Loss of GFP Signal

LauraLynch

Hi, folks.

When we were first getting started with the CLARITY technique, we perfused a lot of brains at once and left them in the hydrogel at 4 degrees until we were ready to embed them. When we switched to brains that had been injected with GFP-labeled PRV and to Thy1-EGFP brains, we did the same thing. What we've been seeing is very good structural preservation, a load of autofluorescence, and no GFP signal at all. We're assuming that overfixation (beyond the 2 days post-fix in hydrogel as stated in the published protocol) is the cause of our signal loss. Has anyone else experienced this?

stuberlab

Yes we did as well and still unable to determine the cause of this. Any thoughts..

nicolasrenier

While it's true that the GFP will definitely get whacked by over-fixation, the auto-fluorescence could be the real problem here. I have also lost my fluorescent signal (not from GFP, but from alexa dyes), but my samples are in the hydrogel for only 1 to 2 days at 4°c. The fixation could also cause some amount of autofluorescence.

I'm working on that problem, and will post if I find any workaround.

AReady

Our lab is also hitting the auto-fluorescence barrier - it seems to be showing up as soon as we add the hydrogel in both transcardial perfusions and PFA/Hydrogel hybridized slices. That being said, I've been able to find GFP signal at times, it's just extremely dim in comparison to what we usually see in standard PFA perfusions.

jberry

Are you guys seeing autofluorescence after incubating the sample in FocusClear? Or are you all using a different RI matching solution? I have seen autofluoresence after incubating with 80% glycerol, but I will try FC later this week and see if there is any difference. It is strange that so many of us are seeing autofluorescence when the D-lab hasn't had this problem.

jmeeks

At this time I cleared 3 brains, all of which have suffered from the standard gripe: autofluorescence with diminished intrinsic fluorescence. I have taken pictures during the process and can confirm that the loss of fluorescence occurs during/after ETC clearing. Working on things to avoid it, as we all are. Hope to see progress soon. Still very skeptical that this major issue was only slightly alluded to in the initial report.

LauraLynch

How long did you fix your tissue before embedding? We let ours go too long in fixative (more than a month!), which we suspect diminished the signal.

jmeeks

Not sure if @LauraLynch your comment was directed to me, but if so: we perfused with hydrogel, followed by post-fixing/embedding in hydrogel for ~48 hours. Intrinsic tdtomato fluorescence was not macroscopically affected. After clearing, intrinsic fluorescence was all but gone. Will post pics once we have identified the stage at which it goes away.

LauraLynch

@jmeeks: Thanks for the info. We took a segment of an overfixed (in hydrogel for 2 months!) Thy1-EGFP brain and went straight to PBS-T and our version of FocusClear. We were unable to detect any GFP signal.

joelrosiene

@jmeeks and @LauraLynch, how long are you running your clearing setups? and at what temperatures?

jmeeks

We have been using a relatively slow peristaltic pump (several hundred mL/min) and running for 50-150 hours between 35 and 45 degrees C (20-30V). We get pretty substantial yellowing under these conditions, and are currently moving towards a higher flow rate system to do a better job of clearing away bubbles and heat.

LauraLynch

I'm using a dehumidifier pump (the sort of thing one finds in one's basement attached to the furnace/air conditioner), holding about 2 L of buffer. I haven't measured the flow rate, but it's fast enough that we've got half of it diverted, and there are no bubbles in the inflow. I aim for 40 V and get around 34-36 V, sometimes fewer. We were running at 42 degrees but I've lowered the temperature to 37 by running about ten feet of tubing through a chiller. Our best clearing has been at about 6 days.

Lea

I am wondering if any of you tried to block autofluorescence with sudan black before running immuno. although clarity is supposed to remove all the lipids, maybe it is worthy to give a try

KeithSiew

@Lea I was just thinking that myself! I regularly use 0.3% /v sudan black in 70% EtOH to quench lipofuscin autofluorescence in conventional tissue sections. I would have suspected that lipofuscins were a major contributor to the "false" signal though not sure what the effects of the CLARITY protocol on these would be... perhaps they're stubborn and refuse to leave... even after ETC!!!

My concern with the Sudan black would be that it may result in a black blob essentially and whether the hydrogel-tissue hybrid would tolerate those concentrations of EtOH for extended periods without resulting in massive shrinkage... you'd also need to have many long washes in PBS to remove the excess Sudan Black stain... and may reduce the overall depth of light transmission by absorbing the laser lines (it has been shown to absorb and fluorescence in the red/far-red spectrum... though I've gotten decent results using Alexa633 in the past with sudan black) and it has also been shown to reduce the "actual" signal from fluorochromes too. So there are some drawbacks... but I think a weaker intensity signal is worth the trade-off for a high signal:noise ratio.

Another potential source of autofluorescence would be from the formaldehyde fixation. I often use 100mM Ammonium Chloride and 100mM glycine in PBS to incubate my slides in for 30mins prior to immunolabelling. This deals with any unreacted aldehyde groups left over from the formaldehyde cross-linking reducing fixative-induced autofluorescence which gives off a green/yellowish colour. I've also found sodium borohydride in the past to be quite effective at dealing with this... but due to the effervescent nature of the reaction I'm not sure it would be useful for CLARITY.

Another suggestion my be to expose the Clarified samples to extended periods of UV light (on ice to prevent overheating) to eliminate autofluorescence though there is potential risk of damaging/degrading your target proteins.

When I've cleared my kidneys I'll try a few autofluorescent quenching techniques and report in. But otherwise I'd suggest something like Alexa 633 as a work around for now.

I've a whole load of papers locked away in a folder somewhere that I'll dig out for ye. I was actually putting together a list of protocols about 2-3 years ago when I stumbled across this wonderful summary of techniques about a year ago that pretty much encapsulated most of what I had discovered after trawling through the web:

http://www.uhnresearch.ca/facilities/wcif/PDF/Autofluorescence.pdf

Enjoy! ;-)

Lea

thank you Keith. We should start troubleshooting this problem next week. We will definitely give a try on your suggestions and report back

Richard_Wetzel

Dear @joelrosiene, @jmeeks and @LauraLynch,

Are there any updates from your side on the "Loss of GFP Signal" discussion? I recently did a couple of ETC clearings on sections and whole brains (30 V, 37°C, 3-6 days) and finally recieved a descent clearing. However, after mounting in FauxcusClear I could not detect any specific fluorescent GFP-signal. The background signal was elevated but acceptable.
So far, sections and brains were kept in PFA-hydrogelsolution for 5 days at 4°C. I need to say, that we are working with promoter-driven expression of GFP. It is usually a strong signal and stable throughout standard PFA-fixation and staining procedures. But could it be washed out of the cells during the ETC application? I assume so...

Thanks for your opinion,
Richard

LauraLynch

So, it turns out that our FauxcusClear doesn't work. I just verified that yesterday and was about to post about it when I read your message.

When we made our FauxcusClear, we were still attempting to get transparent tissue. We weren't looking at fluorescent samples. What we'd seen was that it was no more autofluorescent than FocusClear. However, we didn't have a good, fluorescent control until now. We'd been honing in on clearing the tissue; for that, Faux and Focus are almost the same. However, there's something in our recipe, or lacking in it, that is quenching fluorescence. When I compared thin slices of Thy1EGFP neocortex (passively cleared for 4 days) under an epifluorescent microscope, it was obvious that our FauxcusClear kills the signal. So does DMSO. I got high background with FocusClear, but I could still see GFP-labeled neurons.

I agree with @nicolasrenier: ScaleU2 (4M urea, 30% glycerol, 0.1% Triton X-100) is better than FocusClear, even though it does make the tissue swell to about twice its original size. The ScaleU2 solution has far less background than FocusClear.

I'm so sorry if I've wasted anybody's time. It was worth a shot, though, to try to save money. Fortunately, the ScaleU2 solution is cheap and easy to make.

joelrosiene

Thanks Laura,

Richard - I'm currently the process of clearing a brain with a high amount of native fluorescence. I'll mount sections of it in FauxcusClear, ScaleU2, and ScaleA2 to try to get an idea of what I'm looking at. I'll get back in a few days when the tissue has been processed.

Richard_Wetzel

Dear @LauraLynch and @joelrosiene,

Thanks for your immediate reply. That was a very valuable hint. I gonna run some controls without ETC treatment and with FocusClear and come back to you as well with the results.

Richard_Wetzel

Update: I can confirm @LauraLynch's post. FauxcusClear squenches fluorescent signals and FocusClear prevents them. We tested both clearing reagents on mouse brain patella sections (3 mm thickness). The fluorescent signal for GFP-label neurons was good, background moderate. Images were aquired with 2-photon LSM.

Any news from @joelrosiene?

joelrosiene

@Richard_Wetszel

I have become hung up on the fluorescent signal issue for a few days now. In my fully processed tissues, I have had no success in visualizing native fluorescence, and only limited success with antibodies. To make imaging easier, I have been working with large slices, rather than whole brains.

I have found that the embedding step, for myself at least, produces green and far blue band autofluoresence, to the point of obscuring some (very bright) native signal. The brains in question also contain bright RFP (which did not have background), so I continued with the sections through the ETC process.

After ETC, I found native signal to be severely reduced - only a few cells were still visible in red bands, and of course the green band was obscured with autofluorescence. Notably, after ETC, red band autofluorescence increased dramatically.

I may be running my samples too long in ETC.

More relevant to the topic at hand, I did clear two (post-ETC) pieces of multi-color tumor brain in Scale U2, and Fauxcusclear, but was unable to obtain any signal - likely due to the reasons stated above.

I'm still working on it, will hopefully have better results after I run a more gentle ETC.

KeithSiew

@joelrosiene

Any further luck with your samples.

Also has anyone tried RapiClear?