Signals That Initiate Somatic Hypermutation of B Cells In Vitro1 Sigridur Bergthorsdottir,* Aoife Gallagher,2* Sandra Jainandunsing,* Debra Cockayne,† James Sutton,* Tomas Leanderson,‡ and David Gray2,3* Somatic hypermutation is initiated as B lymphocytes proliferate in germinal centers. The signals that switch on the mutation process are unknown. We have derived an in vitro system to define signals that will initiate mutation in normal, naive splenic B cells. We find that three signals are required to allow detection of somatic mutation in vitro; these are anti-Ig, anti-CD40, and anti-CD38. If any one of these is omitted, mutation remains off. We show that CD40 is obligatory in vivo, as CD40 knockout mice exhibit no Ag-driven mutation. In contrast, CD38 is not, as CD38 knockout mice mutate normally. We believe that, in vitro, CD38, in combination with other stimuli, drives extensive cell division, allowing the detection of mutated sequences. However, in germinal centers in vivo, proliferative activity is instigated by a different molecule. This is the first demonstration of the initiation of hypermutation in vitro with normal splenic B cells using defined stimuli. The Journal of Immunology, 2001, 166: 2228–2234. ffinity maturation of the Ab response occurs largely as a liferation, or what causes differentiation within the GC (e.g., the result of selection of the Ig V gene products that have transition of centroblast to centrocyte). CD40 is implicated in the A undergone a process of somatic hypermutation. The vast first two functions as CD40 knockout mice develop no GC (8, 9); majority of late primary and secondary response Abs exhibit however, its role in GC induction may be an indirect one as rudi- changes from the germline V sequence (1, 2), and most memory B mentary GC can be regenerated in these mice simply by injection 4 cells carry somatically mutated B cell receptors (BCR) (3). In of CD40-Ig (10). However, CD40 signals do seem to be crucial generating the repertoire for affinity selection to act upon, somatic during the final stage of memory differentiation in GC; the final hypermutation forms an integral part of the process of memory B rescue of mutated GC B cells from apoptosis (11) and entry into cell development, and the two processes seem to be intimately the memory pool (12). Signals via the BCR are clearly also crucial, linked, both occurring in the same anatomical site. Germinal cen- as without recognition and uptake of Ag the B cell would be un- ters (GC) are necessary for the generation of memory B cells (4) able to elicit T cell help. Whether the BCR signals generated by and are sites in which B cell hypermutation is switched on (5). small protein Ags have any role to play in driving the cell through Under normal, physiological circumstances GC appear to be the cell cycle is not known. Ag uptake and subsequent processing for only sites of mutation (6), although, if pushed to the extreme by presentation is important at two stages of the B cell response: the repeated immunization with enormous amounts of Ag, mutations can be detected in lymphotoxin-␣Ϫ/Ϫ mice that lack GC (7). initiation and for the selection of mutants in the GC for survival in B cell memory development, GC reaction, and somatic hyper- the memory pool (13). So we know that the processes involved in mutation occur only during responses to T-dependent Ags (gener- memory generation require signals via surface immunoglobulin ally proteins), suggesting that signals from T cells are necessary at and signals derived from T cells; indeed, we have shown that so- some point in all of these processes. The exact molecular identity matic mutation could be maintained (although not initiated) by of the signals driving these processes is still unknown. We do not culturing B cells with anti-Ig and helper T cells (14). This has since know what initiates the GC reaction, what drives GC B cell pro- been supported by similar findings using human B cell lines (15, 16) and tonsillar B cells (17). The exact identity of these T cell- derived signals is still mysterious. *Department of Immunology, Division of Medicine, Imperial College School of Med- In this study we have set out to find the signals that initiate icine, Hammersmith Hospital, London, United Kingdom; †Neurobiology Unit, Roche Bioscience, Palo Alto, CA 94304; and ‡Immunology Unit, University of Lund, Lund, somatic mutation of B cells in vitro. To do this we have stimulated Sweden B cells from a transgenic mouse carrying a V␬Ox1 gene with all of Received for publication June 23, 2000. Accepted for publication November the upstream elements required to target somatic mutation to the V 20, 2000. gene (18). As this V␬ is rearranged to a rat C␬ the transgene is The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance readily identified, avoiding the problem of unequivocal identifica- with 18 U.S.C. Section 1734 solely to indicate this fact. tion of the germline equivalent of a mutated V␬ that is often dif- 1 This work was supported by grants from the Wellcome Trust and the Medical ficult in normal mice due to close similarities within V gene fam- Research Council (U.K.). T.L. was supported by grants from the Swedish Medical ilies. B cells from these mice were stimulated and maintained in Research Council and the Swedish Cancer Society. culture under a number of conditions. These included BCR cross- 2 Current address: Institute of Cell, Animal and Population Biology, University of Edinburgh, Ashworth Laboratories, King’s Buildings, West Mains Road, Edinburgh linking in combination with the various effector molecules of T EH9 3JT, U.K. cell help (i.e., CD40 ligand and cytokines). B cells stimulated us- 3 Address correspondence and reprint requests to Dr. David Gray, Institute of Cell, ing anti-Ig together with anti-CD40, and supernatants from Th2 Animal and Population Biology, University of Edinburgh, Ashworth Laboratories, ␬ King’s Buildings, West Mains Road, Edinburgh EH9 3JT, U.K. E-mail address: clones did not accumulate somatic mutations in the V Ox1 trans- [email protected] gene. However, the replacement of cytokines with an Ab to CD38 4 Abbreviations used in this paper: BCR, B cell receptor(s); GC, germinal center(s). led to the detection of somatic mutation in vitro. Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00 The Journal of Immunology 2229 FIGURE 1. Mutation initiated in vitro. Sequences of the V␬Ox1 transgene from B cells cultured with LPS (A) and anti-␬, anti-CD40, and anti-CD38 (B). The top line shows the germline sequence; only changes from the germline are indicated. Note that similar results were obtained in three other experiments. 2230 SIGNALS THAT INITIATE HYPERMUTATION IN B CELLS FIGURE 1. Continued. The Journal of Immunology 2231 Table I. Frequency of somatic mutations in the V␬Ox1 transgenic in splenic B cells stimulated in vitro No. of No. of Mutated No. of Frequency of Stimulusa Sequences Sequences Mutationsc Mutationsd pe Anti-CD38 ϩ anti-CD40 ϩ anti-Ig 20 9 20 1/290 0.016 Anti-CD38 ϩ anti-CD40 15 6 6 1/725 0.73 Anti-CD38 ϩ anti-Ig 14 4 6 1/676 0.52 Anti-CD40 ϩ anti-Ig 22 9 10 1/627 0.56 Anti-CD38 ϩ anti-CD40 ϩ anti-Ig ϩ Th2 supernatant 6 1 1 1/716 0.75 Anti-CD38 8 0 0 Ͻ1/2320 0.65 LPS 12 3 3 1/1140 – a Cultures are the same as those from which the sequences in Fig. 1 were derived. b Number of independent sequences used for analysis. c Total number of mutations in all mutated sequences. d Number of mutations in all sequences divided by the total number of base pairs sequenced. e Probability using Fisher’s exact test that the difference in mutation frequency between LPS and triple stimulus is due to chance. Materials and Methods L-glutamine (Life Technologies, Paisley, Scotland), 50 mM 2-ME, and Mice and immunizations penicillin-streptomycin (50 mg/ml). The cultures were maintained at 37°C in a humidified 5% CO2 atmosphere. The following stimuli were added ELK mice, carrying a transgene incorporating V␬Ox1 and upstream regu- singly or in combination: LPS (Salmonella typhosa 0901) at 5 ␮g/ml (Difco, latory elements (18), were provided by Dr. Michael Neuberger (LMB, Detroit, MI), anti-CD38 supernatant at 1:200 (NIMR-5, Ref. 22, provided by ϫ Cambridge, U.K.). CD38 knockout mice (B6 129 F2) (19) were provided Dr. Michael Parkhouse; IAH, Pirbright, Surrey, U.K.), anti-CD40 (FGK-45, by Dr. Maureen Howard (Anergen, Redwood City, CA) and CD40 knock- Ref. 23) at 10 ␮g/ml, and anti-␬ light chain (187.1) at 10 ␮g/ml. Supernatants out mice (8), originally made by Dr. Hitoshi Kikutani (Osaka, Japan) have from alloreactive Th2 clone (20) produced as described previously were used been maintained in our laboratory for some time and backcrossed to at a 1:10 dilution. The cells were harvested after 7 days. C57BL/6 for seven generations. Wild-type mice were C57BL/6 strain. All mice were bred and maintained under standard laboratory conditions in the Analysis of cell division animal facilities of Imperial College School of Medicine, Hammersmith The number of cell divisions undergone in culture was measured using Hospital and later at the Ashworth Laboratories, University of Edinburgh. FACS analysis (FACSCaliber running CellQuest; Becton Dickinson, Mice were used at 8–12 wk of age.