| |
|
|
|
|
|
|
|||
|
Blood, 1 September 2005, Vol. 106, No. 5, pp. 1676-1684. Prepublished online as a Blood First Edition Paper on May 12, 2005; DOI 10.1182/blood-2004-10-4047.
IMMUNOBIOLOGY Malaria-specific transgenic CD4+ T cells protect immunodeficient mice from lethal infection and demonstrate requirement for a protective threshold of antibody production for parasite clearanceFrom the Division of Parasitology, National Institute of Medical Research, The Ridgeway, Mill Hill, London, United Kingdom; Division of Molecular Immunology, National Institute of Medical Research, The Ridgeway, Mill Hill, London, United Kingdom; and Max Planck Institute for Immunobiology, Freiburg, Germany.
T cells are important in the immune response to malaria, both for their cytokines and their help for antibody production. To look at the relative importance of these roles, a T-cell receptor (TCR) transgenic mouse has been generated carrying a TCR specific for an epitope of the merozoite surface protein 1 (MSP-1) of the malaria parasite, Plasmodium chabaudi. In adoptive transfer experiments, malaria-specific CD4+ T cells expand and produce interferon (IFN- ) early in infection, but the population contracts quickly despite prolonged persistence of the parasite. MSP-1-specific CD4+ cells can protect immunodeficient mice from lethal infection; however, the parasite is only completely cleared in the presence of B cells showing that T helper cells are critical. Levels of malaria-specific antibody and the speed of their production clearly correlate with the time of resolution of infection, indicating that a critical threshold of antibody production is required for parasite clearance. Furthermore, T cells specific for a shed portion of MSP-1 are able to provide help for antibody to the protective region, which remains bound to the infected erythrocyte, suggesting that MSP-1 has all of the components necessary for a good vaccine. (Blood. 2005;106:1676-1684)
T helper cells are essential for a protective immune response to the blood stages of the rodent malaria parasite, Plasmodium chabaudi chabaudi (AS).1 There is debate, however, over the extent to which T cells protect via Th1 cytokine-mediated mechanisms or by the antibodies that they help to produce. Both mechanisms dominate the T-cell response to P chabaudi in turn, with an early Th1-type cytokine response, which switches later in infection to one that provides effective help for malaria-specific antibody production and produces less interferon (IFN- ).2,3 Increased T-helper cell activity in the later stages3 is clearly beneficial because experiments in B cell-deficient mice demonstrate that B cells and antibodies are required for complete clearance of parasites,4-6 although the requirements and specificity of T-cell help and antibody for rapid clearance or immunity to reinfection are not known. To determine the precise role of malaria-specific T helper cells and the potential of an important vaccine antigen, merozoite surface protein 1 (MSP-1), in the protection from and clearance of malaria infection, we have generated a T-cell receptor (TCR) transgenic mouse with a TCR specific for P chabaudi MSP-1. This molecule is expressed on the invasive merozoite surface,7 and its C-terminal domain can induce a protective immune response.8-11 However, this domain is not efficiently processed by antigen-presenting cells and is a less effective inducer of CD4+ T-cell responses than other parts of MSP-1.10,12,13 For more effective vaccination strategies it would be important to know whether CD4+ T-cell help can be generated from other parts of MSP-1, which are more readily processed and presented on antigen-presenting cells. The MSP-1-specific TCR of our transgenic mouse, which recognizes a peptide within the soluble and readily processed 37/39-kDa fragment (Figure 2A) in the context of I-Ed,12,14 allows us to determine whether such help can be provided. Using this malaria-specific B5 TCR transgenic mouse, we show that anti-MSP-1 T cells can protect mice from a lethal P chabaudi infection. Furthermore, anti-MSP-137/39 T cells provide efficient help for the antibody response to the protective C-terminal fragment of MSP-1 even though they recognize an epitope within another region of MSP-1. Using this new model, we show that a critical level of protective antibody, as well as both effector T and B cells, are required for rapid control of parasitemia.
Generation of transgenic mice
Transgenic mice were generated using V(D)J segments of the TCR- Mice and parasites
Female BALB/c (MRC strain) and DO11.10 TCR Tg mice21 were maintained in the breeding facilities of the National Institute of Medical Research (NIMR). BALB/c V Antibodies and flow cytometric analysis
Single-cell suspensions of thymus, spleen, and lymph nodes were incubated in phosphate-buffered saline (PBS) containing 2% fetal bovine serum (FBS) and 0.1% sodium azide with anti-CD16/32 followed by appropriate combinations of phycoerythrin (PE)-, fluorescein isothiocyanate (FITC)-, biotin-, or allophycocyanin (APC)-conjugated antibodies: V Recombinant malaria proteins MSP-1 fragments (Figure 2A) in the vector pMalCR1 (New England Biolabs, Bishops Stortford, United Kingdom) were expressed and purified as described.26 MSP-11658-1746/MSP-119/21 was cloned into pPIC9K, expressed in Pichia pastoris SMD1168, and purified as described.10 MSP11658-1746 was linked to MSP1900-1507 by incubating 24 µM MSP11658-1746 with 8 µM MSP1900-1507 in 5 mM glutaraldehyde for 30 minutes at 4°C. The reaction was stopped by addition of 1 M ethanolamine, pH 10, to a final concentration of 1% and dialyzed into PBS. Linked protein was observed on a sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel. Recombinant P chabaudi adami apical membrane antigen 1 (AMA-127 was kindly provided by Robin Anders (La Trobe University, Melbourne, Victoria, Australia). In vitro stimulation of transgenic CD4+ T cells Lymph node cells or CD4+ T cells (4 x 104) isolated from spleens and lymph nodes of B5 Tg or BALB/c mice using magnetic-activated cell sorting (MACS) columns (Miltenyi Biotec, Surrey, United Kingdom), were cocultured with different antigens and either 5 x 103 bone marrow-derived dendritic cells (DCs) derived as described,28,29 or 1 x 105 irradiated (3000 rad) normal BALB/c splenocytes in a total volume of 225 µL Iscove medium (Sigma) containing 10% FBS, 2 mM L-glutamine, 0.5 mM sodium pyruvate, 100 U penicillin, 100 µg streptomycin, and 50 µM 2-mercapto-ethanol (Gibco, Invitrogen, Paisley, Scotland) in round-bottom 96-well plates (Nunc, Roskilde, Denmark). The following were used as antigens: (1) the specific (B5) peptide, MSP-11157-1171 (ISVLKSRLLKRKKYI); (2) an irrelevant peptide (B7), MSP-11690-1709 (RCEKDTEATCSINKGGCDPS); (3) recombinant protein fragments of MSP-1; (4) P chabaudi-infected erythrocytes prepared described previously30; or (5) uninfected erythrocytes. CFSE (5-(and-6)-carboxyfluorescein diacetate) (Molecular Probes, Eugene, OR) was used at 1 µM, and cells were labeled for 15 minutes at 37°C and cultured for 5 days. The B5 T-cell hybridoma12 was used as a positive control for the response to MSP-11151-1171.
T-cell proliferation was measured by incorporation of 3H-thymidine for 12 hours after 4 days, and interleukin 2 (IL-2) production was measured in the supernatants of the cultures using CTLL-2 cells.31 Results were expressed as stimulation indices (SIs) where SI = cpm obtained in presence of test antigens/cpm obtained in presence of medium, maltose-binding protein, or normal erythrocytes at equivalent concentrations as indicated. IFN-
Adoptive transfer into V Splenic CD4+ T and CD19+ B cells were purified using anti-CD4 or anti-CD19 microbeads on the autoMACS (> 95% purity, Miltenyi Biotec). For RAG-/- cell transfers, a further purification (> 99%) was performed on a high-speed cell sorter (Cytomation MoFlo, Fort Collins, CO) gating on PE-labeled CD4+ T cells and CD19+ B cells. Cells were washed thrice and resuspended in 0.09% saline at 5 x 107/mL. CD4+ cells (5 x 106) were injected intravenously into mice that were infected with 104 to 105 P chabaudi-infected erythrocytes intraperitoneally. Mice receiving DO11.10 cells were immunized with 50 µg chicken ovalbumin (OVA; Sigma) intraperitoneally at the time of cell transfers. Immune B cells for RAG-/- transfers were generated by infecting BALB/c mice with 105 P chabaudi twice 2 months apart and then, 1 week before transfer immunizing with MSP11658-1746. RAG-/- mice received CD19+ B cells or CD4+ purified cells (or both) intravenously. Some groups were immunized 1 to 2 days later, with 100 µg MSP11658-1746-MSP1900-1507 in 200 µL monophosphoryl lipid A (MPL)+trehalose dicorynomycolate (TDM) RIBI adjuvant (Sigma). Mice were infected with 104 parasites intravenously the next day. Mice were humanely killed at intervals during the infection and spleen and lymph node cells were analyzed by flow cytometry. Survival curves were tested for significance using the log-rank comparison (Prism, GraphPad, San Diego, CA). Malaria-specific and MSP-1-specific ELISA Malaria-specific and MSP-1-specific antibodies were measured by ELISA as described.15 A lysate of P chabaudi blood-stage parasites, recombinant MSP11-672, MSP1581-921, MSP1900-1507, or MSP11508-1766, and recombinant AMA-1 were used as the coating antigens (2-5 µg/mL in PBS).15 Maltose-binding protein at the same concentration was used as a negative control. Results are expressed as arbitrary units (AU) of specific antibody based on a standard hyperimmune plasma (1000 AU).15 Results are plotted as geometric means, with SEM of 7 animals at each point. Statistics were performed in Prism (GraphPad) using the Student t test, with P equal to .05 considered significant.
Generation of TCR transgenic mice recognizing an epitope of MSP-1
Because the 37/39-kDa region of P chabaudi MSP-1 was found to stimulate a strong CD4+ T-cell response in a primary infection of BALB/c mice,15 we selected a T-cell hybridoma specific for a peptide in this region with I-Ed for the generation of a CD4+ TCR transgenic mouse. The MSP1-specific TCR of the B5 hybridoma15 is composed of V
Transgenic T cells have a second TCR but are specific for MSP-1
When B5 TCR Tg mice were crossed onto a RAG2-/- background to eliminate endogenous TCR rearrangement, virtually no single-positive thymocytes or peripheral T cells expressing the transgene were found (Figure 1C), suggesting that the V To confirm that B5 Tg cells respond to the same epitopes as the precursor hybridoma, and that expressing a second TCR does not change their specificity, we stimulated them in vitro with specific peptide in several forms and measured cytokines in the supernatant (Figure 2B-E). B5 TCR Tg CD4+ cells and B5 hybridoma responded specifically to the B5-specific peptide (MSP-11157-1171; Figure 2B,E), a recombinant fragment of MSP-1900-1507 containing the epitope (Figure 2C), and to P chabaudi (AS)-infected red blood cells (RBCs; Figure 2D). Control stimulations with irrelevant peptide (B7, MSP11690-1709), recombinant fragment (MSP11508-1766), and uninfected RBCs (Figure 2B-E) showed minimal responses, and naive BALB/c T cells had only a low response to P chabaudi-infected erythrocytes (Figure 2B-E; BALB/c).We also observed (Figure 2F) B5 TCR Tg cells proliferating by flow cytometry in cells loaded with the fluorescent dye CFSE, which showed dilution of the dye over a period of 5 days' exposure to MSP-1 B5 peptide. The cells also up-regulated CD25 and CD44 and down-regulated CD45RB and CD62L as expected for activated T cells. In vivo response of transgenic T cells to P chabaudi infection
With this transgenic mouse, we were able to follow the immune response of malaria-specific CD4+ cells in vivo. CD4+ T cells from B5 Tg mice were transferred into BALB/c V
The activation marker, CD69, was up-regulated on the majority of transgenic CD4+ cells early in infection at day 8 (Figure 3A), as was the activation marker IL-2 receptor
Both B and T cells are required to control and reduce parasitemia To determine whether MSP-1-specific transgenic CD4+ cells could control a P chabaudi infection, purified CD4+ B5 Tg cells were transferred into T cell-deficient BALB/c nu/nu mice or B and T cell-deficient RAG2-/- mice, which were infected with P chabaudi. Unreconstituted RAG2-/- mice were not able to clear parasitemia, unlike BALB/c wild-type control mice, which cleared infection by day 20 (Figure 4A). Although neither wild-type nor MSP-1-specific CD4+ cells brought about clearance of parasites, transfer of T cells into RAG2-/- mice improved their survival. Ninety percent of the RAG-/- mice that received B5 T cells survived compared with 60% of the recipients of BALB/c T cells, whereas 60% of unreconstituted RAG2-/- mice died by day 15 of infection (Figure 4B). Although the difference between B5 and wild-type CD4+ cells does not reach significance, the survival curves of RAG mice with or without B5 T cells are significantly different using the log-rank comparison of the curves (P < .05). To investigate whether transgenic T cells together with naive B cells reduced parasitemia, B5 Tg and BALB/c CD4+ T cells were transferred into T cell-deficient BALB/c nu/nu mice, which were infected with P chabaudi 1 day later. In contrast to RAG-/- mice, all nu/nu mice receiving B5 Tg CD4+ cells survived and reduced parasitemias to subpatent levels on a par with recipients of polyclonal BALB/c CD4+ T cells (Figure 4C). Unreconstituted nu/nu mice showed a chronic parasitemia throughout the course of infection and a variable proportion (approximately 40%) died between day 9 and day 35 (Figure 4D). Control of infection was not due to nonspecific downstream effects of priming CD4+ T cells in vivo because OVA-specific DO11.10 TCR Tg CD4+ cells that were transferred into nu/nu mice and primed in vivo with OVA protein 1 day before infection with P chabaudi had no effect on the clearance of parasite, despite their activation, and all recipient mice died within 24 days (Figure 4D). Activation of T and B cells reduces parasitemia and increases rate of clearance of infection Transfer of CD4+ T cells into nu/nu mice demonstrated the requirement for T cells in reducing parasitemia. However, parasites were not cleared until 28 to 30 days after infection and there was no effect on magnitude of the peak parasitemia. To determine whether immune B cells would reduce the parasitemia earlier or improve the rate of clearance, naive B cells of the nu/nu mouse were replaced by immune B cells, which contain a higher frequency of malaria-specific cells, as well as activated and memory B cells. In these experiments, B cells isolated from mice that had recovered from 2 P chabaudi infections and been additionally primed with MSP119/21 were cotransferred with naive CD4+ T cells from either B5 transgenic or BALB/c mice into RAG-/- mice, which were infected 2 days later with P chabaudi (Figure 4E).
Immune B cells clearly accelerated the clearance of parasites, and transgenic CD4+ T cells were more effective in promoting this B cell-dependent clearance of parasites. RAG-/- mice receiving T cells along with immune B cells cleared their infections 10 to 14 days faster than was observed in the nu/nu mice. Strikingly, those RAG-/- mice receiving B5 TCR Tg CD4+ cells and immune B cells cleared parasitemia 4 days earlier than recipients of wild-type T cells (Figure 4E). However, the replacement of naive B cells by immune B cells still did not reduce the magnitude of peak parasitemia, suggesting that further activation of CD4+ T and B cells might be required to bring about overall reduction of the challenge infection.
When RAG-/- recipients were additionally immunized at the time of transfer of T and B cells with MSP1900-1507 (contains B5) covalently linked to MSP11658-1746 (induces protective B-cell responses) (antigen, Figure 4F), the time taken to clear parasitemia was further reduced in mice receiving transgenic B5 cells (Figure 4F), and parasitemia was controlled 6 days sooner than in mice receiving wild-type T cells. Activation of transgenic cells and immune B cells after transfer not only reduced parasitemia faster, but also significantly reduced peak parasitemia in recipient RAG-/- mice (Figure 4F, inset), suggesting that an effective immune response was accelerated by the activation step, such that it could now inhibit early parasite growth as well as enhance the clearance. In fact, it can be seen in Figure 5A with CD69 up-regulation that immunization with antigen induces early activation of the B5 T cells by day 7, whereas both groups down-regulate CD45RB by day 15 in RAG mice. Because cytokines are critical in parasite clearance and regulation of pathology in the infection, we have analyzed cytokine production of anti-MSP-1 CD4+ T cells in infection by intracellular cytokine staining and flow cytometry. Transferred B5 T cells can make IFN- Antibody is produced faster by immune B cells, and its production is enhanced by activation of malaria-specific T-cell help The requirement for B cells for elimination of parasites by B5 Tg CD4+ cells infers that these anti-MSP-1 Tg T cells are functional T helper cells for malaria antibody production. However, the epitope recognized by B5 Tg T cells is not within the protective C-terminal fragment of MSP-1 (MSP-119/21). The B5 peptide is part of the p37/39 fragment, which is cleaved off the merozoite surface prior to invasion of the erythrocyte and released into the serum.7 It was therefore important to determine whether B5 Tg CD4+ cells could provide help not only to B cells specific for the cleaved fragment containing the B5 epitope but also for the membrane-bound C-terminal region or other malaria proteins.
To determine the ability of anti-MSP-137/39 T cells to provide help for malaria-specific antibodies, we analyzed the serum antibody production. Immunoglobulin G (IgG) antibodies specific for all fragments of MSP-1 and for crude parasite lysate were generally detected sooner and at higher levels in infected nu/nu mice receiving Tg CD4+ cells than in those receiving BALB/c CD4+ cells (Figure 6A-E). Mice with B5 TCR Tg cells produced significant levels of anti-MSP1 and antiparasite IgG by week 4 of infection compared with mice receiving BALB/c CD4+ cells, which still had significantly lower antibody levels at week 6 (P < .05 for fragments MSP11-672, 581-921, and 900-1507). This suggests that transgenic T cells specific for the B5 epitope in MSP1900-1507 provide T-cell help to B cells producing antibody not only to that fragment but to the other 3 MSP-1 recombinant fragments and whole parasite lysate containing many malarial antigens as well as to another P chabaudi protein, AMA-1, expressed in the micronemes of the merozoite37 (Figure 6F).
Because transgenic T cells express 2 TCR Because RAG-/- mice receiving transgenic CD4+ cells and immune B cells recovered from infection faster than transferred nu/nu mice, and transgenic CD4+ cells and immune B cells cleared parasites faster than wild-type cells, it was important to determine whether improvement in parasite clearance rate was due to a faster and greater antibody response. The differences were indeed reflected in antibody level and speed; at 3 weeks of infection, antibodies to MSP1900-1507, MSP11508-1766, and crude parasite lysate in RAG-/- mice receiving both T and B cells were already more than 10- to 30-fold higher than those observed at the same time in the transferred nu/nu mice (compare units in Figure 7 with Figure 6). Furthermore, antibody responses to MSP1900-1507, and MSP11508-1766 of immune B cells transferred with the B5 transgenic cells were significantly greater than those of immune B cells cotransferred with wild-type T cells. It appears that a threshold level of protective antibodies may have been reached sooner in the infection, which led to faster parasite clearance. Immunization of mice receiving B5 transgenic T cells and B cells activated with coupled MSP-1 antigens resulted in a further 10-fold increase in antibody level. At 2 and 4 weeks after infection there were significantly higher levels of antibodies to MSP1900-1507, MSP11508-1766, and parasite lysate when the cells were primed in vivo (Figure 7D-F; note different scale in panels A,C and B,D). This was not the case for the transferred wild-type T and B cells except the antibody response to MSP11508-1766 at week 4. In addition to higher antibody levels after immunization, antibody production was much faster. Malaria-specific antibody was already present in plasma at the earliest time point measured, at 2 weeks after transfer and infection. This was significantly higher with the transfer of B5 Tg cells than wild-type cells (Figure 7; P < .05). Thus, faster clearance of parasites by activated B5 Tg cells and immune B cells is accompanied by a more rapid and greater antibody response.
We have generated a TCR transgenic mouse carrying CD4+ T cells specific for a peptide within P chabaudi MSP-1, a highly expressed surface molecule of the malaria parasite, and an important vaccine candidate. These transgenic CD4+ T cells respond specifically in vivo with early production of IFN- , are effective helper cells for antibody production, and together with B cells bring about elimination of a blood-stage P chabaudi infection.
Adoptive transfer experiments allowed us to observe the expansion and contraction of a malaria-specific T-cell population during infection. The transgenic CD4+ T-cell population expands and contracts in a manner similar to that described for model antigen systems.38,39 Contraction of T-cell populations is a common feature of CD4+ and CD8+ T-cell responses to nonreplicating antigens and infectious agents, as reviewed by Marsden and Strasser.40 The kinetic shown here is fairly similar to that reported for CD4+ cells in Salmonella infection, which is cleared up more quickly than P chaubadi.41 It would be important to determine whether this early contraction plays any role in immune evasion and persistent parasitemia as proposed for malaria and other infections.40,42-44
The TCR used for generating the transgenic mouse was cloned from an MSP1-specific hybridoma15 that transcribed 2 CD4+ T cells have 2 functional roles in the control or elimination of blood-stage malaria. One of these is a cytokine-dependent effector role that controls but does not eliminate P chabaudi,25,47-51 and the other is as a helper cell for the production of antibodies by B cells, which eliminate the parasite.4,48 Our data provide evidence for both of these roles. Naive and activated Tg CD4+ T cells transferred into RAG2-/- recipients, although not eliminating the parasite, prevent mortality. However, it is clear that the most important role of CD4+ T cells is to act as helper cells for the production of protective antibodies, because parasitemia is cleared only when B cells are included. In adoptively transferred nu/nu mice, this clearance took 28 days. Cotransfer of transgenic CD4+ T cells with immune B cells, which included a higher frequency of malaria-specific B cells, into RAG-/- mice, increased the rate of clearance of parasites and resulted in a significantly faster and several-fold greater antibody response than that seen in the nu/nu, or in RAG-/- mice with transferred BALB/c T cells. Addition of antigen to the B5 CD4+ T and B cells at the time of transfer, activating both cell populations, enhanced the antibody response still further with the kinetics of a classic secondary response. Antigen generates effector cells, which respond quickly and strongly enough that the antibody induced is sufficient not only to decrease the time of clearance but also to reduce the peak parasitemia. Interestingly, in the liver stages of malaria, where CD8+ T cells are the predominant effector mechanism, it has been found that they also must be activated specifically before infection to improve protection.52 Transgenic B5 CD4+ T cells were effective helper cells for malaria-specific antibodies, and they not only provided help for an antibody response to the fragment containing their specific peptide, but also for antibodies to other regions of MSP-1 including the protective C-terminal region. Because MSP-1 is cleaved by the parasite prior to invasion of the erythrocyte into 5 noncovalently associated fragments and one merozoite-bound fragment, which remains attached during erythrocyte invasion,7 it was possible that the various fragments could be taken up and processed by different antigen-presenting cells. Linked recognition is a requirement for specific T-cell help,53 and therefore CD4+ T cells specific for epitopes within soluble fragments may not have been very effective helper cells for antibody production for domains that remain attached to the parasite. This was clearly not the case; recognition of the same MSP-1 fragment by the Tg T cell was not necessary for a specific antibody response. The most obvious explanation for these results is that the MSP-1 protein complex is not always completely cleaved or that it associates noncovalently54 but tightly enough for the fragments to be taken up by the same B cell; however, it is also possible that the transgenic cells provide noncognate help to B cells of diverse specificities. Thus, our data indicate that microbial products, which may only be noncovalently associated, meet the requirements for linked recognition necessary for T-cell help and antibody production. In the case of MSP-1, this means that CD4+ T helper cells specific for epitopes outside of the protective region can still provide help for a protective antibody response. This has distinct advantages for vaccination with MSP-1. The protective region of this molecule (MSP119/21) is difficult for antigen-presenting cells to process,10,12 thus limiting the production of T-helper epitopes. An alternative source of CD4+ T-cell epitopes from more readily processed forms of MSP-1, such as that recognized by B5 T cells,12 would provide a more effective source of T cell help for a protective vaccine. In summary, our data indicate that in malaria, as in a variety of virus infections, high-titer, neutralizing antibody can effectively control a challenge infection.55 To achieve protection, a critical threshold of protective antibody is required, and effector B cells, activated T-cell help, and antigen are necessary for enhanced and rapid clearance (Figure 7G). Thus, vaccination strategies may need to strive for long-term survival or continuous restimulation of effector B cells and T-cell help.
The authors thank Andre Boonstra and Alexandre Potocnik for critical reading and constructive comments on this manuscript. We would also like to thank Anna Sponaas for experimental support.
Submitted October 20, 2004; accepted April 28, 2005.
Prepublished online as Blood First Edition Paper, May 12, 2005; DOI 10.1182/blood-2004-10-4047.
Supported by the Medical Research Council, United Kingdom. S.Q. was supported by a Wellcome Trust Prize studentship, and B.J.P. by a Fulbright Scholarship. R.S. and F.R.A. contributed equally to this study.
The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked "advertisement" in accordance with 18 U.S.C. section 1734.
Reprints: Jean Langhorne, National Institute of Medical Research, The Ridgeway, Mill Hill London, NW7 1AA, United Kingdom; e-mail: jlangho{at}nimr.mrc.ac.uk.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Copyright © 2005 by American Society of Hematology Online ISSN: 1528-0020 | |||||||||