N-acetyl-L-cysteine (NAC) is a thiol antioxidant that stimulates glutathione synthesis in cells. Several studies indicate that NAC possesses immunomodulatory properties in vitro, but both inhibitory and activating effects on immunity have been reported. We observed that allogeneic stem cell transplantation (ASCT) patients who were randomized to receive NAC 100 mg/kg/day (n=73) had an increased prevalence of grade II-V acute graft-versus-host disease (GvHD) compared to patients who did not receive NAC (n=87), indicating that NAC has an immunostimulatory effect in vivo. When studying the effect of NAC on T-cell-mediated immunity in vitro, we found that moderate levels of NAC (0.4–3.2 mM) increased alloantigen-induced proliferation, expression of activation markers CD25 and CD71 on T cells, and production of IFN-γ and IL-10. In contrast, high concentrations of NAC (12.5–50 mM) were suppressive, which may explain previously conflicting data. NAC did not cause an increase in expression of CD86, CD80, and CD83 on mature DCs at any concentration, whereas high concentrations suppressed DC maturation. Furthermore, T cells exposed to suppressive concentrations of NAC in a primary stimulation were highly responsive when re-stimulated in the absence of NAC. To conclude, NAC appears to increase acute GvHD and has an immunostimulatory effect on alloantigen-specific T cells.
Allogeneic stem cell transplantation (ASCT) is an established therapy for various malignancies of the hematopoietic system, but there are still problems with handling transplant-related disorders such as graft-versus-host disease (GvHD), viral infections, and reactivation, as well as relapse of the underlying malignant disease. Furthermore, the conditioning regimen may induce toxicity, such as sinusoidal obstructive syndrome (SOS) of the liver 1.
N-acetyl-L-cysteine (NAC) is a thiol anti-oxidant that stimulates glutathione synthesis in cells and works as a scavenger of reactive oxygen species. NAC is used clinically as an antidote for paracetamol toxicity 2 and as a mucolytic agent in chronic pulmonary disease 3. It has also been evaluated for its potential to restore intracellular deposits of reduced glutathione (GSH) in HIV-infected patients suffering from glutathione deficiency 4, 5. The ability of NAC to protect cells and cellular components against oxidative stress has provided a rationale for using it to treat busulfan-induced liver toxicity in ASCT patients 6, 7.
Several in vitro studies have indicated that NAC exerts effects on the immune system, but there have been conflicting results. Whereas most studies have indicated that NAC has an inhibitory effect on T-cell-mediated immunity and activation of the transcription factor NF-κB 8–11, others have suggested an immunopotentiating effect 12, 13. The discrepancies between these observations have not been thoroughly studied, but can partly be explained by the use of different concentrations of the drug, the nature of the stimulating antigen, and differences in the responding cells.
Here, we show that NAC may have an immunopotentiating effect in immunocompromised patients and we provide in vitro data supporting this observation. We also present data indicating that NAC may have dual and opposing effects on immunity, depending on the dose and kinetics, and also differential effects on different kinds of immune cells.
Administration of NAC promotes alloreactive immunity in ASCT patients
The ability of NAC to protect cells from oxidative stress by inducing glutathione has provided a rationale for using it to prevent SOS and liver toxicity after ASCT. Unfortunately, a prospective randomized study with 160 patients failed to show a positive effect of NAC on hepatotoxicity in ASCT patients 7. Maximum bilirubin levels and recovery of liver enzymes were the same in patients randomized to receive NAC and control patients. SOS developed in two and three patients in the two groups, respectively. However, in this study we observed that patients who received NAC as a therapy to treat liver toxicity had an increased prevalence of grade II–IV acute GvHD compared to patients who had been randomized to not receive NAC (p=0.04) (Fig. 1A). Importantly, the NAC-treated group only had an increased risk of grade II acute GvHD (30%) compared to the control group (10%), whereas corresponding values for severe acute GvHD of grades III–IV in the two groups were 8 and 14%, respectively (data not shown).
Additionally, we analyzed soluble IL-2 receptor (sIL-2R) serum levels, which previously have shown to be a good marker for GvHD and other complications involving immune activation 14, 15. Median levels of sIL-2R on day 3, 10, 17, and 24 after ASCT was 1504, 1597, 1299, and 1682 U/mL, respectively. In multivariate analysis, we found that NAC treatment was correlated to high levels of sIL-2R (>median) on day 10 after ASCT (OR 3.13, p<0.01) (Fig. 1B). No correlation between NAC and sIL-2R was found on day 3, 17 and 24 after ASCT. High levels of serum sIL-2R was also correlated to unrelated donor (OR 3.53, p=0.004), conventional full dose conditioning (OR 2.72, p=0.014) and acute GvHD (OR 1.95, p=0.08). Altogether, these results indicate that NAC may have an immunostimulatory effect in an allogeneic setting in vivo.
NAC has an immunostimulatory effect under certain conditions in vitro
These observations prompted us to study in great detail how NAC affects cell-mediated immunity in an allogeneic setting in vitro. We first examined how different concentrations of NAC influence the proliferative response in mixed lymphocyte reactions (MLRs) in which PBMC were stimulated with irradiated allogeneic PBMC. Interestingly, we found that proliferation increased when low to moderate concentrations of NAC (0.4–3.1 mM) were added to the cultures, whereas high concentrations (12.5–50 mM) essentially abrogated proliferation (Fig. 2A). This pattern was observed in 11 of 12 donors, where the proliferative response in one donor peaked at 0.4 mM, in five at 1.6 mM, and in five at 3.1 mM. When NAC concentrations of 12.5 mM or higher were added to the cultures, all donors showed reduced proliferation compared to control MLRs. These effects were not due to differences in pH, as the pH value in all cultures remained at 7 after 6 days (data not shown).
When the production of cytokines was analyzed in this setting, we found that the production of IFN-γ and IL-10 was increased after addition of low or moderate concentrations of NAC, similarly to that of proliferative responses, whereas the production of IL-17 was reduced in a dose-dependent fashion (Fig. 2B–D). We also measured the levels of IL-2 in these cultures, and found that there were no detectable levels of IL-2 in the control MLRs, but we noted that addition of NAC induced production of IL-2 in 2 out of 5 donors (Fig. 2E).
Likewise, when we examined the effect of NAC on the activation status of T cells, we found that the expression of CD25 and CD71 on alloantigen-stimulated T cells was increased when 1.6 mM NAC was added to the cultures, whereas 12.5 mM NAC had the effect of reducing the surface expression of these markers (Fig. 3A and B). The activation status of the cells was also illustrated by an increased or reduced number of lymphoblasts in the presence of low or high NAC concentrations, respectively. A representative experiment is shown in Fig. 3A and the results of CD25 expression from four experiments are summarized in Fig. 3B.
DCs are central for activation and priming of naive T cells, and to further study the effect of NAC during the activation of naive CD4+ T cells, as well as total CD3+ T cells, we stimulated T cells with allogeneic DCs with or without NAC. We found that moderate levels of NAC significantly promote proliferation in naive CD4+CD45RA+ T cells and in total CD3+ T-cell in response to allogeneic DCs, whereas high levels abrogate proliferation (Fig. 3C and D). A similar pattern was observed when measuring IFN-γ and IL-10 levels in supernatants from these cultures, although it did not reach statistical significance (Fig. 3C and D).
To conclude, these experiments show that NAC may have dual and opposing effects on the immune system, depending on the dose. These observations might explain the conflicting results from previously published reports indicating that NAC can either stimulate or inhibit immune responses.
The immunostimulatory effect of NAC directly affects T cells
Previous studies have shown that high concentrations of NAC have a direct suppressive effect on DCs, which in turn inhibit the activation of T cells 11. We then examined whether NAC has a direct effect on APCs or whether it mainly affects T cells. Immature monocyte-derived DCs were cultured in the presence of different concentrations of NAC during TNF-α and PGE2-induced maturation and the expression of activation markers was analyzed to determine how NAC affects the activation status and antigen-presenting abilities of DCs. We found that high concentrations of NAC suppressed the upregulation of CD83, CD86, and CD80, whereas expression of HLA-DR was not affected by NAC (Fig. 4A and B). In contrast to the results on alloantigen-induced activation of PBMCs described above, we did not observe any activating effect of low to moderate concentrations of NAC on the expression of co-stimulatory and maturation markers on DCs. Cytokine secretion from DCs from three different donors in three independent experiments was examined after maturation with LPS in the absence or presence of 0.8–25 mM NAC. High NAC concentrations inhibited the production of IL-10 and IL-6, but low to moderate concentrations had no effect on cytokine secretion (data not shown).
To examine the effects of NAC on T-cell-mediated immunity, we then used a culture system devoid of APCs. To this end, purified T cells were stimulated with anti-CD3- and anti-CD28-coated beads in the presence of different concentrations of NAC. These experiments showed that low concentrations of NAC substantially increase the proliferation of activated T cells in the absence of APCs, and that high concentrations of NAC only have a marginal suppressive effect on T-cell proliferation (Fig. 5A). In this system, NAC had very similar effects on production of IFN-γ, IL-10, and IL-17. Purified naive CD4+CD45RA+ T cells also showed an increase in response to anti-CD3- and anti-CD28-coated beads when moderate levels of NAC was added to the culture, whereas no suppressive effect was observed at high doses (Fig. 5B). Thus, the activating effect of low doses of NAC appears to directly influence T cells whereas the inhibitory effect mainly affects APCs.
High doses of NAC render T cells more responsive in a secondary MLR
To investigate the fate of T cells that have been exposed to NAC during activation in greater detail, we studied how alloantigen-stimulated PBMCs cultured with different concentrations of NAC respond to a secondary stimulation with the same antigen in a culture where NAC is no longer present. Surprisingly, we found that PBMCs cultured in the presence of high doses of NAC (25 mM) during the primary stimulation (6 days) and then extensively washed, responded vigorously when stimulated again in the absence of NAC (Fig. 5C). In contrast, cells that had been stimulated under the influence of low concentrations of NAC exhibited similar proliferative responses to that of control MLRs when re-stimulated. Thus, this indicates that high doses of NAC render T cells more responsive when exposed again to the same antigen when NAC is no longer present.
Since NAC is rapidly metabolized 16, this observation might reflect the situation when NAC is consumed and, as it appears, has rendered T cells more effective. When performing a time-course titration on the proliferative response in MLRs after addition of various concentrations of NAC, we noted that the proliferation peaked at different time points for different concentrations of NAC. When proliferation was analyzed on day 5 (pulsed on day 4 and harvested on day 5), we found that low doses appeared to have a greater activating effect than on day 9, whereas the proliferation on day 9 peaked at higher doses (Fig. 5D). For example, addition of 6.2 mM NAC had a suppressive effect on day 5, whereas this dose potentiated proliferation compared to control MLRs on days 9, indicating that T cells exposed to NAC still retain their ability to proliferate when NAC is consumed or metabolized.
High doses of NAC suppress NF-κB activation in T cells
One mechanism that has been proposed for the immunosuppressive effect of NAC is inhibition of NF-κB activation 8, 11, possibly by preventing the formation of reactive oxygen species. In line with previous studies, we found that high concentrations of NAC reduced the PHA-induced activation of NF-κB in T cells, as measured by intracellular staining of phosphorylated NF-κB p65 peptide by flow cytometry (Fig. 6A). In our system, low to moderate levels of NAC had no detectable effect compared to stimulated control T cells. This flow cytometry-based method was not sensitive enough to study NF-κB activation in alloantigen-stimulated T cells in our hands.
The activating effect of NAC is not due to increased intracellular levels of GSH
Very little is known about how NAC mediates its immune activating effects. NAC has been reported to increase the intracellular levels of GSH in cells, which is important for T-cell proliferation 17. Next, we examined the effect of NAC on intracellular levels of GSH in our culture system and found that doses of 12.5 mM or higher were required to be able to detect increased intracellular GSH content compared to control cultures (Fig. 6B). Furthermore, when we blocked the de novo GSH synthesis with L-(S,R)-buthionine sulfoximine (BSO), we noted that the activating effect of NAC on T-cell proliferation was retained (Fig. 6C). Thus, this indicates that the immune activating effect of NAC is not due to an increased pool of GSH.
NAC reduces T-cell apoptosis
When examining the effect of NAC on apoptotic markers in activated T cells, we found that NAC induced a dose-dependent decrease in the expression of annexin V (Fig. 6D). This finding indicates that NAC can save activated T cells from apoptosis, which may partly explain the immunopotentiating effect of NAC.
Reactive oxygen species (ROS) are generated as a by-product of normal metabolism in cells during the conversion of oxygen to water. ROS are important for eradication of microorganisms by phagocytes, but they may also cause oxidative stress, leading to damage of cellular compartments. Oxidative stress has been reported to stimulate inflammatory responses by activating NF-κB, leading to expression of various inflammatory cytokines 18. These observations have led to the general assumption that substances with antioxidant capacity, such as NAC, may also have anti-inflammatory effects. Indeed, several studies have shown that high concentrations of NAC (>10 mM) have a suppressive effect on NF-κB activation in both immune cells and other types of cells in vitro 8, 11, but there have been other studies showing that NAC may have a potentiating effect on T-cell-mediated immunity 19. Here, we have presented data that may provide an explanation for these conflicting results in previously published studies.
In our prospective randomized study, NAC significantly increased the risk of acute GvHD of grades II–IV (Fig. 1). The control group consisted of a higher number of patients with non-malignant diseases and as these patients do not benefit from a graft-versus-cancer effect, all patients with non-malignant diseases were treated with Anti-T-cell globulin (ATG) to reduce GvHD. However, the proportion of patients receiving ATG was the same in the NAC group and in the control group. There was a trend for more patients in the NAC group, 17% as opposed to 10% of the controls, to be treated at home, which is associated with a reduced risk of acute GvHD 20. Other risk factors for acute GvHD, such as type of donor, G-CSF prophylaxis, GvHD prophylaxis, were the same in the two groups. Interestingly, NAC appeared to increase the risk of moderate but not more severe acute GvHD.
We also noted that NAC treatment increased systemic levels of sIL-2R 10 days after ASCT, even if corrected for other factors such as GvHD, conditioning and donor. This biochemical marker has previously been shown to be a good indicator for immune activation, both in allogeneic settings as a marker for GvHD 14, 15, 21, and in autoimmunity as sIL-2R levels are increased in relapsing multiple sclerosis 22. On the other hand, sIL-2R may function as a decoy receptor preventing IL-2 from binding to membrane-bound IL-2R and thereby inhibit further T-cell proliferation 23. Whether or not the secretion or shedding of sIL-2R has an inhibiting effect on the inflammatory response in the patient is difficult to interpret in this setting, but it is nevertheless a biological marker for increased T-cell activation 24. Thus, the fact that NAC-treated patients had higher systemic levels of sIL-2R after ASCT further supports the immune activating effect of this drug. Using in vitro experiments, we tried to explore the reason for these observations.
We first noticed that NAC may have an activating or inhibiting effect on alloantigen-induced proliferation, cytokine production, and expression of activation markers on T cells depending on the concentration of the drug. The effect of different concentrations of NAC on the secretion of IFN-γ and IL-10 in MLRs corresponded to the observed effect on proliferation, but IL-17 showed a different pattern as NAC had a dose-dependent suppressive effect on the production of this cytokine. This can be due to that IL-17 was not produced in all experiments and therefore made the putative activating effect undetectable. It is also possible that Th17 cells are affected differently by NAC than are Th1 cells.
When we investigated the effect of NAC on immunity in greater depth, we found that different kinds of immune cells are affected by NAC in distinct ways. Whereas immunity mediated by T cells is potentiated at moderate concentrations of NAC and only mildly suppressed at high doses, DCs only appeared to be affected by the NAC-mediated suppressive effect. Verhasselt et al. have previously shown that high concentrations of NAC inhibit T-cell activation at DC level 11, but the T-cell activating effect of NAC was not explored in that study. We also found that PBMCs exposed to high concentrations of NAC do not proliferate with a primary stimulation, but respond strongly when re-stimulated in the absence of the drug. Thus, the inhibitory effect of NAC is probably only transient in T cells and the ultimate effect appears to be an increased immune activation. In the clinical study, NAC was given due to toxicity during the neutropenic phase. The high dose of NAC given to these patients (100 μg/kg/day) results in plasma concentrations in excess (>500 μM) 25 of what inhibits PBMC proliferation in MLR (Figs. 2, 3 and 5). As NAC is rapidly metabolized 16, we speculate that the remaining low levels of NAC in tissues may have stimulated donor alloreactive T cells and GvHD.
We tried to identify the mechanism(s) behind the activating effect of NAC. GSH levels in T cells are important for the responses of these cells to mitogenic stimuli 19, 26, 27, and it has been suggested that NAC may improve T-cell function by increasing intracellular GSH content. However, this did not seem to be the case in our system, as intracellular GSH levels only increased when 12.5 mM NAC or higher was added. Furthermore, blocking of de novo synthesis of GSH by BSO retained the activating effect of NAC on T-cell proliferation. Thus, this indicates that the immune-activating effect of NAC is not due to an increased pool of GSH. In line with the results of others 28, 29, we found that NAC reduces apoptosis in T cells, which might at least partially account for the observed increase in alloantigen-induced activation of T cells. However, there are probably other as yet unidentified mechanisms that are of greater importance for this phenomenon.
The documented effect of NAC on immunity and inflammation in vivo is also inconsistent, but the parameters analyzed are generally proinflammatory mediators produced by innate immune cells and non-immunological cell types. One study showed no effect of NAC on serum levels of various proinflammatory cytokines, including TNF-α, IL-1β, IL-6, and IL-8, in sepsis patients receiving NAC intravenously 30, whereas another study showed that administration of NAC in patients with severe sepsis was associated with attenuation of NF-κB activation in mononuclear leukocytes and reduced circulating concentrations of IL-8, but had no effect on IL-6 and sICAM-1 31. Animal models have suggested that TNF-α, but no other proinflammatory cytokines, are suppressed by administration of NAC 32. Oral administration of NAC to healthy volunteers increased the ex vivo production of chemokines such as MIP-1α, MIP-1β, and RANTES 33. We observed that patients who received NAC had an increased prevalence of acute GvHD and higher levels of systemic sIL-2R compared to control patients. Although this study was not designed to examine the outcome of GvHD, this observation indicates that NAC may increase T-cell-mediated immunity in vivo, at least in immunocompromised patients.
Even though GvHD can cause life-threatening conditions, it has been noted that patients with low-grade acute GvHD or chronic GvHD have a better long-term relapse-free survival, which is probably due to the graft-versus-leukemia (GvL) effects associated with GvHD 34–36. Thus, patients with high-risk malignancies may benefit from developing mild GvHD. Treatments based on adoptive transfer of donor-derived immunocompetent cells are efficient in promoting immunity in ASCT patients 37, but it would be of great clinical interest if we could develop cheaper and more readily available treatment modalities. Further studies are needed to establish if the immunomodulatory effect of NAC observed could be clinically exploited. A possible approach would be to use it as a therapeutic agent to promote T-cell-mediated immunity in immunocompromised ASCT patients with threatening relapse of the underlying malignant disease.
Materials and methods
The clinical study was designed to evaluate the efficacy of NAC in improving liver toxicity and preventing SOS 7. An open-labeled prospective randomized study was performed, treating patients with early liver dysfunction with NAC or with no treatment after ASCT. A detailed description of this study has been published previously 7. Between October 1999 and February 2007, 160 patients (31%) with early liver toxicity were randomized to be treated with NAC (Astra-Zeneca, Södertälje, Sweden) or not (controls). NAC, 100 mg/kg body weight, was given as a continuous 6 h infusion. The treatment was continued until elevated bilirubin and/or transaminases normalized, or when the patient was discharged from hospital. NAC was given for a median of 15 (range 4–89) days. The study was approved by the Ethics Committee of Huddinge University Hospital and was performed according to the Declaration of Helsinki. All patients included in the study gave informed consent.
As prophylaxis against GvHD, cyclosporine was combined with a short course of methotrexate in most patients 38, 39. ATG was given prior to transplant in patients with non-malignant disorders, or with an unrelated donor, regardless of disease 35, 36. Acute GvHD of grade I was treated with prednisolone, 1 mg/kg twice daily 38. Acute GVHD was scored according to established criteria by Glucksberg et al 40.
Conditioning and supportive care
Myeloablative conditioning consisted of cyclophosphamide combined with fractionated total body irradiation (TBI) or busulfan 41. Reduced intensity conditioning (RIC) consisted of fludarabine combined with busulfan or 2 Gy total body irradiation 41, 42.
Most patients were treated in reversed isolation. Some were treated at home 20. Post-transplant G-CSF was given to some patients, but was discontinued when it was found that G-CSF increased the risk of acute GVHD 43. All patients were treated with ursodiol (12 mg/kg per day) to prevent liver toxicity 44. Details of supportive care have already been published 38, 41.
In 141 of the patients, we analyzed levels of sIL-2R in serum samples. Blood samples were drawn from the patients weekly during their stay at the hospital. Whenever possible, samples were drawn on day 3 (n=55 and 70 in control and NAC-treated group, respectively), day 10 (n=52 and 68), day 17 (n=49 and 64), and day 24 (n=30 and 30) after ASCT. Blood samples were allowed to clot and centrifuged within 1 h and the serum was removed. Analysis was performed on fresh serum. Levels of sIL-2R were analyzed using an automated chemo-luminescence immunoassay (IMMULITE, Siemens, Eschborn, Germany) with a sensitivity of 5 U/mL, intra-assay coefficient of variation (CV) of 3.5% and an inter-assay coefficient of variation of 6.5%.
Preparation of PBMCs and DCs
PBMCs were prepared from buffy coats. DCs were generated from CD14 MACS-bead isolated monocytes (Miltenyi Biotech, Bergisch Gladbach, Germany) by culturing the cells in complete RPMI medium supplemented with 800 U/mL GM-CSF and 1000 U/mL IL-4 (R&D Systems, Minneapolis, MN) for 6 days.
PBMCs were plated in triplicate at 2×105 per well in 96-well plates in RPMI with 5% AB-serum and were stimulated with an irradiated pool of allogeneic PBMCs from five donors (2×105 per well) in the absence or presence of 0.001–50 mM NAC in twofold dilutions. The NAC was purchased from Astra-Zeneca (Södertälje, Sweden) as a premade solution (200 mg/mL NAC (1227 mM) and 0.5 mg/mL disodium edentate in water, set to pH 7 with NaOH). This solution was diluted at least 50 times in culture medium for the experiments. After 6 days, PBMCs were pulsed with 1 μCi 3H-thymidine (Amersham Biosciences, Little Chalfont, UK) per well for 16 h. To study the proliferative response of purified T cells, CD3+ cells or CD4+CD45RA+ were first separated from normal donor PBMCs using the Pan T-cell isolation kit or the naive CD4+ T-Cell Isolation Kit II, respectively (Miltenyi Biotech) according to the directions of the manufacturer. CD3+ T cells or CD4+CD45RA+ T cells were stimulated with anti-CD3- and anti-CD28-coated beads (Miltenyi Biotech) for 3 days or with mature allogeneic DC for 6 days. 3H-thymidine incorporation was measured with a MicroBeta TriLux (Perkin-Elmer Weiterstadt, Germany).
Cytokine levels in supernatants from PBMCs stimulated with irradiated allogeneic PBMCs for 5 days, or CD3+ or CD4+CD45RA+ T cells stimulated with allogeneic DC or anti-CD3- and anti-CD28-coated beads, were analyzed by ELISA with paired capture and biotinylated detection antibodies for IFN-γ, IL-10, IL-17, or IL-6 according to the manufacturer's instructions (R&D Systems) or with IL-2 ELISA MAX Deluxe Sets from BioLegend (San Diego, CA).
For all flow-cytometric analyses, a FACS Sort flow cytometer (BD Biosciences, San José, CA) was used to acquire data and Summit version 4.1 software (Dako, Fort Collins, CO) was used to analyze the results.
Surface staining: PBMCs were stimulated with irradiated allogeneic PBMCs for 5 days in the absence or presence of various concentrations of NAC. The cells were then harvested and stained with the antibodies indicated. Immature DC were matured with TNF-α 10 ng/mL and PGE2 1 μg/mL for 48 h in the presence or absence of NAC before analysis, which is a standard maturation cocktail which has been used for preparation of clinical grade DCs 45, 46. All anti-human monoclonal antibodies were purchased from BD Biosciences.
Intracellular Phosflow NFκB staining: PBMCs were preincubated for 2 h with the concentrations of NAC indicated before they were stimulated with 10 μg/mL PHA for 1 h. The cells were then immediately fixed with 4% formaldehyde to maintain their phosphorylation state, permeabilized with Perm buffer III (BD Biosciences), and stained with anti-CD3 and anti-NF-κB p65 antibodies (both from BD Biosciences).
PBMCs or purified CD3+ T cells were stimulated as above and harvested after 5 or 3 days, respectively. The cells were processed and GSH was measured with an ApoGSH glutathione colorimetric detection kit according to the manufacturer's instructions (Biovision, Mountain View, CA). The data are expressed as the relative levels of intracellular GSH in different samples compared to untreated control cell cultures. To block neosynthesis of GSH, 1 mM BSO (Sigma Aldrich, St Louis, MO) was added to the cultures.
The incidence of GvHD was estimated non-parametrically by cumulative incidence curves taking competing events into consideration, and compared by the Gray test. The competing event was death within 100 days after ASCT without GvHD. Differences in the distribution of patients' characteristics were compared with the χ2 test and corrected with the Yates' method. The Mann–Whitney U-test was used to compare continuous variables. The logistic regression method was used to analyze factors associated to high (>median) levels of sIL-2R after ASCT. Factors analyzed were patient and donor age and sex, diagnosis, disease stage, donor, conditioning, NAC treatment, GvHD and other early complications, G-CSF treatment and graft source. Analyses were performed using the CMPRSK package (developed by Gray, June 2001), SPLUS 6.2 software and STATISTICA software.
Wilcoxon matched pairs test (Prism software; GraphPad Software, San Diego, CA) was used to analyze the proliferation, cytokine levels and expression of CD25 on T cells in control MLRs compared to MLRs after addition of indicated concentration of NAC (Figs. 2, 3B–D, 4, 5A and B).
We thank the Staff at Center for Allogeneic Stem Cell Transplantation and Department of Hematology and Department of Pediatrics for compassionate and competent care of our patients. This study was supported by grants from the Swedish Cancer Society, the Children's Cancer Foundation, the Swedish Research Council, the Cancer Society in Stockholm and Tissue and Motion, Karolinska Institutet.
Conflict of interest: The authors declare no financial or commercial conflict of interest.