The conditions and relative detection limits are quite different for SDS-PAGE/western blotting vs IHC. It is quite common for antibodies not to detect targets particularly in IHC where they might work on western blotting.

In IHC the tissue is most often fixed meaning that epitopes may be obscured/cross-linked and the protein is nearer its native form with parts of the a.a. chain folded inside. Sometimes epitope retrieval can reverse this process, but I personally don't hold much store by results gained solely from IHC. My old PhD supervisor called IHC a voodoo art. But there may be occasions where antibodies are more suited to IHC than western blotting; depending on how the antibody binds. Most often though they are raised to peptide sequences and hence bind better to peptide chains not in their native conformation. In western blotting, the protein is denatured meaning that such epitopes are available for the anibody to bind.

If you are detecting the protein of interest with your specific Ab(Monoclonal?), then just prove it by several ways (may be different Antibodies/tag system/host or cell specificity etc) with proper controls and finalize by Co-IP combined Mass spec. Then just publish what you saw which was not shown before. Cheers.

to make a short answer...YES, definitely possible. As Ananda pointed out, is your Ab mono or poli? and more...is your WB clean? without other bands? the band is at the right MW?if all of this is YES and you are using a monoclonal Ab, then you should try to use another monoclonal Ab, raised against a different part of the protein. If you still get a signal in WB, then you can try to publish these results, maybe confirming it by mRNA expression, like RT-PCR or in-situ hybridization...good luck!!

Yes, it possible because some antibodies are designed for immunohistochemistry, others for western-blotting or for the 2 ones. The problem is the tridimensionnal aspect of your antigen, in immunohistochemistry it is near its native form but in western-blot with SDS, beta-mercaptoethanol and heating before the loading, the antigen has a completly different form. thus, some antibodies recognize the antigen only in the native form or in its reduced form. You must have a look to your antibody datasheet for more information. Cheers.

It's possible.

Antibodies recognizing three-dimensional structures where binding sites like hydrophobic, hydrophilic, charges ... amino acid residues are at defined points. Therefore it's possible that after different treatments those confirmations are not present anymore. Those agents are salt, water, alcohols, detergents, heat, proteases, ....

If you are using monoclonal antibodies, you can see this often. Using polyclonal antibodies, this will be a rare event.

The conditions and relative detection limits are quite different for SDS-PAGE/western blotting vs IHC. It is quite common for antibodies not to detect targets particularly in IHC where they might work on western blotting.

In IHC the tissue is most often fixed meaning that epitopes may be obscured/cross-linked and the protein is nearer its native form with parts of the a.a. chain folded inside. Sometimes epitope retrieval can reverse this process, but I personally don't hold much store by results gained solely from IHC. My old PhD supervisor called IHC a voodoo art. But there may be occasions where antibodies are more suited to IHC than western blotting; depending on how the antibody binds. Most often though they are raised to peptide sequences and hence bind better to peptide chains not in their native conformation. In western blotting, the protein is denatured meaning that such epitopes are available for the anibody to bind.

Absolutely. It depends on the distribution of the antigen within the tissue. With immunohistochemistry, detection depends on the concentration (and often retention) of the antigen on the site. With Western detection depends on the total tissue concentration.

Hi, Mayur, for immuno you need to have epitope in the domain which can be accessibly by antibody in the native protein, ea. try to avoid trans membrane domain, potentially site of phosphorylations etc... Of course, suggestion that is protein not present by testing only one antibody was not accurate. WB will give you primary evidence about antibody specificity (ea. you should have only 1 band) and protein presence in the tissue. However, if only few cells in the tissue express proteins, your WB signal will be quite low. Than you can try immuno, with the positive control (ea. any other antibody which you are sure is work- just to be sure that your immuno procedure is working well). In my hands there are quite a lot of antibodies work fine in WB, but do not work well in immuno because of specific epitope which recognize only denatured protein ... Good luck!.

Hi Mayur, All the points have been covered by friends above. It is good to check the same antibody on a control tissue (positive control). As mentioned the Ab may not work the same way for IHC and WB.

It is hard to compare your results with the no staining founded by the other group that you cited since you two used different antibodies in different techniques (WB vs. IHQ). If you have seen a single WB band on the right MW using a monoclonal antibody, there is a big chance that the other group reported a false negative result, based on the issues that have been already discussed regarding the IHQ technique.

Thanks everyone for your answers. It helped me to gain insight into addressed issue.

you can always check your antibody and that from the other lab by reading the spec sheets for each antibody. also you can see what is published using the two different antibodies. this is always my first step before purchasing an antibody. it saves time and money! you may also be surprised that some antibodies can be used for both Westerns and for immunohistochemistry.

To add to all suggestions above, it's good to check how the other group fixed the tissue or check if the have done frozen section. Then can figure out the antigen retrieval method. IHC is a bit tricky method, as Peter mentioned a voodoo art.

The difference between a WB and immuno is that the immunocytochemical product is "fixed". That means that the protein/molecule is covalently bound to nearby proteins when the solutions that "fix" the tissue are applied. Thus antibodies that only recognize the nascent protein might not have the parts of the molecule available in the fixed tissue to see them. Altermatively, sometimes the antibodies are generated against the fixed protein and the WB will not work. Some will thankfully work for both.

I find that no useful information is achieved by spec sheets since some companies will only test one of the approaches. Often you have do just test it yourself.

It's also worth considering how the antibody was made. People make antibodies using peptides, folded protein, or fairly denatured (sometimes deliberately, sometimes not) protein. In general, you would expect that peptides and denatured protein will give linear epitopes, whereas folded proteins may well give non-linear epitopes (i.e. will detect only folded protein - sometimes you'll get a hit in IHC and not on a Western, too).

As has been discussed above, linear epitopes are often not exposed, or are damaged, in IHC.

It would be best to find a cell line where it is well established that your protein of interest is expressed, and one where it is certainly not expressed. Then use these as a control for both the Western and the IHC. If you have good positive and negative controls, you are much more likely to convince a reviewer. Sadly, it sounds like you are in the position of having to overturn someone else's published result, and the bar for doing this is always higher.

Yes. Western blotting is much more sensitive than IHC which utilizes fixed tissues. Results are very much dependent on the retrieval methods used. You can increase the chances of a positive result using frozen tissue or fixed free floating tissue. The tyramide protocol is also helpful in getting a positive result when the antigen level is low.

I am not sure that WB is more sensitive over IF. This is largely dependent from gene expression level, especially in the case of multi-cellular organism, when gene repression can restricted only to few specific cells. Even if expression is strong in particular cells, you have quite weak signal in WB. One should just use optimal protocol for IHC.. Good luck!

Looking at all the great answers/suggestions - Do you think that it is wrong to conclude that a protein is present or absent just on basis of Immunofluorescence, Immunohistochemistry and immunocytochemistry only and it is essential to perform western blot analysis also to confirm the presence or absence of particular protein.

The reason I brought this question because several papers conclude the presence or absence of the protein just on the basis of not able to see protein staining using Immunofluorescence, Immunohistochemistry and immunocytochemistry .

Thanks everyone for your valuable input. Many of my questions and doubts are clear now. Really appreciate all of your feedback.

Hi, Mayur,

actually, WB and ICH give you quite different information. Both technique (if, of course, you include all positive and negative control) give you evidence of the presence of protein, In the case if protein is highlky cell-type specfic, WB give you rather information about the relatively ampount of the cells which have the protein, while ICH give you level in certain cells. It is very importnat to know which epitope have been used to Ab production! And structure of this domain in native form (Transmembrane, or other). Usually,WB can recongnise more broad amount of epitope, to compare with ICH. So, ask form the company about epitope. Goog luck.

It is possible to detect the protein in western but not immunohistochemistry and vice versa. Need specific antibody made for histochemistry. Sometime in tissue the protein of interest is shielded, need to retrieve. It also depend how you have fixed your tissues? how thick sections were? how long you have incubated with antibody?

I read an article about antibody validation http://www.biotechniques.com/news/Antibody-Validation-Whose-Job-is-it/biotechniques-347556.html?autnID=328597. It points out commercial antibody may not been validated for a particular application. It also raises the question: manufacturers or researchers should be responsible for antibody validation? Apparently, it is best to validate an antibody before any experiments of either Western blot or IHC.

If it so, how much concentration of proteins can IHC can detect

It's difficult to give a concentration for dection limit of the protein because obviously in IHC the antigen can be difuse or concentrated. I suppose the limit would be best stated as an amount of protein per surface area. I haven't seen a method for calculating this but I suppose some kind of mass-spec imaging would be needed to really quantify it.

The detction limit for IHC really depends on the detection method, the sample and the specificity of the antibody becuase all of these will change the signal to noise ratio as well as the sensitivity. For example, using tyramide signal amplification you should in theory be able to detect close to single molecules of protein but you would need a highly specific antibody.

Yes, it is possible, because some antibody work only in IHC, only  in Western-Blot or in both cases. It is possible that the epitope is hidden in IHC and more accessible when the protein is linearised in Western-Blot.

Does the antibodies recognize the same epitope of ERK5 ?

Ha, I have the reverse problem.

Is it possible that one can detect a protein by immunocytochemistry but not in western blotting?

My protein of interest gets detected in IF, but not by western. Have used different tags. GFP and HA. Both works fine in IF. Protein is expressed and localized properly. No cell death or protein aggregation/degradation issues, based on IF info. Absolutely no signal in untransfected cells. Transfection efficiency was fairly good- 50-60%.

But only the GFP fusion gets detected in western, but not the HA-tag fusion. Absolutely blank lanes, no sign of degradation products/smear either. The same insert got cloned into both vectors. Both were N' fusions and were sequenced to confirm their fusion and frame. No question of epitope masking issue as it is an SDS-PAGE-Western.

This reverse issue is more maddening and will be glad if anyone has any thoughts/suggestions.

One must be aware of how an antibody is produced and how extracts are prepared for WBs to understand the relationship between WB and ICC.  For small proteins or molecules (like serotonin, GABA, histamine, for example), the proteins must be linked to larger foreign protein or modified in some way to make the molecule antigenic.  This makes the immune system "think" the molecule is foreign.  In addition the linking process of fixation  changes the availabe epitopes - some normally present are hidden, others exposed in ways that may not be apparent when the nascent protein is in solution.  Thus some antibodies may work for WB (that uses either denatured protein or nascent protein) but NOT work when the molecule is linked by fixation.  Conversely some antibodies will ONLY recognize the bound ("fixed") protein and will never detect protein that is just floating around in a solution.  One also needs to be aware that when WB are used for protein detection MUCH higher concentrations of the antibody are required compared to staining on tissue.  The difference can be more than 100-fold.  Thus for antibodies of low titer, a WB may never provide enough antibody to detect in WB; whereas the tissue can provide antigen that would be detected by the more sensitive methods of immunocytochemistry. Please be aware that standard IF is the LEAST sensitive of immunohistochemical staining methods (see attached document).  Thus I would never use that method for single staining; amplified methods of immunofluorescent techniques alleviate that problem.  Moreover, be also aware that the time used for the antibody-antigen reaction in immunocytochemical methods (and WB) affects the concentration of primary antibody needed for signal detection.  Short times: higher concentration.  The difference between 24 hr in immunocytochemical reactions and 48 hr will require 3x more antibody for the shorter time.  Using only a couple of hrs increases the need to 20-100x! As an example an antibody that stains at 1: 800,000 for a 48 hr incubation would require 100x more antibody (1:8,000) if only a couple of hrs incubation was used.

Thanks Gloria. Yes, thinking back, it just occurred to me that it is possible that the monoclonal antibody may be recognizing a complex 3D epitope, which is intact in fixed cell samples but may have got disturbed in westerns and therefore difficult to detect. And GFP being a bigger tag, may have partially helped to retain the epitope, while the small HA tagged is not able to do so. But this is only an assumption. Now I am thinking of using a polyclonal antibody for the same protein and see if it helps. I am hoping that this would resolve the problem, if everything else about the constructs are fine, as we believe. Can't wait to check it ASAP. Will post my observation soon. Thanks again.