It is estimated that 25% of patients awaiting renal transplantation have significant amounts of preexisting anti-HLA antibodies (1). Reported as percent panel-reactive antibody (%PRA), these result from previous transplantations, pregnancies, and blood transfusions and are identified by testing the reactivity of recipient sera against a panel of well-defined HLA antigens (2,3). Strategies applied to reduce HLA antibodies in potential transplant recipients have had only partial success and include intravenous gamma globulin (4–6), plasmapheresis (7), and combined plasmapheresis and intravenous gamma globulin (8). Without some form of intervention, the typical patient with preformed HLA antibodies waits at least 3 years before receiving a transplant.
Rituximab (RTX), a genetically engineered chimeric murine/human monoclonal antibody, is approved for treatment of relapsed or refractory B-cell non-Hodgkin’s lymphoma (9–12). It binds to the CD20 surface antigen, a hydrophobic transmembrane protein found on both precursor and mature B cells, resulting in elimination of B cells. Possible mechanisms of action include complement-dependent cytotoxicity and antibody-dependent cellular cytotoxicity (13). The principal hypothesis of our studies is that RTX could reduce PRA via B-cell depletion. In the data presented here, we report the safety, pharmacokinetics, and pharmacodynamics of RTX use in patients with end-stage renal failure awaiting renal transplantation.
MATERIALS AND METHODS
Nine subjects were enrolled in a phase I single-center, nonrandomized trial after meeting eligibility criteria. The study was performed under an investigator-initiated IND after approval by the institutional review board. Informed consent was obtained from all subjects. Subjects included in the trial were older than 18 years of age and receiving dialysis while awaiting a renal transplantation for at least 1 year, with persistent PRA >50% for more than 1 year. Patients were excluded from eligibility if they had significant cardiac or pulmonary disease, hepatitis C or human immunodeficiency infection, or malignancy within the last 5 years (except adequately treated basal or squamous cell carcinoma) or if they were receiving any other immunosuppressive drugs. Other than the renal failure, the subjects demonstrated no evidence of inflammatory processes. All subjects were treated with a single dose of RTX (n=3 per group) at 50, 150, or 375 mg/m2. Subject demographics are shown in Table 1.
RTX was administered in the General Clinical Research Center at University Hospital, Indiana University, Indianapolis, IN. Subjects receiving hemodialysis were infused after completion of baseline studies and 1 day after completion of the previous dialysis session. Subjects receiving continuous ambulatory peritoneal dialysis (CAPD) were infused in the morning after completion of the baseline studies. All patients were given acetaminophen (650 mg) and diphenhydramine (25 mg), orally, 30 min before the infusion. RTX was administered intravenously at 50 mg/hr for the first hour. Rates of infusion were increased by increments of 50 mg/hr every 30 min, up to 400 mg/hr, if no side effects occurred. Subjects were followed for 1 year after the infusion. Complete blood count, serum chemistry including electrolytes, pancreatic enzymes, liver function tests, serum IgM and IgG, anti-cytomegalovirus (CMV) IgG, and anti-pneumococcal antibodies, and peripheral blood lymphocyte immunophenotyping were measured during the study. All serum samples were obtained from venous blood samples after centrifugation. Samples were stored at −80° C until testing. During the entire 1-year study, approximately 250 mL of blood was obtained from each subject.
Venous blood samples for lymphocyte immunophenotyping by flow cytometry were collected in EDTA tubes. All subject sera were tested for the presence of HLA antibody by flow cytometry using FlowPRA beads (One Lambda, Canoga Park, CA) as described by Gebel and Bray (3). For each subject, sera from 1 year and from 6 months before RTX infusion and from 0, 3, 6, and 12 months after RTX infusion were tested. The percent of FlowPRA was determined using a FACS IV (Becton-Dickinson Co.) flow cytometer. Results were accumulated into list mode data files and analyzed by online software. Samples were assessed for PRA using undiluted and diluted specimens (1:4 to 1:128).
Human Anti-Chimera Antibody
Human anti-chimera antibody (HACA) was measured by Genentech using a quantitative enzyme-linked immunosorbent assay (ELISA) on sera obtained at weeks 12, 26, and 52. The assay has a lower limit of detection of 5 ng/mL.
Blood samples were obtained before dose administration, immediately after infusion, at 1 hr, and on days 3, 7, 14, 21, 28, 42, 56, and 84. RTX levels were measured by Genentech using ELISA with goat anti-RTX-specific polyclonal serum. Noncompartmental analysis was used to determine the pharmacokinetic disposition and estimation of apparent elimination half-life (t1/2) and area under the concentration curve (AUC). AUC∞ was estimated by trapezoidal approximation and extrapolated to infinite time. Peak concentrations (Cmax) were obtained by direct inspection of the data. AUC and Cmax were divided by the total dose given to provide dose-normalized comparable values for analysis.
Statistical comparisons of results of white cell blood count, lymphocytes, and immunophenotyping before, at 1 month, and at 1 year after infusion were made by repeated measures analysis of variance (RM ANOVA) using SigmaStat 2.0 statistical software (SPSS Inc., Chicago, IL). The comparison of previous 1-year IgG or IgM levels used a paired t test. The pharmacokinetics used analysis of variance to compare AUC of the three groups. Results were expressed as mean ± SD with significance at P less than or equal to 0.05. With the small sample size and multiple comparisons, beta error could have occurred, causing us to miss truly significant differences.
All nine patients tolerated the RTX infusion. No patient withdrew from the study. There were four significant drug-related adverse events. Two patients developed bacterial infections of their peritoneal dialysis catheter site: patient 1 at 3 and 11 weeks, treated with oral and intravenous antibiotics, respectively, with resolution; and patient 6 at 51 weeks with peritonitis that resolved after removal of the peritoneal dialysis catheter. Both patients had a history of gram-positive and gram-negative bacterial peritonitis before entering the study and had infections within 3 months before RTX treatment but were not infected at the time of dosing. Patient 4 developed presumed histoplasmosis at 18 weeks with resolution of symptoms (chest pain and shortness of breath) after an empiric course of itraconazole. Patient 9 developed fever during RTX infusion that resolved after administration of additional oral acetaminophen.
Clinical Laboratory Testing
Hematocrit, hemoglobin, platelet count, glucose, sodium, potassium, amylase, liver function tests, and anti-CMV and anti-pneumococcal antibody titers were not affected by infusion of RTX (data not shown). No significant changes were seen in total white blood cell count (pre-RTX 5.53×103 cells/mm3 ± 0.74 vs. 1 month post-RTX 4.93×103 cells/mm3 ± 0.6 and 1 year post-RTX 5.69×103 cells/mm3 ± 1.87, respectively, P >0.3) or total lymphocyte count (pre-RTX 1,529×103 cells/mm3 ± 610 vs. 1 month post-RTX 1,203×103 cells/mm3 ± 516 and 1 year post-RTX 1,255×103 cells/mm3 ± 424, respectively, P >0.13). IgG levels were unchanged after RTX infusion. IgM levels sustained a gradual decrease (pre-RTX 189 mg/dL vs. 1 year post-RTX 133 mg/dL, P <0.001).
CD19+ cells (pre-RTX 181±137 cells/mm3 vs. day 2 post-RTX 12.4±5.6 cells/mm3, P <0.001) and CD20+ cells (pre-RTX 205±116 vs. day 2 post-RTX 11±12 cells/mm3, P <0.001) were both extensively depleted. Only partial recovery of CD19+ cells (88.9 cells/mm3 ± 88.3) and CD20+ cells was noted at 1 year after RTX infusion (Fig. 1). There was no significant difference between CAPD and hemodialysis patients for CD20 cells at baseline (165.5±129.3 cells/mm3 vs. 225.0±106.9 cells/mm3, P =0.5, CAPD vs. hemodialysis) There was no significant change in CD3+ cells at 1 week, 6 months, or 1 year after infusion (pre-RTX 1,075±509 cells/mm3 vs. 1 week post-RTX 1,044±514 cells/mm3, 6 months post-RTX 889±363 cells/mm3, and 1 year post-RTX 936±394 cells/mm3, P =0.58). The CD4:CD8 ratio remained unchanged after infusion (pre-RTX 1.8±0.7 vs. 1.9±0.7 and 1.7±0.6 at 1 and 12 months, respectively (P =0.4). No significant changes were seen in levels of activated (CD3+ HLADR+) T cells from preinfusion levels (171 cells/mm3 ± 91 cells/mm3) when compared with those at 1 month (127±74) or 1 year (181±123, P =0.23) after infusion.
The pharmacokinetic parameters for RTX for each of the three groups are summarized in Table 2. A plot of the average serum concentrations obtained from the subjects is presented in Figure 2. The AUC was directly proportional to drug dose. When dose-adjusted for each group, the AUC was not significantly different between the three groups (P =0.392) but was significantly higher than that of basiliximab (P <0.001). Half-life between the three RTX groups was not dependent on dose (P =0.375) but was also significantly longer than that reported for basiliximab (14,15). Dose adjusted Cmax was not significantly different among the three groups or when compared with basiliximab. There was no statistical difference between the half-life for CAPD (230.4±96.0) and hemodialysis (321.2±66) (P =0.13) or for dose-adjusted AUC for CAPD-treated (65.01±19.72) or hemodialysis-treated (68.32±21.86) patients (P =0.83).
Human Anti-Chimera Antibody
Two of nine subjects developed HACA. Positive antibodies were detected at 6 months after treatment (5.06 ng/mL) in patient 3 and at 3 months after treatment in patient 8 (12.4 ng/mL), were persistently positive at 6 months (12.9 ng/mL), and converted to negative at 12 months after treatment.
No appreciable changes were seen when comparing the PRA in samples from 1 year or 6 months before RTX treatment to the baseline samples in any subject. In two (22%) of nine patients, there was no appreciable change in PRA percent or titer (patients 7 and 9). Sera from patients 1, 4, 6, and 8 displayed less antibody diversity as well as diminished fluorescence intensity (Fig. 3) using diluted subject sera; undiluted samples seemed unchanged compared with the pre-RTX samples for these subjects. The %PRA in sera from patient 3 decreased by almost half, when assessed on undiluted and diluted (up to 1:8) sera at 3, 6, and 12 months (from 93% to 52–55%). Furthermore, the patient’s 1:16 titer (class II PRA) dropped from 51% to 0%. Patient 5 had a decrease in intensity and a decrease of the PRA from 87% to 51% after 6 months of infusion in undiluted serum. In patient 2, although no appreciable change was seen in %PRA from baseline to 1 year after treatment, a donor-specific flow crossmatch became negative after RTX infusion (Fig. 4), suggesting loss of specificity. A different living donor, with whom the recipient also had a conversion from a positive to negative crossmatch, was used and the patient currently has a functioning allograft with a serum creatinine level of 1.7 mg/dL.
Although RTX has been shown to be safe and effective for the treatment of lymphoma, there are only limited studies on its use in nonmalignant conditions such as rheumatoid arthritis, immune thrombocytopenia, and nephrotic syndrome (16–19). At the time that this study was designed, there were no data on RTX in renal failure patients and only minimal data on its use in nonmalignant conditions; therefore, the study was designed conservatively as a single-dose, dose-escalation study. Although the principal goal is to use RTX to reduce allosensitization, the primary goal of this study was to analyze the safety and pharmacokinetics/pharmacodynamics of TRX for B-cell depletion. Any observed effect on PRA, while desirable, would not be absolutely expected because of trial design.
The study showed that RTX could be safely administered to patients with dialysis-dependent end-stage renal disease. None of the nine subjects developed hypotension, angina, bronchospasm, rash, or life-threatening adverse events. Two of three subjects receiving peritoneal dialysis developed catheter site infections, and one lost her catheter because of peritonitis resistant to antibiotics. These were not felt to be a result of RTX treatment because they were typical complications seen in CAPD patients. As noted, one patient had a suspected case of histoplasmosis. Although it is not possible to entirely exclude the RTX dosing as a contributor to this case, histoplasmosis is controlled by the T-cell arm (20,21), which is not inhibited by RTX (21a). Furthermore, Indianapolis is considered a hyperendemic area for histoplasmosis, in which one study demonstrated that 24% of solid organ transplant recipients had circulating anti-histoplasmosis antibodies at the time of transplantation (22). No viral infections, including CMV infection, were diagnosed nor were increases of CMV titers detected. Fever has been the most common infusion-related reaction, occurring in 43% to 85% of the patients (9–12). Furthermore, most side effects caused by infusion of RTX in cancer patients have been related to volume of tumor cells. Fever was an uncommon side effect (1/9,11%) in this group of dialysis patients. It is possible that the dose of acetaminophen and diphenhydramine was sufficient to abrogate this effect. As well, our study in six of nine subjects used a lower dose than in most lymphoma treatment protocols. Fever occurred in one of the patients receiving the highest dose. Only two other mild reactions were noted in two patients: a headache and an increase in blood pressure. The absence of major side effects and fever in less than 30% of the subjects was also noticed when RTX was used in other nonmalignant states (16,18). This reduced frequency of adverse events could be a result of the absence of a tumor burden and the lower number of target CD20-expressing cells in these dialysis and other nonlymphoma patients versus lymphoma patients.
As expected, RTX was selective in ablating only the B-cell lineage of CD19+, CD20+, and CD19+CD20+ cells. RTX had no effect on the T-cell lineage. The B-cell depletion seemed to be greater than the 70% to 80% decrease in patients with non-Hodgkin’s lymphoma and the rate of recovery was slower (9,10). B cells were fully repopulated by 9 to 12 months after the infusion of RTX in lymphoma patients and in patients with other nonmalignant disease. The mean half-life after the first infusion was longer in dialysis patients and the AUC was greater than that seen with lymphoma patients, perhaps because of the lower number of CD20 cells; but it is unlikely that this explains the slower recovery (11). The method of dialysis, peritoneal versus hemodialysis, did not have an obvious effect on pharmacodynamics (data not shown).
Although there were no significant differences between the groups with regard to the RTX pharmacokinetic parameters, there were numerical differences. With only three members in each group and a relatively large SD and only three patients receiving CAPD, it is not possible to rule out significant differences that might have been apparent in a larger sample size. Group 2, with a longer t1/2 than group 3, had one patient receiving CAPD, whereas all patients in group 3 were receiving hemodialysis. It is therefore not likely that the method of dialysis can be an explanation. There were no differences between patients receiving CAPD versus hemodialysis for AUC, which suggests that RTX does not bind significantly to the dialyzer. The numerically shorter t1/2 in patients receiving CAPD could suggest that there was some washout of drug with dialysis. There are several possible explanations for the differences in pharmacokinetics between structurally related chimeric antibodies, RTX, and basiliximab. First, the target cell and antigen are clearly distinct. Second, the RTX patients were in renal failure and receiving dialysis, whereas the basiliximab-treated patients had just received a renal transplant. Third, the basiliximab-treated patients were receiving other medications, for example, steroids, which could effect metabolism. Which, if any, of these mechanisms explains the differences cannot be established from the data presented here.
Among our nine patients, two developed HACA. For the ELISA assay, exogenous RTX is used as a capture and detection antibody. HACA sampling can only be done once RTX is cleared from the circulation, that is, after about 3 months. It is conceivable that short-lived (<3 months) HACA developed, and the rate was underestimated (23). In previous clinical trials of cancer patients, only 1% of the patients had developed HACA to RTX (9). The rate of HACA formation to a similar chimeric antibody—basiliximab (anti-interleukin 2 receptor)—was only about 1% in a postkidney transplant population that also received immunosuppression (24). In contrast, in patients receiving another chimeric antibody, infliximab, 61% developed HACA, but the percentage was reduced in patients who were given immunosuppression (25). The fact that only 22% developed HACA in our study, despite the lack of additional immunosuppression, is likely a result of the elimination of B cells. It is not clear whether the development of HACA has any effect on efficacy (25).
We speculated that persistent antidonor B cells were closely associated with production of high-titer anti-HLA antibodies, and we reasoned that RTX was a promising drug to eliminate these. Despite the limitations of the study, almost 80% of the subjects sustained some change in their PRA, albeit only minor nonclinically significant changes in four. The three patients with infections (patients 1, 4, and 6) were included in this latter group of minor changes. There are several possible explanations for the incomplete reduction in PRA despite the nearly complete and sustained reduction in B cells. The single-dose infusion may not have completely eliminated residual B cells in solid lymph organs such as the spleen, lymph nodes, and bone marrow. Alternatively, there may have been insufficient time for the circulating anti-HLA antibody to be metabolized. Because the subjects in this study had high-titer as well as broadly reactive (high PRA) anti-HLA antibodies, many half-lives would be required to eliminate all detectable circulating antibody.
Finally, the existence of long-lived plasma cells could explain the incomplete reduction in PRA (26,27). Several surface proteins are down-regulated upon B-cell differentiation into plasma cells, including the major histocompatibility complex class II—CD19, CD20, CD21, and CD22 (27–29). Plasma cells that have undergone down-regulation of CD20 would not be susceptible to RTX. Plasma cells until recently were considered short-lived, with half-lives of days to months. The maintenance of antigen-specific antibody in such a system would require the repopulation of the plasma cell pool from memory B cells. Such physiology would be an ideal target for treatment with an agent such as RTX. The memory B cells are rapidly and completely eliminated by RTX treatment and the plasma cells would be lost from attrition. However, in the murine system, data suggest that plasma cells can last the entire life of the animal (27). If this is true in humans, then the simple elimination of mature B cells would be expected to have little if any effect on the persistence of an alloantibody response. However, such long-lived plasma cells as the source of persistent antibody does not seem to be the case in humans. Bernasconi et al. (30) demonstrated that, after tetanus immunization, specific IgG antibody decreased with a half-life of 40 days, consistent with the presence of a population of plasma cells with a half-life of 40 days. This population of plasma cells was derived by antigen-nonspecific polyclonal activation of antigen-specific memory B cells. If these B cells were eliminated by a drug such as RTX, and no source of alloantigen was present to restimulate newly appearing B cells, then the alloantibody should decrease over time, at rate controlled by the half-life of immunoglobulin and the half-life of plasma cells. This is consistent with the findings in our study, in which reductions in PRA were seen long after the dose of RTX.
The authors thank Paul Brunetta and Mark Benyunes of Genentech for review of the manuscript and for arranging the analyses of RTX blood levels and human anti-chimeric antibody assays.
Carlos A. Vieira, Avinash Agarwal, Christopher M. Bearden, Timothy Taber, Michael A. Kraus, and Mark D. Pescovitz were intimately involved in the clinical aspects of the protocol and collection of clinical data. Carlos A. Vieira, Avinash Agarwal, Christopher M. Bearden, Howard M. Gebel, Anthony L. Roggero, Benita K. Book, and Richard A. Sidner performed and analyzed the laboratory analyses including flow cytometric studies. Naomi A. Fineberg was responsible for statistical analyses. Carlos A. Vieira, Avinash Agarwal, Christopher M. Bearden, Benita K. Book, Richard A. Sidner, and Mark D. Pescovitz were responsible for writing the manuscript. Mark D. Pescovitz proposed the original hypothesis.
Conflict of Interest Statement
Mark D. Pescovitz has been an ad hoc consultant to Genentech, the manufacturer of rituximab. As corresponding author, Mark D. Pescovitz had full access to all data in the study and had final responsibility for the decision to submit for publication. None of the other authors have any reported conflicts.
1. Clinical transplants 2001. Los Angeles, CA, UCLA Immunogenetics Center 2002.
2. Rebibou JM, Cahbod J, Bittencourt MC, et al. Flow-PRA evaluation for antibody screening in patients awaiting kidney transplantation. Transplant Proc 2000; 32: 2745.
3. Gebel H, Bray R. Sensitization and sensitivity, defining the unsensitized patient. Transplantation 2000; 69: 2745.
4. Tyan DB, Li VA, Czer L, et al. Intravenous immunoglobulin suppression of HLA alloantibody in highly sensitized transplant candidates and transplantation with a histoincompatible organ. Transplantation 1994; 57: 553.
5. Jordan SC, Quartel AW, Czer LSC, et al. Post-transplant therapy using high-dose human immunoglobulin (intravenous gammaglobulin) to control acute humoral rejection in renal and cardiac allograft recipients and potential mechanism of action. Transplantation 1998; 66: 800.
6. Glotz D, Haymann JP, Sansonetti, et al. Suppression of HLA-specific alloantibodies by high-dose intravenous immunoglobulins (IVIg). Transplantation 1993; 56: 335.
7. Higgins RM, Bevan DJ, Vaughan RW. 5-year follow-up of patients successfully transplanted after immunoadsorption to remove anti-HLA antibodies. Nephron 1996; 74: 53.
8. Montgomery RA, Zachary AA, Racusen LC, et al. Plasmapheresis and intravenous immunoglobulin provides effective rescue therapy for refractory humoral rejection and allows kidneys to be successfully transplanted into cross-match-positive recipients. Transplantation 2000; 70: 887.
9. Maloney DG, Liles TM, Czerwinski DK, et al. Phase I clinical trial using escalating single-dose infusion of chimeric anti-CD20 monoclonal antibody (IDEC-C2B8) in patients with recurrent B-cell lymphoma. Blood 1994; 84: 2457.
10. Coiffier B, Haioun C, Engert NK, et al. Rituximab (anti-CD20 monoclonal antibody) for the treatment of patients with relapsing or refractory aggressive lymphoma: a multicenter phase II study. Blood 1998; 92: 1927.
11. Tobinai K, Kobahashi Y, Narabayashi M, et al. Feasibility and pharmacokinetic study of a chimeric anti-CD20 monoclonal antibody (IDEC-C2B8, Rituximab) in relapsed B-cell lymphoma. Ann Oncol 1998; 9: 527.
12. McLaughlin P, Grillo-Lopez A, Link BK, et al. Rituximab chimeric anti-CD20 monoclonal antibody therapy for relapsed indolent lymphoma: half of patients respond to a four-dose treatment program. J Clin Oncol 1998; 16: 2825.
13. Cerny T, Borisch B, Introna M, et al. Mechanism of action of rituximab. Anticancer Drugs 2002; 13 (suppl 2): S3.
14. Kovarik JM, Rawlings E, Sweny P, et al. Pharmacokinetics and immunodynamics of chimeric IL-2 receptor monoclonal antibody SDZ CHI621 in renal allograft recipients. Transplant Int 1996; 9: S32–S33.
15. Kovarik JM, Wolf P, Cisterne JM, et al. Disposition of basiliximab, an interleukin-2 receptor monoclonal antibody in recipients of mismatched cadaver renal allografts. Transplantation 1997; 64: 1701.
16. Zaja F, Iacona I, Masolini P, et al. B-cell depletion with rituximab as treatment for immune hemolytic anemia and chronic thrombocytopenia. Haematolgica 2002; 87: 189.
17. Ratanatharathorn V, Carson E, Reynolds C, et al. Anti-CD20 chimeric monoclonal antibody treatment of refractory immune-mediated thrombocytopenia in a patient with chronic graft-versus-host disease. Ann Intern Med 2000; 133: 275.
18. Leandro MJ, Edwards JC, Cambridge G. Clinical outcome in 22 patients with rheumatoid arthritis treated with B lymphocyte depletion. Ann Rheum Dis 2002; 61: 883.
19. Remuzzi G, Chiurchiu C, Abbate M, et al. Rituximab for idiopathic membranous nephropathy. Lancet 2002; 360: 923.
20. Deepe GS Jr. Role of CD8+
T cells in host resistance to systemic infection with Histoplasma capsulatum
in mice. J Immunol 1994; 152: 3491.
21. Deepe GS Jr. Immune response to early and late Histoplasma capsulatum
infections. Curr Opin Microbiol 2000; 3: 359.
21A. Agarwal et al. (in press).
22. Vail GM, Young RS, Wheat LJ, et al. Incidence of histoplasmosis following allogeneic bone marrow transplant or solid organ transplant in a hyperendemic area. Transpl Infect Dis 2002; 4: 148.
23. Kimball J, Pescovitz MD, Book BK, et al. Mycophenolate mofetil reduces human IgG anti-ATGAM antibody formation. Transplantation 1995; 60: 1379.
24. Kovarik JM, Pescovitz MD, Sollinger HW, et al. Differential influence of azathioprine and mycophenolate mofetil on the disposition of basiliximab in renal transplant patients. Clin Transplant 2001; 15: 123.
25. Baert F, Noman M, Vermeire S, et al. Influence of immunogenicity on the long-term efficacy of infliximab in Crohn’s disease. N Engl J Med 2003; 348: 601.
26. Slifka MK, Ahmed R. Long-lived plasma cells: a mechanism for maintaining persistent antibody production. Curr Opin Immunol 1998; 10: 252.
27. Slifka MK, Antia R, Whitmire JK, et al. Humoral immunity due to long-lived plasma cells. Immunity 1998; 8: 363.
28. McHeyzer-Williams MG, Ahmed R. B cell memory and long-lived plasma cell. Curr Opin Immunol 1999; 11: 172–179.
29. Calame KL. Plasma cells: finding new light at the end of B cell development. Nat Immunol 2001; 2: 1103.
30. Bernasconi N, Traggiai E, Lanzavecchia A. Maintenance of serological memory by polyclonal activation of human memory B cells. Science 2002; 298: 2199.