The effect of anti-CD40 ligand in immune thrombocytopenic purpura


James B. Bussel, 525 E. 68th Street, Room P-695, New York, NY 10021, USA.


Immune thrombocytopenic purpura (ITP) is mediated by antiplatelet autoantibodies possibly involving CD40:CD40L co-stimulation. Two humanized anti-CD40L monoclonal antibodies, hu5c8 and IDEC-131, were evaluated in 46 patients with chronic refractory ITP. Fifteen patients were treated with 20 mg/kg of hu5c8: four (27%) had complete responses (CR) and two (13%) had partial responses (PR). Three responders to hu5c8 were successfully retreated with IDEC-131, demonstrating its efficacy. Thirty-one subsequent patients were treated with 5–20 mg/kg per infusion of IDEC-131: five (16%) responded (one CR and four PR). The 24% (11/46) overall response rate demonstrates the potential role of CD40-CD40L in the pathogenesis of ITP.

Autoantibodies in patients with immune thrombocytopenic purpura (ITP) bind to circulating platelets and accelerate their destruction, as well as interfering with megakaryocytopoiesis (Karpatkin, 1997; Chang et al, 2003). The production of anti-platelet antibodies is known to be driven and regulated by cellular mechanisms, thought to involve CD4+ T-cells (Kuwana et al, 2003). In addition, the CD4+ T cells themselves may be primed by these autoreactive B cells.

CD40 is expressed constitutively on antigen-presenting cells (APCs), including B cells. CD40L (CD154, Gp39) is a member of the tumour necrosis factor family of transmembrane glycoproteins and is expressed on activated CD4+ T-cells, platelets, and endothelial cells. The CD40-CD40L interaction is essential for normal T-B cell interactions, including T-cell priming, immunoglobulin class-switching, and the T-cell dependent humoral immune response (Jiang & Chess, 2004).

The primary mechanism by which anti-CD40L is thought to increase the platelet count in ITP is by blocking T-cell based stimulation of autoreactive B-cells (Jiang & Chess, 2004). Alternatively, anti-CD40L may also block expression of CD40L on platelets, thereby preventing autopresentation of platelet glycoprotein antigens to macrophages. Anti-CD40L has been shown to be efficacious in animal models for several autoimmune diseases and in humans with systemic lupus erythematosus (Davis et al, 2001; Kalunian et al, 2002). This report describes the effects of two distinct preparations of anti-CD40L in 46 patients with ITP.


Both Institutional Review Board-approved studies included patients ≥16 years of age diagnosed with ITP for ≥2 months with platelet counts ≤30 × 109/l; all patients were refractory to several conventional ITP therapies.

Both antibodies were used in open-label phase I-II pilot studies between 1997 and 2002. Fifteen patients were infused with hu5c8, (BG9588, AntovaTM; Biogen, Cambridge, MA, USA), a fully humanized IgG1 anti-CD40L monoclonal antibody at 20 mg/kg over 1 h once every 4 weeks for 12 weeks. Responders continued the same treatment schedule for 12 additional weeks whereas non-responders were escalated to treatment every 2 weeks.

Subsequent to relapse, three hu5c8 CRs participated in a compassionate study in which they were treated with IDEC-131 (toralizumab, IDEC, San Diego, CA, USA). The preliminary positive results in these three patients led to the larger multi-centre study using IDEC-131 in 31 patients at doses of 5–20 mg/kg every 2 weeks for a total of 16 weeks. Responders who relapsed were eligible for re-treatment within 36 weeks of their last treatment.

A response was defined as a platelet increment >20 × 109/l from baseline to achieve a platelet count >30 × 109/l sustained for at least 1 month. A PR was considered a platelet count >50 × 109/l whereas a CR was a platelet count >150 × 109/l on two occasions at least 1 week apart. Patients were followed for 24 (hu5c8) or 36 (IDEC-131) weeks following their last infusion.

Safety evaluations of adverse events were based on the National Cancer Institute (NCI) common toxicity criteria ( Analysis of data was descriptive using means, medians, and ranges for platelet increases and for duration of response. Chi-square and Fisher exact analysis were used to determine the relationship of specific variables with response. P values ≤0·05 were considered significant. Incomplete data prevented analysis of splenectomy status and ITP duration in the IDEC study.


Forty-six patients with chronic ITP were enrolled; 15 were treated with hu5c8 and 31 with IDEC-131 (Table I). Age and sex distribution were similar in both studies. Among the patients treated with hu5c8, all but one had previously undergone splenectomy; the median duration of ITP was 12 years. Patients in the hu5c8 study averaged 7·5 previous ITP therapies as compared to only two previous therapies in the IDEC-131 cohort (P < 0·005).

Table I.   Patient characteristics and platelet response to anti-CD40L treatment.
Dose:hu5c8 (Biogen)IDEC-131 (IDEC)
20 mg/kgTotal5 mg/kg10 mg/kg15 mg/kg20 mg/kg
  1. N/A, not applicable.

Patient characteristicsn = 15n = 31n = 6n = 11n = 10n = 4
 Average age (years)42·450·540·857·545·458·3
 Sex [male/female]12/323/84/28/39/12/2
 Number of patients with prior bleeding episodes related to ITP [n (%)]7 (46·7)16 (53·3) 5 (83·3) 1 (9·1) 7 (77·8) 3 (75·0)
 Average number of previous ITP therapies7·522·31·71·92·3
Platelet response
 Response (%)6 of 14 (43%)5 of 31 (16%)
 Response classifications4 CR, 2 PR1 CR, 4 PR
 Median peak platelet response (×109/l)228·595·0
 Response onset <2 months5 of 6 responders5 of 5 responders
 Relapse (platelet count ≤20 × 109/l) prior to final anti-CD40L infusion3 of 6 responders4 of 5 responders
 Re-treatment respondersN/A3 of 4

Platelet counts for all 46 patients were <30 × 109/l within 2 weeks prior to starting anti-CD40L therapy. The median peak platelet count for the six hu5c8 responders was 228 × 109/l and for the IDEC-131 responders was 95 × 109/l (Fig 1). Six of the 14 evaluable hu5c8 treated patients (43%) responded to treatment with a sustained increase in their platelet counts to a level >30 × 109/l above baseline. Five of the 31 IDEC-131 treated patients (16%) responded. Among the six hu5c8 responders, there were four CR and two PR whereas there were only one CR and four PR among the IDEC-131 responders (Table I). The CR rates in the hu5c8 (4/14) and IDEC-131 (1/31) cohorts were significantly different (P < 0·05). Hemorrhagic symptoms disappeared in the responders in both groups.

Figure 1.

 The Y axis represents the platelet count and the X axis represents time in days from the first infusion. Biweekly median platelet counts are depicted for responders and non-responders in the hu5c8 and IDEC-131 studies. Platelet counts for the IDEC-131 responders (n = 5) are denoted by squares and for the non-responders (n = 26) by triangles. Platelet counts for the hu5c8 responders (n = 6) are denoted by diamonds and for the non-responders (n = 8) by X’s. The final days for each of the IDEC-131 and hu5c8 treatments are marked by vertical arrows; note hu5c8 non-responders received their last “extra” infusion 2 weeks later than the line indicates (on day 154). Time to response was <2 months from first infusion in 5 of the 6 hu5c8 responders and all 5 of the IDEC-131 responders. The median platelet count for hu5c8 responders continued to increase for 2 months following the last infusion. Median platelet counts of the IDEC-131 responders began to fall immediately after completion of the last infusion.

Three of the four IDEC-131 treated patients who relapsed responded to retreatment at the same dose and schedule that they had previously received but none were long-term responders once treatment was stopped. In the hu5c8 study, none of the eight non-responders to infusion of 20 mg/kg every 4 weeks for 12 weeks responded to escalated treatment at 2-week intervals.

The dose range from 5 to 20 mg/kg did not appear to relate to response in the IDEC-131-treated patients. The CR was in the 10 mg/kg group and the four PR were in the 5 (one patient), 10 (one patient), and 15 (two patients) mg/kg groups; there were no responders among the four patients treated at the 20 mg/kg dose.

Three of the CRs in the hu5c8 cohort were also treated with IDEC-131 after relapse. All three responded to IDEC-131 almost identically to the way that they had responded to hu5c8 (data not shown). These three patients are not among the 31 patients in the study of IDEC-131. In those patients treated with hu5c8, the primary clinical difference between responders and non-responders was duration of ITP. Five out of six responders had ITP <10 years duration compared to two out of eight non-responders (P < 0·05). Other clinical and demographic variables did not correlate with response in either the hu5c8 or the IDEC-131 cohorts.

Infusions of both anti-CD40L were well tolerated; no premedications were given and no infusion-related symptoms occurred. There were no severe adverse events (AE) attributed to infusion of either anti-CD40L. There were nine AE related to hu5c8 infusion, eight mild and one moderate. One patient died 18 d after her only infusion, from recurrent intracranial hemorrhage that was deemed unrelated to treatment; this patient was not considered evaluable for efficacy or toxicity. In the 31 patients treated with IDEC-131, there were 23 AE; 19 were mild and four were moderate. No thromboembolic phenomena were noted in any of the patients in these two studies.

In the compassionate IDEC-131 study, one patient who was diabetic, overweight and hypertensive had a myocardial infarction when her platelet count increased dramatically to >100 × 109/l. Coronary angiography demonstrated long-standing atherosclerotic changes in her coronary arteries.


The CD40:CD40L interaction has been shown to have a significant role in the production of autoantibodies in animal models and in humans (Durie et al, 1993; Davis et al, 2001; Im et al, 2001; Kalunian et al, 2002). Multiple experiments using animal models demonstrated that temporary blockade can result in suppression of autoimmune and alloimmune T-cell responses (Durie et al, 1993; Im et al, 2001; Xu et al, 2002) for the duration of anti-CD40L treatment. Subsequently, in vitro experiments using glycoprotein (GP)IIb/IIIa-specific T-cell lines generated from ITP patients demonstrated that T-cell-dependent anti-GPIIb/IIIa antibody production can be suppressed by anti-CD40L antibody treatment through induction of GPIIb/IIIa-specific anergic T-cells with potential regulatory function (Kuwana et al, 2003). Thus, it seemed reasonable to hypothesize that repeated doses of anti-CD40L antibody might induce tolerance to platelet autoantigens (Stasi et al, 2007) in addition to simply blocking T-cell stimulation of B cells in patients with ITP.

Prior single-infusion studies in ITP patients at 5 and 10 mg/kg (Kuwana et al, 2004) showed decreases in the number of B-cells producing anti-GPIIb/IIIa antibodies, as measured by enzyme-linked immunosorbent spot (ELISPOT) assay, reduced GPIIb/IIIa-induced T-cell proliferation, and decreased anti-GPIIb/IIIa antibody production in vitro. These single-infusion trials established that a dose ≥5 mg/kg of IDEC-131 demonstrated a biological effect in most patients when assayed to explore helper T-cell function.

The studies described here were undertaken to assess the clinical effects of two different humanized IgG1 monoclonal CD40L antibodies. Both antibodies bind to the trimer of CD40L on T-cells with high binding affinity and specificity; however, it is unknown whether the binding epitopes and clearance rates are comparable.

The platelet response (once a threshold of 5–10 mg/kg was exceeded) to both antibodies was similar and did not appear to be dependent on the initial dose, supporting the hypothesis that the response to anti-CD40L was related to the pathophysiology of ITP in those patients. Three patients were treated with both anti-CD40L antibodies and had virtually superimposable responses. The response rate in both groups was significantly divergent only when CRs (4/14 vs. 1/31) were considered (P < 0·05). Also, hu5c8 responders continued to experience increases in platelet counts after the last infusion was completed. These different responses could possibly reflect a difference in the initial patient populations, such as T cell-mediated disease in the more refractory hu5c8 patients.

No severe AEs were reported. Infusions of anti-CD40L have been linked to an increased occurrence of thromboembolic events, which may depend upon binding of anti-CD40L to platelets with complement deposition on the platelet surface (Kawai et al, 2000). The low platelet counts of ITP patients may protect against development of thrombosis due to these agents.

In summary, these data demonstrate a platelet response in 16–43% of patients. These estimates help to quantitate the likelihood that other modalities developed to target CD4+ T-cells may be effective in patients with ITP. Furthermore, it may prove useful to pursue other anti-T cell strategies in the future, especially for those patients with ITP refractory to conventional treatments.