A glowing cell tells you a label is there, not that it is on the right molecule. Dobrucki warns that specificity of labeling "should not be taken for granted": DRAQ5 binds DNA at low concentrations but also RNA when DNA sites are saturated, as do "most, if not all, fluorescent labels that exhibit affinity for DNA." Antibodies stray through cross-reactivity with homologous proteins, non-specific binding in highly charged tissues, and background from Fc receptor binding on immune cells. Isotype controls and negative tissue controls are needed to validate specificity.
The microscope is the second choice, and Combs and Shroff's rule is to match it to the sample. A 4 to 5 µm section has little out-of-focus blur, so widefield fluorescence is often sufficient, and on a 4 µm FFPE section it gives results equivalent to a confocal's, faster and more gently. Confocal gives better optical sectioning for thicker sections or when axial discrimination matters. Two-photon excites deeper and bleaches less but is typically overkill for standard histology sections.
Amplification is the third. TSA increases signal greatly, but the amplified signal is no longer linear in target abundance: at high target densities it saturates, and doubling the target doesn't double the signal. That is harmless for a positive-or-negative call and limiting for measuring expression levels. Standard indirect IF keeps better linearity, with a weaker signal.
An antibody does not always bind only its target. Cross-reactivity and non-specific binding can cause false positives, so controls are needed. Thin sections (4 to 5 µm) often need only widefield fluorescence; confocal is for thicker samples or when depth discrimination matters. TSA makes the signal much brighter but loses the linear relationship between signal and target abundance.
