Isotype controls—time to let go?

Keeney et al. (1) provide a graphic example of where isotype controls don’t work and can in fact lead to erroneous estimations of the target subpopulation. We would agree that for the accurate enumeration of CD34 positive stem cells, isotype controls are not always helpful and indeed there are other examples of rare event analysis where the isotype controls don’t work. We recognize therefore that it is appropriate for some individuals in some laboratories to eliminate isotype controls for some procedures. Admittedly isotype controls have limitations, many of which were expanded upon in the Keeney perspective. In fact we could probably be convinced that for every example where isotype controls are currently being used there is some other reagent, tool or advanced gating strategy which could replace them. However in response to the perspectives suggesting that isotypes controls be let go, we will provide examples where isotype controls are useful and indeed recommended. Basically, the isotype control provides a ‘‘negative control.’’ Cells stained with these products serve to indicate the amount of fluorescence that is emitted by cells labeled with a monoclonal (in most cases) antibody that is not specific for any protein on those cells. This ‘‘negative control’’ is sometimes called background or non-specific fluorescence. In an ideal world, a sample stained with an isotype control reagent allows the operator to set a cursor or a discriminator at a point where any event which generates a signal above this point is specifically stained with the antibody of interest. Unfortunately isotype controls are not perfect negative controls. Right off monoclonal antibodies are proteins secreted by transformed hybridoma cells, therefore although the immunoglobulins may be matched for subclass, there may be subtle differences because these hybridomas are not normal cells. Furthermore, although procedures for purifying, conjugating and repurifying monoclonal antibodies have improved significantly, the procedures are not entirely consistent from lot to lot or from company to company. In the ideal situation of matched isotype and subclass, matched F/P ratio and concentration, there may still be differences in background binding characteristics of the negative controls. These are limitations inherent in the use of isotype controls. We argue however that if their use is approached with an understanding of these limitations then they can be particularly useful and there will be no unrealistic expectations as to the information they can provide. In addition to these limitations, the argument is made (correctly) that in many cases a separate isotype control is not required because you often have a negative control right in the tube of interest. This is correct when the positively stained population of interest is clearly separated from the ‘‘negatively staining’’ population. In fact when you have a clear positive and a clear negative population, you can move the positive/negative discriminator around quite a bit and you will not affect the percent positive cells of interest. With these points taken into consideration, where would you use an isotype control antibody? Isotype controls are particularly useful for a new technologist, a new laboratory, evaluating a new monoclonal antibody reagent and/or in the establishment of establishing a new procedure. The isotype control generates what should be in most cases baseline or background fluorescence and will allow for the optimization of the cytometer and a reference point for negativity of monoclonal antibodystained cells. For example, if a new laboratory with little experience in flow cytometry wanted to measure a marker, which is present on all leukocytes, for example CD45 or CD11b, it might be difficult to optimize the settings without a negative reference point. Would you adjust the gains to have the dimmest populations in the first decade, second decade or third decade? The use of unstained cells as your control for background fluorescence would not be appropriate because monoclonal antibodies bind nonspecifically in different amounts on different subsets (e.g., lymphocytes vs. monocytes vs. granulocytes). Staining with a non-specific isotype control antibody allows the user to determine the level of negativity in the subset of interest. As an example, the absence or very low expression of CD11b on granulocytes is used in our laboratory to screen for the Leukocyte Adhesion Deficiency Type-1(2). Since neutrophils express receptors which bind the Fc portion of immunoglobulin they bind significantly more antibody simply by virtue of these receptors. Therefore comparing antigen expression on granulocytes vs. lymphocytes is problematic without first acknowledging that the background staining will be significantly higher on granulocytes. It is imperative to have some indication of the amount of ‘‘background’’ staining generated by the anti-