Abstract
The behavior of p-azophenylarsonate (Ars)-specific B cell clones during the primary T cell-dependent splenic response of A/J mice was investigated using an immunohistochemical approach. The earliest Ars-specific B cells were observed as isolated cells in the red pulp by day 3 after immunization with Ars-keyhole limpet hemocyanin, (KLH) and at day 6, large clusters of Ars-specific B cells were first detected in germinal centers, which continued to be observed for an additional 8 to 15 days. Surprisingly, no Ars-specific B cell foci were observed in or near the CD4 T cell-rich periarteriolar lymphoid sheath (PALS) during the entire primary response. Nevertheless, A/J mice immunized with (4-hydroxy-3-nitrophenyl)acetyl-chicken gamma globulin (NP-CGG) or Ars-CGG mounted robust splenic (4-hydroxy-3-nitrophenyl)acetyl or CGG-specific PALS-associated focus reactions, respectively. In contrast, no Ars-specific PALS B cell foci were detected in A/J mice immunized with Ars-CGG. These data add to a growing body of evidence indicating that B cell proliferation and differentiation in CD4 T cell-rich microenvironments are not prerequisites for the GC reaction. Taken together with previous results obtained using other model Ags, the data suggest that the specificity of the B cell Ag receptor may strongly influence the lymphoid microenvironment in which a B cell clone first undergoes Ag-driven clonal expansion and differentiation.
Distinct areas in secondary lymphoid organs are known to support early primary B cell proliferation and differentiation during humoral immune responses (1, 2, 3, 4). The currently accepted model explaining the genesis of these early events in the mouse spleen, and their relationship to the germinal center (GC)3 reaction has been based on work done using model T-dependent (TD) hapten-soluble protein conjugates (1, 2). Elegant studies conducted by Kelsoe’s group using the genetically restricted response to the hapten (4-hydroxy-3-nitrophenyl)acetyl (NP) conjugated to chicken gamma globulin (CGG) have shown that within days of immunization, NP-specific B cells enter the CD4 T cell-rich splenic outer periarteriolar lymphoid sheath (PALS) and establish extrafollicular foci, subsequently differentiating into Ab-forming cells (AFCs), which ultimately reach 100 to 200 cells in size (5). Most low-affinity serum Abs produced early after immunization are presumed to originate from this PALS-associated focus reaction (6). It has been postulated that Ab produced by such foci binds cognate Ag, forming immune complexes that are deposited on follicular dendritic cells (FDC). This deposition is thought to be critical for the initiation of the GC reaction (7, 8). Work conducted by Cerny’s group using another TD Ag, p-nitrophenyl-6-(O-phosphocholine)hydroxyhexonate-keyhole limpet hemocyanin (KLH), has shown the absence of splenic AFC foci in nu/nu mice, clearly demonstrating the T cell dependence of this reaction (9). Jacob and Kelsoe (10) also demonstrated that B cells in neighboring PALS foci and GCs can have common clonal origins. Based on the time of appearance of PALS foci and GC, they postulated that a few activated cells from the PALS develop locally into AFC foci and also seed adjacent GCs.
Recent data have questioned the generality of this model. In an analysis of the primary immune response to vesicular stomatitis virus (VSV), no viral-specific PALS foci were observed, even though virus-specific B cells were routinely found in GC (4). However, the initial stages of this response appear to be T cell independent. Therefore, it remains to be ascertained whether the PALS-associated focus reaction is a necessary step in primary TD B cell proliferation and differentiation, or whether this requirement is limited to immune responses to hapten-protein conjugates.
Another complication in the interpretation of data pertaining to the relationship between the PALS focus and GC reactions has emerged from the studies of Klinman’s group (11), who suggested that separate B cell lineages nucleate early AFC and GC reactions, and give rise to primary Ab and memory B cell responses, respectively. In this regard, in the anti-NP response of C57BL/6 mice primary Abs predominantly bear the λ1 light chain, whereas secondary Abs are predominantly of the κ isotype (12). In other words, the major clonotypes dominating the primary anti-NP response do not dominate the anamnestic responses to this hapten. Not all anti-hapten responses display such a “repertoire shift.” During the primary anti-Ars (p-azophenylarsonate)-KLH response of A/J mice, a single anti-Ars clonotype (expressing a single combination of VH, D, JH, Vκ, and Jκ gene segments, termed “canonical”) becomes predominant, and subsequently continues to dominate anamnestic responses (13). Since this clonotype fulfills the criteria of a true “memory clonotype,” we investigated the development of splenic foci and GCs by this and other anti-Ars clonotypes during the primary anti-Ars-KLH response of A/J mice.
Materials and Methods
Ag and immunization
A/J mice were purchased from The Jackson Laboratories (Bar Harbor, ME). Mice were housed in a specific pathogen-free facility in microisolator cages, and provided autoclaved food and water. Mice (8–10 wk old) were immunized i.p. (100 μg/mouse) with NP13 (Cambridge Research Biochemicals, Cheshire, U.K.)-CGG, Ars8-CGG (Sigma Chemical Co., St. Louis, MO), or Ars-KLH (Calbiochem, La Jolla, CA), all precipitated in alum, for induction of primary responses.
Immunohistochemistry
Spleens were removed at indicated time intervals after immunizations and embedded in Tissue-Tek OCT compound (Fisher Scientific, Bridgewater, NJ) by flash freezing in a 2-methylbutane bath cooled with liquid N2. Frozen spleens were stored at −70°C until sectioned. Six-micron sections were cut on a cryostat microtome, and thaw mounted onto 0.05% polyl-lysine (Sigma Chemical Co.)-coated slides. Sections were air dried, fixed in ice-cold acetone for 10 min, air dried a second time, and stored at −70°C.
Frozen sections were thawed and rehydrated in Tris-buffered saline (TBS) for 20 min. Endogenous peroxidase activity was blocked by immersing sections in 0.3% v/v aqueous H2O2 solution. Sections were then blocked with 5% BSA and 0.1% Tween-20 in TBS and stained with biotinylated mAbs anti-λ1 Ls136 (10), or anti-idiotypic Abs E4 or AD8 (14). Ars-BSA-biotin (1.5 μg/ml) was used for staining Ag-specific cells. NP-CGG-biotin (1 μg/ml) in a solution containing CGG (10 mg/ml) was used to stain NP-specific cells in sections derived from mice immunized with NP-CGG. CGG-specific cells in sections from Ars-CGG-immunized mice were stained with NP-CGG-biotin (1 μg/ml) only. All slides were incubated with biotinylated Abs or Ags for 1 h, washed in TBS-BSA, and then incubated with streptavidin-alkaline phosphatase (Southern Biotechnologies, Birmingham, AL) for 1 h. PNA (peanut agglutinins) coupled to horseradish peroxidase (E-Y Laboratories, San Mateo, CA) was used to identify GCs. Bound alkaline phosphatase and horseradish peroxidase activities were visualized using Napthol AS-MX/Fast Blue BB and 3-aminoethylcarbazole (Sigma Chemical, St Louis) respectively.
Microdissection of GC, DNA amplification, and sequencing
Idiotope-positive GCs were microdissected from different areas of a stained section obtained from a single spleen using a micromanipulator (Carl Zeiss, Thronwood, NY)-controlled capillary pipette, essentially as described earlier by Jacob and Kelsoe (10). Briefly, scraped tissue containing 50 to 100 cells was placed in tubes containing 15 μl of H2O and 5 μl of PBS. To this mixture, 5 μl of 2 mg/ml proteinase K (Fisher Biotech, Bridgewater, NJ) was added and the mixture was incubated at 37°C overnight. The following day, proteinase K was inactivated by heating at 95°C for 20 min, and then the lysate was subjected to two rounds of PCR amplification. The PCR was carried out using Taq DNA polymerase (Perkin-Elmer Corp., Norwalk, CT) as per the manufacturer’s instructions. Both rounds of PCR contained 40 iterative cycles (95°C for 1 min, 56°C for 30 s, 72°C for 3 min). The primers used for the first round were as follows: 5′ VHProm1(5′-GAGCACACTGCTGTCTGACC-3′), which hybridizes to the 5′ flanking region of the VH promoter; and 3′ JH3–4Int (5′-TCACAAGAGTCCGATAGACC-3′), hybridizing in the intron between JH3 and JH4. Two microliters of the first round products were amplified for 40 additional cycles using the following primer combination: 5′ VHProm2EcoRI (5′-GACGAATTCAGTCCTTCCTCTCCAGTT-3′), which is internal to VHProm1; and 3′ HindIIIBack (5′-GACTTCAAGCTTCAGTTCTGGC-3′), internal to JH3–4Int. The amplified products were gel purified, digested, and cloned into EcoRI-HindIII-linearized pBluescript vector (Stratagene, La Jolla, CA). The cloned fragments were sequenced using Sequenase (United States Biochemicals, Cleveland, OH) and the internal primers CDR2, hybridizing in the CDR 2 region (5′-GCCCTTGAACTTCTCATTG-3′) and JH2–3Int, hybridizing in the intron between JH2 and JH3 (5′CCTAGTCCTTCATGACCTGA-3′).
Results
A PALS-associated B cell focus reaction does not form during the primary immune response of A/J mice to the hapten Ars
To investigate the early events associated with the anti-Ars humoral immune response in the spleen, we immunized groups of A/J mice with Ars-KLH in alum, and at different time points processed their spleens for immunohistochemical analysis, as described in Materials and Methods. The majority of clonotypes participating in the primary anti-Ars response express the CRI-A (IdCR) Id. All Abs bearing this IdCR are encoded by a single VH gene segment (VHIdCR) (13). Splenic cryosections were stained for B cells expressing Abs capable of binding Ars-BSA and expressing idiotopes recognized by the anti-IdCR mAbs E4 and AD8. E4 recognizes all somatically mutated and unmutated forms of canonical V regions (14). All Abs encoded by the VHIdCR gene segment in unmutated form are AD8+. A majority of early primary anti-Ars Abs are AD8+ (14).
The earliest Ars-specific B cells observed after immunization were seen as isolated cells in the red pulp at day 3. Large clusters of Ars-specific B cells were first detected by day 6 in GCs, and such GCs continued to be observed through day 14. No PALS-associated foci specific for Ars or expressing the E4 or the AD8 idiotopes could be detected at days 3, 6, 9, 12, or 16 of the primary immune response (at least three mice were examined per time point). In this analysis we observed more than 60 E4+ and 50 AD8+ GCs, averaging approximately five independent GCs per spleen from days 6 through 16. A typical staining pattern observed in a splenic cryosection at day 9 in an Ars-KLH-immunized spleen is shown in Figure 1⇓. The Ag-specific GC stained with Ars-BSA (blue) is shown in Figure 1⇓A, and the same GC in parallel sections stained with anti-idiotypic Abs E4 or AD8 (blue) is shown in Figure 1⇓, B and C, respectively. Ars-specific GCs were usually accompanied by scattered Id-positive cells in adjacent regions of the white and red pulp, some of which were brightly staining and could be AFCs (Fig. 1⇓, B and C). The vast majority of the Ars-, E4-, and AD8-stained cells in the white pulp are confined to the GC, demonstrating the complete absence of a PALS-associated focus reaction.
A typical staining pattern of an A/J spleen section obtained 9 days after immunization with Ars-KLH in alum. A single follicle stained for Ars-binding cells (blue) is shown in panel A. The same follicle stained with anti-idiotypic Abs E4 (blue) and AD8 (blue) in tandem sections are shown in panels B and C, respectively. The GC is stained red in panel C (PNA). The central arteriole (ca) is marked for orientation. Note that Ars-BSA-biotin nonspecifically binds to the marginal sinuses, hence giving a blue border to the follicle in panel A. Original magnification: ×100.
The identity of the B cell clonotypes in the E4- and AD8-positive GCs was confirmed by microdissection-PCR analysis. PCR amplification of the rearranged VH genes from several Id+ microdissected GCs showed that the Abs expressed in these GC were encoded by the VHIdCR gene segment. A detailed analysis of two E4+, AD8+ GCs from different areas of the same spleen section revealed that both contained B cells expressing canonical anti-Ars heavy chain genes (Fig. 2⇓). Moreover, these heavy chain genes contained many somatic mutations, consistent with participation of the resident clonotypes in the B memory cell pathway (data not shown).
Comparison of CDR3 sequences of VH genes recovered from two different microdissected GC from an A/J spleen section obtained 10 days after immunization with Ars-KLH. The sequences shown are compared with a consensus canonical germline sequence composed of sequences of the canonical VHIdCR gene segment, the DFL16.1 gene segment, and the JH2 gene segment in the CDR3 region. Nucleotides that vary among canonical VH genes due to differences in joining site and N region addition are indicated by an “N.” Nucleotide identity is indicated by a dash, and differences are shown explicitly. The numbers specify amino acid codons, starting from the mature amino terminal end. Only one representative sequence is shown for GC 1, as three different clones obtained from GC 1 had identical sequence. Three clones from GC 2 showed three different sequences, which are depicted as GC 2 a, b, and c. The nucleotide differences in the core D region (codons 101–106) are most likely due to somatic mutation. Scattered point mutations were also observed in the VH region (not shown).
A/J mice can develop an NP-specific PALS-associated focus reaction
To assess whether the A/J strain had a generic defect in supporting PALS-associated focus reactions, we immunized groups of mice with NP-CGG in alum. This particular antigenic preparation had been shown to support a vigorous splenic focus reaction at the edges of the PALS in C57BL/6 mice (10). A robust NP-specific splenic focus reaction was observed in A/J mice during the primary response to NP-CGG, as seen by staining with NP-CGG-biotin in the presence of excess unlabeled CGG (Fig. 3⇓A). The Ag-specific B cells in these foci bore the λ1 light chain (Fig. 3⇓B) analogous to what has been reported for the anti-NP response in C57BL/6 mice. To confirm that these focus reactions were in the outer regions of the PALS, parallel sections were stained with an anti-CD4 mAb. As can be seen in Figure 3⇓C, the NP-binding λ1+ foci are located at the edges of the CD4+ T cell-rich PALS region. The intensity of the staining of the cells in these foci suggests that many were AFC.
Representative histologic analysis of parallel spleen sections obtained from an A/J mouse at day 8 after immunization with NP-CGG in alum. Three parallel sections were stained for NP-binding cells (blue, panels A and C), λ1 light chain-bearing cells (blue, panel B), GC cells (red, panels A and B), and CD4+ T cells (red, panel C). ca, central arteriole; F, AFC focus; T, CD4+ T cells. Original magnification: ×100.
Ars-CGG-immunized A/J mice develop CGG-specific PALS foci but not Ars-specific PALS foci
Earlier work has indicated the requirement of CD4 T cells for the PALS-associated focus reaction (9). Most studies on the focus reaction have used the TD carrier protein CGG (9, 10), and it has been proposed that the molecular form of the Ag can affect the immune response (15). Therefore, it was of interest to determine whether a different carrier for the Ars hapten had any effect on the development of the focus reaction. Since we knew that CGG was compatible with focus formation in A/J and C57BL/6 mice, we immunized A/J mice with Ars coupled to CGG. Staining of spleen sections with NP-CGG-biotin revealed the presence of CGG-specific focus reactions (Fig. 4⇓A) in these mice. These foci appeared to be composed predominantly of AFC and were present at the edges of the CD4+ T cell-rich PALS region (Fig. 4⇓B) as revealed by anti-CD4 staining. However, E4+ and AD8+ GCs present in the same histologic section were accompanied by only a few individual Id+ cells in the adjacent follicles, PALS, and red pulp (Fig. 4⇓, C and D).
CGG-specific foci are observed after immunization of A/J mice with Ars-CGG in alum. Results obtained from staining of parallel sections from a spleen taken 9 days after immunization are shown. Panel A shows CGG-specific staining (blue) of foci and GCs (also stained with PNA-red). Panel B shows staining of a parallel section with CGG (blue) and anti-CD4 (red). No Ars-specific foci are detected around E4+ (blue cells, panel C) or AD8+ (blue cells, panel D) GCs (stained red (PNA) in both panels C and D) in a different area of a parallel section from the same spleen. F, AFC focus; ca, central arteriole; T, CD4+ T cells. Original magnification: ×100.
Discussion
Previous models explaining the early events in TD splenic B cell responses segregate the responding B cell population mainly into two distinct spatial compartments: foci containing AFC at the periphery of the PALS, and GCs in the lymphoid follicles (1, 2). Based on the kinetics of appearance of these two locales of B cell proliferation and differentiation, it was suggested that PALS-associated B cell activation and proliferation first occurred in the PALS, and the products of this response locally differentiated into AFC and also seeded the adjacent GC reaction (10). Moreover, the Ab produced by the AFC in PALS foci has been implicated in the formation of immune complexes whose deposition in the follicle is thought to be involved in initiation of the GC reaction (6, 7, 8). In contrast, our data demonstrate that the only locale of extensive focal proliferation and differentiation during the primary TD anti-Ars response of A/J mice is the GC.
The lack of PALS foci in this response is not due to a generic deficiency of A/J mice in mounting a focus reaction. A/J mice immunized with NP-CGG did support a vigorous NP-specific PALS focus reaction, whose timing of appearance and locale were identical to that observed in the anti-NP response of C57BL/6 mice (5, 10). Moreover, the nature of the protein carrier did not influence the focus reaction, as Ars-specific PALS foci were not observed in A/J mice immunized with Ars-CGG. However, in these mice, CGG-specific PALS foci could be readily detected. These CGG-specific PALS foci did not express the λ1 light chain, thus ruling out the possibility that focus formation is restricted to λ1-bearing B cell clones (16).
Interestingly, the primary serum anti-Ars Ab response is much slower to develop than responses to other haptens, such as phosphorylcholine and NP (17, 18, 19). We found that in A/J mice immunized with 100 μg of Ars-CGG in alum, no λ-bearing anti-Ars Abs were detected in the primary response, and total κ-bearing anti-Ars Ab reached a concentration of only 200 to 300 μg/ml by day 13 of this response. In contrast, A/J mice immunized with 100 μg of NP-CGG in alum mounted a rapid serum anti-NP response, in which λ-bearing Ab reached 600 to 700 μg/ml by day 13. Moreover, κ-bearing anti-NP Abs had risen to approximately 5 to 10 times the level of λ-bearing anti-NP Abs by this time (data not shown). This high level of anti-NP serum Ab (κ and λ) during the early anti-NP response is correlated with the large PALS-associated AFC foci present at this time during the response. Conversely, the absence of an AFC focus reaction during the anti-Ars primary response may account for the very low serum Ab titers generated during the early stages of this response. Our results provide a possible explanation for these observations, in that GC development may be a prerequisite for AFC formation in the absence of an early PALS focus reaction (20). Interestingly, anti-Ars AFC cell clusters are observed in the red pulp during the secondary anti-Ars response (K. Vora and T. Manser, manuscript in preparation). However, these clusters are far less compact than the PALS-associated AFC foci observed during the primary anti-NP response.
A review of previously published data suggests that the absence of focal PALS-associated B cell proliferation and differentiation during the primary immune response may not be unique to the anti-Ars response. Bachmann et al., studying the BALB/c splenic response to VSV, also observed that VSV-specific B cell foci were not induced in the T cell areas but rather in the primary B cell follicles and in the red pulp near the marginal zone (4). Moreover, during the primary immune response of rats to alum precipitated DNP-spider crab hemocyanin, Liu et al. (21) observed only occasional hapten binding B cell blasts in the splenic T cell zone during the first week following immunization, even though DNP-specific B cells did populate GC. These observations are consistent with ours in indicating that initial Ag-driven B cell activation leading to GC formation in normal mice need not take place in the PALS.
Differences in locale of proliferation and differentiation, as well as migration patterns of B cell clones, could be influenced by many factors. These include B cell clones arising from distinct lineages (11), differential dependence on T cell help for priming, site of Ag localization (3, 4), surface Ig affinity for Ag (22, 23), surface Ig engagement of self-Ag (24), or prior activation by cross-reactive foreign Ags (25). Moreover, such factors need not be mutually exclusive. With regard to these issues, there are certain interesting similarities and differences between the major B cell clonotypes responding to NP and Ars in C57BL/6 and A/J strains of mice, respectively. The “starting affinities” for their cognate epitopes of the Ag receptors expressed by both these clonotypes (VH186.2-λ1 anti-NP and canonical anti-Ars) in the preimmune repertoire are similar (∼Ka of 105 M−1) (26, 27) and hence would not appear to account for their differential behavior upon Ag stimulation. These differences in clonotype behavior also could not be easily ascribed to differences in Ag localization (4), as A/J mice immunized with Ars-CGG (the anti-NP response was also studied using CGG as carrier) do form CGG-specific foci while still lacking Ars-specific foci.
Both the predominant anti-NP and anti-Ars clonotypes participate in the primary immune response, but only canonical Ars clones dominate memory responses. This “repertoire shift” in the anti-NP response could be due to the fact that most members of the prevalent primary clonotype terminally differentiate as AFC in PALS-associated foci, and hence cannot dominate memory responses. This repertoire shift is not unique to the NP system. Several laboratories have independently observed that a major clonotype present during the early primary immune response does not dominate memory responses (28, 29, 30, 31, 32). Hence, it is possible that the majority of anti-Ars clonotypes is a member of a memory lineage of B cells, and therefore contributes to serum Ab production only late in primary responses, and during memory humoral responses (11).
Perhaps the most powerful immunoregulation brought to bear on responding B cells is mediated by CD4 T cells. Therefore, quantitative and perhaps qualitative differences in the T cell help received by responding B cell clones might drastically alter their “in situ behavior.” That quantitative differences in T cell help can influence responding B cell behavior is supported by several previous studies. During the primary immune responses to the haptens DNP and oxazolone in carrier-primed animals, Liu et al. (21) observed hapten-specific B cell foci in the outer layers of the PALS and extending into the red pulp. These large clusters of B cells were not seen in the primary response of animals that had not been carrier primed. In contrast, the early anti-VSV response is T independent, probably explaining the observation that B cell foci specific for this virus develop in the T cell-poor red pulp (4). Moreover, Wang et al. observed that during the immune response to fluorescein-conjugated α(1—>6) dextran, a T-independent type-II Ag, large PALS foci did not form, while Ag B cells could be found in GCs (16).
Even though development of both GC and AFC foci require T cells, the type and number of T cells required to support these B cell differentiation pathways are clearly different. Small numbers of residual and abnormal T cells in nu/nu mice and γ/δ T cells in TCRα−/− mice are sufficient to support the GC/memory pathway in the absence of a detectable AFC response (9, 33). Several studies on the origin and characteristics of GC CD4 T cells have indicated that these cells may be functionally distinct from other CD4 T cells (34, 35). Additionally, Kelsoe’s group (36, 37) have shown that murine splenic GC T cells are immigrants from the PALS, but have characteristics similar to a novel subset of thymocytes. Our results suggest that the primary Ars response occurs primarily through the interaction of Ag B cells with GC T cell help. This clearly questions the generality of the dogma that all early B cell activation occurs in a T cell-rich microenvironment (38).
Given all of these considerations, it seems clear that factors both intrinsic and extrinsic to responding B cell clones can influence the microenvironmental locale in which they initially proliferate and differentiate during an immune response. However, the observation in our experiments that similar carrier-specific T cell help to both Ars and NP clones is compatible with both GC (Ars) and AFC/GC (NP) pathways clearly points to intrinsic differences in the nature of the B cell clones that participate in these two responses. One such difference not related to their cognate foreign Ags is the common anti-self (DNA and other Ags) reactivity associated with canonical and many other IdCR anti-Ars clonotypes (39, 40, 41). Such reactivity could result in anti-Ars clones progressing to either a “pre-activated” or “anergized” phenotype after emergence from the immature B cell pool. This prior ligand engagement could alter the subsequent behavior of these clonotypes in TD foreign Ag-driven responses.
Acknowledgments
We thank Dr. Garnett Kelsoe and his coworkers for lessons on spleen sectioning and immunohistochemistry. We are grateful to all members of the Manser Laboratory for their many indirect contributions to this work.
Footnotes
-
↵1 This work was supported by National Institutes of Health Grant AI23739 to T.M., and K.A.V. was supported by National Institutes of Health Training Grant 5-T32-CA 09678.
-
↵2 Address correspondence and reprint requests to Dr. Tim Manser, Kimmel Cancer Institute, BLSB 708, 233 S. 10th Street, Philadelphia, PA 19107. E-mail address: t_manser{at}lac.jci.tju.edu
-
3 Abbreviations used in this paper: GC, germinal center; Ars, p-azophenylarsonate; PALS, periarteriolar lymphoid sheath; AFC, antibody-forming cell; KLH, keyhole limpet hemocyanin; CGG, chicken gamma-globulin; NP, (4-hydroxy-3-nitrophenyl)acetyl; TD, T-dependent; VSV, vesicular stomatitis virus; TBS, Tris-buffered saline; PNA, peanut agglutinin.
- Received July 17, 1997.
- Accepted October 2, 1997.
- Copyright © 1998 by The American Association of Immunologists
References
In this issue
