The polymerase chain reaction (PCR) analysis of DNA extracted from tissue s
ections can be applied to a variety of research and diagnostic protocols. T
o analyze selectively the specific areas of tissue, a direct microdissectio
n of histochemically or immunohistochemically stained sections, if satisfac
tory for PCR, is helpful. However, the influence of various staining method
s on PCR has been poorly investigated. In this study, paraffin sections of
formalin-fixed lymph node samples were histochemically stained with Mayer's
hematoxylin, eosin Y, methyl green, or May-Grunwald solution and immunosta
ined for CD45 using 3,3'-diaminobenzidine (DAB), DAB with cobalt ion (DAB-C
o), or new fuchsin as the chromogen. In addition, unstained sections were t
reated with trypsin, microwave, or pressure cooker, the techniques frequent
ly used in immunostains for antigen unmasking. DNA was extracted from each
section, and the PCR efficiency in amplifying a 110 bp portion of the beta-
globin gene was evaluated by two parameters: the cycle count in which the f
irst visible band was obtained (CYCLEmin) and the maximum amount of PCR pro
ducts (CONCmax). The hematoxylin stain showed a significantly prolonged CYC
LEmin (P < .01) and lower CONCmax (P < .05) in comparison with unstained an
d untreated control sections. The May-Grunwald stain showed a prolonged CYC
LEmin (P < .01), although the CONCmax was not significantly different from
that of the control (P = .051). The eosin and methyl green stains showed no
effects against PCR, In immunostains, the DAB-Co method showed a lower CON
Cmax (P < .05), whereas the CYCLEmin was not prolonged. The DAB and new fuc
hsin methods had no untoward effects. Antigen-unmasking treatments showed d
eteriorating effects on PCR. The trypsin treatment significantly prolonged
the CYCLE,, (P < .01), and the PCR amplification did not reach the "plateau
" level with a maximum of 60 cycles. The PCR efficiency was worse in microw
ave or pressure cooker treatment, with neither CYCLEmin nor CONCmax being o
btained. When target areas from sections for subsequent PCR amplification a
re microdissected, methyl green is most suitable as a dye for nuclear stain
ing. The immunohistochemical visualization with DAB or new fuchsin yields n
o unfavorable effects. A successful PCR amplification may not be expected i
n sections that are pretreated in a microwave oven or pressure cooker.