We reported a case of T-cell prolymphocytic leukaemia (T-PLL) in which the peripheral blood cells contained an HTLV-I-related tax sequence ( Kojima et al, 1998 ). PCR analysis revealed only the tax sequence, and other parts of the HTLV-I genome were not detected. The patient was seronegative for HTLV-I, and Southern blot analysis using a full-length HTLV-I probe did not detect any monoclonal bands in DNA from peripheral blood mononuclear cells (PBMC).
We were unable to determine if the leukaemia cells contained the deleted provirus. Southern blotting did not detect the provirus in PBMC even when cellular DNA was digested with the internal cutter PstI. We agree with the comment by Pawson et al (1997 ) that the deleted provirus was not monoclonally integrated in the leukaemia cells, which accounted for the majority of PBMC. It is likely that the provirus was present as a polyclonal infection in < 5% of the cells examined, which may be a subset of leukaemia cells or non-neoplastic cells.
Our findings apparently contrast with those of the largest study on HTLV-I in T-PLL ( Pawson et al, 1997 ), and it is possible that the presence of the deleted HTLV-I genome was coincidental in our case of T-PLL. However, we believe that the amplified product in our case was not a technical artefact but indeed the HTLV-I tax sequence. PCR products of the PBMC from our patients, but not from 10 seronegative volunteers, showed 120 bp bands which co-migrated with that of the positive control. Sample preparation and reactions were performed in separate rooms using separate sets of positive displacement pipetes. PCR analysis of a second set of specimens on a different occasion repeatedly detected the presence of the provirus. Furthermore, slot-blot hybridization analysis of PBMC from the patient obtained on a different occasion was performed using another tax primer pair (SK43–44) and a 32P-end-labelled probe (SK45) at a third laboratory, where our PCR results were not known. The presence of the HTLV-tax sequence was also detected by their assay. The sensitivity of our PCR reactions with the HTLV-I tax primer pairs was determined by serial dilution of the HTLV-I-positive cell line HUT-102 in HTLV-I negative DNA, and positive amplification was obtained at up to one HTLV-I positive cell in 1 × 105 cells. We therefore agree with the comment by Pawson et al (1997 ) that the contradictory PCR results in patients with T-PLL are not attributable to the sensitivity of the reaction.
Although the question remains whether HTLV-I plays a direct role in the pathogenesis of T-PLL, the defective HTLV-I provirus has been described previously in seronegative patients with cutaneous T-cell lymphoma (CTCL) ( Pancake et al, 1995 ). It is possible that in certain cases of T-PLL and CTCL, although integration and presumably alteration of normal cell function occurs, the deletion of various regions of the integrated provirus may subsequently limit or prevent active viral replication. In this situation, HTLV-I genome insertion could trigger T-PLL, and no monoclonal proliferation would be detected. Further studies on large patient populations from different geographical origins should clarify whether the HTLV-I plays a role in the pathogenesis of T-PLL.