See the paper by Shindler and Roth.  Their method is wonderful for staining with two antibodies in the same species. Shindler, K.S. & Roth, K.A. Double immunofluorescent staining using two unconjugated primary antisera raised in the same species. J Histochem Cytochem 44, 1331-1335. (1996).  The amplification of one staining series is found in the attached paper

Thanks a lot Dr. Hoffman for your reply. However, it's not that clear in the paper the precise protocol they used. Does it happen and you have/know someone who has a protocol for this?

The system suggested by Dr. Hoffman seems to be the most feasible solution to the problem. I have not used the technique of Shindler and Roth but the key ingredient they used are tyramides. There is an entire cosmos of tyramide reagents available that are suitable for your problem IF you use FFPE tissues. The tyramide system allows you to "cook" or strip the antibodies off but keep the tyramide conjugated fluorochromes on the tissue. (e.g. Methods. 2014 Nov;70(1):46-58. doi: 10.1016/j.ymeth.2014.08.016. Epub 2014 Sep 19. Multiplexed immunohistochemistry, imaging, and quantitation: a review, with an assessment of Tyramide signal amplification, multispectral imaging and multiplex analysis. Stack EC1, Wang C1, Roman KA1, Hoyt CC2.). For two antigenes, you technically have to do the antibody stripping only once.

An alternative has just been published but requires some extra work:

Methods Mol Biol. 2017;1554:155-160. doi: 10.1007/978-1-4939-6759-9_8.

Hapten-Anti-Hapten Technique for Two-Color IHC Detection of Phosphorylated EGFR and H2AX Using Primary Antibodies Raised in the Same Host Species.

Hagen J1, Schwartz D2, Kalyuzhny AE3..

You did look at the Shindler paper, right? In the Shindler paper the staining of two antibodies is sequential.  First one stains with the TSA amplification protocol (in our titration chapter the method is presented).  Then the second antibody is applied and visualized by use of a direct fluorophore-tagged secondary.  Some features in doing this are important:  1) amplified series should be the weaker of the two antigens and is stained first (you must determine this first by the titrations of each antibody outlined in my paper using NiDAB and ABC staining - if the two antibodies are pretty equal, I usually do an AB BA design to see which looks better) the fluorophore for the amplification should be either Cy3 or the Alexa orange fluorophore or Texas Red. NOT rhodamine not a Green fluorophore.   Why because if you get bright staining with a green fluorophore and switch to filters for red, all the green will appear Red. If you use Rhodamine the problem is that the excitation wavelengths are too broad and then bleed through to other colors; it is also less stable.  2) after conducting the amplified fluorescence method the tissue is rinsed and the second of the primary antibodies is applied.  That antibody is revealed with use of a direct fluorophore-tagged secondary antibody.  Lastly when you know the "right" concentration to use for the amplified staining, you take that concentration, put it on tissue, and apply the direct tagged secondary to a section.  You should see no reaction since the amplification enables such a low primary antibody concentration that a direct-tagged secondary method will be too insensitive to get a signal.  That way you know all color of second reaction had come from the second primary. See the figure in my titration paper of the comparison of methods and fluorescence.  This is called "dilutional neglect" and is the basis of the double labeling strategy in the Shindler paper. If you would like, I can formulate a cookbook method to use in conducting the experiment but tell me what antibodies you are going to use, how the tissue is prepared and processed, and what time/methods you've already attempted.  In the titration paper it should be obvious that just conducting fluorescence without knowing how to determine the "right" antibody concentrations will be much more difficult than using ABC and NiDAB staining.  It is also very very important to use incubation times with the primary of minimally 2 days, and to stain with the peroxidase substrates for 20 minutes (shorter is bad). Shorter antibody incubation times require much more concentrated primary antibodies and will make this method more difficult. If the two molecules are in different cells or different compartments within cells then I would NOT recommend use of a dual fluorescent method. I can provide examples that show what to do that is much better, permanent, and more sensitive than most of the fluorescent methods. We have also learned that the step when a fluorophore labeled secondary is applied, 2 hrs at 37˚ is going to give a stronger label that shorter times at room temperature.

Hi Aala,  if you are using primary antibodies out of the same species ( rabbit) you have to t run two seperatede reactions. The first primary antibodie you have to detect with the Anti-Fab-rabbit antibody labelled with anything,  the second primary ab (rabbit) you schould detect with a Anti-Rabbit (HL) Serum. Important are blocking steps. If you are working with fluorophore labelled secondary abs you have to block unspecific bindings in the second run again.  Are you using streptavidin then, please block again with avidine/biotin block. In the second run it is essential to block the streptavidin of the first run. Do you use DAB staining with and without heavy metal enhancement then you have to block the perxidase of the first detection in the second run. This  sounds  a little bit complicated but these blocking steps are very important to get clear and good results. The advantage of this Fab/HL technique is you can use it in IHC as well as in IF. The priciple of this technique you will find in the attachment. Good luck!

While what Ute indicates is true, the Shindler and Roth paper is much easier to do and amplification of one of the antibody reactions is very cost-effective and enables greater detection that would not be seen with any of the direct-tagged secondary methods when primaries are highly diluted.  That is one reason I find it so powerful.  See this paper. 

There are a number of important considerations: 1) if the antigen has sequences shared among molecules, cross reactivity will not be solved by purification. If you purify an antibody know that sometimes all you are doing is reducing the number of antibody molecules that will react with the antigen. We have learned that two features can easily reduce the changes of cross reactivity. For polyclonal antibodies, cross reactive clones are present in lower concentrations that would be seen with "correct" clones. In a recent paper of ours, Hoffman, GE, Koban, M. Hypothalamic L-Histidine Decarboxylase Is Up-Regulated During Chronic REM Sleep Deprivation of Rats. PLoS One. 2016 Dec 20;11(12):e0152252. doi: 10.1371/journal.pone.0152252. PMID: 27997552, by using a more dilute primary antibody, one can eliminate the cross-reactive clones from being detected, and this is far easier than running the risk of having the high affinity antibodies lost in purification. 2) Another feature that helps in my experience when "cross-reactivity is non specific, is amplifying the signal with tyramine amplification. This enables much lower primary antibody concentrations (up to 50x lower than fluorescence with a direct tagged secondary), and in so doing amplifies only the true bound primary antibodies. see: Hoffman, G.E., Murphy, K.J. Sita L.V. Importance of titrating antibodies. Current Protocols in Neuroscience. 76: 2.12, 12.1-12.37, 2016 PMID: 27367962