correspondence: Expansion of CD8+/perforin+ effector memory T cells in the bone marrow of patients with thymoma-associated pure red cell aplasia


Pure red cell aplasia (PRCA) is characterized by normochronic normocytic anaemia, reticulocytopenia, and severe erythroid hypoplasia of the bone marrow. Regarding its pathogenic mechanism, it has been reported that acquired PRCA is accompanied by clonal T-cell expansion in the bone marrow or peripheral blood (Masuda et al, 1997, 1999, 2005; Fujishima et al, 2006). However, it has not been clarified how the T-cell subset proliferates in the bone marrow or peripheral blood. Using bone marrow mononuclear cells (BM-MNCs) obtained from patients with various forms of acquired PRCA, we examined T-cell clonality by polymerase chain reaction (PCR), and the T-cell subsets by four-color flow cytometry.

We have recently experienced 10 patients with newly diagnosed acquired PRCA (Table I): five associated with thymoma; two with granular lymphocyte-proliferative disorders (GLPD) (Oshimi et al, 1993); one with myelodysplastic syndrome (MDS); one with follicular lymphoma; and one idiopathic. Bone marrow samples were taken from the patients after obtaining informed consent. As controls, we used BM-MNCs from 15 patients with various types of anaemia caused by B-cell lymphoma with bone marrow not infiltrated with lymphoma cells, drug-induced anaemia, renal dysfunction, and collagen diseases.

Table I.   Characteristics of 10 patients with Acquired PRCA.
Case No.Age (years)/SexAssociated DiseaseWBC (×109/l)Hb (g/l)Ret (%)Erythroblast %(BM)CD8+/Perforin+%(BM) (0–20%)CD8+ Perforin+ TEM%(BM)Analysis of T-cell clonalityTreatmentClinical Course
TRG@ CD4+ T-cell in BMTRG@ CD8+ T-cell in BM
  1. BM, bone marrow; TEM, effector memory T cell; CsA, cyclosporin A; ND, not done; PSL, predonisolone; T-GLPD, T-cell granular lymphocyte-proliferative disorders; CPM, cyclophosphamide; FL, follicular lymphoma.

 170/FThymoma4·04550·36034·138·0NegativeOligoclonalThymectomy →CsA 250 mg/dayReticulocytosis 2 weeks after CsA therapy
 249/MThymoma5·57610·09029·737·6NegativeOligoclonalCsA 300 mg/dayReticulocytosis 3 weeks after CsA therapy
 358/FThymoma4·10430·13051·848·1NDNDThymectomy →CsA 250 mg/day →PSL 45 mg/dayReticulocytosis 2 weeks after PSL therapy
 464/MThymoma4·10490·100NDNDNDNDThymectomy →CsA 250 mg/dayReticulocytosis 2 weeks after CsA therapy
 555/FThymoma5·46550·130NDNDNDND1 month observation →anaemia not improved CsA 300 mg/dayReticulocytosis 3 weeks after CsA therapy
 686/FT-GLPD (αβT type)5·65670·090·3NDNDNegativeClonalCsA 200 mg/day →CPM 50 mg/dayNot effective Blood transfusion. Died 1 month after
 776/FT-GLPD (γδT type)5·26650·952·468·629·4NDNDCsA 300 mg +PSL 10 mgNot effective Blood transfusion
 878/MMDS4·51680·911·025·330·2OligoclonalOligoclonalCsA 250 mg/dayNot effective Blood transfusion
 933/FIdiopathic3·59530·381·020·815·5NDNDCsA 250 mg/day →PSL 40 mg/dayCsA not effective Reticulocytosis 2 weeks after PSL therapy
1048/MFL5·05680·404·025·519·0NDNDNo therapyAnaemia improved without therapy

Cells were stained with fluorescein isothiocyanate (FITC) conjugated anti-CD4 antibody and phycoerythrin (PE) conjugated anti-CD8 antibody by a conventional method. Cell sorting was performed using a fluorescence-activated cell sorter [FACS Aria cell sorter (Becton Dickinson (BD), San Jose, CA, USA)], and CD4+ and CD8+ cells were collected. T-cell receptor gamma gene (TRG@) rearrangement (clonality assay) of CD4+ T cells and CD8+ T cells sorted from BM-MNCs was assessed by PCR assays using BIOMED-2 (Van Dongen et al, 2003) (InVivoScribe Technologies, San Diego, CA, USA) conducted by Mitsubishi Chemical Medience (Tokyo, Japan).

Patient BM-MNCs were stained with anti-CD8 antibody conjugated with FITC (BD) and anti-perforin antibody conjugated with PE (BD) after fixation. To clarify the bone marrow T-cell subpopulation, we used the following antibodies: anti-CCR7 (CD197) antibody conjugated with PE (BD), anti-CD62L antibody conjugated with peridinin chlorophyll protein (PerCP) (BD), anti-CD27 antibody-allophycocyanin (APC) (BD). Briefly, for perforin staining, BM-MNCs were stained with anti-CD8 antibody-FITC for 30 min on ice. The cells were washed with phosphate-buffered saline and then fixed with BD Cytofix/Cytoperm fixatives (BD). After fixation, the cells were washed with Perm/Wash buffer in accordance with the manufacturer’s instructions. The fixed cells were incubated with anti-perforin antibody-PE for 30 min on ice. After the cells were washed, they were subjected to flow cytometric analysis and data were acquired using a FACS Calibur flow cytometer (BD) with CellQuest software (BD).

Statistical significance of differences between independent groups was determined using Student’s t-tests. P values <0·05 were considered statistically significant.

To determine which population of T cells proliferates clonally in the bone marrow, we analyzed T-cell clonality for each T-cell subpopulation by a PCR-based method after sorting CD4+ and CD8+ T cells, using a cell sorter, from BM-MNCs in two cases of thymoma-associated PRCA. The oligoclonality was exclusively detected in CD8+ T cells of BM-MNCs, not CD4+ T cells, in these two patients (Fig 1A). Therefore, we focused on CD8+ T cells as an effector for thymoma-associated PRCA. Intriguingly, in all of three thymoma-associated PRCA cases that were assessed, including the above two, the CD8+/perforin+ T-cell population was significantly increased in the bone marrow (38·5 ± 11·7%, n = 3), compared to the controls (16·8 ± 7·2%, n = 15) (= 0·005), as determined by flow cytometry (Fig 1B and Table I). The T-cell subpopulation expressing CD8+ perforin+ CCR7lowCD62Lint and CD27+, which is consistent with the CD8+/perforin+ effector memory T (TEM) cell subset (Decrion et al, 2007), was also increased in the bone marrow from all of the three thymoma-associated PRCA patients that we analyzed (Fig 1B and Table I) (patients 41·2 ± 6·0%, n = 3 vs. control 15·6 ± 7·0%, n = 15) (P < 0·0001). Furthermore, in the three thymoma-associated PRCA patients, the CD8+/perforin+ TEM subpopulation in the bone marrow was also significantly increased, compared with that in the seven aplastic anaemia and seven MDS patients (Table SI).

Figure 1.

 (A) Phenotypic expression of oligoclonal CD8+ T cells in BM-MNCs from thymoma-associated PRCA. Oligoclonal pattern (peak) of TRG@ rearrangement of BM-CD8+ T cells in thymoma-associated PRCA (Case 1 and Case 2). Isolated DNA from sorted CD8+ T cells was amplified by PCR in Tube A (TRGV2, TRGV3, TRGV4, TRGV5, TRGV7, TRGV8 + TRGV10 + TRGJP1/2 + TRGJ1/2 [4 primer pairs]) and Tube B (TRGV9 + TRGV11 + TRGJP1/2 + TRGJ1/2 [4 primer pairs]). The peaks were determined to be oligoclonal peaks. (B) Phenotypic expression of T cells in BM-MNCs from patients with thymoma-associated PRCA. CD8+/perforin+ T cells in BM-MNCs from thymoma-associated PRCA significantly increased in all assessable cases (Case 1–3), compared with that in the controls (Control 1–3). The T-cell subpopulation expressing CCR7low CD62Lint and CD27+, which was consistent with the CD8+/perforin+ TEM subset, increased in the bone marrow (Case 3: 48·1% vs. Control 1: 23·8%).

Interestingly, whereas CD8+/perforin+ T-cell subpopulation in the bone marrow was significantly increased in four cases of assessable acquired PRCA, who did not have thymoma (35·1 ± 22·5%, n = 4), compared with that in the controls (= 0·0115), the number of CD8+/perforin+ TEM cells expressing CCR7lowCD62LintCD27+ did not differ from the controls (23·5 ± 7·4%, n = 4 vs. 15·6 ± 7·0%, n = 15). Furthermore, no increment of CD8+/perforin+ TEM cells was found (18·3%, n = 1) in the T-cell subsets of BM-MNCs from a patient with thymoma, who did not have the complication of PRCA. These results show that the CD8+/perforin+ TEM subset potentially contributed to the pathogenesis of thymoma-associated PRCA. However, we could not confirm whether the CD8+ oligoclonal T cells detected by PCR is consistent with the CD8+/perforin+ TEM subset detected by flow cytometry.

It has been reported that thymoma-associated PRCA accounts for 23% of all acquired PRCA in Japan (Sawada et al, 2007). However, it is not clear whether thymoma directly contirbutes to the pathogenesis of PRCA. T cells reacting with autologous erythroid progenitor cells may be generated by thymoma, and can survive a long time in the peripheral blood or bone marrow (Buckley et al, 2001). When these cells expand to a definite number in the bone marrow, they may cause PRCA (Buckley et al, 2001).

It was reported that CD8+ TEM cells exhibited cytotoxic activity for autologous target cells infected with recombinant vaccinia viruses expressing human immunodeficiency virus (HIV)-peptide or directly loaded with Epstein-Barr virus or cytomegalovirus-specific peptides (Decrion et al, 2007). The direct cytotoxic activity on HIV-presenting target cells from CD8+ TEM cell fraction enriched from an HIV-infected donor increased six-fold in the total CD8+ T-cell fraction tested, whereas the proportion of CD8+ effector cells was stable, and the specific lysis of HIV target cells increased eight-fold (Decrion et al, 2007). If some sort of herpesviruses specifically infects erythroid progenitors, CD8+ TEM cells in the bone marrow may exert direct cytolytic effects on erythroid progenitors alone.

In conclusion, this is the first report describing the expansion of CD8+/perforin+ TEM cells in thymoma-associated PRCA. Further examination is required, however, to identify the mechanism of specific impairment in erythroid progenitor cells in PRCA.


The authors would like to thank Dr Kazuo Oshimi (former professor of the Juntendo University School of Medicine, Tokyo) for critical comments on our manuscript, and Sachiko Fukumoto, and Nanae Nakaju for providing excellent technical support in carrying out these experiments.