Myelodysplastic syndrome (MDS) comprises a group of clonal bone marrow disorders characterized by ineffective hematopoiesis and increased predisposition to acute myeloid leukemia. The causes of MDS remain poorly defined, but several studies have reported the NKT cell compartment of patients with MDS is deficient in number and functionally defective. In support of a central role for NKT cells, a pilot clinical study reported that lenalidomide (an approved treatment for MDS) increased NKT cell numbers in patients with MDS, and several in vitro studies showed lenalidomide specifically promoted NKT cell proliferation and cytokine production. We tested this in a much larger study and confirm a moderate in vitro augmentation of some NKT cell functions by lenalidomide, but find no impact on the NKT cell compartment of patients treated with lenalidomide, despite a consistently positive clinical response. We further show that the frequency and cytokine production of NKT cells is normal in patients with MDS before treatment and remains stable throughout 10 months of lenalidomide therapy. Collectively, our data challenge the concept that NKT cell defects contribute to the development of MDS, and show that a clinical response to lenalidomide is not dependent on modulation of NKT cell frequency or function.
NKT cells are a subset of T lymphocytes that express a CD1d-restricted, semi-invariant T-cell receptor.1, 2 The NKT cell T-cell receptor recognizes glycolipids rather than peptide antigens and their unique ability to rapidly produce Th1, Th2 and Th17 cytokines enables NKT cells to participate in a wide range of immune responses.3, 4 The important regulatory role of NKT cells is evident in the association between NKT cell defects and diseases where immune dysregulation is a factor, such as autoimmune diseases, allergies and various malignancies.5, 6, 7, 8, 9 One prominent example is the association between myelodysplastic syndrome (MDS) and NKT cell deficiency.10, 11, 12
MDS is a collective term for a heterogenous group of clonal stem cell disorders characterized by ineffective production of mature blood cells, blood cytopenias and a predisposition to acute myeloid leukemia.13, 14 Chromosomal abnormalities are detected in approximately 50% of de novo MDS cases and of these, interstitial deletions of chromosome 5q31.1 (del(5q)) are the most common.15 Conventional therapeutic management of MDS has usually been restricted to supportive measures such as red blood cell or platelet transfusion and recombinant growth factors; however, recent clinical trials have showed that the thalidomide analogue lenalidomide (Revlimid; Celgene, Summit, NJ, USA) can induce substantial clinical improvements.16, 17 These trials showed that lenalidomide significantly reduced the transfusion requirements of patients with MDS with a normal karyotype or with del(5q).13, 18 Subsequent trials have shown that lenalidomide also has clinical activity in other patient groups with MDS, including low or intermediate-1 risk patients without del(5q) and intermediate-2 or high-risk patients with del(5q).17, 19, 20
Lenalidomide's mode of action in MDS is not well understood, but its proposed biological effects include inhibition of angiogenesis, promotion of erythroid hemoglobin production and acceleration of apoptotic death of abnormal hematopoietic progenitors,16, 17, 21 and it is also reported to have immunoregulatory activity through modulation of natural killer cell, monocyte and dendritic cell function, inhibition of FOXP3+ T regulatory cells and costimulation of T cells leading to increased proliferation and interferon-γ (IFN-γ) production.16, 22, 23, 24 Interestingly, many of these immunological activities have been attributed in other settings to the prodigious cytokine production of NKT cells.2, 3, 4, 25
A causative link has not been established between NKT cell defects and the development of MDS, but disordered NKT cell function has been reported for patients with MDS,10, 11, 12 and is clearly implicated in the control or progression of a wide range of tumors.3, 8, 26, 27, 28 This led initially to speculation that NKT cell dysfunction may be a contributing factor in the development of MDS,10, 11, 29 and more recently, that lenalidomide's efficacy as a therapy for MDS may result from a direct effect on NKT cells.29, 30
Earlier studies of lenalidomide's effect on NKT cells have been limited to in vitro functional analyses,29, 30, 31 or NKT enumeration in a limited number of patients with MDS.29 In this study, we have conducted a detailed longitudinal assessment of the T-cell and NKT cell compartments of patients with MDS before treatment, and during 10 months of lenalidomide therapy. Our objectives were to determine whether the NKT cell compartment in patients with MDS is deficient or defective for cytokine production, if NKT cell frequency or cytokine production is altered with lenalidomide therapy, and if the characteristics of the NKT cell compartment correlate with clinical responses.
Patients and study design
The data in this study from patients with MDS are derived from a broader and ongoing clinical trial designed to evaluate the safety of lenalidomide in untreated patients with MDS. This study examined the first 13 patients who entered the trial and there was no bias in patient selection. The trial was approved by the Health Sciences Ethics Committee at The University of Melbourne (No. 050409) and the Peter MacCallum Cancer Centre Human Research Ethics Committee, and registered at ClinicalTrials.gov identifier: NCT00434239. All patients provided written informed consent. Patients received 10 mg lenalidomide (Revlimid; Celgene) p.o. daily for days 1–21 of each 28-day cycle until disease progression or intolerance. Patients received recombinant human stem cell factor (Ancestim; Amgen, Thousand Oaks, CA, USA) 10–20 μg/kg subcutaneously for the first 7 days of cycle three only to examine the effect of recombinant human stem cell factor on the promotion of myelopoiesis (except one who had already achieved a complete remission by the completion of cycle two). Criteria for participation were symptomatic anemia, defined as Hb<100 g/l or transfusion-dependent anemia, Eastern Conference Oncology Group performance status of 2 at study entry, absolute neutrophil count 0.5 × 109 per l (without granulocyte colony-stimulating factor support) and a platelet count 25 × 109 per l.
Blood and bone marrow from this study was analyzed at enrollment and at 8 weeks, 10 weeks, 4 months and 10 months after initiation of lenalidomide. Timing of sample collections were chosen to minimize any potential impact of recombinant human stem cell factor on NKT cell number and function, while still allowing full examination of lenalidomide effect. Clinical responses were assessed by monthly full blood examination and bone marrow aspirate and trephine examination and assessed in accordance with published criteria.32
NKT cell isolation and assessment
Peripheral blood mononuclear cells (PBMCs) from anticoagulated blood of enrolled patients or from healthy donors (Australian Red Cross Blood Bank Service, Southbank, Melbourne, Australia) were isolated by gradient centrifugation using Histopaque (density 1.077 g/ml; Sigma-Aldrich, St Louis, MO, USA). Cells were cryopreserved initially at −80 °C (in 10% dimethyl sulfoxide, 90% fetal bovine serum) before transfer to liquid nitrogen storage. Viability of thawed cells was typically >90%.
Antibodies and flow cytometry
All fluorochrome-labeled antibodies used for flow cytometry (FITC-conjugated anti-IFN-γ (4S.B3) and anti-IgG1, PE-Cy7-conjugated anti-CD3 (SK7), APC-H7-conjugated anti-CD8 (SK1), AlexaFluor 700-conjugated anti-TNF (MAb11), Pacific Blue-conjugated anti-CD4 (RPA-T4) and APC-conjugated anti-IL4 (MP4-25D2) and anti-IgG1) were purchased from BD Biosciences (San Diego, CA, USA). PE-conjugated αGalCer-loaded CD1d tetramer is produced in-house by K Kyparissoudis from a construct provided by Professor M Kronenberg. Flow cytometric data were acquired using an LSRII (BD Biosciences) or sorted on a FACSAria (BD Biosciences) and analyzed with FlowJo (Treestar Inc., Ashland, OR, USA). All events were initially acquired, with nonspecific staining from autofluorescent cells, doublets or nonviable cells excluded from analysis on the basis of forward/side scatter characteristics and staining by 7-aminoactinomycinD (Invitrogen Life Technologies, Carlsbad, CA, USA) and vehicle-loaded CD1d tetramer.33
Cells were cultured in RPMI 1640 medium (Invitrogen Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (JRH), 100 U/ml penicillin (Invitrogen Life Technologies), 100 μg/ml streptomycin (Invitrogen Life Technologies), 2 mM Glutamax (Invitrogen Life Technologies), 1 mM sodium pyruvate (Invitrogen Life Technologies), 15 mM HEPES (Invitrogen Life Technologies), 0.1 mM nonessential amino acids (Invitrogen Life Technologies) and 50 μM 2-mercaptoethanol (Sigma-Aldrich) at 37 °C and 5% CO2. Additional cytokines were not added to these cultures.
For in vitro stimulation experiments to examine cytokine production after αGalCer stimulation, 2 million PBMCs were cultured in 12-well plates in 2 ml cell culture medium containing 100 ng/ml αGalCer (Sapphire Biosciences, Redfern, NSW, Australia) and 1 μM lenalidomide (Celgene), or an equivalent volume of the dimethyl sulfoxide vehicle alone (Sigma-Aldrich). The final concentration of dimethyl sulfoxide (0.026%) was significantly lower than concentrations we and others have used previously in assays of human NKT cell activity.34 After 7 days, cells were restimulated with 10 ng/ml phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich) and 1 μg/ml ionomycin (Sigma-Aldrich) for 4 h. For cytokine analysis after short-term stimulation with PMA and ionomycin alone, PBMCs were cultured with 10 ng/ml PMA (Sigma-Aldrich) and 1 μg/ml ionomycin (Sigma-Aldrich) for 4 h. Monensin (2 μM) (Golgistop; BD Biosciences) was added to cultures for the last 3 h of culture and cells were prepared for flow cytometric analysis of intracellular IFN-γ, tumor necrosis factor (TNF) and interleukin (IL)-4 using the reagents (and following the guidelines) provided as part of the Cytofix/Cytoperm staining kit (BD Biosciences).
Statistical analyses were performed with GraphPad Prism V5.0 software (GraphPad Software Inc., San Diego, CA, USA). Statistical significance was determined using the Mann–Whitney, Wilcoxon, Kruskal–Wallis and Friedman tests as appropriate.
Patient demographics and clinical response to lenalidomide therapy
A comprehensive study of PBMCs and bone marrow from 13 untreated patients with MDS was performed. Longitudinal data were available from 12 patients, as 1 withdrew from the study after the baseline analysis and before initiation of lenalidomide therapy. The patient group comprised 6 women and 7 men, with a median age of 61 years (Table 1). Patient characteristics including MDS subtype by World Heath Organization classification,35 international prognostic score,36 baseline bone marrow blasts count, NKT cell frequency at baseline and 10 months time points, and best treatment response are presented in Table 1. A detailed analysis &1QJ;of the clinical efficacy of treatment of MDS with lenalidomide will be prepared for publication after the broader trial is completed.
NKT cell compartments of patients with MDS before therapy
We compared lymphocytes from peripheral blood of untreated patients with MDS to those from a healthy control group, as per earlier studies.10, 11, 12 Where age-matched comparisons are indicated, data from donors whose age was outside that of patients with MDS (these were all aged over 45 years) has been excluded from analysis.
Patients with MDS had a similar frequency of T cells to that of the control group (mean 70±3 and 72±1%, respectively) (Figure 1a). The distribution of the conventional CD4+ and CD8+ T-cell subsets was also similar, with T cells of MDS subjects comprising 64±5% CD4+ T cells and 28±4% CD8+ T cells, compared to 63±1% CD4+ T cells and 31±1% CD8+ T cells for healthy donors (Figure 1a).
NKT cells were identified on the basis of CD3 expression and reactivity with αGalCer-loaded-CD1d tetramer (Figure 1b). We compared the two groups on the basis of NKT cell frequency among total lymphocytes and among CD3+ T cells, the latter providing a more stringent basis for comparison by ensuring variability among non-T cells did not have a secondary impact on NKT cell frequency. In contrast to earlier studies,10, 11, 12 the mean and range of NKT cell frequencies in the peripheral blood of the patients with MDS was similar to that of the control group among total lymphocytes (control group: mean 0.08%, range 0.001–0.6%; MDS group: mean 0.06%, range 0–0.36; Mann–Whitney P-value=0.24) and CD3+ T cells (control group: mean 0.11%, range 0.001–0.8%; MDS group: mean 0.10%, range 0–0.8; Mann–Whitney P-value=0.25) (Figures 1b and c). There were similar NKT frequencies in the blood and bone marrow from patients with MDS (Figure 1d). The NKT cell compartment from bone marrow has not been well characterized and donor-matched bone marrow was not available for blood from the control group; however, two smaller studies have previously reported similar NKT cell frequencies in these tissues.10, 37
Despite there being no significant difference in NKT cell frequency between patients with MDS and healthy donors, we separately compared age-matched control donors to patients with MDS to take into account reports that the frequency and subset distribution of NKT cells can vary between age groups.38, 39, 40, 41 This was not taken into account by earlier studies of patients with MDS;10, 11, 12 however, we found no evidence of a difference in NKT cell frequency between patients with MDS (age range 45–89 years; median 61 years) and age-matched control subjects (age range 45–71 years; median 59 years) (Figure 1c).
We next examined the relative frequencies of the functionally distinct CD4+ and CD4− NKT cell subsets in MDS subjects. In adults, CD4− NKT cells have a Th1 bias and are normally overrepresented, although proportions can vary substantially between individuals.42, 43, 44 Due to the relatively small volume of blood available from patients and the low overall frequency of NKT cells, we examined samples from seven patients with MDS where the number of NKT cell events was sufficiently high to permit accurate gating of CD4+ and CD4− subsets and ensure reliable and statistically significant analysis (Figure 2a). Consistent with the overall NKT cell frequency, the proportions of CD4+ and CD4− NKT subsets were similar for patients with MDS and healthy controls (Figure 2b). Interestingly, the younger cohort of control subjects (age range 18–44 years; median 35 years) had a similar NKT cell frequency and CD4+ and CD4− subset distribution as the older control group and the patients with MDS (Figure 2b).
Longitudinal analysis of the NKT cell compartment during lenalidomide therapy
The NKT cell compartment of patients with MDS was normal before treatment, but previous studies have suggested that lenalidomide could increase NKT cell numbers and cytokine production by NKT cells regardless of their initial status.29, 31 In our study, 11 of 12 patients with MDS maintained a consistent T-cell frequency throughout the treatment course (Figures 3a and b). The one exception was subject 4, where a rapid decline of CD3+ cells coincided with progression to acute myeloid leukemia (Figure 3a). Analysis of NKT cells from peripheral blood showed the treatment course also had no significant effect on NKT cell frequency in 11 of 12 MDS subjects (Figures 3c and d). The exception was subject 12, where NKT cell frequency increased approximately twofold (Figures 3c and d).
Earlier studies also reported significant changes in NKT cell cytokine production in response to lenalidomide in vitro.29 Although we observed no evidence of a change in NKT cell frequency, altered cytokine profiles (and therefore function) could result from changes in the relative distribution of functionally distinct CD4+ and CD4− NKT cell subsets.43, 44 We therefore examined NKT cell subset distribution in patients with MDS receiving lenalidomide therapy, but the frequency of these subsets also remained consistent throughout the 10-month treatment course (Figures 3e and f).
Effect of lenalidomide on NKT cell cytokine production
An earlier study suggested that NKT cells from the blood of patients with MDS have impaired cytokine production in response to glycolipid stimulation,10 although that finding was based on a failure to induce IFN-γ in cultures of PBMCs stimulated with αGalCer, rather than a direct analysis of NKT cells. Lenalidomide has, however, been shown to directly increase cytokine production by NKT cells from normal donors.29 In light of these observations, we examined NKT cells from three MDS subjects where peripheral blood NKT cell frequency was sufficiently high to identify enough NKT cells to directly and accurately measure cytokine production in vitro using intracellular flow cytometry. Our objectives were twofold: first, to determine whether the cytokine response of NKT cells from patients with MDS was abnormal before treatment with lenalidomide; and second, to determine whether treatment with lenalidomide enhanced baseline cytokine production by NKT cells.
We tested the ability of NKT cells to produce cytokines in response to PMA and ionomycin, which normally results in the rapid production of IFN-γ and TNF.43, 45 We found that the NKT cell cytokine response of patients with MDS before treatment with lenalidomide was equally strong as that of NKT cells from peripheral blood of healthy subjects, with similar proportions of NKT cells from both groups positive for IFN-γ (mean, healthy 44% IFN-γ+; MDS, 54% IFN-γ+; Mann–Whitney P-value=0.36), TNF (mean, healthy 45% TNF+; MDS, 63% TNF+; Mann–Whitney P-value=0.23) and IL-4 (mean, healthy 8% IL-4+; MDS, 3% IL-4+; Mann–Whitney P-value=0.28) (Figures 4a and b). Using the same conditions, we monitored the cytokine response from these MDS subjects at five time points throughout 10 months of treatment with lenalidomide and found no evidence of a change in cytokine production over the treatment time course, suggesting that prolonged in vivo exposure to lenalidomide did not alter cytokine production by NKT cells (Figure 4c). In the absence of stimulation, there were no cytokines detected by intracellular staining of NKT cells assayed ex vivo, regardless of whether the assay was conducted before or after the initiation of treatment with lenalidomide (Figure 4a), showing that NKT cells from MDS subjects were not constitutively activated, either at baseline or in response to treatment with lenalidomide.
Our findings seemingly conflict with a report that NKT cells activated in the presence of lenalidomide had a greater ability to secrete IFN-γ.29 However, an important caveat to that study was that it was based only on in vitro analysis of NKT cells from healthy donors after a short exposure to lenalidomide in culture. Crucially, the study did not analyze NKT cells from patients with MDS receiving lenalidomide therapy, as this study has done.
To further explore the poor correlation between the reportedly augmented cytokine response of lenalidomide-treated NKT cells from healthy donors,29 and our own data showing unchanged NKT cell function in patients with MDS treated with lenalidomide, we separately tested PBMCs from 20 healthy control donors exposed to αGalCer in vitro for 1 week in the presence of lenalidomide or vehicle. At the end of this time, the NKT cells were restimulated with PMA and ionomycin to optimize cytokine production before analysis. This approach ensured NKT cells were exposed to lenalidomide for an extended period of time in vitro, as they would have been during the extended treatment course of lenalidomide in vivo.
Our in vitro results were more closely aligned with those of the earlier study, with lenalidomide significantly (albeit moderately) increasing the proportion of T cells and NKT cells that produced cytokines29 (Figures 5a–f). We observed a significant increase in the proportion of lenalidomide-treated NKT cells producing IFN-γ (84 versus 65%, P-value<0.0001), TNF (88 versus 73%, P-value<0.0001) and IL-4 (16 versus 8%, P-value<0.001) (Figures 5d–f), and a similar effect for conventional T cells (Figures 5a–c), with cultures treated with lenalidomide showing increases in the proportion of T cells producing IFN-γ (19 versus 15%, P-value<0.01), TNF (36 versus 26%, P-value<0.005) and IL-4 (0.7 versus 0.5%, P-value<0.01). A moderate increase in the proportion of NKT cells in lenalidomide-treated cultures was observed for some samples, although there was substantial variability between individual cultures (data not shown), despite the viability (measured by 7-aminoactinomycinD exclusion) of recovered cells typically being 70–80%. These in vitro observations suggest that lenalidomide can induce moderately increased cytokine production by NKT cells from healthy donors, providing that the NKT cells are co-stimulated with αGalCer. It is important to note, however, that this response appeared to be blunted in the NKT cells from patients with MDS both at baseline and after the initiation of lenalidomide therapy.
To our knowledge, we report the first extended longitudinal study of NKT cells in patients with MDS throughout treatment with lenalidomide. A central finding is that the frequency of NKT cells in MDS subjects before treatment is similar to that of healthy, age-matched donors. This observation stands in contrast to earlier studies reporting a significant NKT cell deficiency among patients with MDS,10, 11, 12 and probably reflects the stringency of identification criteria for NKT cells used by each study. It is now widely accepted that the optimal technique for identifying the primary (that is, type 1, invariant) NKT cells is by positive flow cytometric examination with either αGC-loaded CD1d tetramer or the 6B11 mAb, which provide similar specificity.1, 33, 46 The three earlier studies linking MDS to NKT cell deficiency used less stringent approaches, including a definition of NKT cells as Vα24+Vβ11+CD4−CD8−11 (excluding legitimate CD4+ NKT cells, which comprised ∼40% of NKT cells in our study), or Vα24+CD161+ cells12 (excluding legitimate CD161− NKT cells, while including Vα24+Vβ11−CD161+ non-NKT cells). The third study identified NKT cells as Vα24+Vβ11+,10 which usually provides a reasonably accurate approximation of NKT cell frequency,46, 47 despite not necessarily identifying all NKT cells.48 Therefore, we suggest a more likely explanation for the discrepancy is that the earlier study pooled results from treated (not with lenalidomide) and untreated patients, whereas we separately assessed the NKT cell frequency of patients before and after treatment with lenalidomide.
Another difference was that the earlier studies did not age-match the patient group with MDS to healthy control donors. This was potentially important because the frequency of NKT cells declines naturally with age and MDS usually develops in patients aged over 50 years.38, 39, 40, 41 Somewhat surprisingly, patients with MDS had similar NKT cell frequency (and NKT cell subset distribution) to a healthy age-matched control group, as well as to a larger control group that was not age-matched (as per previous studies). We note in retrospect, however, that the reports of age-related decline in NKT cell frequency were based on a significantly older group of donors than the patients with MDS in this study (∼80+ years compared to ∼61 years),38, 41 or reported only a weak correlation between age and NKT cell frequency that was not evident among patient subgroups with cancer.39
We also report no change in NKT cell frequency in response to treatment with lenalidomide. Lenalidomide augments NKT cell activation in vitro,29, 30, 31 and an in vivo study reported increased NKT cell frequency in three patients with MDS with the del(5q) by cycle 3 of treatment with lenalidomide, although only one maintained the increase through later cycles.29 The findings implied specific expansion of NKT cells within the del(5q) subgroup, but the small group size and variability between patients suggested a more extensive study was warranted. We analyzed patients throughout a 10-month treatment course of lenalidomide and found the frequency of NKT cells remained stable in 11 of 12 subjects, including 4 of 5 with the del(5q). We can assert that there was no correlation between the quality of the clinical outcome and NKT cell frequency at any time in the treatment course for patients with MDS in general, although it remains possible that downstream studies could identify NKT cell characteristics unique to smaller subgroups of MDS.
We also cannot formally exclude the unlikely scenario that treatment with lenalidomide induced significant NKT cell expansion that was masked by equivalent NKT egress from the blood or NKT cell apoptosis. We suggest this is unlikely because of the improbability that the outcome of these events would coincide at each time point we assayed to generate an unchanged overall NKT cell frequency and CD4+ and CD4− subset distribution, and because we observed similar levels of proliferation (measured by carboxyfluorescein diacetate succinimidyl ester fluorescence dilution), and recovered similar numbers of NKT cells, from cultures of NKT cells that were stimulated with the strong NKT cell agonist αGalCer in the presence, or absence, of lenalidomide (data not shown).
We also report no discernable defect in NKT cell function among patients with MDS. An earlier study had shown PBMCs from patients with MDS responded poorly to stimulation with the NKT cell agonist αGalCer, but NKT cells were not directly tested and the authors acknowledged the limitations inherent in this approach.10 We therefore undertook an intracellular flow cytometry-based measurement of cytokine production from individual NKT cells isolated from patients and showed normal production of IFN-γ, TNF and IL-4 to PMA and ionomycin stimulation for each of the three patients tested at baseline and multiple time points throughout treatment with lenalidomide. Although the numbers of patients tested in this assay were small, the results consistently revealed normal cytokine production, strongly suggesting that NKT cells from patients with MDS are functionally competent and that cytokine production was not affected by treatment with lenalidomide.
We do not exclude the possibility that lenalidomide could potentially exert some effects on NKT cells in vivo. There remains strong in vitro evidence that conventional T cells show a Th1 cytokine bias when cultured with thalidomide analogues,49, 50 and along with others, we report similar findings for NKT cells cultured with lenalidomide.29, 30, 31 There is also evidence that lenalidomide can enhance the activation of conventional T cells in vivo,16, 22, 51 but most studies, including our own, suggest lenalidomide's in vitro effects on NKT cells are co-stimulatory, and would therefore require an additional activation signal—such as that provided by αGalCer in in vitro assays. Hence, NKT cell activity may be augmented by lenalidomide if the NKT cells were independently activated through coadministration of αGalCer, or endogenously by self-agonist glycolipids.
In summary, we consistently found the NKT cell compartment of patients with MDS to be normal in frequency, subset distribution and function, both before and after treatment with lenalidomide. Our data suggest a limited role for NKT cells in predisposition to MDS and indicates that a positive clinical response to lenalidomide in patients with MDS does not rely on modulation of the NKT cell pool.
Conflict of interest
The authors declare no conflict of interest.
We gratefully acknowledge the participation of patients and blood donors in the study, and the contributions of Sharyn Meadows and Amanda Marshall (research nurses), Dirk Honemann and Melita Kenealy (clinical fellows), Rhonda Holdsworth (Australian Red Cross Blood Bank) and Kon Kyparissoudis (research assistant). We acknowledge the generous funding of this study by the National Health and Medical Research Council (NHMRC), Project Grant (No. 454363). We also acknowledge the following funding support: SPB is supported by an NHMRC Career Development Award ; DIG is supported by an NHMRC Program Grant (No. 251608, renewed as No. 454569) and an NHMRC Research Fellowship.
About this article
Nature Reviews Immunology (2011)