Infectious Causes of Cancer
Loss of expression of Epstein-Barr virus nuclear antigen-2 correlates with a poor prognosis in cases of pyothorax-associated lymphoma
Article first published online: 29 DEC 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 118, Issue 11, pages 2782–2789, 1 June 2006
How to Cite
Takakuwa, T., Ham, M. F., Luo, W.-J., Nakatsuka, S.-i., Daibata, M. and Aozasa, K. (2006), Loss of expression of Epstein-Barr virus nuclear antigen-2 correlates with a poor prognosis in cases of pyothorax-associated lymphoma. Int. J. Cancer, 118: 2782–2789. doi: 10.1002/ijc.21725
- Issue published online: 14 MAR 2006
- Article first published online: 29 DEC 2005
- Manuscript Accepted: 31 OCT 2005
- Manuscript Received: 10 JUN 2005
- Ministry of Education, Science and Culture, Japan. Grant Numbers: 16790206, 15406013, 17590343, 16390105
- Epstein-Barr virus;
- Epstein-Barr virus nuclear antigen (EBNA)-2;
- pyothorax-associated lymphoma
Pyothorax-associated lymphoma (PAL) is a non-Hodgkin's lymphoma, which develops in the pleural cavity of patients who have had pyothorax for over 20 years, and is strongly associated with Epstein-Barr virus (EBV) infection. The expression of latent genes, especially EBV nuclear antigen-2 (EBNA-2), influences the growth characteristics and malignant features of EBV-infected cells. Here, the effect of EBNA-2 expression on clinical features was examined in 13 cases of PAL. The EBNA-2 transcript was detected in 8 cases but was absent in 5. There was a significant difference in survival between patients with the transcripts and those without: the 1-year survival rate was 87.5 and 0%, respectively (p < 0.01). There was a discrepancy between EBNA-2 expression and EBNA promoter usage in 6 cases. The Cp/Wp promoter was used in 3 EBNA-2-negative cases, whereas the Qp promoter or multiple promoters were used in 3 EBNA-2-positive cases. Analysis of PAL cell lines provided a clue as to the mechanism underlying the discrepancy between EBNA-2 expression and EBNA promoter usage. Loss of EBNA-2 expression, irrespective of the latency pattern, is correlated with a poor prognosis, suggesting that down-regulation of the EBNA-2 expression could be selection pressure for the progression of PAL. © 2005 Wiley-Liss, Inc.
Malignant lymphoma frequently develops in the pleural cavity of patients with at least a 20-year history of pyothorax.1 This tumor stands as a distinct clinicopathologic entity, and thus the term pyothorax-associated lymphoma (PAL) has been proposed by Aozasa and coworkers.2 PAL is now listed as a distinct disease entity in the World Health Organization's Classification of tumors.3 PAL is a non-Hodgkin's lymphoma, mostly a diffuse large B-cell lymphoma, and is strongly associated with Epstein-Barr virus (EBV) infection. PAL usually responds to chemotherapy. Among the 106 cases of PAL previously reported by us,1 the prognosis varies with survival times ranging from 1 to 178 months. Approximately 46.2% of the patients died within 8 months of being diagnosed with PAL, while the 3-year survival rate in the remaining patients was 47.0%, showing that PAL cases could be subdivided into 2 groups.
EBV-infected cells show 3 patterns of latency-associated viral gene expression known as latency (Lat) I, II and III.4, 5 EBV-infected lymphoblastoid cell lines and immunocompromised hosts with lymphoproliferative diseases express all of the 9 latency-associated viral proteins, including 6 EBNAs (1, 2, 3A, 3B, 3C and LP) and 3 LMPs (1, 2A and 2B).4, 6 This pattern of latent gene expression is designated Lat III. In Lat II, as observed in EBV-associated Hodgkin's lymphoma and nasopharyngeal carcinoma, the EBNA-1, LMP-1 and LMP-2 genes but not other EBNA genes are transcribed.7, 8 In Lat I, EBV-infected cells express only EBNA-1, which is essential for episomal replication of EBV. This pattern is observed in Burkitt's lymphomas9 and gastric carcinomas.10
Immunohistochemical (IHC) study of 106 clinical cases revealed that PAL cells showed a Lat III pattern at the protein level.11 However, the expression of EBNA-2 and LMP-1 varied among 6 PAL cell lines established in our laboratory from 6 patients with PAL.12, 13 A recent IHC study of a large number of PAL cases revealed that 67 and 70% were positive for EBNA-2 and LMP-1, respectively.1 Results of IHC analysis might be influenced by the conditions at fixation or the condition of the tissues themselves, including degeneration, necrosis and degree of inflammatory change. In addition, cross reactivity of the antibodies used should be taken into account. Therefore, the pattern of latency could be defined more accurately through an analysis of the promoter usage than IHC analysis. It is evident from recent studies that EBNA promoter usage differs among the Lat patterns, i.e., the CpWp promoter is used in Lat III and the Qp promoter in Lat I/II.14 In this study, EBNA promoter usage was examined in cell lines and clinical samples of PAL.
The expression of latent genes correlates with the growth characteristics and malignant features of EBV-infected cells. EBNA-2 plays an especially important role in the transformation of EBV-infected B-lymphocytes through switching of the EBNA promoter usage from Wp to Cp,15, 16 induction of LMP-1 and LMP-2 expression,17, 18, 19 and control of key cellular growth-related genes.20 In this context, the effect of EBNA-2 expression on clinical features was examined in 13 cases of PAL. The results showed that loss of EBNA-2 expression was correlated with a poor prognosis. Analysis of PAL cell lines provided a clue as to the mechanism underlying the discrepancy between EBNA-2 expression and EBNA promoter usage. This is the first report demonstrating the relationship between loss of EBNA-2 gene expression and poor prognosis in cases of PAL.
Materials and methods
Patients and clinical samples
Thirteen patients with PAL admitted to hospitals in the Osaka area, Japan, between 1995 and 2003 were selected for the present study. Informed consent was obtained from all patients. Age at first admission for PAL ranged from 65 to 82 (median, 72.0) years, with a male to female ratio of 8:5. Histological specimens obtained by biopsy or surgery were fixed in 10% formalin and routinely processed for paraffin embedding. Histological sections, cut 4 μm thick, were stained with hematoxylin and eosin, or subjected to the immunoperoxidase procedure. Tumors were histologically classified as diffuse large lymphomas (immunoblastic variant: 11 cases, centroblastic variant: 2 cases) according to the World Health Organization's criteria. Follow-up data was available for all these patients.
Five cell lines, OPL-1, -2, -3, -5 and -7, were established from 5 patients with PAL; their characteristics were reported in part elsewhere.12, 13 The original tumors in the 2 patients from whom OPL-2 and -5 were established showed rapid growth, and these patients died within 2 and 6 months after the biopsy, respectively. The patients from whom OPL-1, -3 and -7 were derived showed an indolent onset, and survived for 24, 30 and more than 36 months, respectively. Therefore, OPL-5, OPL-2, OPL-3, OPL-7 and OPL-1 were categorized as Agg-1, Agg-2, Ind-1, Ind-2 and Ind-3, respectively. A B95-8 EBV immortalized reference cell line (R-LCL) was established in our laboratory. Burkitt's lymphoma cell lines, Ramos and Namalwa, were purchased from the Japanese Collection of Research Bioresources (Tokyo, Japan). NAB-2, another Burkitt's lymphoma cell line, was generously provided by Dr. Tucker (University of Texas, Austin). These cell lines were grown in RPMI 1640 medium supplemented with 10% heat-inactivated FCS at 37°C in 5% CO2 in air.
RT-PCR analysis for detection of EBNA-1 promoter usage and latent gene expression
Fresh samples from 13 patients and 5 cell lines were subjected to analysis. Total RNA was extracted from the samples with the TRIzol reagent (Invitrogen, Inc., Rockville, MD), according to the manufacturer's instructions. Five micrograms of total RNA was reverse-transcribed by random hexamer priming using the Superscript first strand synthesis system (Invitrogen).
Patterns of EBV latent gene transcription were determined based on an analysis of gene promoter usage. The EBNA genes are transcribed as single polycistronic RNA molecules, which are spliced to generate the individual EBNA molecules.14 In Lat III, transcription initiates from the promoter located within the BamHI regions C and/or W (Cp, Wp), resulting in a Y/U/K-spliced EBNA-1 mRNA. Usage of Cp and Wp promoters was further distinguished using primer pairs C1/C2F and W2R, and W0/W1F and W2R, respectively. In Lat I or II, EBNA-1 is transcribed with the use of a downstream promoter located in the BamHI F and Q regions (the Q promoter, Qp), generating a Q/U/K-spliced mRNA without other EBNA products.21, 22 The Fp transcript, which is used in lytic cycles, was also analyzed. Expression of the EBV latent gene EBNA-2 was also analyzed. Lat I and Lat II were placed in the same category in the present study, because LMP-1 expression is sometimes very faint and unstable. The primer pairs used in the present study are summarized in Table I. All primer pairs were designed to cross the splice sites, thus one could distinguish cDNA from genomic DNA by size. To evaluate the sensitivity of RT-PCR for the detection of the latent genes, the R-LCL (predominantly latency III, with some cells in the lytic cycle) was tested as an undiluted population or diluted at 10−1 down to 10−5 in the EBV genome-negative Ramos line. cDNA diluted at 10−2 to 10−3 gave a positive signal for each RT-PCR (Fig. 1).
|Primer||Nucleotide sequence (5′–3′)||Ta (°C)||size (bps)||Sensitivity|
|Primers for amplification of the EBV genes|
|B-actin||F||5′ CTTCCTTCCTGGGCATGGAG 3′||55||315||Not determined|
|R||5′ TGGAGGGGCCGGACTCGTCA 3′|
|Primers for inverse PCR|
In situ hybridization and immunohistochemistry
RNA in situ hybridization using the EBER-1 probe was performed as previously described with some modifications.23 Briefly, 30-base oligonucleotide probes; 5′-AGACACCGTCCTCACCACCCGGGACTTGTA-3′, which were the sense and antisense for a portion of the EBER-1 gene, a region of the EBV genome that is actively transcribed in latently infected cells, were synthesized using a DNA synthesizer. As a positive control, the Raji cell line was used. As negative controls, the hybridizing mixture containing sense probe or antisense probe after RNase treatment was used. Immunophenotyping of EBNA-2 protein was performed with the paraffin sections of original tumor specimens as well as established cell lines using the avidin–biotin-peroxidase complex method with anti-EBNA-2 antibody (PE2) (Dakopatts, Glostrup, Denmark).
Southern blot analysis and DNA probes
Aliquots (10 μg) of DNA extracted from the cell lines were digested with BamHI, electrophoresed on 0.7% agarose gels and transferred to a Hybond N+(Amersham) nylon membrane. The filters were hybridized with BamHI-C, W, -Y, -H, -F, -Q and -U of the EBV genome as well as 2 probes that contain the unique DNA sequences at either end of the EBV genome. The BamHI fragments mentioned above and BamHI-Nhet fragments used in the vector pUC119 to make the probes-L and -R were kindly provided by Dr. Kenzo Takada (Hokkaido University). Probes-L and -R are portions of Nhet containing nucleotides 4 to 3957 and 166615 to 169424, respectively (GenBank ACC M80517). Each probe was radiolabeled by the random prime method according to the instructions of the manufacturer (Amersham).
Inverse-PCR for recovery of junction sites in the BamHI-U fragment in Agg-2 lines
The primers, BamHI-UF1; 5′-GCACCACCATCGTCCACGAG-3′ and BamHI-UR1; 5′-CGGGAACGGCTCATGGCCCTACC-3′, were designed to extend outward from the BamHI-U fragments of EBV. DNA isolated from the Agg-2 cell line was digested with Bam HI. The DNA (100–1,000 ng/μl) was circularized by ligation using T4 DNA ligase, and subjected to inverse PCR as a template. Thermocycling conditions for each reaction were 35 cycles of denaturation at 98°C for 5 sec, and annealing and extension at 68°C for 5 min, followed by a final 7 min extension at 72°C. PCR products were purified and cloned into pCR 2.1-TOPO (Invitrogen, Carlsbad, CA). Sequencing was performed by the dideoxy chain termination method, using a DNA sequencing kit (Applied Biosystems, Foster City, CA). The samples were analyzed using a Genetic Analyzer (ABI PRISM 310′, Applied Biosystems).
Fluorescence in situ hybridization
To detect the EBV genome by FISH, the cosmid pJB8 containing an EcoRI fragment (37 kbp) of EBV DNA was used to create a probe labeled with biotinylated 14-deoxyadenosine triphosphate by nick translation. FISH was performed by the method described previously.24 The hybridized probe was detected by incubation with FITC-conjugated avidin. Cy3-labelled probes consisting of repeat sequences of α-satellite DNA that hybridize with the centromere regions of chromosomes were purchased from Vysis (Downers Grove, IL). The hybridization signals were observed with a BX50 epifluorescence microscope (Olympus, Tokyo, Japan), and microphotographs were taken on Provia Fujichrome 100 films (Fuji Film, Tokyo, Japan).
Kaplan–Meier methods with a log-rank test were used to calculate survival rate and differences in survival curves. p-values of <0.05 were considered statistically significant.
EBNA-2 expression and EBNA promoter usage in PAL clinical samples
RNA in situ hybridization using the EBER-1 probes revealed positive signals in the nucleus of tumor cells in all 13 cases (data not shown). EBNA-2 protein was judged to be present in 9 of 12 cases. The rate of positive cases was similar to a previous report.1 The EBNA-2 transcript was detected in 8 cases (Case 1–8) but absent in 5 (Case 9–13) (Table II, Fig. 2). A discrepancy between the expression of the EBNA-2 transcript and protein was detected in 4 cases (Case 7, 9, 12 and 13).
|Case no.||Age (yr)||Sex||Duration of pyothorax (yr)||Age at diagnosis of tuberculosis (yr)||Histology||Follow up (months after diagnosis)||IHC||RT-PCR|
|LMP-1||EBNA-2||EBNA-2||Cp-transcripts||Wp-transcripts||EBNA1-Y3K||EBNA1-QK||Fp-K||Predicted latency||Cell line established|
|3||71||F||52||19||IB||36 < (A)||+||+||+||+||+||+||−||−||III||Ind-2|
|4||74||F||55||19||CB||36 < (A)||+||+||+||ND||ND||+||−||ND||III|
In 5 of 8 cases in which the EBNA-2 transcript was detected, Cp and Wp promoters were used, indicating Lat III. In one case (Case 8), only EBNA1-QK was detected, indicating Lat I/II. In the remaining 2 cases (Case 6, 7), EBNA1-QK and Cp and Wp were present, indicating the usage of multiple promoters. In 2 of 5 cases, in which the EBNA-2 transcript was not detected, only EBNA1-QK was detected, thus indicating Lat I/II. In the remaining 3 cases, Wp and/or EBNA-Y3K were detected, indicating Lat III. In these cases, no switching of the promoter from Wp to Cp was observed. The Fp transcripts used in lytic cycles were not detected in any samples.
Tumor cells in all but 2 cases, in which data on EBNA-2 mRNA expression was available, had an immunoblastic morphology. Therefore, there was no significant difference in histology with EBNA-2 expression.
Prognosis of PAL patients
Follow-up study revealed that the median survival was 14 months. The estimated 1, 2, and 3-year survival rate was 46.2, 30.8 and 15.4%, respectively. There was a significant difference in survival between patients with the EBNA-2 transcript and those without: the 1-year survival rate was 87.5 and 0%, respectively (p < 0.01) (Fig. 3).
Analysis of PAL cell lines
There was a discrepancy between EBNA-2 expression and EBNA promoter usage in 6 clinical samples. To understand the mechanism underlying this discrepancy, 5 cell lines established from PAL clinical samples were analyzed further.
Morphologic features and growth characteristics
Growth characteristics differed between the lines derived from aggressive tumors (Agg-1 and Agg-2) and indolent tumors (Ind-1, Ind-2 and Ind-3) (Fig. 4a). Both Agg-1 and -2 grew as single cells while Ind-1, -2 and -3 grew as aggregates. Ind-1 and -2 tended to attach to the bottom of the culture plate. The doubling time of the cells was shorter in Agg-2 (24 hr) and Agg-1 (36 hr) than that in Ind-1 (72 hr), Ind-2 (108 hr) and Ind-3 (48 hr).
EBNA-2 expression and EBNA promoter usage
In situ hybridization revealed the EBV genome in the nucleus of all 5 cell lines (data not shown). EBNA-2 was expressed in Ind-1, -2 and -3, but not in Agg-1 and -2, consistent with the findings obtained with the biopsy samples from which the cell lines were derived. EBV latent gene transcripts found in each cell line were similar to those found in the clinical samples except for Ind-3, which was derived from Case 3. Besides EBNA1-QK, Wp and EBNA1-Y3K were used in Ind-3 (Fig. 4b, Table III).
|Case no.||Age (yr)||Sex||Duration of pyothorax (yr)||Age at diagnosis of tuberculosis (yr)||Histology||Prognosis (months after diagnosis)||IHC||RT-PCR||Predicted latency|
|EBNA-2 biopsy sample||Cell lines||LMP-1 biopsy sample||Cell lines||EBNA-2 transcripts||Cp-transcripts||Wp-transcripts||EBNA1-Y3K||EBNA1-QK||Fp-K|
EBV genomic structure
There was a discrepancy between EBNA-2 expression and EBNA promoter usage in Agg-1 and Ind-3. In this context, the structure of the EBV genome was investigated in 5 lines. First, Southern blotting was performed with probes at either end of the EBV genome; i.e., probe-L and -R. A single fragment of the same size was detected in all lines but Agg-2 (Fig. 5a). In Agg-2, probe-L was hybridized with a single fragment of 9.5 kbp, while probe-R was not hybridized at all, indicating that the EBV genome may not join at the terminal repeat as a circular form in Agg-2.
The filters were hybridized with probes-C, -W, -Y, -H, -F, -Q and -U, respectively. All probes hybridized with the bands of expected size in Ind-1, -2 and -3, (data not shown), but not in Agg-1 and -2 (Fig. 5b). In Agg-1, probes-Y and -H did not hybridize with any fragments. Probe-W hybridized with aberrant bands of about 15 kbp besides BamHI-W and Y fragments, indicating a structural abnormality, such as deletion or rearrangement of the EBV genome in these regions. The EBNA-2 gene is located in the BamHI-Y and -H fragments, therefore it might not be expressed in Agg-1. In Agg-2, probe-U hybridized with an aberrant fragment of 5.4 kbp in addition to the expected fragment of 3.3 kbp.
Analysis of the aberrant fragment in Agg-2
To analyze the aberrant fragment detected with probe-U, inverse-PCR was designed to extend outward from the BamHI-U fragment of EBV. Through inverse PCR with the primers of BamHI-UF and BamHI-UR, a single band of 4.6 kbp was detected (data not shown). Sequencing revealed that the fragment contained an unknown sequence of ˜4.5 kbp that was sandwiched between the 2 BamHI-U fragments, and arranged in the opposite direction to each other. There was only 1 BamHI cleavage site between the unknown sequence and BamHI-U fragment, indicating that this cleavage site was ligated artificially. Standard nucleotide–nucleotide BLAST analysis revealed that the unknown sequence corresponded to nucleotides 38785 to 43245 of clone RP11-1H20 on chromosome 4 (GenBank accession no. AC139724). This suggests that nucleotide 59690 of clone RP1-104D1 was the integration site of the EBV genome.
To corroborate evidence for viral integration, metaphase chromosomes of Agg-2 cells were examined by FISH. As shown in Figure 6a, symmetrical signals on both sister chromatids of a specific chromosome were observed in essentially all metaphase spreads, indicating integration of EBV DNA in Agg-2 cells. In parallel control experiments, Namalwa cells, known to carry 2 copies of EBV per cell, showed a single site of viral integration on chromosome 1, as reported previously (Fig. 6b).25
The EBV genomic structure in Agg-1 and Agg-2 cells is summarized in Figure 7.
A previous study on a relatively small number of PAL cases found a Lat III pattern of EBV infection at the protein level.11 Theoretically, EBNA-2 is expressed in the Lat III pattern of EBV infection, not Lat I/II.4, 5, 6 However, conflicting results were found between EBNA-2 expression and EBNA promoter usage in 6 of 13 clinical samples in the present series. Qp, Cp and Wp were used in 2 EBNA-2-positive cases (Case 6 and 7). Only Qp was used in Case 8. Multiple promoters were used (Qp and Wp) in Ind-3 derived from Case 8. In another PAL-derived cell line, Deglis, multiple promoters were also used (unpublished data). Multiple usages of EBNA promoters may be relatively frequent in PAL cases.
PAL cell lines derived from cases showing an aggressive clinical course, Agg-1 and -2, possessed EBV structural abnormality, deletion and/or rearrangement in Agg-1 and integration into the host genome in Agg-2. In Agg-1, a structural abnormality around the EBNA-2 gene was suggested by Southern blotting. Because Y3UK transcripts were detected, deletion and/or rearrangements might occur in a part of BamHI-Y and BamHI-H, which could explain the loss of EBNA-2 expression. Kelly et al. reported similar findings in clinical samples and cell lines derived from BL, i.e., deletion and loss of expression of the EBNA-2 gene.26 A similar kind of deletion was also reported in P3HR127 and Daudi.28 Furthermore, the integration of EBV into the host genome of Raji cells was identified in this region.29 Taken together, the BamHI-Y and -H region of the EBV genome might be a hotspot for mutations such as deletions, rearrangements and integration. In Agg-2, integration of EBV into the host genome was demonstrated. Recently, Daibata et al. also reported an aggressive case of PAL, in which the EBV genome was integrated into Chromosome 1.30 Such EBV structural abnormality might be caused by chromosomal instability, as indicated by the complicated karyotypes found in PAL cells.12, 13 An abnormality in the EBV genome, such as deletion, rearrangement and integration into the host genome, might also have caused loss of EBNA-2 expression in the remaining cases (Case 9,10).
The present study showed that the poor prognosis of the PAL patients correlated with loss of EBNA-2 expression irrespective of the Lat pattern. It is well known that EBNA-2 plays an important role in B-cell transformation; EBNA-2 expression induces a Wp-to-Cp switch,15, 16 LMP-1 and LMP-2 expression17, 18, 19 and the activation of key cellular growth-related genes.20 Loss of EBNA-2 gene expression results in the downregulation of other latent genes15, 16, 17, 18, 19, 20 and no Wp-to-Cp switch. Recent study showed that a decrease in endogenous EBNA-2 expression allows lymphoma cells to escape immune pressure from host immune surveillance through a loss of antigen processing capacity.31 This might be important for the progression of lymphoma, whereas EBNA-2 expression could have a positive effect on lymphomagenesis.
In Burkitt's lymphoma with the Lat I pattern of EBV infection, only EBNA-1 is expressed. In this lymphoma, however, cMYC is constitutively active through rearrangement, and thus involved in the malignant transformation of lymphoma cells. EBNA-2 and cMYC are reported to have opposite effects: cMYC contributes to the phenotypes of the BL cells through upregulation of CD 10 and CD38 and downregulation of the adhesion molecules.20, 31 When cMYC was expressed, EBNA-2-negative cell lines grew in single cell suspensions.20, 31 In the present study, EBNA-2-negative lines, Agg-1 and -2, also grew in single cell suspensions. A comparable level of cMYC was detected in Ind-3 and Agg-2 in a previous study,13 but it was not determined whether cMYC is abnormally activated or not. The involvement of the cMYC gene in PAL cases should be clarified.
In the present study, expression of the EBNA-2 transcript in PAL cases correlated with prognosis, suggesting that PAL could be subdivided into 2 groups. However, most PAL cases are histologically categorized as immunoblastic. Further histopathological analysis might provide a clue to distinguishing between the 2 groups.
We thank Dr. K Takada (Hokkaido University, Japan) for providing the EBV BamHI fragments in the vector pUC119. We thank Dr. PW Tucker (University of Texas, Austin) for providing NAB-2 cell lines.
- 3Lymphomas. In: TravisWD, BrambillaE, Mueller-HermelinkHK, HarrisCC, eds. World health organization classification of tumors: pathology and genetics of tumors of the lung, pleura, thymus and heart. Lyon: IARC Press, 2004. p 137–40., , , .
- 12Establishment and characterization of unique cell lines derived from pyothorax-associated lymphoma which develops in long-standing pyothorax and is strongly associated with Epstein-Barr virus infection. Cancer Sci 2003; 94: 858–63., , , , , .