Endogenous CD8⁺ T Cell Expansion During Regression of Monoclonal EBV-Associated Posttransplant Lymphoproliferative Disorder

Case report

A 31-yr-old Japanese man was diagnosed with Philadelphia chromosome positive, t(9;22), chronic myelogenous leukemia. He underwent allogeneic BMT with a conditioning regimen that consisted of etoposide (1.8 g/m² over 2 days), cyclophosphamide (60 mg/kg/day for 3 days), and total body irradiation (10 G in five fractions). He received an infusion of bone marrow (2.62 × 10⁸ cells/kg) from an HLA-matched unrelated donor (donor and recipient HLA typing: A2, B46, DR8, DQ1, DQB1-06, DQB1-08) that was not T cell depleted. This treatment was performed after obtaining informed consent, approved by the Institutional Review Board at Roswell Park Cancer Institute.

Prophylaxis against GVHD was initiated 48 h prior to bone marrow infusion (i.e., day −2) with cyclosporine A (CsA, 5 mg/kg/day). One dose of methotrexate (10 mg/m²) was administered on day +1. Solumedrol was initiated on day +3 at a dose of 1 mg/kg/day and was increased to 9 mg/kg/day on day +21 following the development of skin GVHD. Colonoscopy was performed on day +29 to evaluate high volume diarrhea and showed GVHD. OKT3 mAb was therefore administered at 1 mg/kg/day. Severe GVHD persisted, so two consecutive 8-day courses of 2-chlorodeoxyadenosine (2-CDA; 0.02 mg/kg/day) were administered. By day +56 the severity of the GVHD had decreased.

On day +62 flow cytometric analysis of blood showed a phenotype consistent with B cell excess (see Fig. 1⇓). The patient received IFN-α (2 × 10⁶ U/m²/day for 35 days), i.v. Ig (IVIg, 25 g/day for 3 days), and acyclovir (ACV, 645 mg/day for 17 days). In addition, the immunosuppressive regimen was reduced (CsA, 0.6 mg/kg/day; solumedrol, 1.5 mg/kg/day). Over the ensuing weeks, sequential phenotypic and molecular analyses of blood lymphocytes were performed as described below. Severe GVHD affecting skin, gastrointestinal tract, and liver subsequently reoccurred, and the patient expired from fulminant liver failure on day +113. Permission for full autopsy was obtained and confirmed histological evidence of severe GVHD in these organs. There was no evidence of malignancy in any tissues examined histologically.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Temporal course of B cell lymphoproliferation and subsequent T cell expansion. A, The percent change in B-cells (▵, CD3⁻CD56⁻; ◊, CD19⁺CD20⁺CD3⁻) and T cells (▪, CD3⁺CD8⁺) in the blood of the patient throughout the clinical course. Relevant therapeutic events occurring during the clinical course are indicated below the horizontal axis. B, Temporal course of absolute number of lymphocyte subsets (◊, CD19⁺CD20⁺CD3⁻; ▪, CD3⁺CD8⁺; or ○, total CD3⁺) in peripheral blood (cells/mm³). C, Histograms of lymphocyte subsets on day +62 show that 86% of gated lymphocytes are CD19⁺, whereas 7% express CD3 (left panel). B lymphocytes are predominantly CD20⁺CD23⁺, consistent with a B lymphoblast phenotype (middle panel). Ninety-six percent of CD19⁺ B lymphocytes express surface Ig with λLC (right).

Flow cytometric and Southern blot analyses

Two-color phenotypic analysis of blood using directly conjugated mAbs was performed as described (12) using a lymphocyte gate. Routine methods for DNA isolation, digestion, electrophoresis, Southern blotting, and ³²P radiolabeling of probes were followed (9). DNA probes for the J_H region of the IgH gene (10), for EBV DNA (13), and for the constant region of the TCR β-chain gene (14) have been described.

Identification of B cell and T cell gene rearrangement by PCR

To identify the specific IgH V-D-J gene rearrangement associated with the B cell lymphocytosis, DNA was isolated from PBMC on day +75. The sequences of the seven IgH-V family-specific 5′ primers and the four IgH-J generic 3′ primers, along with PCR amplification conditions, have been described (15). Following amplification, one distinct band appeared which was excised and sequenced. DNA from day +65 and day +68 PBMC were also amplified and sequenced; they were found to be identical. Subsequently, IgH-V5 (5′-TGCGCCAGATGCCCGGGAAAG-3′) and IgH-J4 (5′-GAGGAGACGGTGACCAGGGTTCCCTGG-3′) primers were used to detect this IgH gene rearrangement in DNA extracted from serial blood and tissue samples.

Analysis of the TCR Vβ gene repertoire was undertaken at indicated time points and from indicated tissues. RNA was isolated using RNAzol (Tel-Test, Freindswood, TX) and reversed transcribed into first-strand cDNA following the manufacturer’s recommendations (Clontech, Palo Alto, CA). Five microliters of the cDNA was equally distributed to 26 tubes, and PCR amplification was performed with 25 different TCR Vβ 5′ primers, each paired with a single Cβ-specific 3′ primer as directed (Clontech). An identical aliquot of the cDNA was used for amplification of the TCRα gene constant region using C 5′α and C 3′α primers as an internal control to verify the integrity of the cDNA (data not shown). Water was used as negative control.

Identification of EBV latent membrane protein-1 (LMP-1)

RNA isolation and reverse transcription (RT) were performed as above. Thirty-five cycles of PCR amplification (94°C for 1 min, 58°C for 1 min, and 72°C for 1 min) was next performed with primers specific for LMP-1 mRNA (5′-CCTCCGCACCCTCAACAAG-3′; 5′-GAGATGATGACGACCCCCA-3′). β-Actin was amplified from cDNA samples with commercially available primers (Perkin-Elmer, Norwalk, CT).

Microsatellite PCR analysis

The details of performing PCR amplification of the microsatellite locus 220yh4 for purposes of distinguishing donor versus recipient cells have been described (16).

DNA sequencing

PCR products were first purified using the Qiaquick PCR purification kit (Qiagen, Chatsworth, CA) and sequenced on an Applied Biosystems model 373 Stretch DNA sequencing system (Perkin-Elmer). Sequences were analyzed using GenBank databases (IntelliGenetics, Campbell, CA).

Cytotoxicity assays

Target PBMC (day +62) and effector PBMC (day +113) were collected from the patient and immediately frozen viably without further sensitization in vitro. On the evening prior to the cytotoxicity assay, cells were thawed, washed, resuspended in RPMI 1640 medium supplemented with 10% FCS, and cultured separately overnight for equilibration. A total of 50 pM of IL-2 (10 U/ml) was added to day +113 PBMC. Viability was >80%. Day +113 effector cells were then washed, plated with W6/32 mAb (anti-HLA-A, -B, -C shared determinant; Accurate Chemicals, Westbury, NY) and anti-CD8 mAb (Becton Dickinson, Mountain View, CA), or nonreactive isotype control mAb (mouse IgG2a, negative control; Dako, Carpinteria, CA) all at 1:200 dilution. ⁵¹Cr-labeled day +62 PBMC target cells were then added (E:T = 20:1), incubated for 4 h, and assessed for specific cell lysis as described previously (17).

Characterization of donor-derived PTLD and expansion of CTL in vivo

The temporal course of the B cell lymphoproliferation and the endogenous, autologous cellular immune response that followed were assessed with serial flow cytometric analyses over a 3-mo period (Fig. 1⇑, A and B). Sixty-two days following BMT, the first changes indicating a clonal B cell expansion were noted, with 86.7% of lymphocytes expressing CD19 (Fig. 1⇑C). Ninety-six percent of CD19⁺ B lymphocytes expressed surface Ig with λ light chain (λLC), indicating a clonal population of B lymphocytes. This same clonal population expressed CD20 and CD23, consistent with a B cell lymphoblastic phenotype commonly associated with EBV transformation (Fig. 1⇑C) (18). The monoclonal population of B lymphocytes was confirmed by Southern blot analysis and also was shown to contain the EBV genome (Fig. 2⇓, A and B). PCR amplification of microsatellite DNA isolated from FACS-sorted subsets of lymphocytes from day +62 showed the monoclonal B cell population was of donor origin (Fig. 2⇓C). RT-PCR analysis of RNA extracted from blood drawn the same day revealed an EBV-LMP-1 gene product that was present but somewhat smaller than the B95.8 EBV strain LMP-1. Sequence analysis demonstrated numerous point mutations and deletions between the coordinates 168,611 and 168,225, virtually identical to the LMP-1 transcript variant described in the C15 nasopharyngeal carcinoma isolate (data not shown) (19).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Molecular analysis of the B lymphocyte expansion. A, Southern blot analysis for determination of B cell clonality in patient blood on day +81. Negative control (normal donor blood) shows germline configuration of the IgH chain gene for three different enzyme digests. For the patient there is a single additional (rearranged) band in lanes 4–6 indicated by arrows on the right (number in parentheses indicates the location of the rearranged band for the relevant lane). Restriction enzyme digests include HindIII (lanes 1 and 4), EcoRI (lanes 2 and 5), and BamHI (lanes 3 and 6). B, Southern blot analysis for the EBV genome in patient blood drawn on day +81 using the BamW probe (lane 1). Controls include normal blood lymphocytes (negative, lane 2) and an EBV⁺ lymphoblastoid cell line (LCL, lane 3). C, DNA PCR amplification of microsatellite 220yh4 using sorted population of B lymphocytes obtained from the patient before BMT (lane 1), the donor before BMT (lane 2), and the patient on day +62 following BMT (lane 3). A distinction in the size of 220yh4 can be seen when comparing the patient and donor before BMT. The day +62 sample is of donor origin. D, DNA PCR amplification of microsatellite 220yh4 using DNA isolated from sorted T lymphocytes obtained from the patient before BMT (lane 1), the donor before BMT (lane 2), and the patient on day +113 following BMT (lane 3). The day +113 sample is also of donor origin. In both C and D, the sensitivity of detection is ∼1% (16).

With documentation of monoclonal, donor-derived EBV⁺ PTLD, IFN-α, ACV, and IVIg were administered without effect. 2-CDA, administered earlier as an immunosuppressive agent, has no known activity against EBV-associated lymphoma (20). On day +81, 18 days after the reduction of immunosuppressive therapy, an expansion of CD3⁺CD8⁺ T cells was noted, with a concomitant decrease in CD19⁺ B cells (Figs. 1⇑A and 3A). The CD4:CD8 ratio was 1:13, and nearly all T cells showed an activated phenotype with HLA-DR expression (Fig. 3⇓A). Without further pharmacologic or immunologic manipulations, by day +97 the CD3⁺CD8⁺ T cell surged to 84.4% of peripheral blood lymphocytes and the clonal CD19⁺ population was reduced to 14% (Fig. 1⇑A). Absolute numbers of CD19⁺ and CD3⁺ lymphocytes began to decline in parallel at this time (Fig. 1⇑B). Blood from day +97 was analyzed for TCR gene rearrangement by Southern blot analysis and showed no evidence of clonal rearrangement (data not shown). Microsatellite analysis of DNA isolated from sorted T lymphocytes showed complete donor origin of T lymphocytes (Fig. 2⇑D). By day +103 the clonal CD19⁺ B cell population had decreased to 4%, whereas the CD3⁺CD8⁺ T cells persisted at 78%, still with activation Ag (HLA-DR) expression (Figs. 1⇑A and 3B). Ten days later (day +113), the patient succumbed to fulminant liver failure secondary to GVHD. Phenotypic analysis of blood at that time revealed 66% of the lymphocytes to be CD3⁺CD8⁺ donor T cells, with only 3% CD19⁺ B cells (Figs. 1⇑, A and B, and 3C), with the majority of CD8⁺ T cells still expressing HLA-DR.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

A, Histograms of lymphocyte subsets on day +81. CD19⁺ B lymphocytes comprise 41% of gated lymphocytes, 88% of which demonstrate a λLC restriction (left panel). T cells are predominantly CD3⁺CD8⁺ (52%, middle panel) and express HLA-DR. B, Histograms of lymphocyte subsets on day +103. The persistence of CD3⁺CD8⁺ CTLs is seen despite a marked reduction in the CD7⁻CD19⁺ B cell population. The majority are HLA-DR⁺. C, Histograms of lymphocyte subsets from day 113 show 1.5% of gated lymphocytes to express CD19 (left panel), with the persistence of a distinct population of CD3⁺CD8⁺ lymphocytes (middle panel); 68% of gated lymphocytes express HLA-DR (right panel).

We subsequently amplified the IgH gene rearrangement specifically associated with the monoclonal PTLD. Sequencing demonstrated utilization of V5 as the variable region and J4 as the joining segment. Donor and recipient bone marrow obtained prior to BMT, along with serial post-BMT blood samples, were assessed for evidence of this clonotype by DNA PCR, as were liver, spleen, and lymph node tissues obtained at autopsy. In addition, RNA was obtained from the same tissues and assayed for the EBV-LMP-1 transcript. The DNA PCR for IgH gene rearrangement assay had a sensitivity of detecting 1 cell in 10⁵, while the sensitivity of the RT-PCR for the detection of the LMP-1 transcript was 1 cell in 10⁶ (data not shown).

PCR evidence for the IgH gene rearrangement associated with the donor-derived malignant B cell clone was found only after BMT and showed a decline in blood over time that was coincident with the CD3⁺CD8⁺ autologous T cell surge (Fig. 4⇓A). Persistent evidence of the B cell clone could be found at autopsy by PCR in the spleen, bone marrow, and lymph node, but not in the liver where GVHD had been most severe (Fig. 4⇓A). Despite molecular evidence documenting the persistence of this clonal B cell population in the blood and certain tissues at autopsy, the LMP-1 transcript, associated with EBV-driven B cell proliferation (21, 22) was no longer detected in these cells near or at the time of death (Fig. 4⇓B). This was despite a RT-PCR sensitivity for LMP-1 that was 10-fold greater than that for the B cell clone.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

A, PCR amplification of the specific VDJ gene rearrangement associated with the PTLD in this patient. The top panel shows the relative amounts of β-actin present in blood and tissues that were amplified. The lower panel indicates the presence or absence of the B cell clonotype in blood near the peak of the monoclonal B cell expansion (day +63), near the peak of the T cell expansion (day +113), and in the indicated tissues at autopsy. There was a relative decrease in the clonotype by day +113, and no evidence of the clonotype in the liver, which showed grade IV GVHD at autopsy. The clonotype was not detectable in donor or recipient PBMC prior to BMT when assayed by DNA PCR (data not shown). B, Expression of the variant LMP-1 transcript in blood following BMT and in indicated tissues at the time of autopsy. The LMP-1 transcript was undetectable in blood by day +97 and was absent on day +113, despite the persistence of the clonotype that identified the monoclonal B-cell population in blood (A). The LMP-1 transcript is not expressed in any of the indicated tissues at autopsy. Colo 205 is an EBV-negative tumor cell line and serves as a negative control for the EBV-specific latent LMP-1.

Analysis of TCR Vβ gene repertoire in response to PTLD

We next characterized the TCR Vβ gene repertoire in blood over time. As all cells following BMT were of donor origin (Fig. 2⇑, C and D), the TCR Vβ profile of the donor T cells before BMT were also analyzed. This profile showed a typical polyclonal pattern (Fig. 5⇓A). At day 64, 2 days following the detection of a clonal population of B lymphoblasts, the TCR Vβ repertoire was unchanged with the exception that TCR Vβ1 became undetectable and Vβ3 and -23 became apparent for the first time (Fig. 5⇓B). Twenty-three days after the detection of circulating clonal EBV⁺ B cells (i.e., day +83), the TCR Vβ repertoire was unchanged with the exception of Vβ1 again being detected (Fig. 5⇓C). By day +97, PTLD was diminishing (Fig. 1⇑B), LMP-1 transcript became undetectable in blood (Fig. 4⇑B), and the percent CD3⁺CD8⁺ CTL were at their peak (Fig. 1⇑A). At this time there was a switch from a polyclonal to a restricted pattern of TCR Vβ usage, with a predominance of Vβ1, -2, -3, -4, -6, -8, -9, and -23 in blood (Fig. 5⇓D).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

RT-PCR analysis of the TCR Vβ gene repertoire found in blood and in liver. A, TCR Vβ repertoire of the donor’s peripheral blood prior to BMT reveals a typical polyclonal pattern. B, TCR Vβ repertoire of blood on day +64 at the time of B cell was first noted. C, TCR Vβ repertoire of blood on day +83 at the time of B cell expansion. The repertoire seen at day +64 and +83 is identical to A, except that Vβ3 and Vβ23 is now expressed. D, TCR Vβ repertoire of blood on day +97 at the time of T cell expansion, with a predominance of Vβ1, -2, -3, -4, -6, -8, -9, and -23. Both T and B lymphocyte populations were shown to be of donor origin (Fig. 2⇑, C and D). E, TCR Vβ repertoire in the liver at the time of autopsy shows a predominance of Vβ3, -4, -6, and -8.

As the liver was the site of fatal hepatic GVHD but showed no molecular evidence of B cell tumor, we next determined if the TCR Vβ repertoire of T lymphocytes present in the liver at the time of death was similar to that observed in blood when the EBV LMP-1 transcript was undetectable. The T cells in the liver expressed TCR Vβ3, -4, -6, and -8, all of which were present in the blood during the regression of the PTLD, as well as TCR Vβ7, -22, and -24 that were not in the earlier blood sample (Fig. 5⇑E).

Cytotoxic activity of ex vivo CTL against EBV⁺ B lymphoblasts

To determine if the CD3⁺CD8⁺HLA-DR⁺ CTL present in blood during the resolution of PTLD were indeed MHC class I-restricted cytolytic effectors, a standard ⁵¹Cr-release cytotoxicity assay was performed using PBMC from day +62 (86% CD19, CD20, CD23, surface Ig λLC, and EBV⁺) as targets (T) and PBMC from day +113 (66% CD3⁺CD8⁺HLA-DR⁺) as effectors (E). All cells were donor derived (Fig. 2⇑, C and D). No additional in vitro sensitization was performed prior to the 4-h assay. Cytotoxicity at an E:T of 20:1 in the presence of isotype control Abs was 76.3 ± 4.2%. In the presence of Abs reactive against MHC class I and CD8, cytotoxicity was 6.3 ± 1.3% (Fig. 6⇓). Thus, donor-derived CTL that expanded in peripheral blood during the simultaneous reduction of donor-derived CD19⁺, monoclonal, EBV⁺ B lymphoblasts displayed MHC class I-restricted cytotoxicity against this population.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Autologous MHC class I-restricted cytotoxic activity. PBMC from day +113 (66% CD3⁺CD8⁺ T cells, effectors) were plated against ⁵¹Cr-labeled PBMC obtained on day +62 (76% EBV⁺ B lymphoblasts, targets). The assay was performed at an E:T ratio of 20:1 in the presence of nonreactive isotype control mAbs or anti-HLA class I and anti-CD8 mAbs. Results represent the mean ± SE of triplicate wells.