Jump to content

B cell ELISpot questions


sebbeols

Recommended Posts

Hey there Mabtech!

 

I have a few questions about the B cell ELISpot in general and also some on how I should handle the analysis of my data. Plus a question on my sample material. 

 

General questions:

  • According to U-Cytech's ELISpot protocol, they recommend that one washes the back of the membrane between the detection Ab step and the SA-enzyme step to reduce background staining. I have never tried this and find it a little odd, do you have any thoughts on this?
  • With my B cell ELISpot I am studying the plasmablast response after vaccination in NHPs. I have found mixed opinions about the viability of plasmablasts after freeze/thawing, do you have any experience with this? Nakaya et al., Nat Immunol 2012, write that plasmablasts don't survive freeze/thawing but do not show any data to support this claim. I have yet to try running my B cell ELISpot for plasmablasts from frozen samples, but will do so soon. Do you recommend resting the cells o/n after freeze/thawing?

 

Experiment specific:

  • When I run my B cell ELISpot, I always make a 3-fold dilution down the plate (see attached image). To back-calculate the number of ASCs/million PBMCs, I have had a hard time finding guidance on which dilution well I should use. Is there a certain range of spots I should aim for? Should I be consistent with which column I choose to count from? Can I do an average from multiple dilution wells (i.e. columns)? Each sample is run in duplicates (2 rows). 
  • For my Ag-specific wells I coat them with our vaccinating antigen. I also include an irrelevant antigen (OVA) which I coat a well with. What should I do with the spots that get counted in the irrelevant antigen well? I use this negative control well to optimize the counting algorithm for the AID program, but I will never be able to eliminate all unspecific spots without also losing positive spots. Should I subtract the negative spots that are counted from the Ag-specific spots? 

Thanks in advance for your help and feedback!

/Sebastian

 

post-0-0-01825500-1460457683_thumb.png

Link to comment
Share on other sites

Dear Sebastian,

 

Great that you made it to the Mabtech Forum! Both me and Dr. Jens are deeply impressed by your inclusion of an avatar :)

 

  • According to U-Cytech's ELISpot protocol, they recommend that one washes the back of the membrane between the detection Ab step and the SA-enzyme step to reduce background staining. I have never tried this and find it a little odd, do you have any thoughts on this?

     

     

 We do not agree with Ucytechs recommendation. Washing 5x with PBS from the top is sufficient in our minds for keeping membranes free of background staining. One potential problem with washing the backside of the membrane is that you will introduce leakage during the detection steps. You see, if any drops of liquid land on the backside of the membrane, subsequent detection reagents can get pulled through the membrane by capillary force. The membrane becomes permeable and the whole assay becomes more time consuming to develop. 

 

 

  • With my B cell ELISpot I am studying the plasmablast response after vaccination in NHPs. I have found mixed opinions about the viability of plasmablasts after freeze/thawing, do you have any experience with this? Nakaya et al., Nat Immunol 2012, write that plasmablasts don't survive freeze/thawing but do not show any data to support this claim. I have yet to try running my B cell ELISpot for plasmablasts from frozen samples, but will do so soon. Do you recommend resting the cells o/n after freeze/thawing?

     

     

 

We do not have direct experince with NHP plasmablast after vaccination, but we investigated human plasmablast from frozen samples. It works. We only rested the cells for 1h. We will come back with a more elaborate answer later on. 

 

 

  • When I run my B cell ELISpot, I always make a 3-fold dilution down the plate (see attached image). To back-calculate the number of ASCs/million PBMCs, I have had a hard time finding guidance on which dilution well I should use. Is there a certain range of spots I should aim for? Should I be consistent with which column I choose to count from? Can I do an average from multiple dilution wells (i.e. columns)? Each sample is run in duplicates (2 rows). 

     

     

I am no statistics expert, but in my world it would be best to use a well with a high spot  number to back-calculate the number of ASCs/million PBMCs. If you select a well with low number of spots, the sample size is smaller, and the likelihood of an unfair number purely based on chance is higher.

 

Another aspect to this question surrounds the use of counting algorithm. You say that count settings have been adjusted in consideration of the background seen in the irrelevant control. From just looking at your results, it actually looks to me like spot numbers are abit much on the low side. Empahsis HUGE has a tendency to do that. Take for example D1 and D2: my eyes tell me that D1 has more spots than D2, but the reader says something else. 

 

I totally understand how you have set up the reader and why it is reporting these numbers back. However, when it obviously start reporting back incorrect numbers, you know its time to act. Maybe you should find another count setting where the spot-number titrate more correctly. If you manage to do that I think that back-calculations will make more sense.

 

 

We have talked abit here at Mabtech about your use of Ovalbumin (OVA) as an irrelevant antigen. Maybe these monkeys actually have B-cells that secrete Ig's against OVA? Is it beyond all reasonable doubt? Would it perhaps be better to only use the proteins found in your blocking medium as the negative ctrl? In Mabtech's protocol, we block our plates using cell culture medium. What do you use?

 

If still want to use OVA as the negative ctrl antigen and they are in fact true background spots I think their numbers should be subtracted from the positive count. However, I would play with intensity parameter instead of size. It is quite obvious difference in how strong the spots are. I would also consider changing the emphasis to BIG instead of HUGE.

 

This was a quick response. I will maybe edit my post later to clean up my english. 

Link to comment
Share on other sites

Thanks for your quick and informative reply Christian! I recommend that you and Dr. Jens also get yourselves a more personalized avatar. :) (mine was automatically added by a service called Gravatar)

 

 

 

 We do not agree with Ucytechs recommendation. Washing 5x with PBS from the top is sufficient in our minds for keeping membranes free of background staining. One potential problem with washing the backside of the membrane is that you will introduce leakage during the detection steps. You see, if any drops of liquid land on the backside of the membrane, subsequent detection reagents can get pulled through the membrane by capillary force. The membrane becomes permeable and the whole assay becomes more time consuming to develop. 

 

Your reasoning makes perfect sense! Thanks for clearing up my doubts. Dr. Jens mentioned something about Mabtech not recommending the use of PBS-T, could you elaborate on that? 
 

 

 

We do not have direct experince with NHP plasmablast after vaccination, but we investigated human plasmablast from frozen samples. It works. We only rested the cells for 1h. We will come back with a more elaborate answer later on. 

 

That's great news! I'll try an hour of rest as well. 
 

 

I am no statistics expert, but in my world it would be best to use a well with a high spot  number to back-calculate the number of ASCs/million PBMCs. If you select a well with low number of spots, the sample size is smaller, and the likelihood of an unfair number purely based on chance is higher.

 

Another aspect to this question surrounds the use of counting algorithm. You say that count settings have been adjusted in consideration of the background seen in the irrelevant control. From just looking at your results, it actually looks to me like spot numbers are abit much on the low side. Empahsis HUGE has a tendency to do that. Take for example D1 and D2: my eyes tell me that D1 has more spots than D2, but the reader says something else. 

 

I totally understand how you have set up the reader and why it is reporting these numbers back. However, when it obviously start reporting back incorrect numbers, you know its time to act. Maybe you should find another count setting where the spot-number titrate more correctly. If you manage to do that I think that back-calculations will make more sense.

 

That's true. I have aimed at assessing column 2 which tends to have around 60 spots per well. Column 1 is usually too dark and the spots are not counted correctly. I do have some more work to do with the counting algorithm though. Thanks for your tips on adjustments! I think a mix of improving the camera settings and adjusting the count settings might do the trick. 
 

 

We have talked abit here at Mabtech about your use of Ovalbumin (OVA) as an irrelevant antigen. Maybe these monkeys actually have B-cells that secrete Ig's against OVA? Is it beyond all reasonable doubt? Would it perhaps be better to only use the proteins found in your blocking medium as the negative ctrl? In Mabtech's protocol, we block our plates using cell culture medium. What do you use?

 

If still want to use OVA as the negative ctrl antigen and they are in fact true background spots I think their numbers should be subtracted from the positive count. However, I would play with intensity parameter instead of size. It is quite obvious difference in how strong the spots are. I would also consider changing the emphasis to BIG instead of HUGE.

 

I find the spots in the OVA well really strange too. I highly doubt the NHPs would have been exposed to it previously and could have antibodies against it. It must be some sort of over reactivity/unspecific binding. But then again I don't know. If subtracting the spots, should I only do so from the Ag-specific spots or from the total IgG spots too? It makes more sense to only subtract from the Ag-spec but I don't really have any experience. 
 
For the blocking I also use cell culture media (R10: RPMI 1640 + 10% FBS + 1% L-G + 1% P/S) for 1 hour. Most papers I've read on B cell Elispots block for 2 hours with R10, do you think that can make a significant difference? In my first attempts at the the B cell ELISpot, I included a "No Ag" control well but was later recommended by a more experienced (adapted my protocol from them) group to include an irrelevant Ag instead. Do you think the "No Ag" control is better (I do remember having an average of 1-2 spots there)? 
 
When you write "play with intensity parameter instead" do you suggest that I should add a "maximum intensity threshold"? I find that playing with the minimum only subtracts from my Ag-specific wells as their intensity is usually read as lower because of the high background tint in those wells compared to the other wells.
 
 
Thanks a lot for all your help!  
Link to comment
Share on other sites

Your reasoning makes perfect sense! Thanks for clearing up my doubts. Dr. Jens mentioned something about Mabtech not recommending the use of PBS-T, could you elaborate on that? 

 

 

We are happy to help!

 

You are right, we do not recommend adding tween into wash buffer. The reason for this is simple: in our hands, with our protocol, it leads to darker ELISpot membranes. There is absolutely no benefit. 

 

At the same time, other ELISpot practitioners are adamant on the use of Tween as a critical component in making ELISpot results look better. How can our viewpoints be so different? Well, it all comes back to how the protocol is setup. We always recommend that you should EtOH treat your plates prior to coating. This increases the binding capacity of the PVDF membrane, and combined with a good amount of capture antibody (1-1.5 ug/well), will always produce better results compared to not using EtOH pre-treatment. With our protocol, Tween provides no benefit and will actually "hurt" your results by making membranes noticeably darker after the plate has been developed. 

 

However, in cases where you decide not to perform the EtOH pre-treatment, results will actually benefit from the use of Tween. Spots will appear more clear and the "dark-membrane effect" seises to exist, or atleast becomes much less noticeable.  

 

 

I find the spots in the OVA well really strange too. I highly doubt the NHPs would have been exposed to it previously and could have antibodies against it. It must be some sort of over reactivity/unspecific binding. But then again I don't know. If subtracting the spots, should I only do so from the Ag-specific spots or from the total IgG spots too? It makes more sense to only subtract from the Ag-spec but I don't really have any experience. 
 
For the blocking I also use cell culture media (R10: RPMI 1640 + 10% FBS + 1% L-G + 1% P/S) for 1 hour. Most papers I've read on B cell Elispots block for 2 hours with R10, do you think that can make a significant difference? In my first attempts at the the B cell ELISpot, I included a "No Ag" control well but was later recommended by a more experienced (adapted my protocol from them) group to include an irrelevant Ag instead. Do you think the "No Ag" control is better (I do remember having an average of 1-2 spots there)? 

 

 

I agree, it is unlikely that NHP have been exposed to OVA, but what about their diet? Could there be eggs in there? One of my collegues pointed out that a certain percentage of Igs are very "sticky" and will attach to almost anything. This is especially true for IgM, but we occasionally see it for IgG as well in some human donors. We believe that the "No Ag" control is better and more relevant. If the OVA spots are due to specific IgGs that actually exist in these monkeys, the spots numbers should not be subtracted. In our minds the "No Ag" control is a relevant one that is enough for these experiments.

 

The blocking time of either 1h or 2h should not make any difference. 

 

When you write "play with intensity parameter instead" do you suggest that I should add a "maximum intensity threshold"? I find that playing with the minimum only subtracts from my Ag-specific wells as their intensity is usually read as lower because of the high background tint in those wells compared to the other wells.

 

 

No, not maximum. I would increase the minimum intensity threshold to something like 50. What if you increase camera setting quite abit so that elispot images become much lighter, and then change count setting to:

 

Intensity minimum: 50

Size minimum: 20

Gradient minimum: 1

Algorithm C, emphasis BIG

 

That will maybe differentiate the dark antigen specific spots when you coat with antigen, compared to the really faint background spots in you OVA control. 

 

By the way, have you seen our tutorial video on how to setup the AID reader?: 

 

It might be helpful. 

 

 

Another aspect we have not talked about is the possibility of changing your experimental setup where you label your antigen with biotin, instead of using it for coating. We describe it abit here: https://www.mabtech.com/knowledge-center/assay-principles/elispot-assay-principle/b-cell-elispot

 

In addition, you can read this paper from 2009 where it was first introduced: http://www.ncbi.nlm.nih.gov/pubmed/19696434

 

By reversing the system you consume much less antigen and the system many times becomes much more sensitive. It is easy to biotinylate antigens these days, many ready made kits are available.

 

Furthermore, we have just recently demonstrated in a JIM publication that the same approach can be done using FluoroSpot with several peptide labeled antigens at the same time. You are then able to look at cross-reactivity at the single cell level: http://www.ncbi.nlm.nih.gov/pubmed/26930550

 

 

It is the ultimate revenge of the nerds!

Link to comment
Share on other sites

 

 

You are right, we do not recommend adding tween into wash buffer. The reason for this is simple: in our hands, with our protocol, it leads to darker ELISpot membranes. There is absolutely no benefit. 

 

Ok, I'm testing this out today. Will see how it works for me. Do you have a detailed ELISpot protocol to share? I'd very much like to compare its contents to my own. 

 

 

 

I agree, it is unlikely that NHP have been exposed to OVA, but what about their diet? Could there be eggs in there? One of my collegues pointed out that a certain percentage of Igs are very "sticky" and will attach to almost anything. This is especially true for IgM, but we occasionally see it for IgG as well in some human donors. We believe that the "No Ag" control is better and more relevant. If the OVA spots are due to specific IgGs that actually exist in these monkeys, the spots numbers should not be subtracted. In our minds the "No Ag" control is a relevant one that is enough for these experiments.

 

I'll have to investigate this further. Thanks for your insights. 

 

 

 

The blocking time of either 1h or 2h should not make any difference.

 

Great! Saves me some time :) 

 

 

No, not maximum. I would increase the minimum intensity threshold to something like 50. What if you increase camera setting quite abit so that elispot images become much lighter, and then change count setting to:

 

Intensity minimum: 50

Size minimum: 20

Gradient minimum: 1

Algorithm C, emphasis BIG

 

That will maybe differentiate the dark antigen specific spots when you coat with antigen, compared to the really faint background spots in you OVA control. 

 

I'll try this out tomorrow! Will have lots of plates to run and time to fiddle with the settings. I'll get back to you with how it works out.

 

 

 

By the way, have you seen our tutorial video on how to setup the AID reader?

 

Yes, I have! Very useful. Recognised your voice from the video when you presented yourself at the ELISpot Resource Group Workshop last week! :)

 

 

Another aspect we have not talked about is the possibility of changing your experimental setup where you label your antigen with biotin, instead of using it for coating. We describe it abit here: https://www.mabtech..../b-cell-elispot

 

In addition, you can read this paper from 2009 where it was first introduced: http://www.ncbi.nlm....pubmed/19696434

 

By reversing the system you consume much less antigen and the system many times becomes much more sensitive. It is easy to biotinylate antigens these days, many ready made kits are available.

 

Unfortunately we do not have enough of our antigen to attempt biotinylation with a commercial kit. And that 2009 paper was one of the first I read actually, Nilla (the PI of the paper you sent) and her group have helped me setup my ELISpot protocol in our lab. She recommends around 1 mg of antigen to attempt biotinylation. If we manage to get ahold of more antigen in the future from the vaccine manufacturers then we will definitely attempt it as we see many benefits with this. 

 

 

 

Furthermore, we have just recently demonstrated in a JIM publication that the same approach can be done using FluoroSpot with several peptide labeled antigens at the same time. You are then able to look at cross-reactivity at the single cell level: http://www.ncbi.nlm....pubmed/26930550

 

I have yet to read this paper but I have it printed on my desk. For the current vaccine trial we're only interested in one conserved epitope and IgG isotype only. But it might be an interesting technique to implement in other studies we may conduct! 

 

 

 

Link to comment
Share on other sites

Do you have a detailed ELISpot protocol to share? I'd very much like to compare its contents to my own. 

 

 

I have attached our ELISpot Basic protocol for human/NHP IgG. It goes through detection using both coated antigen and biotinylated antigen. There is no tween used anywhere in the assay. 

 

I think you will see that your membranes will become much less dark after you skip the tween.

 

 

best,

Christian

Mabtech_IgG_protocol_ELISpotBasic.pdf

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...