CD8+ T cells suppress autologous megakaryocyte apoptosis in idiopathic thrombocytopenic purpura

Authors


Ming Hou, MD, PhD, Haematology Oncology Centre, Qilu Hospital of Shandong University, 107 West Wenhua Rd, Jinan, Shandong 250012, China.
E-mail: houming@medmail.com.cn

Summary

To investigate the effect and mechanism of the CD8+ T cells in bone marrow on autologous megakaryocytopoiesis in idiopathic thrombocytopenic purpura (ITP) patients, we prepared bone marrow mononuclear cells (MNCs) from 15 chronic ITP patients and 13 controls. MNCs were cultured in vitro directly (MNC group) or after depleting CD8+ T cells (CD8+ T-dep group) or adding purified autologous CD8+ T cells to CD8+ T-dep MNCs (Coculture group) or adding dexamethasone to the coculture (DEX group) all in semi-solid and liquid culture systems. The quantity and quality of megakaryocytes were measured. The megakaryocyte count was increased in the presence of autologous CD8+ T cells of patients with chronic ITP, while platelet production was reduced. In addition, lower percentages of polyploidy and apoptotic megakaryocytes, and higher levels of soluble Fas (sFas) in supernatant were observed. Dexamethasone successfully corrected this effect of CD8+ T cells on autologous megakaryocytopoiesis. These studies provide evidence that activated CD8+ T cells in bone marrow of patients with chronic ITP might suppress megakaryocyte apoptosis, leading to impaired platelet production. Megakaryocyte apoptosis would be a novel target for the management of ITP.

Idiopathic thrombocytopenic purpura (ITP) is an organ-specific autoimmune disorder. Since the 1950s, it has been believed that the thrombocytopenia of this disorder was caused solely by autoantibodies that were directed against various platelet membrane receptors, including platelet glycoproteins, such as glycoprotein (GP) IIb/IIIa (GPIIb/IIIa) or GPIb/IX complexes (Kiefel et al, 1991; Fujisawa et al, 1992). Binding of autoantibodies to these target antigens eventually results in platelet destruction (McMillan, 2000; Cines & Blanchette, 2002). However, these mechanisms cannot account for all observations made in ITP. It has become evident that T-lymphocytes abnormalities might have pathogenetic importance in ITP (Hedlund-Treutiger et al, 2004; Panitsas et al, 2004). Most recently, in vitro studies suggested that cytotoxic T-lymphocytes (CTL) (CD8+) might be involved in the pathogenesis of chronic ITP through the cell-mediated destruction of autologous platelets (Olsson et al, 2003; Zhang et al, 2006).

As megakaryocytes express GPIIb/IIIa and GPIb/IX on their surface during maturation, it is possible that megakaryocytopoiesis and thrombopoiesis are also impaired in ITP. This hypothesis has been supported by platelet survival studies and morphological alterations of ITP marrow megakaryocytes, which suggested that platelet production may be disrupted in ITP (Ballem et al, 1987; Isaka et al, 1990; Houwerzijl et al, 2004). Recently, studies have demonstrated that ITP plasma and purified antiplatelet autoantibodies inhibit megakaryocytopoiesis in vitro (Chang et al, 2003; McMillan et al, 2004). However, whether CD8+ T cells in bone marrow affect megakaryocytopoiesis in ITP remains to be elucidated.

In the present study, bone marrow mononuclear cells (MNCs) were collected and classified into different subgroups according to the percentage of CD8+ T cells and the presence of dexamethasone, then plated in megakaryocytic culture systems. The quantity and quality of megakaryocytes was determined in order to determine the effect and mechanism of ITP CD8+ T cells on autologous megakaryocytopoiesis.

Materials and methods

Patients and controls

Fifteen newly diagnosed patients (10 females and five males) with chronic ITP were studied, their platelet counts were lower than 100 × 109/l (range, 2–58 × 109/l; median, 33 × 109/l). The patients were aged between 15 and 78 years (median, 34 years). They all met the diagnostic criteria for chronic ITP as evidenced by platelet counts less than 100 × 109/l for more than 6 months, normal or increased number of megakaryocytes in otherwise normal bone marrow, no other cause for peripheral thrombocytopenia, and absence of splenomegaly. Thirteen control subjects (five females and eight males) between 23 and 60 years (median, 30 years) whose platelet counts ranged between 136 × 109/l and 312 × 109/l (median, 207 × 109/l) were non-haematological patients undergoing cardiac surgery with no history of blood transfusion or pregnancy. The study was approved by the Medical Ethical Committee of QiLu Hospital, Shandong University. All patients and controls received written information about the study according to the Declaration of Helsinki and gave informed consent.

Cell isolation and group division

Both MNCs and peripheral blood mononuclear cells (PBMNCs) were isolated from heparinized bone marrow and blood from 15 patients with ITP and 13 control subjects by Ficoll-Hypaque (Pharmacia Biotech, Uppsala, Sweden) gradient centrifugation respectively. CD8+ T cells were removed by CD8+ magnetic microbeads (Miltenyi Biotec, Bergisch, Germany), according to the manufacturer's recommendations. The CD8+ T-depleted (CD8+ T-dep) MNCs contained <1% CD8+ T cells and the purified CD8+ T-cell fraction contained >90% CD8+ T cells by flow cytometric analysis. The prepared bone marrow cells were divided into four groups: MNCs were cultured directly (group MNC); CD8+ T-depleted MNCs were cultured (group CD8+ T-dep); purified CD8+ T cells were 1:1 added to autologous CD8+ T-dep MNCs in coculture (group Coculture); and dexamethasone was added to coculture (group DEX) and adjusted final concentration to 1·0 × 10mol/l. Fresh PBMNCs were irradiated (4000 rad) and used as autologous antigen-presenting cells (APCs) in proliferative responses by CD8+ T cells to autologous platelets as described by Ware and Howard (1993).

Characteristics of CD8+ T cells

Using a FACScalibur flow cytometer (Beton Dickinson, Mountain View, CA, USA), phenotypic analysis of T cells was determined with fluorescein isothiocyanate (FITC)-conjugated CD3 monoclonal antibody (mAb) (BD Biosciences, San Jose, CA, USA), allophycocyanin-conjugated CD4 mAb (BD Biosciences) and phycoerythrin (PE)-conjugated CD8 mAb (BD Biosciences). Appropriate isotype matched control antibodies were used.

Proliferative responses by CD8+ T cells to autologous platelets were performed by the modified methods of Ware and Howard (1993) and Zhang et al (1997). Briefly, CD8+ T cells (1 × 10cells/ml) were cultured with autologous platelets (100 × 106 cells/ml) and irradiated PBMNCs (0·2 × 106 cells/ml) in RPMI 1640 medium (Gibco, Hefei, China) containing 10% pooled inert human group AB serum in 96-well plates (200 μl volumes) at 37°C in a humidified atmosphere of 5% CO2. After 6 d of incubation, 0·037 MBq [3H]thymidine (Beijing Atomic Energy Research Institute, Beijing, China)/well was added to the 96-well plates, the plates were incubated for an additional 24 h at 37°C, and the cells were harvested and counted for incorporated radioactivity.

Megakaryocyte culture

Megakaryocyte colony-forming units (CFU-MK) were quantitated using a commercially available kit (MegaCult®-C; Stem Cell Technologies, Vancouver, BC, Canada), according to the manufacturers instructions. The four groups of cells, prepared as described above, were plated at a density of 2·2 × 106 cells/ml in serum-free medium containing thrombopoietin (TPO, 50 ng/ml), interleukin-3 (IL-3, 10 ng/ml), IL-6(10 ng/ml) and collagen(1·1 mg/ml) for 12 d at 37°C in a humidified atmosphere of 5% CO2, followed by dehydration, fixation and immunocytochemical staining with mouse anti-human GPIIb/IIIa antibody and biotin-conjugated goat antimouse IgG to confirm the CFU-MK. Megakaryocyte colonies were defined as groups of three or more GPIIb/IIIa-positive cells.

In liquid suspension culture, the four groups of cells, prepared as described above, were cultured at a concentration of 1 × 106 cells/ml in 24-well plates in 1 ml serum-free medium (SFEM; Stem Cell Technologies) supplemented with 100 ng TPO (Sansheng Pharmacy, Shenyang, China), 100 ng stem cell factor (SCF; Diao Pharmacy, Chengdu, China), and 10 ng IL-3 (Biosouth Research Laboratories, New Orleans, LA, USA) at 37°C in a humidified atmosphere of 5% CO2. After 14 d of culturing in liquid suspension media, the quantity and quality of megakaryocytes were assessed by flow cytometry.

Megakaryocyte analysis

Megakaryocytes were recognized as CD41a+ events with phycoerythrin-Cyano dyes 5(PEcy5)-conjugated CD41a mAb (BD Biosciences) . The PEcy5-conjugated IgG1 (BD Biosciences) was used as isotype control. The number of megakaryocytes was determined [cell number (determined by direct counting of each culture) × %megakaryocytes (determined from FACS data)].

Detection of platelets produced in culture was performed as previously described (Mattia et al, 2002). Briefly, cultured cells were centrifuged at 350 g for 15 min and incubated with PEcy5-conjugated CD41a mAb. After incubation, every sample was diluted to the same volume (300 μl) and collected at median rate for 50 s by flow cytometry. An analytical gate was determined according to scatter properties of normal blood platelets treated similarly using a log scale for forward scatter (FSC) and sideways scatter (SSC). This gate excluded large contaminating cells (megakaryocytes) and small debris or microparticles.

To measure ploidy distribution, the megakaryocytes were identified after labelling with PEcy5-conjugated CD41a mAb and incubated with 500 μl propidium iodide (PI; BD ParMingen, San Diego, CA, USA) containing RNase. CD41a+ cells were gated and ploidy distribution was assessed by the intensity of the PI fluorescence (McMillan et al, 2004).

Megakaryocyte apoptosis was measured using the Annexin V-FITC Apoptosis Detection Kit (Jingmei biotech co., Ltd., Beijing, China) according to the manufacturers instructions. Cells were labeled with PEcy5-conjugated CD41a mAb, incubated with FITC-conjugated Annexin VmAb and PI, and then analysed by flow cytometry. CD41a+ cells were gated and apoptotic megakaryocytes were Annexin V+ PI cells within that population.

Modulators of apoptosis measurement

Intracellular staining of megakaryocytes was performed by first incubating the cells with PEcy5-conjugated CD41a mAb as described above. After staining, cells were fixed in 1% paraformaldehyde, permeabilized with 0·1% saponin (Sigma, St Louis, MO, USA), and incubated with either FITC-conjugated Bcl-xl (Southern Biotech, Birmingham, AL, USA) or FITC-IgG3 (Southern Biotech) as an isotype control. CD41a+ cells were gated and Bcl-xl expression was shown as mean fluorescence intensity (MFI) within that population.

Cells were incubated sequentially with PEcy5-conjugated CD41a mAb and FITC-conjugated Fas mAb. Appropriate isotype-matched control antibodies were used. CD41a+ cells were gated and Fas expression was reported as MFI within that population.

Cell-free supernatants of megakaryocytic cultures were harvested after 14 d of incubation and stored at −80°C. The levels of soluble Fas (sFas) and soluble tumour necrosis factor (TNF)-related apoptosis-inducing ligand (sTRAIL) were quantified by the respective enzyme-linked immunosorbent assay (ELISA) kits (MedSystems Diagnostics, Vienna, Austria) according to the manufacturer's instructions.

Statistical analysis

Results were expressed as mean ± standard deviation (SD). The Student's t-test of paired samples was used to compare the differences between differently treated groups within patients and within controls. Differences between patients and controls were determined by the Student's t-test for analysis of completely randomized 2-group design. The correlation between two sets of variables was determined by linear correlation test (Pearson test). P-value less than 0·05 was considered statistically significant.

Results

Characteristics of CD8+ T cells

Table I lists the phenotypes of T cells for the ITP patients and normal controls. The percentage of CD8+ T cells in ITP bone marrow was significantly higher than that in ITP peripheral blood (P < 0·01) and in normal bone marrow (P < 0·01). Moreover, as shown in Table II, CD8+ T cells derived from both peripheral blood and bone marrow of ITP patients showed vigorous proliferation after incubation with autologous platelets and APCs, but CD8+ T cells from controls did not show high reactivity.

Table I.   Phenotypic analysis of T cells in bone marrow and peripheral blood of ITP patients and controls (mean ± SD).
 ITP (n = 15)Control (n = 13)
Bone marrowPeripheral bloodBone marrowPeripheral blood
  1. *P < 0·01 compared with peripheral blood and control.

  2. P < 0·01 compared with control.

CD4/CD340·06% ± 5·15%*50·67% ± 3·70%†61·3% ± 3·43%63·05% ± 4·27%
CD8/CD359·94% ± 5·15%*49·33% ± 3·70%†38·70% ± 3·43%36·95% ± 4·27%
CD4/CD80·68 ± 0·14*1·04 ± 0·16†1·61 ± 0·171·63 ± 0·15
Table II.   Proliferation analysis of CD8+ T cells from bone marrow and peripheral blood of ITP patients and controls (mean ± SD).
 Bone marrowPeripheral blood
  1. ITP, idiopathic thrombocytopenic purpura.

  2. *P < 0·001 compared with control.

ITP (n = 15)13526 cpm ± 1037 cpm*13764 cpm ± 1049 cpm*
Control (n = 13)1934 cpm ± 319 cpm1725 cpm ± 338 cpm

Effect of CD8+ T cells on the quantity of megakaryocytes

As shown in Table III, CD8+ T cells from normal bone marrow did not affect the quantity of megakaryocytes. More megakaryocytes were observed in the ITP MNC in comparison with control MNC (P = 0·01), as shown in Tables III and IV. Moreover, in ITP cultures (Table IV), compared with group MNC, the depletion of CD8+ T cell procedure caused a significant reduction in megakaryocyte count (P < 0·001), while the more addition of ITP CD8+ T cells increased the number of megakaryocytes (P < 0·001), showing a trend that the megakaryocyte increment was dose-dependent with ITP CD8+ T cells (Fig 1A). However, the CFU-MK formation of chronic ITP patients was not influenced by the depletion or addition of autologous CD8+ T cells, suggesting that increased megakaryocyte count was not mediated by accelerated proliferation of megakaryocyte progenitors. Meanwhile, as shown in Fig 2, the megakaryocyte counts of all ITP patients were not influenced by the depletion or addition of autologous CD8+ T cells. Moreover, the platelet count did not influence the degree to which megakaryocytes were affected (group MNC, r = 0·057, P = 0·839).

Table III.   Megakaryocytopoiesis of controls (n = 13, mean ± SD).
GroupsCFU-MKMK (×105)PLT (×103)Polyploidy (%)Apoptotic (%)Bcl-xlFas
  1. CFU-MK, megakaryocyte colony-forming units; MNC, mononuclear cell; PLT, platelet.

MNC59·23 ± 16·204·57 ± 0·5823·86 ± 5·6826·46 ± 3·8423·07 ± 3·9323·97 ± 4·5236·82 ± 5·14
CD8+ T-dep61·00 ± 16·004·61 ± 0·5324·06 ± 5·4225·76 ± 4·0523·68 ± 3·8223·17 ± 4·1136·07 ± 5·88
Coculture60·77 ± 16·484·60 ± 0·5722·99 ± 5·3026·05 ± 3·9622·75 ± 3·6524·39 ± 4·0635·96 ± 5·82
Table IV.   Megakaryocytopoiesis of ITP patients (n = 15, mean ± SD).
GroupsCFU-MKMK (×105)PLT (×103)Polyploidy (%)Apoptotic (%)Bcl-xlFas
  1. CFU-MK, megakaryocyte colony-forming units; MNC, mononuclear cell.

  2. *P < 0·001 compared with MNC.

  3. P < 0·001 compared with Coculture.

MNC55·87 ± 18·405·75 ± 0·705·21 ± 1·048·20 ± 3·3616·58 ± 3·3737·66 ± 5·1921·99 ± 4·11
CD8+ T-dep59·40 ± 12·765·01 ± 0·58*19·78 ± 8·68*23·04 ± 4·61*20·99 ± 4·21*25·97 ± 5·97*34·14 ± 6·46*
Coculture54·47 ± 21·486·18 ± 0·79*3·16 ± 0·96*13·73 ± 3·32*13·33 ± 3·40*47·33 ± 6·21*15·67 ± 3·42*
DEX56·93 ± 14·545·26 ± 0·70†16·36 ± 6·86†21·37 ± 4·23†19·04 ± 3·41†30·70 ± 5·88†26·57 ± 4·63†
Figure 1.

 Dose response of CD8+ T cells on megakaryocyte counts and apoptotic megakaryocytes. The percentage of CD8+ T cells in cultures of CD8+ T-dep, MNC and Coculture groups was 0·51% ± 0·25%, 19·89% ± 3·69% and 43·29% ± 2·77% respectively. The megakaryocyte counts (A) and apoptotic megakaryocytes (B) in cultures were determined by flow cytometry. A significant difference was found between any two groups(P < 0·001) by Paired-samples t-test. Data are presented as mean ± standard deviation.

Figure 2.

 Effect of CD8+ T cells on megakaryocyte counts. Megakaryocyte counts from individuals indicated on the x-axis in MNC (bsl00046),CD8+ T-dep (bsl00001), and Coculture (bsl00066) groups are shown. The horizontal line represents the mean of the MNC group. Some (10/15) patient’ megakaryocyte counts were influenced by the depletion or addition of autologous CD8+ T cells, the other (5/15) patient’ megakaryocyte counts were not influenced.

Effect of CD8+ T cells on the quality of megakaryocytes

The depletion or addition of normal CD8+ T cells did not influence megakaryocytic characteristics (Table III). As shown in Tables III and IV, the platelet count and the percentage of polyploidy and apoptotic megakaryocytes in ITP MNC were significantly lower than those in control MNC (P = 0·01). In the ITP CD8+ T-dep, MNC and Coculture groups, decreased platelet production and reduced percentage of polyploidy and apoptotic megakaryocytes were observed as the number of autologous CD8+ T cells increased in cultures (P < 0·001)(Fig 1B). Despite the maximal count of megakaryocytes in the Coculture group, platelet production and percentage of polyploidy and apoptotic megakaryocytes were minimal, suggesting that decreased apoptosis of megakaryocytes may result in increased megakaryocyte count and reduced platelet production.

Further study of the modulators of apoptosis in megakaryocytes showed that, as more ITP CD8+ T cells were added to cultures, a higher expression of Bcl-xl and a lower expression of Fas were detected, in accordance with the decreased apoptosis of megakaryocytes.

sFas level in supernatant

As shown in Fig 3, the level of sFas increased as CD8+ T cells numbers increased in cultures, which was consistent with the decreased apoptosis of megakaryocytes, indicating that sFas may be responsible for the megakaryocytic apoptosis. However, sTRAIL was below the detection limit.

Figure 3.

 Level of sFas in supernatant. The level of sFas was determined in the corresponding cell-free supernatants by ELISA. A significant difference was found between any two groups (P < 0·001) by Paired-samples t-test. Data are presented as mean ± standard deviation.

Intervention with dexamethasone

The percentage of CD8+ T cells in coculture was 43·29% ± 2·77%, and was significantly reduced to 3·75% ± 1·30% (P < 0·001) after incubation with dexamethasone. The characteristics of megakaryocytes in the DEX group were compared with the Coculture group (Table IV). The presence of dexamethasone reduced the megakaryocyte count, increased platelet production, and enhanced the percentage of polyploidy and apoptotic megakaryocytes (P < 0·001).

Discussion

It has been demonstrated that ITP plasma and purified antiplatelet auto antibodies inhibit megakaryocytopoiesis in vitro (Chang et al, 2003; McMillan et al, 2004). However, the role of cellular regulation on megakaryocytopoiesis in ITP is largely unclear. Some studies suggested that CD8+ T cells may be involved in the pathogenesis of acquired amegakaryocytic thrombocytopenic purpura by exerting suppression on megakaryocyte differentiation (Gewirtz et al, 1986; Benedetti et al, 1994). In the present study, we performed a similar culture system to the methods of Benedetti et al (1994) to preliminarily measure the quantity and quality of megakaryocytes. Clonal analyses of CD8+ T cells and further studies of the modulators of apoptosis and maturation in megakaryocytes are ongoing in our laboratory.

Recent studies have demonstrated that CD8+ T cell-induced lysis of platelets in chronic ITP may be involved in the pathogenesis of this disorder (Olsson et al, 2003; Zhang et al, 2006). As the progenitors of platelets, megakaryocytes express the same antigens as platelets. We speculated that activated CD8+ T cells in chronic ITP should attack autologous megakaryocytes leading to impaired megakaryocytopoiesis and thrombocytopoiesis. Our study showed bone marrow CD8+ T-cell proliferation in ITP was platelet-specific, which was similar to a previous study on peripheral CD8+ T cells (Ware & Howard, 1993), suggesting that the marked increase of activated CD8+ T cells in bone marrow are functional and may be responsible for the pathogenesis of ITP.

However, our present work showed unexpected results: as more CD8+ T cells were present in ITP cultures, megakaryocyte count was boosted while maturation and ability to produce platelets were suppressed. Unlike the hypothesis that CD8+ T cells may equally lyse megakaryocytes, our finding was compatible with early morphological studies of ITP bone marrow, which showed normal or increased numbers of megakaryocytes (Dameshek & Miller, 1946), and with the findings that plasma thrombopoietin (TPO) levels in ITP patients were normal to slightly elevated, indicating that the total megakaryocytic mass was not decreased in ITP (Porcelijn et al, 1998; Kappers-Klunne et al, 2001; Sun et al, 2006). According to these studies, we believe that CD8+ T cells of ITP patients might affect megakaryocytopoiesis through mechanisms other than cell-mediated lysis exerted on peripheral platelets.

It is now evident that platelet formation is closely associated with megakaryocyte apoptosis (De Botton et al, 2002; Clarke et al, 2003; Patel et al, 2005). Houwerzijl et al (2006) have reported that the inhibited megakaryocyte apoptosis may contribute to thrombocytopenia and augment dysfunctional megakaryocytes. Bcl-xl, an anti-apoptotic member of Bcl-2 family, is known to inhibit platelet formation when overexpressed in megakaryocytes (Sanz et al, 2001; Kaluzhny et al, 2002; Zhang et al, 2004). Several reports have also indicated that the activated Fas/FasL pathway increases proplatelet formation and production of functional platelets in normal megakaryocytes (De Botton et al, 2002; Clarke et al, 2003; Coppola et al, 2006).

Our data showed that the percentage of apoptotic megakaryocytes and the expression of Fas in megakaryocytes were decreased, while the expression of Bcl-xl was increased as more CD8+ T cells were added to the cultures. Furthermore, decreased megakaryocyte apoptosis was followed by an increase in megakaryocyte count and decreased platelet production, suggesting that CD8+ T cells of ITP patients might influence megakaryocytopoiesis and thrombocytopoiesis by suppressing megakaryocyte apoptosis.

It is well known that decreased apoptosis of activated T lymphocytes play an important role in the pathogenesis of ITP, and alteration of the Fas/FasL pathway, particularly increased level of sFas, which inhibit activated T lymphocytes apoptosis by means of blocking FasL, would be responsible for the immunological pathogenesis (Yoshimura et al, 2000; Olsson et al, 2005). A potential mechanism involved in the suppression of megakaryocyte apoptosis in ITP may be a result of anti-apoptotic molecules, such as sFas, which contribute to the accumulation of activated T lymphocytes in ITP, acting as a paracrine regulator of the megakaryocytic apoptosis. Further studies should be performed to define the complete mechanism for decreased megakaryocyte apoptosis in patients with ITP.

Also, besides the well-known inhibitory effects of glucocorticoids on phagocytosis and antibody production, the deletion of CD8+ T cells by DEX-induced apoptosis might explain the correction of dysmegakaryocytopoiesis in our in vitro culture system. This intriguing mechanism might also explain the beneficial effect of glucocorticoids in ITP management and therefore, merits further studies.

In conclusion, our results provide the first evidence that CD8+ T cells in bone marrow may suppress megakaryocyte apoptosis, leading to impaired platelet production in chronic ITP. Blocking the interactions between anti-apoptotic molecules and megakaryocyte might provide a reasonable therapeutic strategy for ITP.

Acknowledgements

This work was supported by grants from 973 Program 2006 CB 503803, National Natural Science Foundation of China (No. 30300312, No. 30470742, No. 30579779 and No. 30600259), and Key Clinical Research Project of Public Health Ministry of China 2004–2006.

Ancillary