Role of HTLV‐1 proviral DNA load and clonality in the development of adult T‐cell leukemia/lymphoma in asymptomatic carriers

Human T-lymphotropic virus type 1 (HTLV-1) is the causative agent for adult T-cell leukemia/lymphoma (ATL).1, 2, 3 Only a small proportion of HTLV-1 carriers eventually develop ATL after a long latency.4 However, the critical events in the leukemogenic process remain unclear. In general, viral load is an important factor affecting the outcome of virus-associated disease. The HTLV-1 proviral DNA load in the peripheral blood mononuclear cells (PBMCs) of carriers exhibits a wide range of values.5 Enhanced expression of HTLV-1 Tax, which transactivates transcription of viral mRNA and of host genes that control cell proliferation, induces amplification of infected cells (i.e., the number of proviral copies).6 One of the host genes encodes the IL-2 receptor (IL-2R), which is overexpressed on the surface of ATL cells.7 We have observed that subjects with detectable tax/rex mRNA have a higher number of IL-2Rα-positive T cells.8 We also have found a positive association between HTLV-1 proviral load and the level of soluble IL-2R in asymptomatic carriers,5 as well as the number of morphologically abnormal lymphocytes on a peripheral blood smear among carriers.9, 10 In addition, the HTLV-1 proviral load within a carrier is stable over many years.5 It has been postulated that clonal proliferation of HTLV-1-infected cells likely is responsible for maintaining the proviral load level in a carrier.11, 12 HTLV-1 Tax may contribute to the clonal proliferation of HTLV-1-infected cells by promoting their abnormal growth.6 Tax also exerts dysregulation of the cell cycle by binding to its inhibitors and inhibiting some tumor-suppressor proteins.6 The accumulated data support the hypothesis that increased HTLV-1 proviral load and clonal expansion of HTLV-1-infected cells are key to the process of leukemogenesis in HTLV-1 carriers. However, proviral DNA levels and clonality of HTLV-1-infected cells have not been directly evaluated as predictors of the development of ATL in asymptomatic carriers. Given an ATL incidence rate as low as 1 case per 1,000 person-years among HTLV-1 carriers, such a study would require extensive follow-up of a relatively large number of HTLV-1 carriers. The Miyazaki Cohort Study is a population-based prospective study of the natural history of HTLV-1, which was established in 1984.13 During follow-up of the cohort through December 2000, 6 ATL cases were identified through an annual census or reports from next-of-kin.14 Four of the 6 cases had available prediagnostic samples of PBMCs and were included in the present analysis (Table I). Diagnosis was confirmed by medical records for 3 of the 4 cases (cases 1, 3, 4). Lymph node biopsy for case 4 and the PBMCs at ATL diagnosis for cases 1 and 4 were available for analysis. Any symptoms suggestive of smoldering- or chronic-type ATL, such as skin lesions or lymphadenopathy, were not observed at the annual health examinations prior to the diagnosis of malignancy in the cases studied. Routine blood tests performed at the annual health examinations also did not reveal any abnormal findings suggestive of subclinical ATL. Therefore, it is unlikely that these cases had smoldering- or chronic-type ATL before the onset of acute- or lymphoma-type ATL. For comparison of HTLV-1 proviral load, age- and sex-matched HTLV-1+ controls were randomly selected from asymptomatic carriers within the cohort. Informed consent was obtained from all study participants, and the study protocol was approved by the Human Subjects Committees of Miyazaki Medical College and Harvard School of Public Health. HTLV-1 proviral DNA in PBMCs was quantitated using the AmpliSensor assay (AcuGen HTLV-1 Quantitation Test; Biotronics, Lowell, MA).5, 15 Prediagnostic HTLV-1 proviral loads of the 4 ATL cases are shown in Table I. Two cases were male (cases 1, 2), and 2 were female (cases 3, 4). Median age at onset of ATL was 73.5 years (range 64–83). Median HTLV-1 proviral load (copies per 100,000 PBMCs) at the earliest prediagnostic PBMC sample (collected 3–8 years prior to ATL onset) was 4,930 (range 830–10,560). Median proviral load of the 37 age- and sex-matched HTLV-1+ control subjects was 820 (10–9,790), which was significantly lower than the earliest prediagnostic proviral load of cases (Wilcoxon rank sum test, p = 0.03). In conditional logistic regression analysis, which accounted for the matched design (LogXact 4.1; Cytel, Cambridge, MA), there was a significant association between higher viral load and case status; the odds ratio of being an ATL case associated with each 1,000 copies (per 100,000 PBMCs) increase in viral load was 1.42 (95% exact confidence interval 1.04–2.10, p = 0.03). Although the number of cases analyzed was small, the difference in proviral load between cases and controls was striking and statistically significant. Our prospective analysis, thus, demonstrates an association between higher proviral DNA load and the subsequent development of ATL. HTLV-1 proviral DNA levels were stable over several years in asymptomatic carriers.5 In cases 3 and 4, who had PBMC samples from more than 5 years, proviral levels were increased in samples obtained closest to the time of ATL diagnosis (1 and 2 years before onset, respectively). These increased proviral loads may reflect the clonal expansion of certain HTLV-1-infected cells. Because the primers of the pol region were used for the quantitative assay of proviral DNA load, HTLV-1-infected cells with pol-defective provirus could not be detected by the current method. It is possible that the HTLV-1-infected cells with such defective virus were seen more commonly in carriers who developed ATL rather than in the control group because the defective provirus is sometimes found in patients with ATL.16 As a result, the difference in proviral DNA load between cases and controls might have been larger if we had used a more conserved region, such as tax, for the primers. We also examined whether the preleukemic HTLV-1-infected clone can be found prior to diagnosis in the peripheral blood of carriers who go on to develop ATL. To amplify the genomic DNA adjacent to the integration sites of the HTLV-1 provirus, inverse long-PCR was used.12 All 4 cases had clonal bands of various sizes and intensities, indicating oligoclonal or polyclonal expansion of HTLV-1-harboring cells prior to diagnosis (Fig. 1). For case 1, a single band indicating the leukemic clone was detected in the peripheral blood by inverse long-PCR at diagnosis (Fig. 1a). Case 3, a carrier who developed lymphoma-type ATL, had several bands that were consistently strong over several years (arrows in Fig. 1c). These bands appear to represent major clones, consisting of large numbers of HTLV-1-infected cells with the same proviral integration site. One of the bands increased in intensity in the samples obtained close to the onset of ATL (large arrow in Fig. 1c). However, it was not possible to determine whether this clone was the preleukemic one because a sample from the time of ATL diagnosis for this carrier was not available. Inverse-long PCR analysis of HTLV-1-infected cells in 4 cases before and at diagnosis of ATL. (a) Case 1. Triplicate analysis of PBMCs in the sample obtained 5 years before the onset of ATL and single analysis of ATL cells at diagnosis. (b) Case 2. Triplicate analysis of PBMCs in the sample obtained 3 years before the onset of ATL. (c) Case 3. Triplicate analysis of PBMCs in samples obtained 1–7 years before the onset of ATL. Arrows indicate major clones. (d1) Case 4. Southern blot analysis of lymph node biopsy specimen at diagnosis. EcoRI, EcoRI digestion; PstI, PstI digestion. (d2) Case 4. Triplicate inverse-long PCR analysis of PBMCs in samples obtained 2–8 years before the onset of ATL and from PBMC and lymph node (LN) samples obtained at diagnosis. MWM, m.w. marker (1 kb ladder); H, HUT102 cells as HTLV-1-positive control; M, Molt4 cells as HTLV-1-negative control; PC, leukemic cells obtained from an ATL patient, who was not a member of the study cohort, as a control. For case 4, monoclonal integration of the HTLV-1 provirus in the leukemic cells from the lymph node was demonstrated by Southern blot assay at diagnosis (Fig. 1d1). However, the leukemic clone was undetectable by inverse long-PCR, and only some faint bands were seen in the lymph node sample obtained at the time of ATL diagnosis (Fig. 1d2). This observation is not surprising since leukemic clones are reported to be detectable only in about half of patients by inverse long-PCR due to the lack of adequate sites for restriction enzymes or to sequence deficiencies of the HTLV-1 provirus in leukemic cells.12 To identify the preleukemic clone in the peripheral blood of case 4 prior to diagnosis, the nucleotide sequence of the 3′ flanking region of the integrated provirus in the ATL cells first was determined using the inverse PCR method (Fig. 2a,b).17 We then synthesized a primer (5′-AACTGAGGCCTACAGGTCCA-3′) that would specifically amplify the junction site of the 3′ long terminal repeat (LTR). Using this leukemic clone-specific primer, the HTLV-1-infected cell clone with the same integration site of the provirus as the leukemic cells was detectable 2, 5 and 8 years before the onset of ATL, with 2 different primers for the HTLV-1 LTR (primer 4 used in inverse PCR) and gag sequences (primer SG295) (Fig. 2c).18 The intensity of this clone was increased at 2 years prior to ATL diagnosis compared to the intensity at 5 and 8 years before diagnosis (Fig. 2c). Detection of the preleukemic clone in PBMC samples obtained from case 4 before diagnosis of ATL. (a) Inverse PCR analysis of lymph node (LN) specimen obtained at ATL diagnosis. (b) Schema showing subcloned and sequenced DNA of the inverse PCR amplicon derived from case 4 and the locations of primers used for leukemic clone-specific PCR. (c) Leukemic clone-specific PCR analysis using primer 4 of the LTR (upper panel) and primer SG295 (lower panel) before and at diagnosis of ATL. MWM, m.w. marker; H, HUT102 cells as HTLV-1-positive control; M, Molt4 cells as HTLV-1-negative control. In addition, case 4 was diagnosed as lymphoma-type, and leukemic cells were rarely observed in the peripheral blood smear at diagnosis. However, clone-specific PCR was strongly positive not only in the lymph node but also in the peripheral blood of case 4. This observation suggests that preleukemic cells without morphologic abnormality, which belonged to the original clone, existed in the peripheral blood even at diagnosis of ATL. At some point during the 8 years before ATL onset in this case, a subclone could have acquired additional malignant potential, presumably due to genetic and epigenetic abnormalities. Indeed, the presence of multiple clones of ATL cells based on the methylation patterns of the CDKN2A gene, despite having the same HTLV-1 integration site, has been reported.19 It also has been reported that multiple clones with simple chromosomal abnormalities already exist in HTLV-1 carriers and that selected clones with more complex chromosomal abnormalities can be found in progressed chronic ATL, based on the clonal culture method.20 Therefore, HTLV-1 carriers are hypothesized to vary with respect to the genetic and epigenetic abnormalities of their HTLV-1-infected cells. The risk of developing ATL might be related to these differences among HTLV-1 carriers. In case 1, a clone with the same integration site of the provirus in the leukemic cells was not detectable in the PBMC sample obtained 5 years before the onset of ATL (data not shown). This finding could be due to insufficient sensitivity of the PCR assay to detect that particular clone. In conclusion, the present prospective study provides evidence that a higher proviral DNA load strongly predicts the development of ATL in asymptomatic HTLV-I carriers. Moreover, the preleukemic clone can be observed in peripheral blood as many as 8 years before onset of the malignancy. This particular finding supports the belief that persistent clonal expansion is important in leukemogenesis. Thus, measuring the proviral DNA load and determining the clonality of HTLV-1-infected cells may provide important information for identifying carriers at increased risk of ATL. Further studies to characterize the level of genetic and epigenetic abnormalities accumulated in HTLV-1-infected cells will be necessary for the development of prophylactic therapies. Yours sincerely, Akihiko Okayama, Sherri Stuver, Masao Matsuoka, Junzo Ishizaki, Gen-ichi Tanaka, Yoko Kubuki, Nancy Mueller, Chung-Cheng Hsieh, Nobuyoshi Tachibana, Hirohito Tsubouchi