DAB, once precipitated cannot be removed. As long as you are not looking for proteins that do not localize in a similar expression pattern, you can repeat IHC, even if you are using the same secondary antibody, just look for the appropriate staining pattern. If your mounting medium is non-aqueous, you can remove the coverslip by soaking for a few minutes in xylene, histoclear or clearite, rinse in graded ethanols followed by DDW and then PBS or TBS and proceed with your 2nd primary antibody. You might have to repeat blocking for a shorter period of time before that.

But, you should optimize using different slides before doing it on your samples. 

Hope this helps. 

Thank you very much for the quick answer=)!

As Priyadharshini mentioned, DAB which appear brownish is irreversible and so the staining will be permanent. You can continue second immunostaining by adding pinch of nickel sulphate while preparing the DAB solution at the end so that it imparts a blueish black color on the area of interest (provided the two proteins are localised on differently on the same cells). We can easily identify the brown and bluish black staining on the same sections. Otherwise one can perform multiple labelling using fluorophores. 

Luckily, we didn't perform DAB staining, just worked with fluorophores like Alexa488. Therefore, we will go on with immunostainings using antibody labelling with fluorophores.

But, I would like to know if dipping in graded ethanol is necessary when the slices have been fixed in 4% PFA before... Can't I just add new antibodies after removing the coverslip and washing?

If your tissues are parafin embeded, you will need to deparafinized by going through xylene and grades of ethanols. For aldehydes fixed samples there is no need to do this steps. Begin IHC staining with blocking of unspecific epitopes with 3% BSA (prepare in washing buffer) for 2-3 hours prior to addition of the primary antibody and follow the remaining steps. Repeat the same steps for multiple labeling but make sure the secondary flourophores (say cy3, c5 or FITc) have no overlapping emission range when capture in confocal/fluorescent microscope.