epifluorescence viewed through a TRITC filter. (C) A merged image using blue for A (brightfield) and red for B (TRITC). (D) Example of a scaled drawing of ELa and ELp including major blood vessels (red lines) that can be used as landmarks to identify the exact location of labeled axons visible only under high magnification (the exact location of blood vessels varies from fish to fish). Dotted line indicates the border between ELa and ELp. (E) Sample images acquired using a TRITC filter from 5 different preparations illustrating a range of successful labeling patterns of small cell axons and somas in ELa. Click here to view larger figure.

Figure 4. Single-unit extracellular recording from a labeled axon. (A) Recording electrode (arrow head) under positive pressure placed adjacent to a labeled small cell axon in ELa (top) records only edge artifact (arrow heads) in response to a 100 msec 20 mV/cm monophasic, contralateral-positive, transverse square pulse (bottom). (B) Releasing outward pressure from the electrode (arrow head) causes the axon to move slightly towards the electrode (top) but there are still no responses to the stimulus (bottom). (C) Slight negative pressure pulls the axon into the electrode (top, arrow head) and action potentials in response to stimulus onset are now visible (bottom, asterisk). Bottom portions of all three panels are overlaid responses to 20 repetitions of the stimulus. Click here to view larger figure.

Figure 5. Representative results using this technique. (A) 5 sample traces showing action potentials evoked by a 0.1 msec 6 mV/cm monophasic, contralateral-positive, transverse square pulse stimulus. (B) Raster plot showing spike times during 20 repetitions of a 75 msec recording window for the same unit stimulated at time 0 with 6 mV/cm stimuli at the range of durations listed on the right. (C) Duration tuning curve quantifying the responses displayed in the raster as spikes per stimulus repetition. Click here to view larger figure.

Table 1. Optimal weight, length and body depth ranges for fish. Optimal weight, fork length (tip of snout to fork of caudal fin) and body depth (maximal dorso-ventral distance in the transverse plane) ranges allowing fish to fit in the recording chamber illustrated in Figure 1. Fish that are too small may be less likely to survive the surgery and will have a small ELp, making dye placement challenging. Fish that are too large will have a large, over-reaching cerebellum that will reduce access to ELa and ELp and may prevent lowering of the high power, water-immersion objective close enough to focus on ELa and ELp.

Table 2. Success rates for each dye injection method. Success rates for each dye injection method. Methods are divided based on injection site and dye type. For each method, the total number of fish attempted and the percentage of these experiments that resulted in successful labeling in ELa and ELp is shown. Note that the targeted recording area is the opposite of the application site (bolded boxes). For the injection site, dye uptake was considered successful with labeling of both somas and axons. In contrast, at the recording site, only preparations with labeled axons were counted as successful labeling experiments. The total number of attempted units and the percentage of these units that resulted in successful recordings are also shown.

Table 3. Quantities of dye applied for each of the three methods used to inject Alexa Fluor into ELa Quantities of dye applied for each of the three methods used to inject Alexa Fluor into ELa (first method from Table 2). Dye injection into ELa was performed with both a nanoinjector and a microinjector using either a 33 gauge Hamilton syringe needle or a pulled glass capillary pipette. We varied the number of injection sites, the volume per injection and the total volume of dye injection.