By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Nichola Cooper MRCP, Department of Haematology, Great Ormond Street Hospital for Children NHS Trust, Great Ormond Street, London WC1N 3JH, UK. E-mail: firstname.lastname@example.org
James B. Bussel MD, P – 695, Department Pediatric Hematology/Oncology, Weill Medical College, Cornell University, New York Presbyterian Hospital, New York, NY 10021, USA. E-mail: email@example.com
Intravenous (IV) anti-D and IV immunoglobulin (IVIG) slow the Fcγ receptor (FcγR)-mediated destruction of antibody-coated platelets in patients with immune thrombocytopenic purpura (ITP). This pilot study explored the mechanism of these immunoglobulin preparations by measuring interleukin-10 (IL-10), monocyte chemoattractant protein-1 (MCP-1), IL-6 and tumour necrosis factor α (TNFα), before and after infusion and by assessing the effect of FcγRIIa and FcγRIIIa polymorphisms on both cytokine and haematologic responses to anti-D. Following IVIG, only IL-10 was increased at 2 h and MCP-1 on day 7 (P < 0·05). In contrast, 2 h after anti-D infusion, plasma levels of all four cytokines were increased (P < 0·01); five of six patients with the highest MCP-1, IL-6 and TNFα levels had chills. Higher IL-10 levels correlated with platelet increases at 24 h and haemoglobin decreases at day 7 (P < 0·025). Patients with the FcγRIIa-131HH genotype had significantly higher MCP-1, IL-6 and TNFα levels. Patients with the FcγRIIIa-158VF genotype had higher platelet increments at day 7 (P < 0·05). Soluble CD16 (sCD16) was increased 2 h after IV anti-D; day 7 levels correlated with day 7 haemoglobin decreases (P < 0·01). In conclusion, the relationship of FcγRIIa and FcγRIIIa polymorphisms with both cytokine levels and platelet increments implicated these receptors in responses to anti-D and supported different mechanisms of FcγR interaction to those seen with IVIG.
Intravenous (IV) anti-D and IV immunoglobulin (IVIG) both substantially increase the platelet count in the majority of Rh+, non-splenectomized patients with immune thrombocytopenic purpura (ITP) (Imbach et al, 1981; Salama et al, 1984; Bussel et al, 1988; Scaradavou et al, 1997). The immediate platelet increase seen after these infusions results primarily from interference with the Fcγ receptors (FcγR)-mediated destruction of antibody-coated platelets (Bussel, 2000). Precise details of the interactions of either treatment with specific FcγR have remained poorly understood.
Despite the obvious similarities of IVIG and IV anti-D in the treatment of ITP, important differences exist. Clinical differences, which have become apparent with recent studies, are that IVIG is generally effective in splenectomized and Rh+ patients, while anti-D generally is not (Bussel et al, 1988; Cardo et al, 1991; Smith & Boughton, 1992; Scaradavou et al, 1997). In contrast, IV anti-D is more effective than IVIG in patients with human immunodeficiency virus (HIV)-related thrombocytopenia (Scaradavou et al, 1997) and may have longer-lasting effects on the platelet count (Newman et al, 2001). Furthermore, individual patients may respond better to one treatment than the other (Bussel et al, 2001). These differences suggest that the mechanisms of effect of these two treatments are distinct.
The central hypothesis of this study is that IVIG and IV anti-D have different interactions with mononuclear phagocyte system (MPS) FcγR. The different FcγR interactions are postulated to underlie the clinical differences of these two treatments. Unfortunately, splenic macrophages are inaccessible to study these receptors and their changes with treatment in human beings. In vivo assessment of the interaction of the FcγR system with IV anti-D and IVIG was therefore pursued by measuring serial cytokine levels and haematological parameters following their infusion. Three cytokines, interleukin (IL)-6, IL-10 and tumour necrosis factor α (TNFα) and one chemokine, monocyte chemoattractant protein-1 (MCP-1) were measured. IL-6 and TNFα were measured because they are pro-inflammatory cytokines that are produced early in the immune response. They are important in regulating the release of other cytokines and serum levels have been shown to increase after IVIG treatment in patients with common variable immunodeficiency (CVI). MCP-1 is also a pro-inflammatory chemokine, which plays an important role in leucocyte activation, recruitment and phagocytosis. In contrast, IL-10 is an anti-inflammatory cytokine, which possesses cytokine regulatory activity.
To explore the role of specific FcγR in these interactions, FcγR polymorphisms were also determined and related to cytokine and haematologic responses. This pilot study focussed on the bi-allelic polymorphisms of FcγRIIa 131: H/R and FcγRIIIa158: V/F. These polymorphisms are known to affect the handling of immunoglobulins and immune complexes, influence the response to infections, and alter susceptibility to immune complex diseases (van de Winkel & Capel, 1993; Bredius et al, 1994; Wu et al, 1997; Koene et al, 1998; Williams et al, 1998).
Patients and methods
Samples were obtained from adults with ITP who were enrolled in three Institutional Review Board approved treatment studies of IV anti-D and IVIG. Consent for the blood sampling was obtained from all patients.
The IV anti-D study included 27 Rh+, HIV− and non-splenectomized, newly diagnosed (1–11 months) adults with ITP. They were a median of 3·5 months (range 1–11 months) from diagnosis and had platelet counts <30 × 109/l at enrolment. They were randomized to receive their initial dose as either 50 μg/kg (n = 14) or 75 μg/kg (n = 13) of IV anti-D (WinRhoSDF; Cangene Corp., Winnipeg, Canada). After 16 patients had been enrolled, prednisone 20 mg was adopted as a standard premedication, along with the continued administration of 650 mg of acetaminophen, to reduce the side-effects associated with IV anti-D. As a result, six of the patients treated with 50 μg/kg and five of the patients with 75 μg/kg of IV anti-D received prednisone premedication. Specimens were obtained pre-infusion and 2 h, 24 h and 7 d after infusion. More than 95% of eligible specimens from the 27 patients were available. This group of patients has been previously reported both to describe the clinical differences between infusions of 50 and 75 μg/kg of IV anti-D (Newman et al, 2001) and to assess their long-term outcome (Cooper et al, 2002). One patient participated in the assessment of long-term outcome but not in the current ‘cytokine’ study or the comparison between 50 and 75 μg/kg.
Sixteen HIV− adults with chronic ITP were included from two separate IVIG studies for comparison. These patients had chronic ITP and platelet counts <30 × 109/l and received 1 g/kg of IVIG infused over 3–4 h on two consecutive days. Patients had samples taken before infusion, 2 h into the infusion, immediately after day 1 infusion, and before infusion on day 2 (at 24 h). A set of five patients had specimens also obtained on days 3 and 7. The IVIG preparations were Endobulin (Baxter Corp., Glendate, CA, USA) and Gamimmune SD and Gamimmune C (Bayer, West Haven, CT, USA).
The patients receiving IVIG had a longer duration of disease than the patients receiving IV anti-D. Background demographics (Table I) were otherwise comparable.
Table I. Background characteristics of patients in each treatment group.
50 μg/kg n = 14
75 μg/kg n = 13
Baxter n = 11
Bayer n = 5
IVIG, intravenous immunoglobulin.
Age (years) [median (range)]
Months from diagnosis [median (range)]
Previous steroid treatment
Previous IVIG treatment
Median platelet count, day 0 (×109/l)
Range platelet count, day 0 (×109/l)
Blood for cytokine assays was collected in sodium heparin, placed immediately in ice water, centrifuged and the resulting plasma stored in aliquots at −80°C within 1 h of phlebotomy. MCP-1, IL-6, IL-10 and TNFα levels were measured in batches using an enzyme-linked immunosorbent assay (ELISA) technique with kits obtained from R&D Systems (Minneapolis, MN, USA) (these three cytokines and the one chemokine, MCP-1, will hereafter be referred to as ‘cytokines’).
The FcγRIIa and FcγRIIIa polymorphisms were evaluated as previously described using a polymerase chain reaction (PCR)-based allele-specific restriction analysis assay in 22 of the patients receiving IV anti-D (Koene et al, 1998). Twenty-five of the 27 patients receiving IV anti-D were serologically assessed for their Rh phenotype. Data from Rochna and Hughes-Jones (1965) was used to estimate the number of D sites on their red cells.
Soluble CD16 (sCD16, FcγRIII) levels were measured using an ELISA technique (Fleit et al, 1992) in 22 patients prior to IV anti-D treatment, in 12 patients 24 h after treatment and in 18 patients 7 d after treatment. The laboratory standard of pooled donors was 100 U. This assay does not distinguish between sFcγRIIIa and sFcγRIIIb.
The mean, standard deviation and range were used to describe the levels of cytokines over time. The signed rank test was used to determine whether the levels of cytokines increased at different times after infusion when compared with baseline. The Wilcoxon's rank sum test was used to compare the difference at each time point between the anti-D treatment groups (50 vs. 75 μg/kg) and to compare the anti-D data to the IVIG data. Bonferroni-adjusted pairwise multiple contrasts were used to determine which measured time point differed from baseline. The Student's t-test was used to compare sCD16 levels at different time points. The Kruskal–Wallis and Mann–Whitney U-tests were used to assess the effect of FcγR polymorphism on cytokine levels at 2 h after IV anti-D and platelet increase and haemoglobin decrease following treatment. A two-tailed P-value <0·05 was considered statistically significant.
The study was powered to follow 28 patients for 1 year from study entry as described (Cooper et al, 2002). Therefore, it was not powered for the secondary end points considered here but was rather designed as a pilot study to explore cytokine levels and FcγR polymorphisms.
The background characteristics of the patients are shown in Table I. Both IVIG groups received 1 g/kg in the first 24 h. Therefore, they were combined for all analyses. None of the IVIG patients developed chill reactions, although headaches were noted in several patients after infusion.
Cytokine levels following IV anti-D and IVIG
The serial cytokine levels following IV anti-D and IVIG treatment are depicted in Fig 1(A–D) and summarized below.
Two hours after infusion of both 50 and 75 μg/kg doses of IV anti-D, the plasma levels of MCP-1, TNFα, IL-10 and IL-6 were substantially increased compared with baseline (P < 0·01). IL-10 and MCP-1 were still increased compared with baseline 24 h after 50 μg/kg dose of IV anti-D and MCP-1 was increased compared with baseline 7 d after 75 μg/kg dose of IV anti-D (P < 0·01). Otherwise, the cytokines were indistinguishable from baseline at 24 h and 7 d. Two hours after IV anti-D infusion, patients who received the 75 μg/kg dose of IV anti-D had a higher level of IL-6 (P = 0·03) and a trend towards a higher level of MCP-1 (P = 0·08) when compared with those patients who received 50 μg/kg of IV anti-D.
Following IVIG treatment, only two significant changes were observed: IL-10 increased 2 h after infusion (P < 0·01) and MCP-1 was increased 7 d after infusion (P < 0·05).
Comparison of cytokine levels following IV anti-D and IVIG
Tumour necrosis factor α, MCP-1 and IL-6 levels were all significantly higher 2 h after both of the IV anti-D doses when compared with levels following IVIG (P < 0·01). There was no significant difference in the IL-10 levels at 2 h between the two treatments and no significant difference between the levels of any of the cytokines at the other two time points.
Correlations among the cytokines following IV anti-D
There was a strong positive correlation among IL-6, MCP-1 and TNFα levels 2 h after infusion of both doses of IV anti-D (P < 0·01), i.e. the patients who had the highest levels of IL-6 also had the highest levels of TNFα and MCP-1. IL-10 levels were distinct; these levels only correlated with IL-6 and MCP-1 levels at 2 h after infusion of the 50 μg/kg dose (P < 0·01).
Cytokine levels following IV anti-D and haematologic response
At 2 h after the 50 μg/kg dose of IV anti-D, IL-10 levels were strongly correlated with the platelet increase 24 h after treatment (P = 0·002). There was also a correlation between IL-10 levels 2 h after the 75 μg/kg dose of IV anti-D and the decrease in the haemoglobin level on day 7 (P = 0·013). There was no association between levels of IL-6, TNFα or MCP-1 at any time point and either platelet increase or haemoglobin fall.
Adverse reactions following IV anti-D
Six patients experienced chill reactions following IV anti-D infusion: three after 50 μg/kg and three after 75 μg/kg. These six reactions occurred in the patients who had six of seven highest TNFα, IL-6 and MCP-1 levels 2 h after infusion (Fig 2A–D). The patients who had high levels of cytokines but did not have an infusion-related reaction had received prednisone premedication. Higher levels of IL-10 were not associated with chill reactions.
Effect of prednisone on cytokine levels and side-effects following IV anti-D
A previous report describing these patients showed that prednisone premedication had significantly reduced both the incidence and severity of reactions following the 75 μg/kg dose (Newman et al, 2001). IL-10 levels were significantly higher in those patients who received prednisone premedication (P = 0·015). There was no difference in the other cytokine levels in patients receiving prednisone compared with those who did not. Only one of the six patients who experienced chills following IV anti-D infusion received prednisone premedication; this patient had the highest TNFα level (965 pg/ml).
Cytokine levels following IVIG, toxicity and haematologic response
In the patients treated with IVIG, there was no correlation between cytokine response and either platelet increase or haemoglobin decrease. No chill or fever reactions were observed following IVIG infusion in this study.
FcγR polymorphisms and IV anti-D: cytokine levels and haematologic responses
Seven (30%) patients had the HH polymorphism, 14 (61%) the HR polymorphism and two (9%) the RR polymorphism (four patients were not assessed). Patients with the HH polymorphism had significantly higher IL-6 (P = 0·001), MCP-1 (P = 0·002) and TNF (P = 0·001) levels 2 h after anti-D (Fig 2A–D). Four of the five patients who experienced chill reactions (and whose FcγRIIa polymorphism was known) had the HH polymorphism. This polymorphism had no relationship to either IL-10 levels or platelet or haemoglobin changes after treatment.
Six patients (26%) had the FF polymorphism, 12 (52%) the VF polymorphism and five (22%) the VV polymorphism (four patients were not assessed). Those patients with the FcγRIIIa VF polymorphism had a greater mean platelet increase on day 7 than those with the FF polymorphism, (P = 0·022). This may have been influenced by the fact that nine of these 12 patients had received the 75 μg/kg dose of IV anti-D, whereas only two of six received 75 μg/kg in the FF group. This polymorphism of FcγRIIIa did not significantly influence cytokine levels or haemoglobin changes.
Soluble CD16 (FcγRIII)
Median sCD16 levels increased significantly from a pre-infusion level of 53 U to a 24 h level of 100 U in the 11 patients measured at both time points (P = 0·008). Higher day 7 sCD16 levels strongly correlated with a greater haemoglobin decrease on day 7 (P = 0·002).
In patients receiving the 50 μg/kg dose of IV anti-D, there was a trend for the patients with the highest estimated number of D sites per red cell to have the highest IL-6 levels 2 h after infusion (P = 0·061). When all patients were combined, the patients with the highest estimated number of D sites per red cell also tended to have the highest MCP-1 levels at 2 h after infusion (P = 0·084). There was no significant effect of ABO blood group on cytokine levels.
The primary mechanism of acute platelet increase following both IVIG and IV anti-D is thought to be the slowing of the destruction of opsonized platelets via effects on the FcγR system. Infusion of monoclonal antibodies to FcγRI (Wallace et al, 1997) and FcγRIII (Bussel et al, 1990) have each resulted in transient but substantial increases in the platelet count in patients with ITP and platelet increases have also been observed following infusion of the Fc piece of IgG (Tovo et al, 1984). These studies support the feasibility of blocking specific Fcγ as a way to increase the platelet count in patients with ITP and also indicate that multiple Fcγ may be involved. The effect of IVIG to slow the clearance of antibody coated red cells (Fehr et al, 1982) and alteration in phagocytic function after IVIG (Kimberly et al, 1984), and the greater efficacy of intact IgG compared with F(ab)′2 fragments of IgG (Tovo et al, 1984) all support the fact that the primary mechanism of the immediate IVIG effect in patients with ITP is by interfering with the FcγR-mediated clearance of opsonized platelets. A recent study in which a mouse model of ITP was created by infusion of anti-mouse platelet antibodies suggested that IVIG slows the destruction of antibody-coated platelets by increasing the expression of the inhibitory FcR, FcγRIIb (Samuelsson et al, 2001). The lack of a monoclonal antibody distinguishing FcγRIIa from FcγRIIb has prevented more definitive study of this hypothesis in human beings. Other proposed FcγR-related mechanisms of IVIG include macrophage interaction with IgG dimers contained in IVIG (Teeling et al, 2001) and increased clearance of autoantibodies by saturating FcγR(n) (Hansen & Balthasar, 2002). However, the ability to generalize effects observed in these mouse models to human beings with ITP remains uncertain. Inflammatory and other cytokine levels have been shown to increase following infusion of IVIG in a variety of diseases such as CVI (hypogammaglobulinaemia) (Aukrust et al, 1994; Farber et al, 1994; Skull & Kemp, 1996; Sewell et al, 1999; Sharief et al, 1999). However, cytokine response to IVIG has not been previously studied in patients with ITP. The findings reported here showing a lack of pro-inflammatory response, especially in comparison with the effects of IV anti-D, are compatible with an anti-inflammatory effect of IVIG, which could be mediated by the up-regulation of FcγRIIb.
The mechanism of action of IV anti-D has not been as extensively examined. Unlike IVIG, it is not effective in patients who are Rh−, demonstrating that the initial response to IV anti-D requires the attachment of (hundreds of) IgG molecules to individual red blood cells (Bussel et al, 1991; Cardo et al, 1991; Smith & Boughton, 1992). The anti-D-coated red cell is thought initially to slow the destruction of opsonized platelets by direct competition for available FcγR binding sites. Subsequent effects may depend on other, immunomodulatory factors such as secretion/release of IL-10 as described here, or secretion of other anti-inflammatory mediators such as IL-1 receptor antagonist (IL-1ra) (Coopamah et al, 2003) by splenic and other macrophages whose FcγRs are ligated by anti-D-coated red cells.
The dramatic increase in the plasma levels of IL-6, TNFα, MCP-1 and IL-10 2 h after IV anti-D in this pilot study (Fig 1A–D) demonstrates the intensity of the interaction of the anti-D coated red blood cells with the MPS especially when compared with the lack of any substantive effect following infusion of IVIG. Two previous studies of IV anti-D also showed increased plasma levels of inflammatory and other cytokines immediately after the infusion of IV anti-D. IL-6, IL-10, TNFα, MCP-1 IL4, IL-1RA, IL-8 and macrophage inflammatory protein-1α levels were increased 3 h after IV anti-D infusion in seven children with ITP (Semple et al, 2002) and IL-6, IL8 and TNFα, 1–20 h after infusion in 18 children with ITP (Malinowska et al, 2001). The two previous studies and the findings reported here overlap as to which cytokine levels are increased (TNFα, IL-6 and MCP-1) and the temporal nature of the increase, i.e. all increase within 1–3 h of the infusion and return to baseline within 7–8 d. The clear association of chill reactions in this study with the highest levels of IL-6, TNFα and MCP-1 is consistent with the known effects of these cytokines and implicates them in the pathophysiology of these chill reactions. The generally higher cytokine levels seen following 75 μg/kg of IV anti-D than following 50 μg/kg are also consistent with the increased frequency and severity of reactions seen following the higher dose of anti-D in these same patients (Newman et al, 2001). The substantially higher levels of the pro-inflammatory cytokines following IV anti-D compared with IVIG described in this report seems to explain why chills occur more frequently following IV anti-D than IVIG in patients with ITP and why there may be important differences in the mechanisms of effect of these superficially similar treatments.
A central finding of this study is that IV anti-D appears to interact directly, but not necessarily exclusively, with FcγRIIa and FcγRIIIa. These interactions were identified through the impact of the responses to IV anti-D infusion by the polymorphisms of these FcR. The HH polymorphism of FcγRIIa was associated with a clear, dramatic and highly significant increase in levels of inflammatory cytokines 2 h after infusion (Fig 2A–C). Patients with the VF polymorphism of FcγRIIIa had greater platelet increases than VV patients. However, by chance, more patients with the VF polymorphism received the 75 μg/kg dose of IV anti-D (Newman et al, 2001). While the impact of the HH polymorphism of FcγRIIa on cytokine levels is compelling, instead of a direct interaction of this FcγR with anti-D-coated red cells, another surface receptor may be linked to the H/R polymorphism of FcγRIIa or the V/F polymorphism of FcγRIIIa, with which the anti-D-coated red cells actually interact.
The association between higher sCD16 levels and increased haemolysis seen here in the ITP patients has also been seen in patients with sickle cell disease in crisis (Lard et al, 1999). It is speculated that this relationship is based on a central role for neutrophils in red cell adhesion to endothelium in sickle crisis. The strong relationship observed in anti-D-induced haemolysis suggests that neutrophils may have more of a role in this setting than had previously been thought. The role of granulocytes in the clinical effect of IV anti-D in patients with ITP was highlighted in a recent study (Coopamah et al, 2003), which described a decrease in phagocytosis, particularly by granulocytes, when peripheral blood leucocytes are incubated with anti-D opsonized red blood cells. The decline in phagocytosis was significantly correlated with the production of IL-1ra, suggesting that beyond its direct FcR blockade effect, anti-D may have an immunomodulatory role affecting leucocytes and particularly granulocytes (Coopamah et al, 2003).
While TNFα, IL-6 and MCP-1 were linked to each other and to the toxicity of IV anti-D, IL-10 levels were not tightly linked to those of the other cytokines and only IL-10 appeared to be linked to the haematologic response. Furthermore, IL-10 levels were not associated with the polymorphisms of FcγRIIa or FcγRIIIa, but were found to be higher in the patients who received a single dose of prednisone premedication prior to IV anti-D. Therefore, it would appear that IL-10 secretion depends upon a different mechanism of interaction of anti-D-coated red cells. As IL-10 was also increased following IVIG and prednisolone, this cytokine, along with IL-1ra, may play an important immunomodulatory effect in the platelet response seen following these therapies.
This pilot study has reported novel information describing the interaction of IV anti-D, and IVIG, and FcγR, but requires additional study to confirm the findings. If individual cytokines were confirmed as being responsible for the efficacy or toxicity of a specific treatment such as IV anti-D or IVIG, then influencing their levels could become a strategic way to optimize the management of ITP.
Supported in part by a clinical research grant from the Cangene Corporation, Winnipeg, Manitoba, Canada and by the ITP Society of the Children's Blood Foundation, New York, NY.