A novel latent membrane 2 transcript expressed in Epstein-Barr virus-positive NK- and T-cell lymphoproliferative disease encodes a target for cellular immunotherapy

Abstract

Therapeutic targeting of virus-encoded proteins using cellular immunotherapy has proved successful for Epstein-Barr virus (EBV)-associated posttransplant lymphoproliferative disease. However, the more limited repertoire and immunogenicity of EBV-encoded proteins in other malignancies such as Hodgkin lymphoma and extranodal natural killer (NK)/T lymphoma has been more challenging to target. The immunosubdominant latent membrane protein 2 (LMP2) is considered the optimal target in such Latency II tumors, although data relating to its expression in T/NK malignancies are limited. In addressing the validity of LMP2 as an immunotherapeutic target we found that LMP2-specific effector CD8(+) T cells recognized and killed EBV-positive NK- and T-cell tumor lines, despite an apparent absence of LMP2A protein and barely detectable levels of LMP2 transcripts from the conventional LMP2A and LMP2B promoters. We resolved this paradox by identifying in these lines a novel LMP2 mRNA, initiated from within the EBV terminal repeats and containing downstream, epitope-encoding exons. This same mRNA was also highly expressed in primary (extra-nodal) NK/T lymphoma tissue, with virtually undetectable levels of conventional LMP2A/B transcripts. Expression of this novel transcript in T/NK-cell lymphoproliferative diseases validates LMP2 as an attractive target for cellular immunotherapy and implicates this truncated LMP2 protein in NK- and T-cell lymphomagenesis. This study is registered at clinicaltrials.gov as NCT00062868.

Figure 1

Antigen processing and presentation capability of EBV-positive T and NK tumor lines. (A) Flow cytometric analysis of cell surface HLA class I. Representative data from 2 lines (SNT 16 and SNK 6, black shading) compared with a LCL (dark gray shading) is shown. (B) Western blot of whole cell lysates probed with antibodies to transporter associated with antigen processing (TAP) 1 and 2 and total cellular class I heavy chains (HC 10) in NK- and T-cell lines, shown relative to levels in a LCL. Calregulin serves as the loading control. (C) EBV antigen expression in NK- and T-cell lines. Western blot of whole cell lysates of EBV-positive NK- and T-cell tumor lines probed with antibodies to EBNA 1, EBNA 2, LMP1, and LMP2A compared with a LCL and an EBV-negative NK line (NKL). Note the different molecular weight of EBNA 1 in the different lines is attributable to the variable size of the Gly-Ala repeat domains. (D) Schematic indicating the exon structure of the LMP2 gene and splicing patterns of conventional LMP2A and LMP2B transcripts, across the TR region of the genome. Each transcript contains a unique 5′ first exon (1A for LMP2A, shaded gray and 1B for LMP2B, shaded white). (E) Real-time PCR for LMP2A and LMP2B mRNA. Gray and white shaded bars represent quantities of LMP2A and LM2B transcripts, respectively, relative to a LCL. Error bars indicate SDs from the mean.

Figure 2

Specific killing of EBV⁺ malignant NK and T cells by polyclonal CTLs, containing LMP2-specificities, from ENKTL patients. (A) Dual-color flow cytometry of CTLs derived from 2 ENKTL patients prepared by ex vivo stimulation with LMP2/LMP1 transfected antigen-presenting cells. CTLs were stained with a PerCP (peridinin-chlorophyll protein)–conjugated anti-CD8 antibody and with the phycoerythrin-conjugated HLA-peptide pentamers: HLA-A*0201-CLGGLLTMV, HLA-A*0201-FLYALALLL, HLA-A24-TYGPVFMSL as previously described. Numbers within the top right quadrant of each plot indicate the percentage of viable peptide/pentamer-specific CD8⁺ cells. (B) The CTLs shown in panel A were tested in standard ⁵¹Chromium-release assays for their ability to kill EBV⁺ NK- and T-cell tumor lines. Results are expressed as the mean percentage of specific chromium release from the target cells at effector:target ratios titrated from 5:1 to 40:1, tested in triplicate. Autologous and HLA-mismatched LCLs were positive and negative controls, respectively, for each assay.

Figure 3

Recognition and killing of SNT 16 by LMP2-specific CD8⁺ T-cell clones through HLA A2 and A24. (A) CD8⁺ clones specific for endogenously processed LMP2-derived peptides FLYALALLLL and TYGPVFMCL were cocultured for 18 hours with 10⁵ target cells (gray and black bars indicate 5000 and 10 000 CD8⁺effectors per well, respectively). HLA-matched and -mismatched LCLs were positive and negative controls, respectively. Other controls included SNT 16 and CD8⁺ effectors cultured alone to assess spontaneous IFN-γ release. Error bars indicate 1 SD from the mean. Results are representative of those seen in 3 separate experiments. Supernatant IFN-γ was quantitated by ELISA. Relative to the maximal values seen using the same target cells preloaded with 5μM cognate peptide, recognition of the unmanipulated cells were as follows: FLY (SNT 16 = 14%, LCL = 26%), TYG (SNT 16 = 5%, LCL = 15%). (B) Lysis of SNT 16 and LCL targets by the LMP2-specific FLY and TYG CD8⁺ clones shown in panel A. Five-hour ⁵¹Chromium-release assays were conducted with SNT 16, HLA class I–matched and –mismatched LCL targets. Spontaneous release of ⁵¹Cr was < 30% of maximum release. Results are expressed as the percentage of specific chromium release from the target cells at the indicated effector-to-target ratios. Gray lines indicate results after preincubation for 1 hour with 5μM cognate peptide. Results are representative of those seen in 3 different experiments. Error bars indicate 1 SD from the mean.

Figure 4

LMP2-specific T-cell recognition of SNK 6 exceeds that of a LCL. Dual-color flow cytometry: the x-axis represents fluorescence from CFSE-stained CD8⁺ effectors, and the y-axis denotes IFN-γ release as determined by intracellular staining. CFSE-negative cells (ie, LCL or SNK 6 targets) have been gated out of this analysis. HLA class I–mismatched LCLs and SNK 6 transduced with a control retrovirus (+ control RV) are negative controls. The top plots within panels A and B indicate T cells expressing IFN-γ in response to endogenously processed peptide, while the bottom plots of each panel show IFN-γ expression by T cells after preincubation of targets with exogenous cognate peptide (+ peptide). Numbers in the right quadrants indicate the percentage of cells positive (top quadrant) or negative (bottom quadrant) for IFN-γ release. (A) SSC-specific CD8⁺ clones tested against HLA-matched LCLs and SNK 6-A11. (B) Effector T cells expressing a recombinant SSC/A11 TCR tested against A11+ LCL and SNK 6-A11.

Figure 5

Expression of a LMP2 transcript containing exons 2 to 6 in EBV positive NK- and T-cell tumor lines. (A) End point PCR products (35 cycles) visualized on a 1% agarose gel. Amplification of cDNA using a forward primer located in exon 1 and reverse primer within exon 3 (F1 and R3; see supplemental Table 1) yields a product of expected size in the LCL but no product in the NK- and T-cell lines. PCR performed with a forward primer in exon 2 and a reverse primer located within exon 6 (F2 and R6) produces a product of the anticipated size in all 4 NK- and T-cell lines. The EBV negative NK leukemia cell line, NKL, and a no reverse transcriptase (NO RT) sample from a LCL served as negative controls. (B) Quantitative RT-PCR assays, using primers and probes designed to amplify LMP2 transcripts containing exons 2 and 6, respectively (avoiding exon-exon junctions in the event of uncharacterized splice variations). Data are expressed using cycle threshold (Ct) values, where Ct₀ represents the assay threshold of detection. LMP2 transcript Ct values are normalized to cellular glyceraldehyde 3-phosphate dehydrogenase (GAPDH) Ct values. The left-hand y-axis represents Δ(Ct₀-Ct), while the right-hand axis shows the relative transcript levels.

Figure 6

Identification of 5′ ends by RACE and quantitation of novel LMP2-TR transcripts in NK- and T-cell tumor lines. (A) The schematic represents the exon structure of the LMP2 gene spanning the EBV TR. Numbers indicate the exact 5′ ends of 4 of 7 cDNAs identified by RACE (EBV genome coordinates: NC007605). Three other 5′ cDNA ends identified in SNK 6 (coordinates 19, 80, and 83) correspond to the intervening intron and beginning of exon 2; having not reached the TR region. (B) Quantitative RT-PCR results: Ct₀ represents the assay threshold of detection. LMP2 transcript Ct values are normalized to cellular GAPDH Ct values. The left y-axis represents Δ(Ct₀-Ct), while the right-hand y-axis shows the equivalent fold difference in transcript levels.

Figure 7

LMP2-TR, but not LMP2A or LMP2B, are expressed in primary ENKTL tissue. (A) Immunohistochemistry of formalin-fixed, paraffin-embedded tissue sections with antibody clone 15F9, specific for LMP2A. Images were captured with a Nikon Coolpix E995 digital camera, via a Nikon Eclipse E400 microscope (optical magnification × 400). A case of EBV⁺ B cell PTLD is a positive control. Sections from 3 representative cases ENKTL (i-iii of 7) shown. (B) Quantitative RT-PCR data using cDNA derived from 7 primary ENKTL tissue biopsies and 1 EBV⁺ aggressive NK leukemia sample. QRT-PCR assays were performed to detect conventional LMP2A and LMP2B, LMP2-TR and transcripts containing exon 5-6 splice; the latter representing total LMP2 mRNAs. Ct₀ represents the assay threshold of detection. LMP2 transcript Ct values are normalized to cellular GAPDH Ct values. The left-hand y-axis represents Δ(Ct₀-Ct), while the right-hand axis shows the equivalent fold difference in transcript levels. Note: input cDNA was 2 μL in a reaction volume of 20 μL, compared with 5 μL in 25 μL for the cell line PCR studies for Figures 5 and 6.