SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

The accumulation of viral proteins in endoplasmic reticulum (ER) may cause ER stress responses and lead to either apoptosis or survival depending on the driving signals. The strong expression of latent membrane protein-1 (LMP1) in Epstein–Barr virus (EBV)-positive Hodgkin lymphoma (HL) cells raises the question whether LMP1-induced ER stress response is associated with the characteristic tumor biology in HL. In this study, we investigated the expression of ER stress signals (glucose-regulated protein 78 [GRP78], X-box binding protein 1 [XBP1], activating transcription factor 6 [ATF6], CCAAT enhance-binding protein homologous protein [CHOP] and phospho-apoptosis signal-regulating kinase 1 [pASK1]) on 156 cases of HL. Furthermore, LMP1 transfection on EBV-negative HL cell lines was used to explore the regulation of ER stress signals by EBV-LMP1. Interestingly, we demonstrated that the survival signals of ER stress response (GRP78, 62%; XBP1u [unspliced], 55%; XBP1s [spliced], 38%; ATF6, 91%) were dominantly expressed over the ER death signals (CHOP, 10%; pASK1, 7%) in all histological subtypes of HL with a similar level in both EBV-positive and EBV-negative cases. However, expression of ER signals did not bear prognostic significance. In vitro, LMP1 transfection increased the expression of GRP78 and XBP1, but attenuated the expression of death signals, CHOP and pASK1. These data indicate that EBV-LMP1 may play a role in shifting EBV-infected cells towards the survival pathway in the presence of ER stress in EBV-positive HL cases. (Cancer Sci 2011; 102: 275–281)

Hodgkin lymphoma (HL), first recognized in 1832 by Thomas Hodgkin, is characterized by scattered tumor cells against an inflammatory background.(1) In the vast majority of HL, the tumor cells are derived from germinal center B cells with defective surface B-cell receptors and/or epigenetic silencing of the B-cell transcriptional program.(2–5) The molecular cell biology of HL has been well studied,(6) and involved mainly the activation of pro-proliferative and anti-apoptotic signals, including nuclear factor κB (NFκB),(7) signal transducer and activator of transcription (STAT) and cytokine pathways,(8,9) and expression of cellular FADD-like interleukin-1 beta-converting enzyme (FLICE)-inhibitory protein (cFLIP) with inhibition of caspases.(10)

Although there is great advance in our understanding of the molecular biology of HL, the characteristic morphology of HL such as Reed–Sternberg (RS) cells and apoptotic mummified cells remains a mystery. The association of Epstein–Barr virus (EBV) with HL provides an important clue to unravel the underlying mechanism of HL tumorigenesis and tumor biology.(11,12) Germinal center B cells with a deficient B-cell receptor, from which the HL cells originate and should be destined to apoptosis, are rescued by EBV infection most likely through the oncogenic effect of latent membrane protein-1 (LMP1).(13,14) Previously, we demonstrated that aberrant expression of cyclin A was associated with RS-like morphology.(15) The presence of characteristic mummified cells in HL suggests that apoptosis signals are prevalent and associated with HL tumorigenesis, and the underlying tumor biology deserves to be clarified.

Recently, the accumulation of viral proteins in endoplasmic reticulum (ER) has been found to induce overloaded or stress-related unfolded protein responses in EBV-infected B cells.(16) The ER stress can initiate either pro-survival or pro-apoptotic signals, depending on whether the rescue effort is effective or not.(17) The ER adaptive or survival response is mediated by three ER proximal sensors: protein kinase-like ER-resident kinase (PERK); inositol requiring enzyme 1 (IRE1); and activating transcription factor 6 (ATF6).(18) ATF6 upregulates transcription of X-box binding protein 1 (XBP1) and provides XBP1 mRNA for IRE1-mediated splicing.(19) Glucose-regulated protein 78 (GRP78) is the target of both XBP1 and ATF6. GRP78, also known as BiP (immunoglobulin heavy chain binding protein), is a critical ER chaperone for tumor survival and resistance to anticancer therapy.(20,21) On the other hand, severe and prolonged ER stress eventually leads to cell death through the pro-apoptotic pathway mediated by CCAAT/enhance-binding protein homologous protein/growth arrest DNA damage-inducible gene 153 (CHOP/GADD153), caspase-12, caspase-7 and apoptosis signal-regulating kinase 1 (ASK1).(17,22,23) CHOP (GADD153) is a target of the PERK-eIF2α-ATF4 pathway; activated ASK1 is a mediator of the apoptotic pathway downstream of IRE1.(17) Therefore, the interaction of the above-described ER stress signals may determine the fate of either apoptosis or survival in cells under ER stress conditions. Given that HL is closely associated with EBV and LMP1-induced ER stress response may occur in EBV-infected cells,(16) it is intriguing to study the expression of ER stress signals in HL.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Hodgkin lymphoma cases.  We enrolled 156 cases of HL, including 46 cases of childhood HL from the First Children Hospital in Ho-Chi-Minh City, Vietnam,(12) 34 cases from the National Cheng Kung University Hospital and 76 cases from the Veterans General Hospital-Taipei. The latter were studied using the tissue array technique as described previously.(24) The diagnosis of HL was confirmed by two pathologists (KCC and IJS) and the immunophenotype of HL tumor cells was CD30+/CD15 ± /CD45/CD3/CD20. For comparison, tissues of lymphoid hyperplasia (= 8) and infectious mononucleosis (= 5) were also stained. The study protocol was approved by our institutional review board and was in accordance with the Helsinki Declaration of 1975, as revised in 2000. The tissue specimens were fixed in 10% neutral formalin solution and paraffin embedded. All the lymphoma cases were classified according to the World Health Organization (WHO) classification scheme.(2) Clinical data, including sex, age, serum level of lactate dehydrogenase (LDH), tumor site, Ann Arbor stage, treatment modality and overall survival in months, were obtained by reviewing patient charts. All patients with the available information were followed up and the duration ranged from 3.4 to 184 months. All HL cases were typically treated with the curative ABVD (epirubicin, bleomycin, vinblastine, dacarbazine) or BEACOPP (bleomycin, etoposide, epirubicin, cyclophosphamide, oncovin, procarbazine, prednisolone) regimen. For selected patients, surgical intervention and/or radiotherapy preceded chemotherapy.

EBV detection. In situ hybridization studies were performed in all cases to detect EBV-encoded RNA (EBER) using a PCR-derived digoxigenin-labeled DNA probe.(12,25) The test for nucleotide integrity was performed using a RNA positive control probe (Ventana Medical System, Inc., Tucson, AZ, USA). The intended target was the poly A tail in mRNA. Positivity was defined as more than 50% of tumor cells showing unequivocal EBER signals on low-power magnification (×40).

Immunohistochemical staining.  Immunohistochemical staining was performed on deparaffinized tissue sections of formalin-fixed material following microwave-enhanced epitope retrieval. Detection was done with streptavidin–biotinylated peroxidase-conjugated reagents (LSAB+ kit; Dako, Carpinteria, CA, USA) with 3-amino-9-ethyl carbazole (AEC) as the chromogen and hematoxylin for counterstain. The staining was graded as positive when more than 10% of tumor cells were stained, as defined previously.(26) The primary antibodies and working dilution were as follows: LMP1 (CS1-4, monoclonal, 1:20; Dako, Glostrup, Denmark), PERK (N-18, polyclonal, 1:25; Santa Cruz Biotechnology, Santa Cruz, CA, USA), GRP78/BiP (N-20, sc-1050, 1:50; Santa Cruz Biotechnology), XBP1 (M-186, polyclonal, 1:20; Santa Cruz Biotechnology, for spliced and unspliced forms), XBP1s (Poly6195, 1:100; BioLegend, San Diego, CA, USA; specific for the spliced form), CHOP/GADD153 (F-168, rabbit polyclonal, 1:50; Santa Cruz Biotechnology), phospho-ASK1 (pASK1 [Ser 83], sc-101633, 1:50; Santa Cruz Biotechnology) and ATF6 (70B1413.1, full-length and active/cleaved forms, IMG-273, 1:100; IMGENEX, San Diego, CA, USA). Appropriate tissues were used as a positive control, including breast carcinoma for CHOP and colon carcinoma for pASK1. Negative controls were performed with primary antibody absent or irrelevant isotype-matched monoclonal antibodies. The study population for XBP1s, ATF6, CHOP and pASK1 stains included cases from Vietnam (= 46) and the National Cheng Kung University Hospital (n = 34). Images were photographed using an Olympus DP12 Digital Microscope Camera (Olympus Co., Tokyo, Japan) and processed using Adobe Photoshop 7.0 (Adobe Systems Incorporated, San Jose, CA, USA).

Transfecting LMP1 into HL cell lines.  EBV-negative HL cell lines, L-428 and KM-H2 (DSMZ, Braunschweig, Germany) were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA). The vector expressing B95.8 EBV-derived LMP1 was constructed in the pSG5 plasmid encoding the simian virus 40 promoter as previously described.(27) Another vector, pEGFP-LMP1, and its control vector, pEGFP-C3, were gifts from Dr Hao-Ping Liu (Chang Gung University, Taoyuan, Taiwan). The N-terminally GFP-tagged LMP1 was generated by ligation of PCR-amplified DNA fragments to HindIII/BamHI-treated pEGFP-C3.(28) LMP1 expression construct was transfected by an electroporation machine Microporator (Digital Bio Technology Co., Ltd, Suwon, Korea) with 6 μg of DNA at 1100 V for 30 mS. The transfection efficiency was determined by flow cytometric analysis on fluorescent cells with a flow cytometer (FACSCalibur with CellQuest software; Becton Dickinson, Franklin Lakes, NJ, USA). Expression of LMP1 was detected by immunoblotting. The L-428 and KM-H2 cells were cultured and collected 48 h after transfection for further analysis. Cell viability was determined by the trypan blue exclusion test.

Immunoblotting assay.  The HL cell lysates were prepared by lysis in the sample buffer (3% sodium dodecyl sulfate, 1.6 M urea, 4%β-mercaptoethanol) and differential subcellular fractions were separated into cytosol and nucleus using a protein extraction kit (ProteoExtract Subcellular Proteome Extraction Kit; EMD Biosciences, Inc., La Jolla, CA, USA). Polyacrylamide gel electrophoresis and immunodetection of LMP1 were performed as described previously.(29) Other antibodies for western blots included GRP78 (1:5000; BD Biosciences-Pharmingen, San Diego, CA, USA), XBP1 (Ab37152, polyclonal, 1:2000; Abcam, Cambridge, UK), XBP1s (Poly6195, 1:1000; BioLegend), PERK (1:500), phospho-PERK (pPERK, [Thr 981], sc-32577, 1:500; Santa Cruz Biotechology), CHOP/GADD153 (1:1000), phospho-ASK1 (1:1000), ATF6 (1:500), α-tubulin as a cytosol marker (Ab-2, DM1A, 1:5000; NeoMarkers, Fremont, CA, USA) and histone H1 as a nuclear marker (AE-4, 1:1000; Millipore Corporation, Billerica, MA, USA). The ratio of ER stress proteins in cytosolic and nuclear fractions was expressed as the amount in cytosol or nucleus divided by the corresponding amount of cytosolic (α-tubulin) or nuclear (histone H1) markers, respectively. The band density for the target proteins in each sample were captured by an imaging analyzer (White Light Transilluminator; Bio-Rad Laboratories, Inc., Hercules, CA, USA), and were measured by Quantity One 4.4 software (Bio-Rad Laboratories, Inc.) according to the Quantity One user guide. The value of each protein was normalized to levels of control proteins (α-tubulin or histone H1). Immunoblotting was carried out in duplicate.

Reverse transcriptase–polymerase chain reaction (RT-PCR) and quantitative real-time PCR.  RNA extraction from L-428 and KM-H2 cells, reverse transcriptase reaction and PCR analysis were done as previously described.(30) The PCR primers for XBP1 were 5′-CCT TGT AGT TGA GAA CCA GG-3′ (forward) and 5′-GGG GCT TGG TAT ATA TGT GG-3′ (reverse); the predicted product was 442 bp for the unspliced form or 416 bp for the spliced form. The PCR primers for GRP78 were 5′-TTG CTT ATG GCC TGG ATA AGA GGG-3′ (forward) and 5′-TGT ACC CTT GTC TTC AGC TGT CAC-3′ (reverse); the predicted product was 1000 bp. The PCR primers for actin were 5′-TGG AGA AAA TCT GGC ACC AC-3′ (forward) and 5′-GAG GCG TAC AGG GAT AGC AC-3′ (reverse); the predicted product was 190 bp.

To further quantify the changes in transcriptional activity mediated by LMP1 transfection, we performed quantitative real-time PCR for mRNA of GRP78 and XBP1. The quantitative real-time PCR was performed by an intercalator-based method (LightCycler FastStart DNA Master SYBR Green I; Roche Applied Science, Mannheim, Germany) and the conditions included 95°C for 10 min for pre-incubation; 40 cycles of 95°C for 10 s, 57°C for 10 s and 72°C for 10 s; 65°C for 10 s for melting curves; and 40°C for 30 s for cooling. The mRNA of the actin gene was in parallel reverse transcribed and amplified as an internal control. Experiments were performed in triplicate and the results were analyzed by LightCycler Software Version 4.0 (Roche Applied Science).

Statistical analysis.  Appropriate statistical tests (χ2 test and Kendall’s tau correlation) were used to examine the associations between variables. Overall survival was measured from initial diagnosis to death from any cause; follow-up data of the surviving patients were assessed at the last contact date. The survival data were available from the Taiwan cases. Estimates of overall survival distribution were calculated using the Kaplan–Meier method.(31) Time-to-event distributions were compared using the log-rank test.(32) Cox proportional-hazard model was used to test the simultaneous influence on overall survival of all covariates found to be significant in the univariate analysis.(33) The P value referred to was two-sided. The analyses were carried using SPSS 13.0 statistical software (SPSS, Inc., Chicago, IL, USA).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Relative frequency and clinicopathological features of HL.  There were a total 156 cases of HL with a male to female ratio of 111:45 and a mean age at presentation of 22.8 years. The subtype distribution of 156 HL cases is summarized in Table 1. Nodular lymphocyte predominant (NLP) HL accounted for 5% of cases (= 8), nodular sclerosis (NS) HL for 62% (= 96), mixed cellularity (MC) HL for 20% (= 31), lymphocyte-rich classic (LRC) HL for 8% (= 12), lymphocyte depletion (LD) HL for 3% (= 4) and unclassified for 3% (= 5).

Table 1.   Results of EBER in situ hybridization and immunohistochemical stains of HL
SubtypeEBERLMP1GRP78XBP1 (M-186)XBP1sATF6CHOPpASK1
  1. †HL unclassified. ATF6, activating transcription factor 6; CHOP, CCAAT enhance-binding protein homologous protein; EBER, Epstein–Barr virus-encoded RNA; GRP78, glucose-regulated protein 78; HL, Hodgkin lymphoma; LD, lymphocyte depletion; LMP1, latent membrane protein-1; LRC, lymphocyte-rich classic; MC, mixed cellularity; NLP, nodular lymphocyte predominant; NS, nodular sclerosis; pASK1, phospho-apoptosis signal-regulating kinase 1; XBP1, X-box binding protein 1; XBP1s, spliced.

NLP85% 3/838% 2/825% 4/850% 3/743% 2/450% 3/3100%0/40%0/40%
NS9662%53/9258%31/9134%61/9366%55/9160%19/4641%44/4892%4/3711%2/494%
MC3120%26/3087%20/2871%16/3152%12/3040% 4/1527%13/1587%0/120%1/166%
LRC128% 4/1233% 6/1250% 8/1267% 6/1250% 3/838% 8/989%2/729%2/825%
LD43% 2/450% 1/425% 2/450% 4/4100% 1/250% 2/2100%0/20%0/20%
HL†53% 4/580% 3/560% 4/580% 2/450% 0/10% 1/1100%0/10%1/333%
Total156100%92/15161%63/14843%95/15362%82/14855%29/7638%71/7891%6/6310%6/827%

Except for five HL cases that were inevaluable, all other cases yielded valid results for EBV detection. EBER positivity was seen only in those areas of tissues involved by HL, and the tumor cells containing EBV genome showed positive signals in nuclear regions. The overall EBER positivity rate was 61% (92/151), with 38% in NLPHL (three positive cases from Vietnam), 58% in NSHL, 87% in MCHL, 33% in LRCHL, 50% in LDHL and 80% in unclassified (Table 1).

Immunohistochemical staining features.  In lymphoid hyperplasia and infectious mononucleosis, the germinal center lymphocytes and EBER-positive cells were all negative for ER stress markers.

The results of immunostaining for LMP1 and ER stress markers for HL are summarized in Table 1. The overall positive rates for LMP1, GRP78, XBP1 (M-186), XBP1s, ATF6, CHOP and pASK1 were 43% (63/148), 62% (95/153), 55% (82/148), 38% (29/76), 91% (71/78), 10% (6/63) and 7% (6/82), respectively. The PERK antibody used in this study did not work on paraffin sections. The staining patterns of survival markers, GRP78 and XBP1 (M-186), were mainly in the cytoplasm and Golgi-ER area, respectively, in HL cells (Fig. 1A,B), but mummified cells were usually unstained. Since the XBP1 (M-186) expression pattern did not follow the nuclear staining in plasma cells observed previously,(26,34) we performed additional staining with an XBP1s (spliced form, Poly6195; BioLegend)-specific antibody. Interestingly, we found weak or dim nuclear expression (Fig. 1C) in tumor cells in 38% of HL cases, a frequency higher than previous ones.(26,34) The antibody, XBP1 (M-186, polyclonal; Santa Cruz Biotechology), against the amino terminus may bind both unspliced and spliced XBP1, while a spliced form-specific antibody, XBP1s (Poly6195; BioLegend), binds the nuclear fraction of XBP1s in a subset of HL cases. Thus, the cytoplasmic expression pattern of XBP1 (M-186) is consistent with unspliced XBP1 (XBP1u). In contrast, HL cells, including mummified cells, rarely expressed CHOP or pASK1 (apoptotic markers), while some inflammatory cells showed weak staining in the cytoplasm (Fig. 1D,E). Nuclear ATF6 staining was found in the majority of HL cases (91%), including NLP (Fig. 1F) and NS (Fig. 1G) types, and in most tumor cells in each case.

image

Figure 1.  Staining patterns of endoplasmic reticulum (ER) stress markers in Hodgkin lymphoma (HL) cells. (A) Glucose-regulated protein 78 (GRP78) shows a uniform cytoplasmic staining (inset, ×400). (B) X-box binding protein 1 (XBP1) (M-186) is expressed mainly in the ER-Golgi area (inset, ×400). (C) XBP1s (spliced) expression is exclusively in the nuclei and usually weak. (D,E) In contrast to the dim/weak staining of inflammatory cells, HL cells, including the mummified cell, are negative for CHOP (D) or phospho-apoptosis signal-regulating kinase 1 (pASK1) (E). (F,G) Activating transcription factor 6 (ATF6) is constitutively expressed in each subtype of HL, including nodular lymphocyte predominant (F; inset, ×400) and nodular sclerosis (G; inset, ×400). (A,B, ×100; C–E ×400; F, ×40; G, ×100; hematoxylin stain.)

Download figure to PowerPoint

GRP78 and XBP1 (M-186) showed similarly positive rates for HL (62%vs 55%), and both markers had a significant correlation for each case (Kendall tau correlation coefficient r = 0.235, = 0.036). The staining correlation between CHOP and pASK1 was also significant (r = 0.306, = 0.015) with similarly low positive rates (10%vs 7%). There was also strong correlation between EBER in situ hybridization and LMP1 stain (r = 0.752, < 0.001). The presence of ER stress signals (GRP78, XBP1 [M-186], XBP1s, ATF6, CHOP and pASK1) was not associated with EBV or LMP1 overexpression. The expression of GRP78, XBP1 (M-186), XBP1s and ATF6 was almost equally distributed in each HL subtype and in EBV-positive (GRP78, 64%; XBP1 [M-186], 57%; XBP1s, 43%; ATF6, 92%) and EBV-negative cases (GRP78, 61%; XBP1 [M-186], 54%; XBP1s, 35%; ATF6, 89%). Taken together, we found that the majority of primary HL cases expressed survival but not death signals of ER stress, and the distribution was present in all histological subtypes and in both EBV-positive and EBV-negative cases in a similar frequency.

LMP1 transfection in HL cell lines increases expression of GRP78 and XBP1.  EBV-positive HL cases provide a model to dissect the relationship of EBV infection with ER stress. We used EBV-negative HL cell lines L-428 and KM-H2 to study the role of LMP1 in ER stress induction. Since the transfection efficiency of KM-H2 (5–40%) was much lower than that of L-428 (mean 69%, Fig. S1), only the data from L-428 are shown. Western blot analysis showed that GRP78 (cytoplasmic fraction, from 0.7 to 1.1) and XBP1 (cytoplasmic fraction, from 0.6 to 0.8; nuclear fraction, from 0.8 to 0.9) were increased by LMP1 expression at 48 h post-transfection (Fig. 2A). In contrast, death signal proteins were attenuated (CHOP, nuclear fraction, from 1 to 0.5; and pASK1, cytoplasmic fraction, from 0.8 to 0.5, Fig. 2A).

image

Figure 2.  Latent membrane protein-1 (LMP1) transfection in Hodgkin lymphoma (HL) cell lines increases protein expression and transcriptional activity of glucose-regulated protein 78 (GRP78) and X-box binding protein 1 (XBP1). (A) LMP1 mildly enhances protein expression of GRP78 and XBP1 but attenuates death signal expression of endoplasmic reticulum (ER) stress. Western blotting for the ratio of tested proteins in cytosolic and nuclear fractions was performed on LMP1-transfected L-428 cells 48 h post-transfection. The parental and GFP-transfected L-428 cells were negative controls. After normalized with cytosolic (α-tubulin) and nuclear (histone) markers, LMP1-transfected cells expressed more GRP78 and XBP1 in cytosol and nucleus than the control group. Activating transcription factor 6 (ATF6) (GFP, 1.2; LMP1, 0.8) and mediators for the pro-apoptotic pathway (PERK [from 1.1 to 0.7], CHOP [from 1 to 0.5] and phospho-apoptosis signal-regulating kinase 1 (pASK1) [from 0.8 to 0.5]) were decreased in expression by LMP1 transfection. C, cytosol; N, nucleus. (B) RT-PCR shows a modest increase in the transcriptional activity of GRP78 and XBP1 by LMP1 transfection. After normalization of the actin gene amount, the intensity of GRP78 and XBP1 mRNA increased after LMP1 transfection (GRP78, from 1.1 to 1.5; XBP1, from 1.2 to 1.4). Studies were performed on both L-428 and KM-H2 cell lines; the results shown are on L-428 cells. XBP1u, unspliced; XBP1s, spliced.

Download figure to PowerPoint

LMP1 transfection induces increase in transcriptional activity of GRP78 and XBP1.  RT-PCR analysis demonstrated a mild increase in the amount of GRP78 (from 1.1 to 1.5) and XBP1 mRNA (from 1.2 to 1.4) by LMP1 at 48 h post-transfection (Fig. 2B). The increase in transcriptional activity of GRP78 and XBP1 was also assessed by quantitative real-time PCR, which showed greatly increased results (GRP78, from 2.03 to 3.52; XBP1, from 1.98 to 3.55, Table 2 and Fig. S2).

Table 2.   Results of relative quantification of GRP78 and XBP1 by real-time PCR
SetSampleCt value (± standard deviation)Concentration ratioNormalized ratio
  1. Ct value, value of threshold cycle; GRP78, glucose-regulated protein 78; LMP1, latent membrane protein-1; XBP1, X-box binding protein 1.

Actin and GRP78GRP78-none27.20 ± 0.090.02 ± 0.001.00 ± 0.00
Actin-none21.30 ± 0.00  
GRP78-GFP26.56 ± 0.170.03 ± 0.002.03 ± 0.31
Actin-GFP21.68 ± 0.00  
GRP78-LMP125.77 ± 0.110.06 ± 0.013.52 ± 0.04
Actin-LMP121.68 ± 0.00  
Actin and XBP1XBP1-none24.49 ± 0.070.11 ± 0.001.00 ± 0.00
Actin-none21.30 ± 0.00  
XBP1-GFP23.89 ± 0.090.22 ± 0.011.98 ± 0.24
Actin-GFP21.68 ± 0.00  
XBP1-LMP123.05 ± 0.030.39 ± 0.013.55 ± 0.24
Actin-LMP121.68 ± 0.00  

Correlation of survival with clinicopathological factors in HL patients.  The survival of HL patients affected by clinicopathological features is summarized in Table 3. In the univariate analysis, the significant parameters related to worse prognosis included old age (>60 years, < 0.001), EBV association (= 0.046) and LMP1 expression (= 0.010). The other factors were insignificant. In the multivariate analysis, the only significant factor related to worse prognosis was old age (= 0.001, hazard ratio 0.259, 95% confidence interval 0.12–0.555). Since EBER and LMP1 were strongly associated with old age, they were insignificant for prognosis (0.611 and 0.183).

Table 3.   Clinicopathological parameters affecting survival of HL patients
FactorsPoorer prognosisPercentage†Survival analysis (P value)‡
  1. †Taiwan cases. ‡Kaplan–Meier method. ATF6, activating transcription factor 6; CHOP, CCAAT enhance-binding protein homologous protein; EBER, Epstein–Barr virus-encoded RNA; GRP78, glucose-regulated protein 78; LDH, lactate dehydrogenase; LMP1, latent membrane protein-1; NS, nodular sclerosis; pASK1, phospho-apoptosis signal-regulating kinase 1; XBP1, X-box binding protein 1; XBP1s, spliced.

Age>60 years38.1<0.001
SexMale vs female68.00.608
StageHigh (3–4)52.90.742
LDH>200 IU/L50.00.601
B symptomsPresence34.40.15
SubtypeNS vs Non-NS64.90.43
EBERPresence47.90.046
LMP1Presence40.00.01
GRP78Presence68.80.930
XBP1 (M-186)Presence52.70.154
XBP1sPresence30.30.345
ATF6Presence87.50.913
CHOPAbsence6.30.473
pASK1Absence3.20.555

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

In this study, we documented for the first time that the expression of survival signals of ER stress (GRP78, XBP1 and especially ATF6) was a universal phenomenon in all histological subtypes of HL and was found in both EBV-positive and EBV-negative cases at a similar level. In vitro, the HL cell line studies showed a mild increase in transcriptional activity and protein expression of ER stress survival signals, GRP78 and XBP1, and decreased expression of death signals, by EBV-LMP1 transfection. These findings, in parallel with a previous report that LMP1 expression drives signaling through NFκB to rescue the defective germinal center B cells from apoptosis,(13,14,35) suggest that LMP1 can also drive the survival signals to rescue the EBV-infected cells from apoptosis under ER stress conditions.(16)

It is intriguing that HL cells expressed survival but not death signals of ER stress. Bax and Bak, the pro-apoptotic Bcl-2 family members, are pivotal in the triggering and mediation of ER stress-induced apoptosis.(36) The HL cells express both anti-apoptotic and pro-apoptotic Bcl-2 family proteins and the former counteract the latter, resulting in the survival of HL cells.(37,38) These reports are consistent with our finding that survival but not apoptotic signals of ER stress predominate in HL cells. It is noteworthy that the expression of ER stress signals was common to both EBV-positive and EBV-negative HL. Constitutive activation of the NFκB pathway is also common for both EBV-positive and EBV-negative HL. In the EBV-positive cases, LMP1 overexpression appears to be critical, whereas in the EBV-negative cases, other unidentified viral proteins or similar mechanisms involving NFκB activation or ER stress may play a role. The significance of expressing ER stress survival signals in EBV-negative HL cases might additionally contribute to the occurrence of hypoxia, which has been found to induce expression of vascular endothelial growth factor (VEGF) in HL cells along with increased tumor cell density.(39,40)

The interaction among survival and death signals of ER stress response may determine the fate of EBV-infected B cells. Among the ER stress signals, the expression of XBP1 is interesting. XBP1 is not only important for the stress response but also for B cell differentiation. XBP1 is expressed throughout B cell development, governs late events in plasma cell differentiation and is not required for antigen-specific memory B cell commitment.(41,42) XBP1 is widely expressed and is not a marker of ER stress per se. It is the expression of the spliced form of XBP1 (XBP1s), which is a marker of the unfolded protein response.(43) XBP1s has a unique carboxyl terminus, which mediates transcriptional activation. Increased expression of XBP1u might otherwise reflect a cell adapted to secretory activity with an expanded ER apparatus, not responding to ER stress. However, GRP78 is also a target of ATF6, which can mediate transcriptional regulation of XBP1 mRNA.(44) Hodgkin lymphoma showed expression of ATF6 in almost all (91%), GRP78 in most (62%) and XBP1s in minor (38%) cases. Taken together, as depicted in Figure 3, HL cells constitutively express ATF6, which triggers the major survival pathway of ER stress through GRP78 activation and the minor pathway through XBP1s activation. In addition, LMP1 shifts the HL cells towards the ER stress-induced survival pathway in EBV-positive cases (Fig. 3). Expression of XBP1u may alternatively reflect HL cells in the recovery phase of ER stress.(43) Notably, we found weak nuclear expression of XBP1s in tumor cells in 38% of HL cases, a frequency higher than the previous ones.(26,34) This might be due to different epitopes of antibodies and/or different methodology.

image

Figure 3.  Model for dominant expression of endoplasmic reticulum (ER) stress survival signals in Hodgkin lymphoma (HL). On ER stress, the protein kinase-like ER-resident kinase (PERK)-eIF2α pathway mediates translational attenuation for pro-survival regulation. Alternatively, phosphorylated eIF2α can translate activating transcription factor 4 (ATF4), upregulate CHOP and lead to cellular apoptosis. Both inositol requiring enzyme 1 (IRE1) and ATF6 mediate transcriptional activation for survival. ATF6 upregulates transcription of X-box binding protein 1 (XBP1) and provides XBP1 mRNA for IRE1-mediated splicing. Glucose-regulated protein 78 (GRP78) is the target of both XBP1 and ATF6. In contrast, prolonged and severe ER stress activate apoptotic signals through induction of CHOP, caspase-12 and the IRE1-ASK1 pathway. We found here that HL cells express dominant survival signals of ER stress, including GRP78, XBP1 and ATF6. The latter is constitutively expressed in almost all cases and triggers the survival pathway of ER stress through GRP78 (major) and XBP1 (minor) activation. EBV-LMP1 transduction shifts ER stress signals toward the survival end by increasing expression of GRP78 and XBP1. ASK1, apoptosis signal-regulating kinase 1; EBV, Epstein–Barr virus; LMP1, latent membrane protein-1.

Download figure to PowerPoint

The oncogenic role of EBV in HL remains to be uncovered. One model suggests that EBV initially infects naïve B cells, which become activated and transform into germinal center B cells by antigen selection, but then persist due to establishment of the EBV type II latent transcription program.(24,45) Constitutive expression of EBNA1 and LMP1/2A may block further differentiation of these infected B cells allowing time for accumulation of additional acquired mutations that leads to neoplastic transformation.(45,46) Thus, it is intriguing that expression of ER stress survival signals was present in the HL cells but absent or dim in the germinal center lymphocytes. Because LMP1 expression regulated by an autocatalytic process in EBV-infected B cells induces the ER stress signaling pathway, and infected cells with lower levels of LMP1 would be more likely to escape the host’s immune attack,(16) it appears that the HL precursor cells first survive the immune selection and then escalate ER stress signaling along with the increasing LMP1 level. We reason that surviving the ER stress may be the essential and decision-making step for the rare development of HL in the large population of latent EBV infection.

In conclusion, our study suggests that the majority of HL cases adapted and survived ER stress and the expression of ER stress survival signals may be an additional mechanism in cooperation with other pathways, such as NFκB, to prevent HL cells from stress-induced apoptosis. In EBV-positive cases, overexpression of LMP1 may play a role in the induction of ER stress survival signals, while hypoxia or other events may contribute to the expression of ER survival signals in EBV-negative cases.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

This work was supported by grants (NCKUH-97-02008 from the National Cheng Kung University Hospital, and NSC 97-2320-B-006-022-MY3 from the National Science Council, Taiwan) to K.C. Chang, and grants from the National Health Research Institute, Taiwan, to I. J. Su. This study was presented in poster format at the 98th United States and Canadian Academy of Pathology (USCAP) Annual Meeting, 7–13 March 2009, Boston, MA, USA. The authors would like to thank Hao-Ping Liu (Chang Gung University, Taoyuan, Taiwan) for providing the pEGFP-LMP1 and pEGFP-C3 vectors, Mi-Chia Ma (Department of Statistics, National Cheng Kung University, Tainan, Taiwan) for advice on statistical interpretation, and Dan Jones (Department of Hematopathology, University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA) for critical review of the manuscript.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References
  9. Supporting Information

Fig. S1. Transfection rates of KM-H2 and L-428.

Fig. S2. Representative results of real-time PCR for mRNA quantification.

FilenameFormatSizeDescription
CAS_1765_sm_FigS1.jpg765KSupporting info item
CAS_1765_sm_FigS2.jpg659KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.