CD30 Characterizes Polylobated Lymphocytes and Disease Progression in HTLV-1–Infected Individuals

Adult T-cell leukemia/lymphoma (ATL) is a T-cell neoplasm with a poor prognosis that is caused by human T-cell leukemia/lymphoma virus type I (HTLV-1) infection. HTLV-1 is transmitted mainly through breast-feeding, and it is estimated that 20 million carriers exist worldwide. ATL develops after about 50 years of clinical latency in 2% to 5% of HTLV-1 carriers. The emergence of malignant cells with polylobated nuclei (the typical appearance is termed “flower cells”) characterizes ATL; however, these cells are already present in HTLV-1 carriers without development of ATL (1).

CD30, a member of the tumor necrosis receptor superfamily, activates prosurvival signals by ligation of its ligand CD30L or by CD30 overexpression (2, 3). CD30 activates nuclear factor (NF)-κB, ERK MAPK, and p38 MAPK (4–6). Previous reports showed CD30 expression in some ATL cells (7, 8). However, the clinicopathologic significance of CD30 in HTLV-1–infected cells is unknown.

Transformation of HTLV-1–infected T cells in vivo is a multistage process, which reflects the status of HTLV-1 infection, i.e., asymptomatic, smoldering, chronic, or acute (9). These cells acquire a transformed phenotype independent of HTLV-1 during this process (1, 10), indicating the importance of molecular mechanisms that occur during this step for tumorigenesis.

In this report, we aimed to elucidate the clinicopathologic significance of CD30 in HTLV-1–infected cells. We present a series of supportive evidence suggesting that CD30 is involved in the generation of abnormal lymphocytes with “flower cell,” which expands during the progression of carrier status and features an actively cycling virus-infected cell population, which causes chromosomal aberrations in HTLV-1–infected cells. We further discuss the significance of our finding in understanding the tumorigenic process of HTLV-1–infected cells and the therapeutic implication for the treatment of HTLV-1–infected individuals.

Primary cells, which were collected with informed consent as part of a collaborative study with the Joint Study on Predisposing Factors of ATL Development (JSPFAD), were used (11). The experiments and their analyses using the primary cells were performed at the University of Tokyo and were approved by its research ethics committee. Separation of mononuclear cells from the peripheral blood (PB) was performed as described (12). Clinical data of HTLV-1–infected individuals used are presented in Supplementary Table S1. Additional samples used for Supplementary Data were listed in Supplementary Table S2. All patients were categorized according to Shimoyama's criteria (9). MT-2 and HUT-102 are HTLV-1–infected T-cell lines (13, 14), and Jurkat and CEM are HTLV-1–uninfected T-cell lines. These cells were obtained from the RIKEN BioResource center. Karpas299, an anaplastic large cell lymphoma (ALCL) cell line, was obtained from DSMZ (German Collection of Microorganisms and Cell Cultures GmbH). Cells were cultured in RPMI1640 with 10% FCS, unless indicated. CHO cells were obtained from DS Pharma Biomedical and were cultured in Ham's F12 (Wako) with 10% FCS. TIG-1 cells and fetal lung fibroblast were obtained from Japanese Collection of Research Bioresources Cell Bank. These cells were used for the primary culture of HTLV-1–infected cells (15).

Inhibitors used to identify the pathway involved in the CD30-mediated generation of cells with polylobated nuclei were as follows: Ly294002 (Cell Signaling Technology) and duvelisib (Selleck Chemicals) for PI3K, UO126 for ERK kinase (MEK), SB203580 for p38MAPK (both from Cell Signaling Technology), and BMS-345541 for NF-κB (Calbiochem). All inhibitors were dissolved in dimethyl sulfoxide (DMSO) prior to use. Phorbol 12-myristate 13-acetate (PMA) was purchased from LC Laboratories and dissolved in DMSO.

For analysis of the expression of cell surface molecules, primary cells (3 × 10⁴ cells) were analyzed as previously described after incubation with fluorochrome-conjugated primary antibodies for 30 minutes at 4°C, using FACS Aria or Verse (both from BD Biosciences) and Flowjo software (TreeStar; ref. 16). To detect the expression of Ki-67, the cells were fixed using IntraPrep (Beckman Coulter) according to the manufacturer's instructions before incubation with antibodies as described (17), and the cells (3 × 10⁴ cells) were analyzed using FACS Verse. Cell-cycle and ploidy analyses of primary cells were performed as previously described, and the number of the cells analyzed by flow cytometry was indicated in the figure legend (18). The following fluorochrome-conjugated primary antibodies were used: anti–CD4-FITC, anti–CD4-Pacific Blue, and anti–CD25-APC (all from Biolegend), anti–CD30-PE, control IgG-FITC, and control IgG-PE (all from Beckman Coulter). For detection of Ki-67, anti–Ki-67 antibody (Santa Cruz Biotechnologies) and anti-rabbit secondary antibody conjugated with Alexa488 (Thermo Scientific) were used. For detection of HTLV-1–infected cells, anti-cell adhesion molecule 1 (CADM1) antibody (CM004-3, MBL) was used.

The PB of randomly selected individuals with different status of HTLV-1 infection, i.e., asymptomatic (Asy#1-22), smoldering (Smo#1-8), chronic (Chr#1-7), or acute (Acu#1-7; N = 44; Supplementary Table S1), was examined by flow cytometry as described. Their clinical data are presented in Supplementary Table S1. The CD30⁺CD25⁺CD4⁺ subpopulation was calculated as either a percentage of CD4⁺ cells or as an absolute number per μL of blood. The values between the different statuses were statistically evaluated. The scatter plot was presented, and Pearson's product moment correlation coefficient was calculated.

Peripheral blood mononuclear cells (PBMC) were separated using flow cytometry according to the expression of cell surface molecules and were sorted using FACS Aria as described (12). The primary cells or cell lines were cytospun onto a slide glass, fixed with 4% paraformaldehyde, and immunostained with anti–Ki-67 antibody (Santa Cruz Biotechnologies) and anti-rabbit secondary antibody conjugated with Alexa488 (Thermo Scientific) as previously described (12). The nuclei were stained with DAPI. The staining was analyzed using fluorescence microscopy (BX50F, Olympus).

Measurement of the HTLV-1 provirus load was performed as described (12). Briefly, genomic DNA was extracted from sorted cells using a DNA blood mini kit (Qiagen), and multiplex real-time PCR was performed with TaqMan probes. Primer pairs for HTLV-1 pX and RNase P enzyme (control) (Applied Biosystems) were used. The calculated provirus loads represent copy numbers per 100 cells. pX-probe(FAM) 5′-CTGTGTACAAGGCGACTGGTGCC-3′, pX2-sense 5′-CGGATACCCAGTCTACGTGTT-3′, pX2-antisense 5′-CAGTAGGGCGTGACGATGTA-3′, Taqman RNase P Control Reagents Kit (Product Number-4316844).

CD30L cDNA was inserted into the CSII-EF1 vector (kindly provided by Dr. H. Miyoshi, RIKEN BioResource Center) for the expression of mCherry-CD30L fusion protein. This plasmid or its control without CD30L cDNA was transduced into CHO cells as described (12), and mCherry-positive cells were sorted using FACS Aria (BD Biosciences). The expression of transfected CD30L in CHO cells and endogenous CD30 in T-cell lines was confirmed using flow cytometry (Supplementary Fig. S1). The resultant transformants were indicated as CD30L/CHO and mock/CHO, respectively. For fluorescent detection of the nucleus, Histone H2B-GFP construct in the CSII-EF1 vector was transduced into HUT-102 and MT-2 cells as previously described (12), and GFP-positive cells were sorted using FACS Aria.

CHO cells with or without CD30L expression (5 × 10⁴ cells each per well of a 12-well plate) were precultured in Ham's F12 (Wako Pure Chemical Industries) containing 10% FBS and were used for stimulation of CD30-expressing cells after 24 hours of culture. CD30-expressing cells (1.5 × 10⁵ cells) were cultured with the CHO cells in media containing an equal volume of RPMI1640 plus Ham's F12. After incubation, the cells were harvested, washed with PBS, and were then further analyzed. Unless indicated otherwise, media were supplied with 10% FBS, and the stimulation time was 96 hours. When the inhibitors were used, they were added to media containing an equal volume of RPMI1640 plus Ham's F12 and cultured for 48 hours. CD30-expressing cells were harvested and reseeded into the new CHO cells already cultured for 24 hours, and fresh media with corresponding inhibitors were used for additional culture for 48 hours.

PBMCs (2.5 × 10⁵) separated from HTLV-1–infected individuals were cultured with 200 μL of RPMI1641 containing 10% calf serum unless indicated, antibiotics, and 10 ng/mL recombinant human IL2 (R&D Systems, Inc.). The cells were treated with 10 μg/mL brentuximab vedotin (kindly obtained from Millennium Pharmaceuticals, Inc., an owned subsidiary of Takeda Pharmaceutical Company Limited) or vehicle (Milli-Q water) for 10 days. The harvested cells were analyzed as previously described, following incubation with fluorochrome-conjugated primary antibodies for 30 minutes at 4°C, using the FACS Aria or Verse (BD Biosciences) and FlowJo software (TreeStar; ref. 16).

Differences between mean values were assessed using a two-tailed t test. A P value < 0.05 was considered to be statistically significant. Correlations were evaluated using Pearson correlation coefficient.

HTLV-1 infects CD4⁺ lymphocytes, and these cells express CD25. To elucidate the expression of CD30 in HTLV-1–infected individuals, we first examined CD30⁺ cells in the PB of individuals with different status of HTLV-1 infection, i.e., asymptomatic (Asy#1-22), smoldering (Smo#1-8), chronic (Chr#1-7), or acute (Acu#1-7; N = 44; Supplementary Table S1). The CD30⁺CD25⁺CD4⁺ subpopulation within PB, calculated as either a percentage of CD4-positive cells or as an absolute number per μL of blood, increased according to progression of the status of HTLV-1 infection (Fig. 1A). A previous report indicated that changes of the serum soluble interleukin-2 receptor (sIL2R) correlate with disease progression in HTLV-1–infected individuals (19). Therefore, we evaluated the relationship between the CD30 expression and sIL2R (Supplementary Fig. S2). The result indicated correlation between CD30 expression and sIL2R. The expansion of the CD30⁺CD25⁺CD4⁺ subpopulation correlated with the expansion of abnormal lymphocytes according to progression of the status of HTLV-1 infection (Fig. 1B).

To characterize CD30⁺CD25⁺CD4⁺ cells, we next examined the morphologic features of these cells in the HTLV-1–infected individuals shown in Supplementary Table S1 (Asy#1-3, Smo#8 and 13, Chr#15, and Acu#5 and 6). Cells with polylobated nuclei, some of which possessed flower-like features, were frequently found in the CD30⁺CD25⁺CD4⁺ subpopulation of asymptomatic carriers. Although these cells were also found in the CD30⁻CD25⁺CD4⁺ subpopulation, their frequency was very low. Representative cell morphology and the proportion of each subpopulation in an asymptomatic carrier (Asy#1) are presented in Fig. 2A (left). In contrast, although there was a minor subpopulation of CD30⁺CD25⁺CD4⁺ cells in healthy donors, polylobated nuclei were not observed in these cells (Fig. 2A, right). Representative cellular morphologies of CD30⁺CD25⁺CD4⁺ cells in an additional 2 asymptomatic carriers (Asy#2 and 3) are shown in Fig. 2B (left most and second from the left). Such cells were also enriched in the CD30⁺CD25⁺CD4⁺ subpopulation of smoldering (Smo#8) and acute (Acu#6) cases whose representative cellular morphologies are also shown in Fig. 2B (second from the right and right most). The percentage of cells with polylobated nuclei in each subpopulation of the three asymptomatic carriers (Asy#1–3) and six advanced stages (two smoldering: Smo#8 and 13, one chronic: Chr#15, and two acute: Acu#5 and 6) is shown in the bar graph and beeswarm graph in Fig. 2C (left and right). The results indicate that the cells with polylobated nuclei characterize CD30⁺CD25⁺CD4⁺ subpopulation.

Although HTLV-1 infects CD4⁺ cells and these cells express CD25, the CD25⁺CD4⁺ fraction is a mixture of infected and uninfected cells. We therefore examined the HTLV-1 provirus load of each subpopulation, and the result for 5 asymptomatic carriers (Asy#4-6, 35, and 36) and 4 advanced stages (Acu#1, 2, 6, and 7) listed in Supplementary Table S1 is shown as a bar graph and beeswarm graph (Fig. 2D, left and right). Because amount of sorted cells from CD30⁺CD25⁺CD4⁺ fraction was limited, we could not measure virus load of samples used for morphological analyses in Fig. 2B and C except for Acu#6. Therefore, we used additional samples for further analyses. The result indicated that CD25⁺CD4⁺ cells bear a high virus load and that, within this cell population, CD30⁺ cells have higher virus load than CD30⁻ cells in asymptomatic carriers, whereas in acute phase, most CD25⁺CD4⁺ cells are virus-infected cells including CD30⁺ cells. To obtain further support for this result, we examined the expression of CADM1, a marker for HTLV-1–infected cells (20). As shown in Supplementary Figs. S3 and S4, almost all of the CD30⁺ cells are CADM1-positive and polylobated cells.

Because the CD30⁺CD25⁺CD4⁺ subpopulation is highly enriched in cells with polylobated nuclei, this subpopulation appears to represent virus-infected cells with polylobated nuclei during the progression of the disease status.

To further confirm the biological significance of CD30 in HTLV-1–infected individuals, we used flow cytometric analysis to examine the state of the cell cycle in the CD30⁺CD25⁺CD4⁺ subpopulation and in other cell subpopulations in the HTLV-1–infected individuals listed in Supplementary Table S1 (Asy#23-30, Smo#9-11, Chr#8-10, and Acu#8-10). Representative results for an asymptomatic carrier (Asy#23) and an acute case (Acu#8) are shown in Supplementary Fig. S5. The CD30⁺CD25⁺CD4⁺ subpopulation was characterized by an increased proportion of S–G₂–M phase cells. The percentages of S–G₂–M phase cells within CD4⁺ cell subpopulations of all the cases examined are shown in bar graphs in Fig. 3A (left). The beeswarm graph of each fraction is presented in Fig. 3A (right). To further confirm these results, we next examined the expression of the cell proliferation marker Ki-67 in the CD30⁺CD25⁺CD4⁺ subpopulation and in other subpopulations of cells from individuals with different status of HTLV-1 infection. Representative results of a smoldering (Smo#12) and a chronic case (Chr#10) are shown in Supplementary Fig. S6. The results of eight cases (Asy#31 and 32, Smo#12 and 13, Chr#10 and 11, and Acu#10 and 11) are shown as a bar graph in Fig. 3B (left). The beeswarm graph of each fraction is presented in Fig. 3B (right). The CD30⁺CD25⁺CD4⁺ fraction contains significantly higher amount of Ki-67⁺ cells compared with other fractions, indicating Ki-67–positive cells are enriched in CD30⁺CD25⁺CD4⁺ subpopulation. These results show that the CD30⁺CD25⁺CD4⁺ subpopulation contains cells with active proliferation, whereas cells in other subpopulations without CD30 expression mostly stay in a resting state.

The results so far indicate that, among HTLV-1–infected cells, CD30 expression is closely related with polylobation of the cells and active cell cycling, whereas most other cells without CD30 expression do not. Because CD30-positive fraction increases according to the progression of the disease (Fig. 1A), the number of these actively cycling cells expands according to progression of the HTLV-1 carrier status.

To elucidate roles of CD30 in HTLV-1–infected cells, we examined the effect of CD30 signals using T-cell lines with HTLV-1 infection (MT-2 and HUT-102). The T-cell lines without HTLV-1 infection (Jurkat and CEM) served as control. We firstly examined the effect on the morphologic features of these cells. Stimulation of CD30 expressed on T-cell lines with HTLV-1 infection by CD30L expressed on CHO cells increased the number of T cells with polylobated nuclei, but this effect was not obvious in the cells without HTLV-1 infection (Fig. 4A, left). A bar graph of the percentage of polylobated cells induced by stimulation with CD30L clearly shows this difference between the response of HTLV-1–infected and –uninfected cells to CD30 stimulation (Fig. 4A, right). Furthermore, the stimulation of PBMCs from an HTLV-1–infected individual by CD30L actually induced polylobated cells (Supplementary Fig. S7). Conversely, the stimulation of Karpas299, an ALCL cell line with CD30 expression unrelated to HTLV-1 by CD30L/CHO, did not generate polylobated cells, indicating the effect of CD30 stimulation in the generation of polylobated cells is different between ALCL cells and HTLV-1–infected T cells (Supplementary Fig. S8).

For clarification of CD30 signals in the cells with or without HTLV-1 infection, we examined the activation of known downstream pathways of CD30, i.e., NF-κB, JNK, p38 MAPK, ERK, and PI3K after CD30 stimulation in the cell lines used in this study by Western blot analysis (4–6, 21). The results indicated that the response for CD30 stimulation was generally undetected in the cells without HTLV-1 infection compared with those with HTLV-1 infection (Supplementary Fig. S9). These results suggest that cellular status might modify the responses of the cells for CD30 stimulation. JunB and IRF4 are considered to involve in CD30 induction (22, 23). In addition to JunB, IRF4, a molecule regulated by HTLV-1, is differently expressed between HTLV-1–uninfected cells and infected cells (Supplementary Fig. S10).

We next examined whether stimulation of CD30 with CD30L triggers promotion of the cell cycle in HTLV-1–infected cells. Stimulation of CD30 expressed on the HTLV-1–infected cell lines HUT-102 and MT-2 by CD30L promoted cell cycling, inducing an increase in the percentage of cells in S–G₂–M phase as shown in representative DNA histograms and in a bar graph (Fig. 4B, left and right). CD30 stimulation by CD30L also increased the percentage of Ki-67–positive cells in HTLV-1–infected cell lines (Fig. 4C, left and right). The expansion of CD30⁺ cells in the PBMCs of HTLV-1–infected individuals by CD30 stimulation was also indicated in the experiments using primary cells (Supplementary Figs. S11–S13).

These results indicate that CD30 signals corroborating with HTLV-1 infection involve both the generation of polylobated nuclei and the induction of an actively cycling phenotype.

To identify the signaling pathway involved in the CD30-mediated generation of cells with polylobated nuclei, we used inhibitors for known downstream effectors of CD30 signaling, i.e., NF-κB, ERK, MAPK, PI3K, and p38 MAPK (4–6, 21), and examined the effect on the generation of cells with polylobated nuclei in the HTLV-1–infected cell line HUT-102. The concentration of the inhibitors was optimized based on past studies (4–6, 21). As shown in Fig. 4D, inhibition of the PI3K pathway by Ly294002 showed significant inhibition of the generation of cells with polylobated nuclei, in the HTLV-1–infected cell line. Similar results were observed using another PI3K inhibitor duvelisib (Supplementary Fig. S14). Stimulation of PKC, downstream of PI3K by PMA, generated polylobated cells in HUT-102 cells (ref. 24; Supplementary Fig. S15). Furthermore, duvelisib inhibited polylobation by CD30L in primary cells from HTLV-1–infected individuals (Supplementary Fig. S7).

Because the results so far raised the possibility that CD30 signals provide impacts on chromosome by accelerating cell division, we next investigated the effect of CD30 signaling on this process. Flow cytometric analysis of DNA histogram indicated the possibility that the stimulation of T-cell lines with HTLV-1 infection by CD30L increases the fraction of 4N<. The representable result of HUT-102 is shown in Fig. 5A. Stimulation of T-cell lines with HTLV-1 infection (MT-2 and HUT-102) by CD30L increased 4N< fractions, whereas the stimulation of those without HTLV-1 infection (Jurkat and CEM) did not, suggesting CD30 signals increase hyperploid cells dependent on HTLV-1 infection (Fig. 5B).

To elucidate the cause of generation of hyperploid cells, we examined the effect for cell division by the stimulation of CD30 with its ligand in HTLV-1–infected cell lines. The nuclei of the T-cell lines with HTLV-1 infection were visualized using fluorescence microscopy following stable transfection of the cells with GFP fused to H2B. These cells were stimulated by CD30L expressed on CHO cells, and its effect on cell division was captured using time-lapse microscopy (Biostation IM-Q, Nikon). This analysis showed that CD30 stimulation frequently triggered abnormal cell division, most of which is abortive mitosis and cytokinesis. The representative cases of abnormal cell division observed in HUT-102 cells with stimulation by CD30L are presented in Fig. 5C (left and middle). Normal cell division observed in HUT-102 cells without stimulation by CD30L is presented in Fig. 5C (right). The original time-lapse videos of the cells are described in the legend. The percentage of abnormal cell division with CD30L stimulation of HUT-102 and MT-2 cells is shown as a bar graph in Fig. 5D.

We next examined the ploidy of each CD4⁺ cell fraction in the cells of individuals with different status of HTLV-1 infection, who are listed in Supplementary Table S1 (Asy#33 and 34, Smo#14 and 15, Chr#12 and 13, and Acu#12) by flow cytometric analysis of DNA histogram. Except for asymptomatic cases that did not contain a subpopulation with hyperploid cells (4N<), the CD30⁺CD25⁺CD4⁺ subpopulation in all other cases showed increased hyperploid cells (4N<) compared with other CD4-positive fractions (Fig. 5E). These results support the notion that induction of abnormal cell division and hyperploidy by CD30 actually occurs in HTLV-1–infected individuals.

The features of CD30⁺ cells presented so far indicated that this fraction may play a role in the transformation of HTLV-1–infected cells. Brentuximab vedotin, a monoclonal antibody for CD30 conjugated with the tubulin inhibitor monomethyl auristatin E, shows a potent effect in classical Hodgkin lymphoma and anaplastic large-cell lymphoma (25). A recent study also showed the effectiveness of brentuximab vedotin on peripheral T-cell lymphoma (26). Therefore, we examined whether brentuximab vedotin can deplete the CD30-positive fraction of PBMCs isolated from HTLV-1–infected individuals. When PBMC samples were treated with brentuximab vedotin, the CD30-positive fraction (CD30⁺CD25⁺CD4⁺) was shown to be reduced compared with those treated with vehicle alone (Fig. 6A). Examples of original flow cytometric data (Chr#18 and Acu#13) are presented in Supplementary Fig. S16. The reduction ranged from approximately 20%, to more than 90% (mean ± SD, 79.6 ± 21.79%; Fig. 6A). To clarify that the effect of brentuximab vedotin is selective for CD30⁺ fraction, we compared reduction rate in the fractions of CD30⁺ and CD30⁻ in the cases treated by brentuximab vedotin. The result showed the significant reduction of the number of viable cells in CD30⁺ fraction, whereas CD30⁻ fraction was relatively not affected (Supplementary Fig. S17). The results, which further support the effect of brentuximab vedotin on CD30⁺ cells, were presented in Supplementary Figs. S18 and S19.

We also examined the relationship between the expression level of CD30 and the response to brentuximab vedotin (Supplementary Table S3; Supplementary Fig. S20). The result suggested that the reduction rate of CD30⁺ cells by brentuximab vedotin was relatively low in cases with low CD30 expression (Asy#39 and Chr#15). These results suggest that brentuximab vedotin can selectively reduce the CD30⁺ cell fraction of PBMCs from HTLV-1–infected individuals.

Previous reports indicated the expression of CD30 in a part of ATL (7, 8). However, the clinicopathologic significance of CD30⁺ cells in HTLV-1–infected individuals including acute phase is entirely unknown. This lack of knowledge might be due to differences in experimental designs for the evaluation of CD30 positivity such as high cutoff value for CD30 positivity. In addition, there have been no reports that have provided a detailed description of CD30⁺ cells in the pre-ATL status. Roles of CD30 signals in HTLV-1–infected cells are also entirely unknown. Analyses of CD30⁺ cells in the CD25⁺CD4⁺ subpopulation enabled us to more sensitively detect and evaluate CD30⁺ cells. CD30 is induced in a portion of HTLV-1–infected cells, and these cells expand according to the progression of carrier status. A previous report indicated that sIL2R correlates with progression of the disease in HTLV-1–infected individuals (19). Therefore, we examined relationship between the expression of CD30 and the serum sIL2R. The result indicated that CD30 expression correlated with sIL2R (Supplementary Fig. S2).

The expansion of the CD30⁺CD25⁺CD4⁺ subpopulation correlated with the expansion of abnormal lymphocytes according to the progression of HTLV-1 carrier status, suggesting impact of CD30 on the status of HTLV-1 infection and its progression. In support of this hypothesis, the experiments using HTLV-1–infected cell lines showed that CD30 is involved in the generation of abnormal lymphocytes with “flower cell” and features an actively cycling, which coincides with abnormal mitosis and chromosomal aberrations in corroboration with HTLV-1 infection. CD30 stimulation also triggered expansion and polylobation of primary cells. We could link these results to the characteristics of CD30⁺ cells in HTLV-1–infected individuals and indicated involvement of CD30 in the progression of the disease status.

In this report, we indicated that the stimulation of HTLV-1–infected cell lines with CD30 could induce polylobation through the activation of PI3K. CD30 stimulation could not induce polylobation in HTLV-1–uninfected cell lines used in this study. This might be explained by the result that the effect of CD30 stimulation was different between HTLV-1–infected and –uninfected cells. Our finding is supported by those of a previous study, which indicated that activation-inducible lymphocyte immunomediatory molecule (AILIM/ICOS)–mediated activation of PI3K is responsible for the polylobation of lymphocytes in ATL (27).

PI3K regulates cytoskeletal rearrangements through the regulation of Rho-family cascades (28). However, we could not show the activation of these pathways by stimulation of CD30 in HTLV-1–infected cell lines (data not shown). The previous report indicated the involvement of microtubule rearrangement in polylobation (27). We could show that PMA, an activator of PKC, triggered polylobation in an HTLV-1–infected cell line, HUT-102 (Supplementary Fig. S15). PKC is a downstream molecule of PI3K and commits microtubule rearrangement (24, 29). Therefore, further study to clarify involvement of CD30 in PKC-mediated rearrangement of microtubule will provide important insights into mechanisms underlying polylobation in HTLV-1–infected cells.

A recent report demonstrated that AP-1 family transcription factors, including JunB, are responsible for the induction of activation-inducible lymphocyte immunomediatory molecule/inducible co-stimulator (AILIM/ICOS; ref. 30). Because CD30 can activate AP-1 by inducing JunB through the ERK pathway (31), it is possible that the expression of AILIM/ICOS is dependent on CD30 signaling and that CD30 can induce polylobated cells not only directly, but also indirectly via the induction of AILIM/ICOS, although further analyses are required to confirm this hypothesis.

Emergence of CD30⁻ polylobated cells appears to increase according to the progression of disease status. Because most of these cells are HTLV-1–infected cells (Supplementary Fig. S4), alteration of the cellular status might mimic the signals for polylobation instead of CD30. Because CD30 is transiently downregulated after stimulation by CD30L, this process might partially be responsible for the existence of CD30⁻ polylobated cells (32). Molecular mechanisms underlying the emergence of CD30⁻ polylobated cells require further study.

We indicated that CD30 stimulation of HTLV-1–infected cells could induce active cycling. In fact, CD30 expression in HTLV-1–infected cells is also closely related with actively cycling cells, whereas most other cells without CD30 expression are quiescent in lymphocytes of HTLV-1–infected individuals. CD30⁺ cells showed an increased number of cells in the S–G₂–M phase and with Ki-67 expression compared with CD30⁻ cells. Ki-67 is a cell-cycle–related nuclear protein that is expressed by cells in a proliferative state, but not by cells in a quiescent state (33). Combined with the close relationship between CD30 expression and polylobated cells, these results imply that CD30 signals trigger cell-cycle promotion of HTLV-1–infected cells, which coincides with the generation of characteristic polylobated cells, and these cells expand according to the progression in HTLV-1–infected individuals.

We also indicated that CD30 stimulation triggers abnormal cell division and hyperploidy in HTLV-1–infected cell lines but does not in unrelated cell lines. Although further studies are required, cellular signaling status caused by HTLV-1 infection might also be responsible for this result (34). Based on the observations of time-lapse microscopy, abortive mitosis and its subsequent effects might involve the generation of hyperploid cells, which might be related with a possible cause of chromosomal instability and cellular transformation of ATL cells (35). We indicated that the CD30⁺ cell fraction actually contains cells with chromosomal aberrations and that these cells expand according to progression of the status of HTLV-1 infection. A previous report indicated that changes in chromosome copy-number and gene expression profile occur according to progression of the disease status in ATL (36, 37). Ki-67 positivity correlates with high-grade malignancy (38). CD30-positive cells might be able to serve as a reservoir for malignant transformation and clonal expansion of HTLV-1–infected cells in HTLV-1–infected individuals. This notion is supported by very recent report, which identified CD30 as one of super enhancers and suggested its involvement of pathogenesis of ATL (39).

CD30 appears to be induced in a portion of HTLV-1–infected cells depending on their cellular status. CD30 expression appears not to be related to virus replication, because HTLV-1 is not replicating in PB (40). CD30 stimulation triggers growth-promoting effects, abnormal mitosis, and chromosomal aberration in HTLV-1–infected cell lines. Furthermore, it was shown that the stimulation of PBMCs from HTLV-1–infected individuals by CD30L actually expanded CD30⁺CD25⁺CD4⁺ cells (Supplementary Figs. S11 and S12). These results suggest that the CD30 expression in HTLV-1–infected cells is important as a cause for the expansion of CD30-positive cells and tumorigenic process. However, further study is necessarily to clarify detailed mechanisms of involvement of the CD30⁺ cells in these processes.

CD30L is present on T cells, B cells, granulocytes, and monocytes. These cells are widely distributed in the body, e.g., PB, bone mallow, and lymphoid system. Especially granulocytes, including eosinophils, are increased in ATL and modulate its prognosis (41, 42). Interaction between CD30L and CD30 triggers local gathering and subsequent internalization of this complex to CD30-expressing cells for signaling (32). ATL is closely associated with exteriorization of parasitic diseases such as strongyloidiasis (43). A previous report indicated that five steps are required for the final transformation of onset of ATL (44). We showed that CD30 expression correlated with sIL2R, a marker for disease progression in HTLV-1–infected individuals (19). Previous reports suggested serum-soluble CD30 is associated with poor prognosis in HTLV-1–infected individuals including ATL (45, 46). These results indicate the possibility that CD30 positivity is associated with poor prognosis in HTLV-1–infected individuals including ATL. CD30 signals might involve the expansion of HTLV-1–infected cells and trigger the transformation or clonal evolution of HTLV-1–infected cells that coincides with abortive mitosis, polylobation of the nuclei, and chromosomal aberrations. It is possible that CD30, which is induced in a portion of HTLV-1-infected cells, commits progression of asymptomatic, pre-ATL, and ATL status through clonal evolution of these cells.

We could show that brentuximab vedotin can purge CD30⁺ cells in PBMCs from HTLV-1–infected individuals. This effect appears to depend on the expression level of CD30 in the cells (Supplementary Fig. S20). Our results by primary cells are almost in accordance with a previous report using HTLV-1–infected cell lines (47). In vivo effectiveness of brentuximab vedotin in mice bearing HTLV-1–infected cell lines was already reported (47). In this context, diminishing CD30-expressing cells with anti-CD30 monoclonal antibody therapies (48) might be considered for both pre-ATL status such as smoldering and chronic status, and ATL (Fig. 6B).

In conclusion, the results of this study indicate a biological link between the expression of CD30, which serves as a marker for HTLV-1 disease progression, and the actively proliferating fraction of HTLV-1–infected cells, which display chromosomal aberrations and polylobation. The findings also highlight the role of CD30⁺ cells as a potential reservoir for transformation and clonal expansion in HTLV-1–infected individuals, and indicate that these cells may thus be a therapeutic target in this disease. Further examination of the CD30⁺ cell fraction may improve our understanding of the tumorigenic processes of HTLV-1–infected cells.

A. Utsunomiya reports receiving speakers bureau honoraria from Bristol-Myers Squibb, Celgene, Chugai Pharma, Daiichi Sankyo, Eisai, JIMRO, Kyowa Hakko Kirin, Mundi Pharma, Nippon Shinyaku, Novartis Pharma, Ono Pharmaceutical CO, Otsuka Pharmaceutical, Pfizer, and Siemens. R. Horie is a consultant/advisory board member for Takeda. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T. Watanabe, R. Horie

Development of methodology: M. Nakashima, M. Watanabe, R. Horie

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Nakashima, T. Yamochi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Nakashima

Writing, review, and/or revision of the manuscript: M. Nakashima, A. Utsunomiya, T. Watanabe, R. Horie

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Watanabe, K. Uchimaru, A. Utsunomiya, R. Horie

Study supervision: K. Uchimaru, M. Higashihara, T. Watanabe, R. Horie

This study was performed as part of a collaborative study with the JSPFAD. Flow cytometric analysis was supported technically by the FACS Core Laboratory, Center for Stem Cell Biology and Regenerative Medicine, The Institute of Medical Science, and The University of Tokyo. This work was supported in part by a MEXT/JSPS KAKENHI grant to R. Horie (26460440 and 17K08728), M. Watanabe (26460439), and T. Watanabe (221S0001).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.