So as I near the end of my grad school career, I’m going back over all the stuff I’ve done in the last 4 years that won’t make it into a publication but that should be made known so that others can benefit. I’ve done a lot of tinkering with tools like viral vectors, or different behavior paradigms. I know just about everyone has had similar experiences, where they try stuff, it works or it doesn’t work, but it doesn’t go further.
I want to open this thread to that. Over the next months, I’ll keep updating with particular things I’ve tried. I encourage you to do the same if you feel comfortable. For the trainees, please make sure you’re supervisor is OK with your sharing. If you’re not sure, ask. I don’t want anyone getting in trouble.
What I tried: I injected AAV9-hSyn-EGFP (from Penn) into the plantar hindpaw of adult mice. The virus was 10^13 vg/ml. I injected 10 ul. I saw some (~10%) transduction in the L4 DRG after injection 3 weeks later. It wasn’t impressive.
Thoughts: AAV9 is known to retrogradely transport, but the efficiency is probably low. There is likely more infection that I can detect with fluorescence microscopy, but the expression is too low to be noticeable. Other serotypes such as AAV6 may be better for certain DRG neuron types (small).
Behavior assay
Here is something I tried a while back and talked about here. I’d love to develop a more objective assay for mechanical pain where the mouse reports on its aversion/preference without having to evoke some kind of response that needs to be scored (I’m looking at you Von Frey)
I crossed Gad2-Cre (https://www.jax.org/strain/010802) with some other R26-LSL reporter lines and found expression in mostly inhibitory neurons (as expected) but also saw some IBA1+ microglia with reporter expression. I also saw this via RNA-seq. I haven’t seen this reported before, but it’s not entirely unexpected. Macrophages have been shown to produce GABA (http://www.jleukbio.org/content/78/2/393.abstract).
When using this Cre line, this is something to keep in mind. Not every reporter will show up in microglia, but some will. Check for it.
I crossed it with some R26-LSL reporters and saw expression neurons (some, not all) but also glia in the white matter. They were most likely oligodendrocytes (I didn’t go further). So, if you were looking for a pan-neuronal Cre, this is not a good one.
Here are some more images. Overall, this FosGFP line is good for activity-dependent labeling. It’s not perfect. There are some Fos+ (IHC) neurons that don’t get labeled, but overall, it’s useful. Could be good for electrophysiology.
The stimulus here was 55ºC heat for 30s, followed by perfusion 2 hours after the stimulus.
I saw a report that achieved high levels of labeling in the DRG by applying virus to the plantar hindpaw after hypertonic saline and scarifying (nail file). I tried it with AAV8-DREADDs from UNC. I didn’t see anything. I didn’t give it another try. I won’t say it doesn’t work, because it clearly does in the paper, but when I tried to do it myself, it didn’t. It could be the virus. Maybe my technique. The hypertonic saline also created some pressure necrosis in some mice.
In Spring 2016, I had an idea for a way to gain permanent genetic access to activated neurons in a calcium- and light-dependent manner. I pitched at the idea at a neuroscience event at Duke.
I assigned the project to a talented undergrad that I supervised, and naively thought that (1) no one else was working on this (2) me and the undergrad had the time/resources to bring this to completion.
Here is the idea video:
We made some preliminary constructs and tried some assays in cell culture. It didn’t work as planned.
In June, two excellent papers came in Nat Biotech from two groups (Ting and Kwon groups) that basically achieved what I was hoping for. It was a ton of work for them, and the work was impressive. I’m glad that this tool was created and hope to apply it at some point. I grossly underestimated the amount of work it would take to execute this project. That’s an important lesson for me going forward. You really need to think about (1) can I do this? (2) do I have the time to do this? (3) who else might be doing this?
I tried to cross this PV-2a-Cre line with some R26-LSL reporters and found much wider expression than expected. This could be real (developmental) expression of PV, or spurious. I didn’t go much further with it to characterize. It could be the Cre, or it could be the reporter line I used. Not every R26-LSL is the same, even though the design is identical for most reporters.
Intrathecal AAV9 works excellently to get DRG neurons in the lumbar region. I injected the same AAV-hSyn-EGFP (10^13), 10 ul, and got very high expression. In light of the recent work looking at promoters in DRGs, I bet CAG would have been even better.
However, it also hits spinal neurons and probably higher CNS areas (as reported)
Inhibitory neuron staining for GAD2 (IHC) in the spinal cord
I tried several GAD2 antibodies in the spinal cord to see if I could see the somata of inhibitory neurons. I used two different GAD2 antibody from Cell Signaling.
GAD2 is known to be preferentially expressed in terminals, and so it’s difficult to see somata. Nevertheless, I gave it a try.
This antibody showed very nice terminal staining. It was difficult to see somata.
Hey Alex, I have done a bit of primate behavior work, and I find this kind of ‘box roaming behavior testing’ a bit intriguing. What if the animals learn to completely avoid the chamber which is harmful to them? These tests are usually long enough to learn aversion. Is there a possibility that the animal will roam in the ‘pain’ chamber for a ‘moderate’ amount of time is the pain is ‘bad’, but not ‘excruciating’? How do we assess the ‘amount’ of pain? I have no rodent behavior experience, but monkeys will absolutely refuse to explore any situation where it has been hurt before. They only work with rewards.
These are good questions. My goal was to create a real-time assay rather than one where the mouse has a chance to learn and make associations (which in some tests, like CPA, is the goal). So basically, if you put a naive mouse in, it has no associations with either side of the chamber. but then, when it starts exploring, if it experiences something painful (aversive), it will avoid that area as it explores. That’s my thought. Amount of aversion (can’t say it’s pain) would be measured by time spent in one side or the other.
Is there a possibility that the animal will roam in the ‘pain’ chamber for a ‘moderate’ amount of time is the pain is ‘bad’, but not ‘excruciating’?
Sure and maybe in that case, you’d miss an effect if it’s not large enough.
For pan-neuronal labeling in the DRG, I’ve moved to Neurotrace, which is a fluorescent Nissl stain. The different colors have variable staining qualities. Here is Neurotrace Blue. I used 1:200 diluted in PBS for 10 mins at room temp. I do a quick wash in PBS and then mount.
The Green Neutrotrace is very strong, but it also stains non-neuronal nuclei too (like real Nissl). I’ve found the blue and Far red to do that less.
This is 1:100 in spinal cord Nissl Green
20 mins at RT. 1x wash with PBS
Blue is nice because you don’t have to take up the green channel. That’s good since we very rarely use a secondary fluor that is blue. It’s generally FITC/488, Cy3 and Cy5 in our lab. So you can do 4 colors this way, assuming you exclude DAPI.
I’ve been trying to do some labeling with the retrograde tracer Fluorogold, and have been running into some issues with the color being a little too close to a DAPI co-stain. Do you think far red Neurotrace would be a reasonable DAPI alternative for co-staining?
Get onto a scope that can refine the different spectra. A confocal should be able to do this
Use another nuclear stain
Use the far red Nissl to mark cells and nuclei
I’ve had great results with the far red Neurotrace. So yes, I’d use that, as long as you didn’t also want to stain with an antibody in Far red.
The Neurotrace dyes also have wide sprectra, and on our scope, the green bleeds into red, and the red bleeds into far red. That’s why I don’t use. Now I’m using either the blue neurotrace or far red.
You could also use another retrograde tracer. I’ve used Fast Blue a lot.
Good luck. Let us know how it goes.
Here is a figure panel I made with Far Red Neurotrace (pseudocolored Green). It’s very nice.
This behavior is 3 weeks post infection. At 1 week, there was no behavior.
Also, at same time period, retro-AAV-Cre injected into the same mouse line did not produce any behavior. So, I’m not optimistic about retro-AAV for mouse paw.
HSV129 is a strain of replication-competent herpes virus that is used for anterograde transsynaptic tracing.
Since HSV naturally infects sensory neurons, I thought this would be a useful tool to trace connections from DRGs to the CNS.
I tried the Cre-dependent HSV129 from the paper above, as well as an HSV-129 GFP that is constitutive. Neither infected any DRG, sadly. I didn’t try a wide range of conditions. I just injected into the paw. This is surprising and I don’t really have an explanation.
In the CNS, this stuff is potent, and the mice eventually die after injection due to the spread and death of neurons. But the DRG neurons seem impervious.
RNAscope probes are really expensive, and it’s a shame to have to dump them after one use. The last time I did ISH, I saved the probe mix by carefully pipetting off the slide.
About 1 month later, we reused the probes. The images look just as good as the first time!
I think the important thing here is to get as much wash buffer off the tissue before applying the probes. If you have residual wash buffer, you will dilute the probe mixture each time you use it, and then over time, the concentration will reduce, potentially affecting your signal. So, if you’re going to re-use, make sure you quickly but completely remove the wash buffer on your tissue (without drying the tissue).
I’m sure this won’t be supported by ACD, but it seems to work.
Certainly a very creative way to reuse the probes - agreed, not recommended - but as long as it works and you’re aware of the outlying factors that may contribute to a poor result…
As always, nice post and beautiful photot @achamess
