CD30 induces Reed-Sternberg cell-like morphology and chromosomal instability in classic Hodgkin lymphoma cell lines

Abstract

Classic Hodgkin lymphoma (cHL) is characterized by multinucleated cells called Reed-Sternberg (RS) cells and genetic complexity. Although CD30 also characterizes cHL cells, its biological roles are not fully understood. In this report, we examined the link between CD30 and these characteristics of cHL cells. CD30 stimulation increased multinucleated cells resembling RS cells. We found chromatin bridges, a cause of mitotic errors, among the nuclei of multinucleated cells. CD30 stimulation induced DNA double-strand breaks (DSBs) and chromosomal imbalances. RNA sequencing showed significant changes in the gene expression by CD30 stimulation. We found that CD30 stimulation increased intracellular reactive oxygen species (ROS), which induced DSBs and multinucleated cells with chromatin bridges. The PI3K pathway was responsible for CD30-mediated generation of multinucleated cells by ROS. These results suggest that CD30 involves generation of RS cell-like multinucleated cells and chromosomal instability through induction of DSBs by ROS, which subsequently induces chromatin bridges and mitotic error. The results link CD30 not only to the morphological features of cHL cells, but also to the genetic complexity, both of which are characteristic of cHL cells.

Abbreviations
• CD30L
• ligand for CD30
• CGH
• comparative genomic hybridization
• cHL
• classic Hodgkin lymphoma
• DSB
• DNA double-strand break
• EBV
• Epstein–Barr virus
• GO
• Gene Ontology
• H
• Hodgkin
• LCL
• lymphoblastoid cell line
• LNAC
• N-acetyl-L-cysteine
• PCA
• principal component analysis
• ROS
• reactive oxygen species
• RS
• Reed-Sternberg
• TNF
• tumor necrosis factor
• TNFR
• tumor necrosis factor receptor
• TRAF
• TNFR-associated factor

INTRODUCTION

CD30, a member of the TNFR superfamily, is strongly expressed on the surfaces of large mononuclear H and multinuclear RS cells of cHL. CD30L, a member of the TNF superfamily, is expressed relatively broadly on the cell surfaces of granulocytes, monocytes, macrophages, mast cells, and activated lymphocytes.¹

Ligation of CD30L to CD30 (CD30 stimulation) triggers oligomerization of CD30 and is thought to initiate signaling events on the cell surface by recruiting TRAFs to the cytoplasmic domain of CD30.¹ A previous report also suggested that highly expressed CD30 can self-oligomerize and transduce signals in a ligand-independent manner.² A recent report showed that CD30 stimulation triggers the internalization of CD30L-CD30 complexes into a cHL cell line and transduces signals, suggesting that signaling events by CD30 are more dynamic than previously thought.³

The biological effects of CD30 stimulation on cHL cells have been debated for many years and are still not fully understood. Previous reports using cHL cell lines have shown that CD30 is involved in signaling pathways such as the activation of nuclear factor-κB, ERK1/2, JNK, and AKT, and might play a key role in antiapoptosis and the expression of cytokines, committing cells to the characteristic features of cHL.^2,4–7 However, the biological effects of CD30 stimulation in cHL cell lines have not always been consistent among reports, and one report has argued that CD30 signals are absent in cHL cell lines.⁸

Most previous reports are based on transient stimulation of CD30. However, cHL cells are surrounded by the infiltrate of reactive cells, which express CD30L and appear to be repeatedly stimulated by CD30L under physiological conditions.⁹ Therefore, we stimulated CD30 on cHL cell lines by CD30L for a longer period and observed its effects. We found that CD30 is involved in the generation of RS cell-like multinucleated cells, chromosomal instability, and altered gene expression. Based on our findings, we discuss their significance in the biology of cHL cells.

MATERIALS AND METHODS Cell lines and cell cultures

Classic Hodgkin lymphoma cell lines (L428, L540, and L591) were purchased from the German Collection of Microorganisms and Cell Cultures. Another cHL cell line, AM-HLH was kindly gifted from Dr. Ohno H at Tenri Institute of Medical Research.¹⁰ Among cHL cell lines, L591 and AM-HLH are infected with EBV. A T cell line, Jurkat, was obtained from the Japanese Cancer Research Resources Bank. We used two LCLs from two donors, LCL2 and LCL3, which had previously been established from PBMCs purchased from Cellular Technology Limited by the method based on the past report.¹¹ CHO cell lines stably expressing CD30L and its control without CD30L expression were described previously.¹² L540 and L591 were cultured in RPMI-1640 (Sigma-Aldrich) with 20% FBS. AM-HLH was cultured in RPMI-1640 with 13% FBS. CHO cells were cultured in Ham's F-12 (Fujifilm Wako Pure Chemical) medium with 10% FBS. Other cell lines were cultured in RPMI-1640 with 10% FBS.

Chemicals

LY294002, a PI3K inhibitor, and doxorubicin, a topoisomerase II inhibitor that produces DNA damage, were obtained from Cell Signaling Technology. LY294002 was used for experiments at the specified concentrations.¹² Hydrogen peroxide, a ROS, was purchased from Nacalai Tesque.¹³ We used DAPI (Dojindo) for staining DNA, and LNAC (Sigma Aldrich), a ROS inhibitor, was used based on a previous study.¹⁴

Giemsa stain

Giemsa solution (Muto Pure Chemicals) was diluted 20 times by PBS (Muto Pure Chemicals) of pH 6.4 and a final concentration of 1/150 M. The cells were cytospun onto glass slides and fixed with methanol for 2 min. The cells were stained with the Giemsa solution for 20 min at room temperature, washed, and analyzed by microscopy (BX53; Olympus).

Flow cytometry

After incubation with phycoerythrin-conjugated anti-CD30 Ab or its isotype-matched control (Beckman Coulter) for 30 min at 4 °C, cells were analyzed using FACSVerse (BD Biosciences) and FlowJo software (TreeStar).

Stimulation of cells with CD30L

CHO cells with or without CD30L expression (2 × 10⁴ cells each per well of a 24-well plate)¹² were precultured overnight in Ham's F-12 containing 10% FBS. Cell lines (5 × 10⁴ cells) were overlaid and cultured on CHO cells (2 × 10⁴) in RPMI-1640 with 10% or 20% FBS. After coculture, the cells were separated from CHO using the method described previously¹⁵ and served for analyses.

Detection of DNA damage

DNA Damage Detection Kit-phosphorylated histone H2AX (γ-H2AX)-Green (Dojindo) to detect DSBs was used for analyses. Cells treated with 100 ng/mL doxorubicin overnight according to the manufacturer's instructions served as a positive control. Antibodies used were as follows: anti-γ-H2AX Ab supplied with the kit and Alexa Fluor 488 goat anti-mouse IgG (H + L) secondary Ab (Invitrogen). After staining of nuclei by DAPI (Dojindo) the cells were observed by a fluorescence microscope (BZ-X810; Keyence).

Array CGH

Genomic DNA was isolated using a QIAamp DNA mini kit (Qiagen). The extracted whole genomic DNA was used as a template for array CGH analysis with SurePrint G3 Human CGH microarray 8 × 60 K (Agilent Technologies). Briefly, 200 ng DNA was digested and labeled with cyanine 5-dUTP (Cy5-dUTP) or cyanine 3-dUTP (Cy3-dUTP) using a SureTag Complete DNA Labeling Kit (Agilent Technologies). Equal amounts of sample and reference DNA were mixed and used for hybridization. Glass slides were scanned using an Agilent C microarray scanner. The images were quantified using Feature Extraction software, and the data were analyzed by cytogenomics software (Agilent Technologies). Aberrant regions were determined using the ADM-2 algorithm at the threshold of 3.

RNA sequencing

Total RNA was isolated using a Direct-Zol RNA miniprep kit (ZYMO Research). Pretreatment, sequencing, and analyses were carried out at GENEWIZ (https://www.genewiz.com/ja-JP). In short, the clean data were mapped to the human reference genome (hg38) using Hisat2 (http://daehwankimlab.github.io/hisat2/main/). Gene expression levels were estimated by HTSeq (https://htseq.readthedocs.io/en/master/index.html). Differential expression analysis used the DESeq2 (http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html). Bioconductor package. Only genes with a fold change >2 or <1/2 and padj <0.05 were selected as differentially expressed genes. Principal component analysis to visualize sample-to-sample distances was calculated using R language. Data are available at the DDBJ Sequence Read Archive, accession number DRA012670.

Measurement of intracellular ROS

CellRox Oxidative Stress Reagent green (Thermo Fisher Scientific) was added to the media at the final concentration of 5 μM and cultured for 30 min. The cells were cytospun onto glass slides and fixed with 4% paraformaldehyde PBS (Nacalai Tesque) for 15 min. After permeabilization with 0.5% Triton X-100 and staining with DAPI, the cells were observed by fluorescence microscopy (BZ-X810; Keyence).

Statistical analysis

Differences between mean values were evaluated using two-tailed t-tests. Statistical significance was defined as p <0.05.

RESULTS CD30 stimulation induces multinucleated cells and suppresses proliferation of cHL cell lines

We recently showed that that physiological ligation of CD30L on CD30 triggers internalization of CD30L-CD30 complexes into cHL cell lines, resulting in signaling processes.³ Classic Hodgkin lymphoma cells are surrounded by reactive cells expressing CD30L and appear to be repeatedly stimulated.⁹ Therefore, using the experimental system above, we stimulated CD30 on cHL cell lines L428 and L540 by CD30L for 1 month, unless otherwise described, and examined its effect.

We first confirmed the expression levels of CD30 in these cHL cell lines (Figure 1A). Photographs of cHL cell lines being cultured with CHO cells with or without CD30L expression are shown in Figure 1B. The cHL cells interacted closely with CHO cells expressing CD30L. We next evaluated the morphological changes of these cHL cells by Giemsa stain. We found that large cells and multinucleated cells resembling RS cells stood out by CD30 stimulation (Figure 1C, D). Cell size of cHL cell lines became larger (Figure 1E) and percentages of the multinucleated cells were significantly increased (Figure 1F) by CD30 stimulation. We further evaluated the effect on cell proliferation. Both cell lines showed significant suppression of proliferation by CD30 stimulation (Figure 1G). These effects by CD30 stimulation on cHL cell lines were more potent in L540 than L428. Interruption of CD30 stimulation on day 7 recovered the proliferation of cHL cell lines (Figure 1H).

FIGURE 1. Effects of CD30 stimulation on morphology and cell growth in classic Hodgkin lymphoma (cHL) cell lines. After confirmation of the expression of CD30, cHL cell lines were stimulated with (+) or without (−) CD30L for 1 month unless otherwise described. (A) Expression of CD30. cHL cell lines were labeled with phycoerythrin (PE)-conjugated anti-CD30 Ab (anti-CD30) or its isotype control (Cont.). (B) Microscopic observation of cells under stimulation. (C) Morphological features by Giemsa stain. Representative results from triplicate experiments are shown. Scale bar, 20 μm. (D) Multinucleated cells. Representative multinucleated cells observed in cHL cells after CD30 stimulation. The magnification of each panel differs depending on the size of the cells. (E) Cell size. The area of the cells was measured using ImageJ software. Relative area sizes of 50 cells each were measured in triplicate samples, and the results are presented as mean ± SD. (F) Percentage of multinucleated cells (those containing two or more nuclei). Fifty cells each were measured in triplicate samples, and the percentage results are presented as mean ± SD. (G) Effects on cell growth. The number of viable cells was counted. Experiments were carried out in triplicate, and the fold growth compared to the initial cell number before stimulation is presented as mean ± SD. (H) Effects of the interruption of CD30 stimulation on the cell growth of cHL cell lines. Cells (2 × 105) were stimulated with ligand for CD30 (CD30L) for 7 days (red line). These cells were further cultured with (orange line) or without (blue line) stimulation with CD30L for an additional 10 days. Number of viable cells was calculated at day 7 and day 17. *p [less than] 0.05.

Connection of nuclei by chromatin bridges and generation of DSBs by CD30 stimulation in cHL cell lines

The nuclei of multinucleated cells of cHL cell lines, as observed in Figure 1C, overlapped or contacted with each other. By careful observation, we found that in cells whose nuclei existed very closely, each nucleus was connected by bridges, although this phenomenon is not typical in cHL cell lines with CD30 stimulation. Based on this observation, we undertook the additional experiment and confirmed this result. Representative photographs of the bridges are shown in Figure 2A. The bridges connecting the nuclei could be stained by DAPI (Figure 2B) and were also observed in mitotic cells of anaphase, suggesting the existence of chromatin bridges (Figure 2C).

FIGURE 2. Bridges observed among the nuclei of multinucleated cells, and induction of DNA double-strand breaks (DSBs) by CD30 stimulation in classic Hodgkin lymphoma (cHL) cell lines. (A,B) cHL cell lines were stimulated with or without ligand for CD30 (CD30L) for 10 days. (A) Giemsa stain. Arrows indicate bridges. (B) DAPI stain (top panels) and cell morphology (bottom panels). Arrows indicate bridges. Multinucleated cells presented were from the cells stimulated with CD30L, except for the left panel of L428 in Figure 2B, which is from the cells without stimulation by CD30L. Magnification of each cell differs depending on their size. (C) Anaphase of the cHL cell line L428 stimulated with CD30L featuring the bridge (arrow) observed in the experiment presented in Figure 2B. (D,E) Staining of cHL cell lines by anti-γ-H2AX Ab, which detects DSBs. Cells were stimulated with (+) or without (−) CD30L for 14 days. (D) The cells were stained by anti-γ-H2AX Ab and DAPI as described, and observed by fluorescence microscopy. cHL cells treated with doxorubicin (Dox) served as control. Scale bar, 20 μm. (E) The percentage of γ-H2AX-positive cells was determined by counting 50 cells in three different regions in each sample and presented as mean ± SD. Cells showing staining, as was observed in most of the Dox-treated cells, were counted as positive. (F,G) Effects of CD30 stimulation on chromosomes in cHL cell lines. cHL cell lines were stimulated with or without CD30L for 1 month. Array comparative genomic hybridization analyses were carried out. (F) Gain and loss sites compared to untreated cells without coculture with CHO are indicated by blue and red whiskers, respectively. Representative results from triplicate experiments are shown. (G) The number of copy number alteration (CNA) regions is presented. Results from triplicate experiments are presented as mean ± SD. *p [less than] 0.05. BF, bright field.

We previously reported that CD30 stimulation induces abortive mitosis.¹² Previous reports suggested that chromosome segregation error is one cause of abortive mitosis.¹⁶ Therefore, it is possible that CD30-mediated increase of multinucleated cells is triggered by the acceleration of chromosome segregation error, which causes chromatin bridges and abortive mitosis. As chromosome segregation errors are caused by critical damage to chromosomes such as DSBs,^17,18 we examined whether DSBs were increased after CD30 stimulation in cHL cell lines. The results of cHL cell lines stained with an Ab for γ-H2AX, a marker for DSBs, are presented in Figure 2D. Anti-γ-H2AX Ab staining was significantly increased after CD30 stimulation (Figure 2E).

CD30 stimulation also increased generation of large cells and multinucleated RS-like cells (Figure S1A–C), and induced DSBs (Figure S1D, E) in cHL lines with EBV infection. However, these effects were not observed in immortalized B cells (LCLs) or a T cell line (Figure S2). The expression of CD30 in these cells was confirmed by flow cytometry (Figure S3).

CD30 stimulation triggers chromosomal imbalances in cHL cell lines

As CD30 stimulation has critical impacts on chromosomes in cHL cell lines, we further examined whether CD30 stimulation for 1 month in cHL cell lines triggers genomic copy number variations by array CGH. Representative results of triplicate experiments are shown in Figure 2F. The chromosomal imbalances by CD30 stimulation appear to span the entire regions of the chromosomes. The number of affected regions was significantly increased by CD30 stimulation (Figure 2G). The effect of CD30 stimulation appears to be more potent in L540 cells compared to L428 cells. We could extract two genes commonly gained by L540 and L428 after CD30 stimulation: MAPT (microtubule associated protein tau) and KANSL1 (KAT8 regulatory NSL complex subunit 1).

Taken collectively, the results so far suggest that CD30 stimulation increases the generation of multinucleated RS-like cells in cHL cells, which might be caused by abscission error due to chromatin bridges. The results also suggest that CD30 stimulation induces DSBs, which trigger chromatin bridges and chromosomal imbalances. DNA damage and subsequent abscission error might be responsible for retardation of cell proliferation.

CD30 stimulation alters gene expression in cHL cell lines

As CD30 stimulation triggered chromosomal imbalances in cHL cell lines, we next addressed alterations of gene expression in cHL cell lines after CD30 stimulation for 1 month. For this purpose, we used RNA sequencing. The PCA revealed that, although L428 and L540 cell lines have been shown to be bona fide cHL cell lines,¹⁹ triplicate samples of these cell lines without CD30 stimulation clustered at distant regions, suggesting the existence of biologically different backgrounds between these cell lines. CD30 stimulation caused a shift in each cell line and the formation of new clusters at distant regions (Figure 3A). Heat map analyses as shown in Figure 3B indicated changes in gene expression and clusterization by CD30 stimulation compared to the control in each cHL cell line, supporting the results shown in Figure 3A. Genes that were significantly changed are shown in volcano plots (Figure 3C). The number of genes that significantly changed (more than 2-fold or less than 1/2-fold) is expressed as a bar graph (Figure 3D). These results suggest that CD30 stimulation in cHL cell lines causes alterations of gene expression, although it is more potent in L540 cells.

FIGURE 3. Effects of CD30 stimulation on gene expression profiles in classic Hodgkin lymphoma (cHL) cell lines. cHL cell lines were stimulated with or without ligand for CD30 (CD30L) for 1 month in triplicate samples. RNA was extracted and subject to RNA sequencing. Changes after stimulation with CD30L were calculated by comparing the results of cHL cell lines with stimulation by CD30L compared to those without CD30L. Genes significantly changed by CD30 stimulation (≥2-fold and ≤1/2-fold) were evaluated. (A) Principal component analysis (PCA) plot. CD30L and C indicate stimulation with and without CD30L, respectively. Numbers added to CD30L and C indicate replicates in each experiment. (B) Heatmap. Differences in mRNA expression in each sample are shown. The labeling of the samples is the same as those in (A). (C) Volcano plot. Genes significantly changed by stimulation with CD30L are colored red (up) and blue (down). (D) Number of genes significantly changed by stimulation with CD30L is shown and colored red (up) and blue (down). Number of genes is indicated at the top of each bar. (E) Venn diagram of the genes significantly changed by stimulation with CD30L. Number of genes is indicated.

Analyses of genes commonly changed between L428 and L540 by CD30 stimulation

Next, we examined the genes commonly changed by CD30 stimulation in L428 and L540 cell lines. A Venn diagram for the number of genes significantly up- or downregulated by CD30 stimulation is presented in Figure 3E. Among 23 genes commonly changed between L428 and L540, 21 were commonly upregulated; however, two genes showed opposing changes (Table S1). Six genes selected from these commonly changed genes were further validated by quantitative RT-PCR, and we could validate the results by RNA sequencing (Figure S4). Primer pairs used are listed in Table S2.

Two genes commonly gained in L428 and L540 by CGH analyses are not included in Table S1, suggesting that changes in copy numbers do not directly affect the expression profile by modifications such as epigenetics.²⁰ We showed that CD30 stimulation triggers DSBs and found chromosomal bridges in L428 and L540 cells. These results suggest the existence of replication stress by CD30 stimulation.^17,18 Genes directly involved in replication stress are not listed in Table S1, and this is supported by GO analyses of the genes commonly changed between L428 and L540 (Figure S5). The top six GO terms based on the p value were as follows: immune response, protein secretion by platelet, regulation of cell diameter, phospholipase C-inhibiting G-protein coupled receptor signaling pathway, positive regulation of inositol-polyphosphate 5-phosphatase activity, and negative regulation of inositol phosphate biosynthetic process.

CD30 stimulation induces ROS production, which triggers chromatin bridges, multinucleated cells, and DSBs in cHL cell lines

A previous report showed that forced dimerization of the cytoplasmic tails of CD30 by fusion to the extracellular and transmembrane domain of CD28 induces ROS from mitochondria through activation of TRAFs.²¹ As ROS are critical factors, which trigger DNA damage including DSBs,²² we examined changes in the level of ROS production in cHL cell lines by CD30 stimulation. For this purpose, we used CellRox Oxidative Stress Reagents (Thermo Fisher Scientific), which bind to cellular DNA and produce green fluorescence upon oxidization. The results are shown in Figure 4A. The green fluorescence per cell (Figure 4B) and per unit area in the cell (Figure S6) were significantly increased after CD30 stimulation, suggesting an increase in ROS production by CD30 stimulation in cHL cell lines.

FIGURE 4. CD30 stimulation triggers reactive oxygen species (ROS) production, and ROS induces multinucleated cells and double-strand breaks (DSBs) in classic Hodgkin lymphoma (cHL) cell lines. (A,B) cHL cell lines were stimulated with (+) or without (−) ligand for CD30 (CD30L) for 5 days. (A) Cells were stained with CellRox Oxidative Stress Reagent green and DAPI. Scale bar, 20 μm. (B) Fluorescence intensity by CellRox was determined by measuring more than 50 cells with ImageJ software in three different regions; relative fluorescence intensity per cell is presented as mean ± SD. Mean fluorescence intensity of the cells without CD30L was set to 100. (C,D) Morphological features of cHL cell lines treated with 25 μM H2O2 for 3 days. (C) Giemsa stain. Scale bar, 20 μm. (D) Percentage of multinucleated cells. Fifty cells each were measured in three different regions and the percentage results are presented as mean ± SD. (E) Bridges (arrows) observed in multinucleated cHL cells after treatment with H2O2 for 3 days and stained with Giemsa. Concentration of H2O2 was 25 μM, except for the right panel of L540, which was treated with 50 μM H2O2. Magnification of each cell differs depending on their size. (F) Anaphase of cHL cell lines treated with 25 μM H2O2 for 3 days featuring the bridge (arrows). (G) Staining of cHL cell lines treated with ROS by anti-γ-H2AX Ab, which detects DSBs. cHL cell lines were treated with H2O2 (25 μM) for 3 days and cytospun onto glass slides. Cells were stained with anti-γ-H2AX Ab and DAPI, and observed by fluorescence microscopy. Scale bar, 20 μm. *p [less than] 0.05.

To examine generation of multinucleated cells and chromatin bridges by ROS in cHL cell lines, we treated cHL cell lines with hydrogen peroxide (H₂O₂), a ROS. This treatment increased the percentage of multinucleated cells (Figure 4C,D) and we could find chromatin bridges (Figure 4E,F) as observed in cHL cell lines with CD30 stimulation.

As ROS are well known to trigger DNA damage, including DSBs,²² we wanted to examine the expression of γ-H2AX after treatment with H₂O₂ in cHL cell lines. We found increased expression of γ-H2AX in most cells by H₂O₂ treatment (Figure 4G), as observed in cHL cell lines with CD30 stimulation.

To confirm that ROS actually involves CD30-mediated generation of multinucleated cells and induction of DNA damage, we undertook the experiments using ROS inhibitor, LNAC. The results showed that ROS inhibition decreased morphological changes (Figure 5A-C) and induction of γ-H2AX (Figures 5D,E and S7) by CD30 stimulation in cHL cell lines.

FIGURE 5. Effects of reactive oxygen species (ROS) inhibitor, N-acetyl-L-cysteine (LNAC), on CD30-mediated generation of multinucleated cells and double-strand breaks in classic Hodgkin lymphoma (cHL) cell lines. cHL cell lines were stimulated with ligand for CD30 (CD30L) for 14 days with (+) or without (−) the addition of 1 mM LNAC. Media was replaced by fresh media with or without 1 mM LNAC twice a week. (A) Giemsa stain. Scale bar, 20 μm. (B,C) Cell size and percentage of multinucleated cells were measured. (B) The area of the cells was measured using ImageJ software. Relative area sizes of 50 cells each were measured in three different regions, and the results are presented as mean ± SD. Mean value of the area of the cells without LNAC was set to 100. (C) Fifty cells each were measured in three different regions, and the percentage of multinucleated cells is presented as mean ± SD. (D) Staining by anti-γ-H2AX Ab and DAPI. Scale bar, 20 μm. (E) Fluorescence intensity by anti-γ-H2AX Ab was determined by measuring 50 cells with ImageJ software in three different regions. Relative fluorescence intensity per cell is presented as mean ± SD. Mean fluorescence intensity of the cells without LNAC was set to 100. *p [less than] 0.05.

Taken together, these results suggest that ROS produced by CD30 stimulation involves generation of DSBs, which causes chromatin bridges and formation of multinucleated cells with RS cell-like morphology in cHL cells.

CD30 mediated generation of ROS and multinucleated cells were induced by PI3K pathway in cHL cell lines

We previously found the involvement of the PI3K pathway in abnormal cell division by CD30 stimulation in human T-cell leukemia virus type 1 (HTLV-1)-infected cells.¹² We next examined the involvement of PI3K in the production of ROS and the generation of multinucleated cHL cells by CD30 stimulation. The PI3K inhibitor LY294002 reduced the generation of large and multinucleated cells in cHL cell lines with CD30 stimulation (Figure 6A-C). The effects of LY294002 on ROS production by CD30 stimulation are shown in Figure 6D. The green fluorescence per cell (Figure 6E) and per unit area in the cell (Figure S8) were significantly decreased by LY294002 treatment.

FIGURE 6. Phosphatidylinositol 3-kinase inhibits CD30-mediated generation of multinucleated cells and reactive oxygen species (ROS) production in classic Hodgkin lymphoma (cHL) cell lines. cHL cell lines were stimulated with ligand for CD30 (CD30L) for 5 days with or without the addition of the PI3K inhibitor LY294002 (10 μM). Media was replaced by fresh media with or without 10 μM LY294002 every 2 days. (A) Cells were stained by Giemsa. Scale bar, 20 μm. (B,C) Cell size and number of multinucleated cells were measured. (B) The area of the cells was measured using ImageJ software. Relative area sizes of 50 cells each were measured in three different regions, and the results are presented as mean ± SD. Mean value of the area of the cells without LY294002 was set to 100. (C) Fifty cells each were measured in three different regions, and the percentage of multinucleated cells is presented as mean ± SD. (D) Cells were stained with CellRox Oxidative Stress Reagent green and DAPI. Scale bar, 20 μm. (E) Fluorescence intensity by CellRox was determined by measuring more than 50 cells with ImageJ software in three different regions; relative fluorescence intensity per cell is presented as mean ± SD. Mean fluorescence intensity of the cells without LY294002 was set to 100. *p [less than] 0.05.

The results suggest that CD30 stimulation induces ROS through the PI3K pathway and induces DSBs and RS-like multinucleated cells in cHL cell lines.

DISCUSSION

Classic Hodgkin lymphoma is characterized by the emergence of giant mono- and multinucleated cells, called H and RS cells. Previous reports indicated that these giant cells are hyperploid cells and are not generated by cell fusion, but by abnormal cell division of smaller cHL cells.^23,24 This has also been shown to occur in the final phase of cell division due to abscission failure and has also been reported as a refusion of dividing cells.²⁵ However, its mechanisms are still under debate.²⁶ In this study, we found chromatin bridges in multinucleated cHL cells and the generation of DSBs by CD30 stimulation. These findings suggest that the generation of DSBs and subsequent formation of chromatin bridges triggered by CD30 stimulation appears to be a cause of abscission failure and the generation of RS cells. Subsequent segregation error by DSBs might also be involved in the generation of not only RS cells but also H cells.

The results of the present study link CD30 not only to the morphological features but also to the genetic complexity of cHL cells. Previous reports showed numerical chromosomal aberrations and dynamic chromosomal rearrangements in cHL cells.^27,28 The number of aberrant regions of chromosomes was significantly increased in cHL cell lines by CD30 stimulation. Our results showed that the increase of chromosomal imbalances after CD30 stimulation appears to occur not at specific sites but over wide regions spanning entire chromosomes. Most of the chromosomal imbalances induced by CD30 stimulation consist of a gain of chromosomes. These features are consistent with a previous report,²⁹ further suggesting that CD30 is involved in the genetic characteristics of cHL cells.

A previous report indicated that various chromosomal rearrangements already exist in mononuclear cHL cells, and that these rearrangements increase in multinucleated cHL cells.²⁸ Therefore, we prefer the notion that chromosomal instability by CD30 stimulation accumulates chromosomal rearrangements in mononuclear cHL cells, and causes abnormal cellular division by chromatin bridge, which generates multinucleated cHL cells. Detailed molecular mechanisms underlying this process await further clarification.

As chromosomal instability promotes clonal evolution,³⁰ our results also indicate the possibility that CD30 is involved in the clonal evolution of cHL cells through chromosomal instability. This might take place not only in mononuclear cHL cells, but also in multinucleated cHL cells. However, multinucleated cHL cells are disadvantaged for proliferation, and most of these cells are destined to die³¹; therefore, the commitment of these multinucleated cells to clonal evolution might be limited. CD30 stimulation rather suppressed the growth of cHL cell lines. This could explain the existence of a small amount of cHL cells in the large amount of reactive cells as a typical histological feature of cHL. However, the interruption of CD30 stimulation triggered regrowth of cHL cell lines. Therefore, CD30 stimulation and its interruption might possibly contribute to clonal evolution of cHL cells, although further studies for its mechanisms are required.

We showed that CD30 stimulation increased generation of large cells and multinucleated RS-like cells, and induced DSBs not only in cHL lines without EBV infection, but also those with EBV infection. However, these phenomena were not observed in immortalized B cells or unrelated T cell line Jurkat. These results suggest that intracellular statuses, which attenuate CD30-mediated DNA damage and subsequent formation of multinucleated cells, affect the response of the cells against CD30 stimulation, although further studies are required.

Recent work has verified that L428 and L540 cell lines are bona fide cHL cell lines.¹⁹ However, our results by PCA and heatmap analyses suggest that the biological backgrounds, as indicated by the RNA expression profile, differ among cHL cell lines. CD30 stimulation showed more potent effects on L540 compared to L428 cells. Morphologically, L428 appears to be much more diverse compared to L540. Differences in the background of each cHL cell line appear to affect differences in the response by CD30 stimulation.

We indicated that CD30 stimulation induces ROS and DSBs in cHL cells. We also indicated that the PI3K pathway is involved in ROS production by CD30 stimulation in cHL cells. Reactive oxygen species has been reported to be a major cause of DSBs.¹⁷ In fact, we showed induction of DSBs by ROS in cHL cell lines. Double-strand breaks provide critical damage to chromosomes by the breakage-fusion-bridge cycle and trigger chromosomal instability.³² Therefore, it is possible that CD30 is involved in the induction of DSBs and chromosomal instability through production of ROS by activation of the PI3K pathway in cHL cells.

We previously indicated that morphological change of small to large mononuclear and giant mono- or multinucleated cells, is regulated by increase of intracellular ROS.¹³ This observation suggests that generation of multinucleated cells by ROS is not specific for CD30 as we observed multinucleated cells and chromatin bridges in cHL cell lines without CD30 stimulation. The results in this study suggest that CD30 stimulation accelerates a series of morphological change of cHL cells. The morphological changes of cHL cells from an immature stem cell-like population to larger cells, which we named “differentiation”, appears to be synchronized with accumulation of DNA damage and chromosomal abnormalities.

Collectively, the results of this study link CD30 not only to the morphological features but also to the genetic complexity characteristic of cHL cells. A schematic representation that summarizes these results and their implications is shown in Figure 7. The schema suggests involvement of CD30 in chromosomal instability triggered by DSBs through the generation of ROS, which induces the generation of RS cell-like morphology and possible clonal evolution in cHL cells.

FIGURE 7. Schematic representation of the possible roles of CD30 signaling in chromosomal instability, generation of Reed-Sternberg cells, and clonal evolution. cHL, classic Hodgkin lymphoma; DSB, double-strand break; ROS, reactive oxygen species.

AUTHOR CONTRIBUTION

M.W. performed the experiments and wrote the manuscript. H.H. performed the experiments and coordinated the research. K.N. performed flow cytometry and supported the research. M.N. performed CGH experiment. K.U, M.O., and K.M. discussed the results and provided comments. R.H. planned, designed, coordinated the research and wrote the manuscript. All authors read the manuscript and approved its contents.

ACKNOWLEDGEMENTS

This work was supported in part by MEXT/JSPS KAKENHI grants to R.H. (17 K08728, 20 K07379, and 23 K06429), M.W. (19 K07442), and M.N. (19 K16580 and 22 K15420).

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest.

ETHICS STATEMENTS

Approval of the research protocol by an institutional review board: N/A.

Informed consent: N/A.

Registry and registration no. of the study/trial: N/A.

Animal studies: N/A.