Survey of the use of platelet transfusions in centres participating in MRC leukaemia trials

Kojima et al (1998) recently reported the finding of HTLV-I tax sequences in fresh peripheral blood mononuclear cells (PBMCs) from a patient with T-prolymphocytic leukaemia (T-PLL) who was HTLV-I seronegative. They were unable to amplify other parts of the HTLV-I genome and suggested that such findings reflected the presence of a deleted HTLV-I provirus in T-PLL cells. However, we note the failure to detect HTLV-I by Southern blot in this case despite a peripheral blood lymphocyte count of 7·9 × 10/l (Kojima et al, 1998). This argues strongly against the presence of the apparently deleted HTLV-I genome in the tumour cells and therefore against an involvement in the pathogenesis of T-PLL. We also note that the one nucleotide deletion in the tax sequence would have caused a frameshift, arguing against tax being involved in the transformation process. The findings of Kojima et al (1998) contrast with those of the largest study on HTLV-I and -II in T-PLL to date (Pawson et al, 1997). 36 patients from HTLV-I non-endemic areas were examined by our group using PCR on DNA from fresh PBMCs and from PBMCs after short-term culture (STC) using primers against all parts of the HTLV-I genome (tax/rex, LTR, gag, env, pol ) and against HTLV-II pol and gag. Rigorous precautions against contamination of the PCR were taken and several negative and positive controls per PCR run were included. Reverse transcriptase (RT) activity was measured on supernatants from STCs using a sensitive PCR-based technique. No HTLV-I or -II sequences were found by PCR and no RT activity was detected in any of the 36 cases (Pawson et al, 1997). The sensitivity of the PCR in our study ranged from the detection of one HTLV-I/II positive cell in 15 000 cells in simple PCR to one in 3 × 10 cells in nested PCR. The application of nested PCR to cultured PBMCs and of a highly sensitive PCR-based RT assay should allow detection of any culturable retrovirus with a RT of similar functional characteristics to that of HTLV-I/II. If HTLV-I/II proviruses were present at a level below the lowest limit of detection here, i.e. one HTLV-I/II positive cell in 300 000 cell equivalents, a direct role of the virus in the pathogenesis of T-PLL would be unlikely. Our findings are in agreement with those documented in 48 T-PLL cases studied previously (Matutes et al, 1991; Pandolfi et al, 1985; Pombo de Oliveira et al, 1995; Fouchard et al, 1995). 27 cases tested previously by our group were HTLV-I seronegative but DNA was not analysed (Matutes et al, 1991). All 13 cases of ‘T-helper phenotype T-CLL’ reported by Pandolfi et al (1985), some of whom corresponded to T-PLL, were HTLV-I/II seronegative and had no evidence of HTLV-I or -II infection by Southern blotting. Five cases of T-PLL from Brazil, an area in which HTLV-I is endemic, were also negative for HTLV-I by serology, Southern blotting and PCR (Pombo de Oliveira et al, 1995), and three French cases with an unusual CD8 phenotype were negative for both HTLV-I and -II by Southern blotting and PCR (Fouchard et al, 1995). These and our findings provide strong evidence against a role for HTLV-I/II infection in T-PLL in patients from HTLV-I non-endemic areas. Positive reports such as that of Kojima et al (1998) may represent technical artefacts or the coincidental presence of a deleted HTLV-I genome in cells other than the leukaemic population.