Reprints Available Directly from the Publisher Photocopying Permitted by License Only a Mouse with a Monoclonal Primary Lmmunoglobulin Repertoire Not Further Diversified by V-gene Replacement

We have generated a monoclonal B-cell mouse by introducing homozygous, nonfunctional RAG-2 alleles and a)1 light-chain transgene into the quasi-monoclonal (QM) mouse, which contains a "knocked-in" VHDJH rearrangement. Thus, this mouse, which we call MonoB, is devoid of T cells and contains preformed heavy-and light-chain genes encoding immunoglo-bulin with an anti-NP specificity. The MonoB mouse allows us to examine immunoglobulin diversity 'in the absence of processes mediated by V(D)J recombination and T cells. Here we report that not only is the MonoB's primary immunoglobulin repertoire monoclonal, but also that its secondary repertoire is not further diversified by V-gene replacement or gene conversion. Among 99 heavy-chain and 41) light-chain genes from peripheral B cells of the MonoB mouse, there were no V-gene replacements. When compared to the QM mouse, which has RAG activity, and for which V-gene replacement is the major diversifying mechanism , these data suggest that V-gene replacement is mediated by V(D)J recombination and not by other recombination systems.

d in acute lymphoblastic leukemia (Steenbergen et al., 1993), can generate a heavy (H) chain with a different specificity (Klein- field et al., 1986; Reth et al., 1986; Cascalho et al.,   1996) or destroy a productive gene (Taki et al., 1995).V H replacement is thought to be mediated by a hep- tamer recombination signal sequence (RSS) embed- ded near the 3' end of many VH'S (Covey et al., 1990;   Usuda et al., 1992: Chen et al., 1995; Fanning et al.,  1998) and by a putative nonamer sequence found ca. 12 bp upstream in the transgene (Chen et al., 1995).Because V-gene replacement shares the same RSS as V(D)J recombination, it has been proposed that the two share the same mechanism (Covey et al., 1990;  Usuda et al., 1992; Chen et al., 1995; Fanning et al.,   1998).

In order to determine whether V-gene replacement is dependent on V(D)J recombinase activity, we bred a mouse with preformed IgH and IgL genes onto a RA -knockout background.For the heavy-(H-) chain locus, we used the quasi-monoclonal (QM) mouse, which is heterozygous for two gene-targeted alleles: On one homologue is a targeted insertion in which the stretch of DNA containing the JH segments has been replaced by a rearranged VHDJH exon encoding an H-chain V region bearing the 17.2.25 idiotope (id) (Cascalho et al., 1996); on the other homologue is a nonfunctional "knockout" in which all JH gene segments have been deleted (Chen et al.,  1993a).The J: stretch has been knocked out on both homologues (Chen et al., 1993b), and, as a conse- quence, QM mice produce only -type light (L)

chains.We bred the QM mouse to the RAG-2-defi- cient strain (Shinkai et al., 1992) to generate a mouse homozygous for a disrupted RAG-2 allele and con- taining the inserted 17.2.25 VHDJH exon.As the absence of RAG expression also prevents V to J rear- rangement at the L-chain locus, we introduced a 1 transgene (Young et al., 1994).The absence of RAG-2 expression alo means that the resulting mouse lacks T cells.This mouse, which we call MonoB, has a nominally monoclonal primary B-cell repertoire with a B-cell receptor (BCR) consisting of an H chain encoded by an inserted VHDJH (with a constant region) and the ,1 transgene; this BCR is specific for the hapten NP [(4-hydroxy-3-nitrophenyl) acetyl] (Cascalho et al., 1996).

We report here experiments to investigate whether the secondary repertoire of the MonoB mouse becomes diversified by V-gene replacement, gene conversion, or any other process involving transfer of donor sequences.RESULTS (Cascalho et al., 1996) specific for the H chain encoded by the inserted 17.2.25 VHDJH and either anti-B220 (Figs.1A-1C) or anti-t a (Figs.1D-1F),

Figure 2 shows flow-cytometric profiles of peripheral blood lymphocytes from MonoB, QM, and RAG-2 -/- mice stained with an anti-chain antibody (Cascalho  et al., 1996) and either anti-B220 (Figs.2A and 2B) or anti-CD 19 (Figs.2C to 2E).

As can also b

seen in
ig. 1, there are no apparent ta-negative, id-positive (switched) cells in the MonoB mouse (Fig. 1D), although as expected, the QM mouse has switched cells (Fig. 1E).Analysis of serum Ig also failed to detect H-chain isotypes other than IgM (data not shown).

As shown in Fig. and 2, ell over 20% of the peripheral B cells in MonoB mice do not have the idiotypic BCR.About 30% of the B220-positive (Fig. 1A) as well as the a-positive (Fig. 1D) cells are id-negative; about 25% of the B220-positive (Fig. 2A) and CD19-positive (Fig. 2C) cells are -negative.In QM mice, about 15% of the B220-pos- itive (Fig. 1B) and ta-positive (Fig. 1E) cells are id-negative, and only a few (<3%) of the B220-posi- tive (Fig. 2B) or CD19-positive (Fig. 2D) are -negative.As the number of B220-positive cells in the periphery of MonoB is about one-third that of QM mice, the frequencies given in Figs. 1 and 2 mean that there are more L1-negative but about the same or fewer id-negative cells in the PBL of MonoB than in those of QM mice.

As shown in Fig. 3, 7.6% of MonoB and 1.4% of QM PBL were B220-positive, CD43-positive.These cells are immature B-lineage cells, and they stain nly dimly for B220 in both strains; as they mature, they will lose CD43 and increase the amount of B220 on their surface.But the id-negative and L-negative pop- ulations of Figs. and 2 stain brightly with anti-B220, so they are not likely to be immature B lymphocytes that have not yet expressed their BCR.,:. . .i"-"-"" i F Anti-ld FIGURE Flow-cytometric analysis of peripheral blood cells from MonoB (A, D), QM (B, E), and RAG-2 -/-mice (C, F). y axis; PE-conju- gated anti-B220 monoclonal antibody (A-C) or anti-t (D-F) x-axis: FITC-conjugated anti-idiotype 17.2.25 (anti-id) monoclonal antibody from ,-negative cells as well as Ig cDNA sequences cloned from LPS-stimulated cells.All 140 sequences were clearly the 17.2.25 VHDJH or X1, the knocked-in H-chain and transgenic L-chain genes present in the germline of MonoB mice, respectively.That is, there were no V-gene replacements.For the expressed X-chain transgene, th s is not surprising, for there are no likely donors for its replacement.But the fre- quency of V-gene replacements in the id-negative fraction of the QM mouse was previously found to be almost 100% (Cascalho et al., 1996).

There was also no evidence of gene conversion among the 140 variable-region gene segments sequenced.Moreover, in the same cDNA preparation and using the same reagents (except for the primers), there were more mutations in the VHDJH segment than in a similarly sized Ctl-2 segment-41 nucle- otide changes in 6496 bases of the VHDJH segment vs. 5 nucleotide changes in 7935 bases of the Cll-2 segment.Nevertheless, because of the large number of PCR cycles used, an accurate assessment of the mutant frequency would require more extensive methodological controls...,, ...._ :::....;...-(#:,;...... :.. :;.,.:, .,.-.. t.',-"'.'.:-" "... " ,'. ; ::.:4.::,?" ,.,;:: .....'.

.#_tg,r:.,.[:.:...m._''3_.-. . .
-fi i RAG2-/- C D E Anti-Lambda FIGURE 2
Flow-cytometric analysis of peripheral blood cells from MonoB (A, C), QM (B, D), and RAG-T/mice (E).y axis; PE-conju- gated anti-B220 monoclonal antibody (A-B); or anti-CD19 (C-E).x-axis" FITC-conjugated anti-2 monoclonal antibody, which reacts with ,1 and 2 L chains


DISCUSSION

What Are the Idiotype-Negative B Lymphocytes in MonoB Mice?

The high frequency of B-lineage cells lacking the idiotypic BCR of MonoB mice remains a puzzle.We have experimentally excluded that their lack of idio- type results from its loss by V-gene replacement, gene conversion, or any other process involvin

transfer of donor sequences.Moreover, we have experimentally excluded that thes
cells are immature B cells that

have not yet expressed their BCR.Although we have not formally excluded the possibility that the lack of reaction with anti-idiotypic and anti-2 antibodies is due to soma ic hypermutation, the mutation frequency is not high enough to make this a likely possibility unless the loss of idiotype per se confers a selective advantage.

It is possible that the relevant epitopes in these cells are masked by cross-reactive antigen(s), perhaps endogenous, perhaps environmental.Under normal circumstances, B cells with bound antigen home to the secondary lymphoid organs and, thus, are taken


CD43

RAG2-/-FIGURE 3 Flow-cytometric analysis of peripheral blood cells from MonoB, QM, and RAG 2-/-mice.y-axis; PE-conjugated anti-B220 mon- oclonal antibody, x axis: FITC-conjugated anti-CD43 out of the peripheral lood.Because T cells are important in promoting the germinal center reaction (Jacob and Kelsoe, 1992; MacLennan, 1994), the absence of T cells in MonoB mice might compromise the homing of lymphocyt s that have bound antigen, thus leaving many of t,hem circulating in the periph- eral blood with masked ) and id epitopes.Because reacting MonoB PBL with the cognate hapten bound to a carrier abrogated the binding of the anti-id anti- body (M.Madan, unpublished), the masked epitope notion is our working hypothesis.

Consistent with this hypothesis is the lack of cells with switched isotypes and the lack of circulating antibody other than IgM in MonoB mice, because switching also takes place in germinal centers.Interestingly, there are switched isotypes in a similar strain with another inserted VHDJH, B 1-8, and t e same )1 transgene on a RAG-2-deficient background (Lansford et al., 1998).As the only known significant dif- ference between the two strains is the sequence of the inserted VHDJH gene segment and, as a consequence, the fine specificity of the BCR (White-Scharf and Imanishi-Kari, 1982; David et al., 1992), it is credible that the difference in switching behavior is due to antigen binding.

Why

on by Transfer of Donor Se
uences in MonoB Mice ?

In the QM mouse and other gene-targeted mice (Chen   et al., 1995; Taki et al., 1995; Cascalho et al, 1996;   1997), V-gene replacements are strongly selected and most new BCRs are generated by this mechanism.

The selective forces must be at least as great in the MonoB mouse, which has an even smaller primary repertoire.Here we have shown that the primary rep- ertoire of MonoB mice is not further diversified-by V-gene replacement, gene conversion, or any other process involving transfer of donor sequences at a frequency detectable by current methods, even in a system with strong selective pressures.The interpre- tation of this result is complicated by the fact that the absence of RAG-2 expression in MonoB mice affects B cells in two ways-directly through the absence of V(D)J rearrangement and indirectly through the absence of T cells.

Whatever the nature of the cells lacking the idio- typic BCR of MonoB mice, any cells with V-gene replacement at the H-chain locus would be found in the id-negative fraction (Cascalho et al., 1997).

Among 83 VHDJH segments sequenced from the id-negative population, there were no r

lacements.As the id-negative pop
lation used was 25% to 30% of all B cells, we have effectively sampled about 300 cells, which means that at the 95% confidence level, there are less than 1% replacements.That is, the fre- quency of V-gene replacement is at least 100-fold lower than in the QM mouse.As it is not likely that the absence of T cells could result in such a large depression of the replacement frequency, the absence of V-gene replacement must be, at least in part, a direct consequence of the absence of V(D)J recombi- nase.

Gene conversion, another diversifying mechanism that is common in chickens and rabbits (Reynaud et  al., 1987; Knight and Becker, 1990; David et al.,  1992; Takeda et al., 1992), was also not observed.If it occurs at all in mouse germinal-center B cells, it is very rare (Ford et al., 1994).From our results, it is not clear whether its absence is due to dependence on RAG-2, T cells, or its rarity.


MATERIALS AND METHODS


Mice Strains and Breeding

The MonoB mouse was generated by breeding thr

strains: the QM mouse (Cascalho et
l., 1996), a L1 transgenic (Young et al., 1994), and a RAG-T/-(Shin- kai et al., 1992).After appropriate breeding, mice were screened by flow cytometry for the absence of T cells in peripheral blood lymphocytes (PBL) with an anti-CD3 antibody, a kind gift of Phillipe Poussier