Mitoxantrone induces cell death in peripheral blood leucocytes of multiple sclerosis patients

Abstract

Mitoxantrone (MX) is a cytotoxic drug with proven clinical efficacy in active multiple sclerosis (MS). In this ex vivo study we investigated the immunological effects of MX on peripheral blood leucocytes (PBL) from MS patients. PBL were isolated from 46 patients with active MS (mean age 42 years, female : male 1·4 : 1) before and immediately after 1 h MX infusion. Isolated PBL were cultured and stimulated with phytohaemagglutinin (PHA), T cell receptor stimulating monoclonal antibody (MoAb) X35 or kept in culture medium alone. Proliferation was measured by [³H]-thymidine incorporation. MX-uptake and cell death in PBL subpopulations was analysed by flow cytometry using antibodies against cluster of differentiation (CD)-surface antigens, annexin V (AnnV) and propidium iodide (PI). MX was incorporated rapidly into PBL. After only a 1-h in vivo exposure, MX reduced proliferative responses in unstimulated and stimulated PBL (PHA: − 17%, MoAb X35: − 13%). MX-exposed PBL showed an increase of AnnV⁺/PI⁺ cells (unstimulated: 12%, PHA: 15%), which was even more pronounced 2 weeks after infusion. No difference was observed between de novo MX-treated patients and those on long-term MX treatment. In T cell receptor stimulated PBL, cell death was induced preferentially in CD19-positive B cells and to a lesser extent in CD8-positive T cells. MX is incorporated rapidly in circulating PBL of MS patients and induces a pronounced suppression of proliferative responses. This suppression appears to be mediated at least partly by the induction of late apoptotic/necrotic cell death with a preferential susceptibility of B cells.

Introduction

The cytotoxic drug MX is a highly effective immunoactive agent in the treatment of active multiple sclerosis (MS) [1]. MX is currently recommended as an escalation therapy if other immunomodulatory treatment options have failed [2,3]. Despite proven clinical efficacy, only few data are available to understand its superior mode of action. MX inhibits T and B cell proliferation in vitro and in vivo [4–6], decreases antibody synthesis and deactivates macrophages. The anthracenedione MX intercalates with the DNA and causes DNA single- and double-strand fragmentation via interaction with topoisomerase-2 [7,8]. In chronic leukaemias, mitoxantrone cytotoxicity is mediated by induction of apoptosis in vitro [9,10]. MS is currently the only established non-oncological indication for MX. In contrast to malignancies, untransformed immune cells are the presumable therapeutic target of MX in MS.

In this study we have investigated immunological short-term effects on peripheral blood leucocytes (PBL) isolated from MS patients immediately before and after MX administration.

Materials and methods

Patients

Forty-six patients (female : male 1·4 : 1; 42 ± 11 years, mean ± s.d., min. 24, max. 64 years) were included. Of these, 14 patients had relapsing–remitting MS (RRMS), 27 secondary progressive MS (SPMS) and five primary progressive MS (PPMS). Disability as measured by the expanded disability status scale (EDSS) was 5·1 ± 1·4 (mean ± s.d., min. 2·5, max. 8). Patients with highly active disease were suitable for treatment with MX as an immunosuppressive escalation therapy following established criteria [1–3]. Thirty-nine patients had been pretreated with immunomodulators or immunosuppressive agents (two interferon (IFN)-β1a, 34 IFN-β1b, one glatirameracetate, four cyclophosphamide, seven azathioprine, two methotrexate; multiple pretreatments in 11 patients). The study was approved by the local ethics committee and informed consent was obtained from the patients. A total of 60 blood samples were taken; however, in some patients not enough material could be obtained to perform both proliferation and cell death assays. Thus, for proliferation studies, 53 probes and for cell death assays 56 probes were investigated. While 25 patients had received MX previously, 21 patients received MX de novo. Ten patients could be followed over repetitive cycles (min. two times, max. four times), usually administered in 3-month intervals. Four patients were additionally followed up 2 weeks after infusion. This heterogeneity reflects daily clinical practice.

Intravenous mitoxantrone therapy

After antiemetic drugs (i.e. ondansetron, granisetron or dimenhydrinate) MX was given intravenously during a standardized infusion time of 1 h. At 24 visits, the standard dose of 12 mg/m² body surface, at 14 visits a dose of 5 mg/m² was infused [1]. At 22 visits, patients received a dosage between 5 and 12 mg/m² tailored individually to leucocyte counts [1]. In 14 of 60 examinations patients had received glucocorticosteroids around the time-point of MX infusion. However, in 11 of these patients, glucocorticosteroids were given with an interval of at least 1 day before MX; 20–40 ml EDTA blood was taken from each patient before and immediately after MX infusion.

Cell culture

All culture media and supplements were purchased from Gibco BRL (Eggenstein, Germany). Immediately after blood samples had been taken, PBL were separated by gradient centrifugation using Ficoll-Paque (Nycomed AS, Oslo, Norway). For  proliferation  assays,  PBL  were  seeded  at  4 × 10⁵ cells per  well  in  96-well  round-bottomed  microtitre  plates (Nunc, Wiesbaden, Germany) in 100 µl RPMI-1640 (2 m m glutamine, 100 U/ml penicillin G, 100 µg/ml streptomycin sulphate, 5% fetal calf serum). For immunoflow-cytometry (FACS) analyses, PBL were cultured at 2 × 10⁶ cells/ml in Petri dishes. Cell cultures were either kept untreated or stimulated with the mitogenic lectin phytohaemagglutinin (PHA) (5 µg/ml; Sigma, Deisenhofen, Germany) or the T cell receptor-activating anti-CD3 MoAb X35 (0·1 µg/ml, Coulter Immunotech, Unterschleißheim, Germany) [11]. Positive controls for cell death were treated with methylprednisolone (0·1 mg/ml, Aventis Pharma, Frankfurt am Main, Germany).

Proliferation assays and immunoflow cytometry

After 48 h, cultures were pulsed with [³H]-thymidine (0·009 MBq/well, Amersham-Buchler, Braunschweig, Germany, 16 h). Cells were collected on fibreglass filter paper in a Betaplate 96-well harvester (Pharmacia Biotec), and incorporated radioactivity was quantified using a Betaplate 96-well liquid scintillation counter (Pharmacia Biotec).

After overnight culture, cell death was analysed by FACS using fluorescein-conjugated annexin V (AnnV-FITC, Roche Diagnostics Boehringer Mannheim, Mannheim, Germany) and propidiumiodide (PI), characteristic of early apoptotic or late apoptotic/necrotic changes, respectively [12]. Subpopulation analyses were performed by double-staining using phycoerythrin (PE)- or cychrome (CyC)-conjugated monoclonal antibodies against human CD-surface antigens (BD Biosciences, Heidelberg, Germany, Table 1) and AnnV-FITC as described previously [13]. A minimum of 5000 events was counted (FACSCalibur, CellQuest software, Pharmingen).

Table 1.

Fluorochrome and cell specificity of antibodies used for cytofluorometric analyses.

Antibody	Fluorochrome	Antigen/cell specificity
Anti-human CD3	CyC	T cell receptor
Anti-human CD4	CyC	Helper T cells
Anti-human CD8	CyC	Cytotoxic T cells
Anti-human CD14	PE	Monocytes/macrophages
Anti-human CD19	CyC, PE	B cells
Anti-human CD25	PE	IL-2 receptor α-chain
Anti-human CD56	PE	Natural killer cells
IgG1κ, IgG1 γ1, IgG2bκ	CyC, PE	Isotype controls

CyC: cychrome, PE: phycoerythrin.

Quantification of incorporated MX was performed using a 635-nm red-diode laser and a 670 nm bandpass filter [14]. For direct ex vivo analysis of incorporated MX, PBL from nine patients before and after MX infusion were subjected to FACS-analysis immediately after isolation.

Statistical analysis

After checking for a symmetrical distribution of data, the t-test for grouped data or, when applicable, Wilcoxon tests were used for non-parametric calculations using StatView software (SAS Institute, Inc.).

Results

Rapid incorporation of mitoxantrone into PBL can be detected ex vivo reliably by flow cytometry

Autofluorescence of MX can be detected by flow cytometry using an 635-nm red-diode laser. Figure 1 depicts a representative FACS histogram analysis of PBL isolated before and directly after MX infusion over a short and standardized infusion time of 1 h. Incorporated MX could be detected reliably in PBL after MX in vivo exposure in an applied concentration range of 5–12 mg/m² body surface. Moreover, by adding MX to PBL of the same patient in vitro, a concentration-dependent increase of the fluorescence intensity could be demonstrated (Fig. 1).

Fig. 1.

Mitoxantrone (MX) is incorporated rapidly into circulating PBL. Representative FACS histogram of PBL before or after mitoxantrone in vivo exposure (5 mg/m², 1 h), or after MX addition in vitro in different concentrations as indicated (1 h). Detection of autofluorescence using a 635 nm red-diode laser/670 nm bandpass filter. x-axis: fluorescence intensity, y-axis: cell count.

A short in vivo mitoxantrone exposure decreases proliferative responses of PBL

PBL isolated after 1 h MX infusion exhibited clearly reduced cellular proliferation in comparison to their intra-individual control before infusion. In PHA-stimulated PBL (Fig. 2), inhibition was observed in blood samples obtained at  44 of 53 visits (83%) with a mean decrease of 17·3% [23·089 ± 11·960 counts per minute (cpm, mean ± s.d.) before infusion versus 19·449 ± 10·858 cpm after infusion, P < 0·0001, paired t-test]. In MoAb X35-stimulated PBL an inhibition occurred in 21 of 23 (91%) visits with a reduction of 12·7% (18·114 ± 10·886 cpm before infusion versus 15·810 ± 10·061 after infusion, P = 0·003). Unstimulated PBL also showed an inhibition; however, total cell counts were below 1·000 cpm. In six patients with an initially decreased cellular proliferation who were accessible for follow-up examinations inhibition persisted at the next MX cycle, usually given in 3-month intervals (Fig. 2. e.g. examinations no. 2/43, 8/46, 15/27, 16/41, 33/42 and 38/47).

Fig. 2.

One hour in vivo exposure to MX decreases PBL proliferation. Proliferation-index: ratio of PBL [³H]-thymidine incorporation (cpm) of individual MS-patients after MX exposure/before exposure (PHA-stimulation, 5 µg/ml). Eight MS-patients were observed longitudinally as indicated by resembling bars (2/43, 8/46, 15/27, 16/41, 31/39, 33/42, 38/42, 5/20/36/45/52). Asterisks indicate de novo MX-treated MS patients; x-axis: examination number.

Decreased cellular proliferation could be observed in patients treated with MX de novo and in patients receiving maintenance therapy (Fig. 2). Especially under PHA-stimulated conditions, a small increase in proliferation was observed in 17% of the examinations (Fig. 2) both in patients treated with MX de novo [four of 20 (20%) visits] and in patients receiving maintenance therapy [five of 33 (15%) visits]. Two patients with an initial increase in cellular proliferation could be followed and eventually demonstrated a decrease in PBL proliferation. Of these, one patient receiving MX maintenance therapy demonstrated a decrease after 3 months (visits 31 and 39). The other de novo-treated patient (visit 5) still had an augmented proliferation at follow-up after 3 months (visit 20) and finally showed a decrease after 9 months (visit 36), which persisted in subsequent examinations after 12 and 16 months (visits 45 and 52). Interestingly, this female patient (RRMS, EDSS 5·0) experienced a relapse and an increase of lesion volume in the T2-weighted cranial MRI just prior to month 9, whereas she remained clinically stable during the rest of the observation period.

High-dose methylprednisolone pulse-therapy decreases proliferative responses by rapid induction of apoptosis in PBL of MS patients [11]. The MX-induced decrease of PBL-proliferation was still apparent in those patients who had not received glucocorticosteroids previously, indicating that glucocorticosteroid effects were not responsible for the results observed here (cpm before versus after MX; unstimulated: 763 ± 582 versus 454 ± 385, P = 0·0001; PHA-stimulated: 22·497 ± 11·140 versus 18·845 ± 10·539, P < 0·0001; MoAb X35-stimulated: 19·370 ± 12·555 versus 16·485 ± 11·583 P < 0·05).

PBL proliferation of four patients was also examined 2 weeks after MX infusion and compared to baseline levels before MX administration. While stimulated PBL showed decreased cellular proliferation directly after MX infusion, this inhibition could be observed in only two patients after 2 weeks.

Short in vivo exposure to mitoxantrone induces late apoptotic/necrotic cell death

In PHA-stimulated PBL, the rate of AnnV⁺/PI⁺-cells after MX was augmented in 36 of 51 examinations (P < 0·0007, Wilcoxon test) with a mean increase of 15·4% (Fig. 3a) and 16·1 ± 8·2% (mean ± s.d.) before infusion versus 18·7 ± 8·5% after infusion, P = 0·015, t-test). In unstimulated PBL, an increase of AnnV⁺/PI⁺-cells could be observed in 37/52 examinations (P < 0·005, Wilcoxon test) with a mean increase of 11·8% (mean 6·7 ± 4·8% before infusion versus 7·5 ± 5·4% after infusion, P = 0·003, t-test). With MoAb X35 stimulation there was a trend only to increased cell death (7·5 ± 3·4% before infusion versus 7·8 ± 3·8 after infusion (P = 0·38, t-test), due presumably to lower numbers of examinations (n = 31).

Fig. 3.

(a) One-hour in vivo exposure to MX induces late apoptotic/necrotic cell death. Bars indicate proportion of AnnV/PI-positive PBL of individual MS patients before (white bars) and after i.v. MX administration (coloured bars, PHA-stimulation, 5 µg/ml). Asterisks indicate de novo mitoxantrone-treated MS patients. (b) Time-course of mitoxantrone-induced late apoptotic/necrotic cell death. Bars indicate proportion of AnnV/PI-positive PBL of individual MS patients before, 1 h after and 2 weeks after i.v. MX administration (MoAb X35-stimulation, 0·1 µg/ml); x-axis: examination number.

The increase of late apoptotic/necrotic cell death after MX occurred both in de novo-treated patients and those receiving maintenance therapy (Fig. 3a). An increase of cell death could still be observed, omitting patients who had received glucocorticosteroids around the time-point of MX infusion (PHA-stimulation, increase in 31/37 examinations P < 0·0001 Wilcoxon, mean increase 21·5%, P < 0·0001), indicating that glucocorticosteroid-induced apoptosis was not responsible for the MX effects observed here.

Whereas after a 1-h MX in vivo exposure an augmentation of cell death was already detectable, this effect was even more pronounced at the time-point known to correspond with maximal leucocyte suppression (Fig. 3b). Thus, in four patients investigated 2 weeks after infusion, there was a strong increase of AnnV⁺/PI⁺-cells under unstimulated and stimulated conditions. Especially in the MoAb X35-stimulated PBL (Fig. 3b) late apoptotic/necrotic cells increased 1·2–4·8-fold compared to baseline values before infusion (AnnV⁺ PI⁺ cells, mean 6·5 ± 1% before infusion versus 20·6 ± 8% 2 weeks after infusion).

In CD3-stimulated PBL, mitoxantrone-induced cell death occurs preferentially in CD19-positive B cells

After PHA stimulation, cell death was distributed evenly among all subpopulations investigated (for investigated cell types see Table 1; data not shown). However, after MoAb X35-stimulation, which acts on CD3-positive T-cells only, MX induced cell death preferentially in the CD19-positive B cell compartment. As shown in Fig. 4, an increase in the rate of apoptotic CD19-positive B cells after MX administration could be observed in 27 of 34 visits (79%, P = 0·006, Wilcoxon test), with a mean increase of 12·5% (P = 0·02, t-test). This predominant susceptibility of CD19 cells appeared to be most pronounced in the patients who received MX de novo, with an increase of apoptotic CD19 cells in 12 of 14 visits (86%, P = 0·01, mean increase 18·3%, P = 0·01, Fig. 4). Also in CD8 T cells an increase of cell death after MX could be observed in the majority of the visits (61%). The CD56 positive natural killer cell subpopulation appeared to be least susceptible to MX (increase in cell death in 41% of the visits). All other investigated PBL subgroups showed a heterogeneous response (increase of cell death in percentage: CD3: 44%, CD4: 50%, CD14: 52%, CD 25: 48%).

Fig. 4.

Mitoxantrone-induced cell death in MoAb X35-stimulated, CD19-positive B cells. Ordinate: % change of AnnV/CD19-positive cell fraction after MX administration versus intra-individual control before MX; positive values indicate increase of cell death, negative values indicate decrease. MoAb X35-stimulation, 0·1 µg/ml; x-axis: examination number. Asterisks indicate de novo MX-treated MS patients.

Discussion

The delayed decrease of peripheral blood leucocytes is a well known indicator of a haematological response to MX which can be used clinically for dose adjustments. Our principal finding is a profound early immunological effect of MX on PBL as short as 1 h after in vivo exposure. MX is incorporated rapidly into circulating PBL of MS patients and this can be detected readily by flow cytometry ex vivo in a dose-dependent manner due to its autofluorescence. Rapid MX incorporation underscores the notion but does not prove that the early immunological effects described here may be of therapeutic relevance.

MX induces pronounced early antiproliferative effects that appear to be mediated at least partly by late apoptotic/necrotic cell death, as demonstrated ex vivo. CD19-positive B cells and to a lesser extent CD8-positive T cells are affected predominantly by this mechanism. Cell death increases dramatically 2 weeks after MX therapy. No major difference of de novo-treated patients and those receiving maintenance therapy could be observed, suggesting that the effects of MX are not reduced by tachyphylaxis.

Immediate ex vivo analyses yield only negligible amounts of apoptotic PBL ([11], own unpublished observations), due presumably to the rapid elimination of apoptotic cells, e.g. in the reticuloendothelial system [13]. Therefore PBL were first cultured overnight for the assessment of proliferation and cell death, according to our previously published data [11].

Inhibition of proliferative responses after MX was lower than that seen after glucocorticosteroids, where a decrease of 65·2% with RRMS, 56·8% with SPMS and 20·3% with PPMS patients has been reported after PHA stimulation [11]. By comparison, in our study population with a majority of chronic-progressive patients, an inhibition of 17·3% (PHA) or 12·7% (MoAb X35) could be observed. Also in our study population the most pronounced inhibitory effect was found in RRMS patients (22%, SPMS 15%, PPMS 5%, PHA). Twenty-three per cent of the MX visits were performed following a glucocorticosteroid-pulse therapy. However, all but three patients had received glucocorticosteroids at least 1 day prior to MX infusion, making a major bias highly unlikely when comparing comparing MX-mediated effects immediately after infusion with the patient control before infusion. In addition, omission of glucocorticosteroid-pretreated patients showed no major difference. To our surprise, in one small subgroup especially after PHA stimulation, an increase of proliferative responses could be observed after MX. Two of these patients could be followed and eventually showed a decreased proliferation. Of these, one patient showed this decrease in proliferation after MX after 9 months of MX/four treatment cycles while demonstrating signs of clinical and paraclinical (cMRI) disease progression. In the European dose comparison trial a clinical response with a reduction in relapse rate and prolongation of intervals between relapses was observed after approximately 3 months on MX therapy. However, as the short-term effects of mitoxantrone as described in this study cannot be distinguished clinically from potential long-term effects, we did not attempt to correlate these effects to overall clinical/paraclinical outcome parameters.

Antiproliferative effects of MX on PBL can, in principle, be mediated by anergy or cell death. In leukaemic cells, MX cytotoxicity is mediated by apoptosis in vitro [9,10]. Here we were able to demonstrate a mitoxantrone-induced increase of AnnV/PI-double-positive PBL, corresponding to late apoptotic/early necrotic cells [12]. This relatively late stage of cell death can be explained presumably with the overnight cell culture in our experimental design. Apoptosis of autoreactive as well as bystander T cells in situ has been identified as a major mechanism in the termination of autoimmune inflammation in the rodent and human central nervous system (CNS) [15]. Glucocorticosteroids increase apoptosis in PBL of MS patients as well as encephalitogenic T cells in situ in the model disease experimental autoimmune encephalomyelitis (EAE) [16]. The increase of cell death in PBL in this study was lower than that reported after glucocorticosteroid pulse therapy using a different, yet comparable cell death detection technique [11]. The increase of PBL cell death after in vitro addition of methylprednisolone argues against a reduced overall susceptibility towards cell death as a reason for the quantitative differences observed. Glucocorticosteroids mediate their anti-inflammatory effects partly via rapid, non-genomic pro-apoptotic mechanisms directed at cellular membranes [17]. In contrast, MX-induced cell death in vitro is mediated via DNA-intercalation with subsequent conformational changes or interactions with DNA and topoisomerase II [18]. Thus, the rapid non-genomic glucocorticosteroid effects may explain, at least in part, the higher apoptosis rate observed at similar time-points. Although the absolute magnitude of MX-induced PBL cell death appears to be small, this might still have a major quantitative impact on the entire cell population. Given the rapid completion of the apoptotic cell death programme within 4–5 h in certain tissues in vivo, even a small increase of cell death will lead to a major cumulative loss of cell numbers over 24 h [19]. Moreover, taking into account the long-terminal plasma half-life of up to 215 h major effects on PBL cell death may occur. Following this line, the profound increase of cell death in circulating PBL 2 weeks after MX infusion observed here could be responsible for the observed decrease of leucocyte counts rather than myelosuppression.

Recent studies argue for an important role of autoantibody-producing B cells in the immunobiology of active MS [20,21]. Our data with a predominant susceptibility of CD19-positive B cells to mitoxantrone-induced immediate cell death is well in line with quantitative longitudinal observations after MX therapy in MS patients. Thus, after three cycles of MX a significant decrease could be observed only in the number of B cells [22], and after 3 years therapy B cells were reduced by approximately 60% [23]. Cell death of CD19-positive B cells was most pronounced in PBL treated with a T cell receptor-activating antibody. The mechanisms by which T cell activation lead to an increased susceptibility of B cells towards mitoxantrone-induced cell death are currently under investigation.

In conclusion, in addition to the long-term effects, the short-term immunological effects of MX on circulating PBL may be clinically relevant. MX dose limitations due to cardiotoxicity warrant further investigation on the mechanisms of action and potential therapeutic surrogate parameters.