Polymerase chain reaction/denaturing gradient gel electrophoresis (PCR/DGGE): sensitivity, band pattern analysis, and methodologic optimization.

Abstract

Polymerase chain reaction/denaturing gradient gel electrophoresis (PCR/DGGE) has been gaining popularity as a preferred method to determine the clonality of T-cell populations in small or sparsely infiltrated specimens such as skin biopsies. T-cell receptor (TCR)-gamma gene rearrangements are amplified using nested consensus primers in two rounds of PCR and then are separated by DGGE. Sensitivity is better than with conventional Southern blot analysis but not fully defined. In addition to a discrete primary band resulting from a monoclonal TCR-gamma gene rearrangement, there are often weaker secondary bands of unknown origin. Our goals were to define the PCR/DGGE clonal detection threshold, determine the genesis of the multiband pattern, and optimize methods to minimize extraneous bands. Titration studies showed that the sensitivity of PCR/DGGE for detecting clonal T-cell DNA is affected by the polyclonal T-cell content of the background DNA. The detection threshold is 0.001% using keratinocyte DNA as diluent but only 1% with tonsil DNA. Analysis of monoclonal T-cell lines showed that multiple bands can be produced by a single TCR-gamma gene rearrangement. Mixing of inner and outer primer pair PCR products showed that this is an artifact resulting from different sized PCR products produced during the two rounds of nested PCR required for optimal specificity. Repeat DGGE of isolated bands ruled out variable mobility of partially melted PCR products. Reduction of first round PCR product used as second round target from 10 microl to 1 microl, or a decrease of first round primers from 40 pmole to 5 pmole, resulted in diminished secondary bands without compromising primary band intensity. These results show that: 1) PCR/DGGE has a realistic clonal detection threshold of 0.1% to 1%, 2) multiple bands are consistent with a monoclonal T-cell population, and 3) conditions can be optimized to minimize artifactual secondary bands.