Reprints Available Directly from the Publisher Photocopying Permitted by License Only Antigen-independent Selection of B-cell Ontogeny in Mice T15 Idiotype During

Precursors of B cells capable of responding to a T-independent form of phosphorylcholine (PC) in splenic focus assays were detected in the spleens of neonatal mice as early as 4 days after birth. The earliest anti-PC B cells were T15-. T15 foci-forming B cells were first detected 6 days after birth and expanded rapidly to constitute greater than 80% of the total PC-specific foci by day 10. Injection of heat-killed S. pneumoniae (R36A) into neonatal mice resulted in priming of the antibody response to PC, with an idiotype profile reflecting that of precursors of foci-forming B cells at the time of antigen administration. Priming of 2-day-old mice with 2 X10 and 2 X10 R36A induced a five-and tenfold increase in the antibody response to phosphorylcholine 6 to 8 weeks later. However, only 10 to 15% of the serum antibodies expressed the normally dominant T15 idiotype. Doses below 2 xl0 R36A showed no detectable priming activity. PC-specific hybridomas derived from mice injected with 2 X10 R36A 2 days after birth lacked the idiotypic and molecular characteristics typical of T15 antibodies. Antibodies to phosphorylcholine, raised by immunization of 6-week-old mice are normally protective against pneumococcal infection. However, serum antibodies from mice treated with R36A 2 days after birth and responding to phosphorylcholine following challenge with R36A at 6 weeks of age failed to protect against deliberate infection with virulent S. pneumoniae. These observations imply that the antigen phosphorylcholine does not play a role in the selective expansion and dominant expression of the T15 idiotype.


and Cohn, 19
2; Hansberg et al.,  1977).The antibody response to phosphorylcholine (PC),-which is a major immunogenic determinant expressed by a variety of microorganisms, is a striking example of this phenomenon (Potter and  Lieberman, 1970; Claflin et al., 1974).In many inbred strains of mice such as BALB/c and C57BL/6, the majority of antibodies in the primary immune response to PC express the TEPC15 (T15) idiotype (Gearhart et al., 1977; Claflin and Cubberly, 1980).*Corresponding author.Present address: 263 Tumor Institute, University of Alabama at Birmingham, UAB Station, Birmingham, Alabama 35294.

Furthermore, it has been shown that monoclonal antibodies expressing this idiotype provide optimal protection against infection by pathogens expressing PC (Briles et al., 1982).Dominant expression of J606 and J558 idiotypes is likewise observed in the immune responses to bacterial levan and czl,3 dextran, respectively (Blomberg et al., 1972;  Lieberman et al., 1976; Hansberg et al., 1977) in appropriate responder strains of mice.

In the few cases where this phenomenon has been investigated at a molecular level, idiotypic dominance appears to reflect the clonal expansion of one or a few B cells (Claflin and Cubberley, 1980).Given that B cells have the capacity to generate an extremely diverse repertoire, the recruitment of only a handful of B-cell clones in response to these T-independent antigens is intriguing and suggests developmental regulation of a high order.Environ- mental antigens have often been suggested as responsible for the preferential expansion during development of B cells that express high-affinity receptors for antigen, thus leading to idiotype domi- nance (Etlinger et al., 1982).However, it is cle r that even in germfree mice T15 + antibodies dominate the responses to phosphorylcholine (Sigal et al., 1977).

In our previous studies we have characterized the specificities of ntibodies in the perinatal B-cell repertoire in BALB/c mice and in vivo activities of their antibody products (Vakil and Kearney, 1986).

Our studies focused primarily on the development of T15 and J558 idiotypes and led to the hypothesis that autoantiidiotypic B cells and/or their antibody products in perinatal mice play a fundamental role in the selection or expansion of target-idiotypebearing B cells that no.rmally arise later in ontogeny.

This view predicts that antigen is not required for the establishment of dominant clones of B cells and that the role of antigen in clonal expansion is secondary to the effects of autoantiidiotypic B cells or their antibody products during ontogeny.This argument is supported by the observation that even in germfree mice T15 emerges as the dominant idio- type in response to challenge with R36A (Sigal et  a! , 1977; Etlinger and Heusser, 1986).A corollary to this hypothesis is that premature exposure to antigen will interfere with establishment of clonal dominance.

Here we report a set of experiments that examines the effects of exposure to PC early during B-cell development on the magnitude and characteris ics of the anti-PC response upon reexposure to the same antigen as adults, with emphasis on the idiotype profile and protective capacity of anti-PC antibodies.

These studies show that extrinsic antigen is not responsible for the dominant expr

sion of
15 idio- type in the response to PC and suggest, instead, a key role for developmental idiotypic regulation of this germline-encoded antibody response.


RESULTS


Precursors of PC-Reactive and T15 + B Cells During


Ontogeny

It has been previously shown in a functional assay for B-cell precursors involving an organ-fragment .culturesystem that splenic B cells capable of responding to PC could not be detected in mice until 4-5 days after birth (Sigal et al., 1977).We have employed a modification of this assay (Hurwitz et  al., 1982) to enumerate splenic B cells in neonatal mice that can respond to a T-independent form of PC (a heat-killed vaccine of S. pneumoniae [R36A]) to further assess development of the T15 idiotype as a function of age (Table 1).Until 3 days after birth no detectable splenic B cells responded to PC in fragment cultures.Spleen cells from 4-day-old mice con- tained precursors of foci producing PC-specific antibodies, but these did not express the T15 idio- type.Precursors capa le of developing into T15 / foci did not appear until 6 days of age.The frequencies of T15 / foci increased with age of the donor mice and rapidly outnumbered non-T15 PC-specific foci.

By day 10, 80% of the PC-responsive foci expressed the T15 idiotype, which reflects the contribution of Age of donor mice in days  T15 / B cells in adult mice.These results suggest that following initial expression of T15-B cells, T15 / B cells rapidly expand during the second week of neo- natal development, leading to the dominant expres- sion of this idiotype later in life.

Dominance of Non-T15 Antibodies in Mice

Prematurely Exposed to PC

Based on the results obtained before, we chose to explore possible mechanisms leading to this idio- typic dominance by investigating the priming capacity of heat-killed R36A to influence the PC response as a function of age.Mice were injected with 2 X10 R36A at intervals between 2 to 21 days after birth.A group of control mice was injected with saline at 2 days of age.At 7 weeks of age, all mice were injected with 2x108 R36A, and bled 7 days later.As shown in Fig. 1, in all mice pre- exposed to antigen, the levels of PC-binding IgM antibodies were elevated.Furthermore, greater than 95% of these antibodies expressed a V. $107 associ- ated idiotype defined by monoclonal antibody TC54 (Desaymard et al., 1984).However, in mice pretreated with R36A at 2 or 4 days after birth, only 10 to 15% of the PC-binding antibodies expressed the T15 idiotype as determined by binding to the mono- clonal anti-T15 antibody AB1-2 (Kearney et al.,  1983).This antiidiotype antibody reacts with the prototype idiotope expressed on anti-PC antibodies dependent on the combination of VH S107-encoded heavy chains and V K 22-encoded light chains.In mice pretreated at 6 days of age, the T15 idiotype constituted 50 to 60% of the total anti-PC response, and in mice pretreated with R36A at days 9, 12, 15 or 21 after birth, greater than 85% of the antibodies were T15 positive, typical of adult values.These results show that PC-specific B cells in neo- natal mice are primed by neonatal exposure to R36A to generate an augmented immune response later in life.Further, the idiotype profile of the neonatally primed anti-PC responses in adults reflects the ratios of non-T15 and T15 foci-forming B cells present at the time of treatment.In addition .totreat- ment with whole R36A, injection of 5/zg of pneumococcal polysaccharide extracted from cap- sules of type III pneumococci into 2-day-old mice also resulted in priming of non-T15 anti-PC anti- bodies (results not shown).These experiments sup- port the hypothesis that during normal development, selective expansion of T15 / B cells by endo- genous elements rather than exogeneous antigen leads to the dominant expression of the T15 idioty

in adult mice.age in days at treatment FIGURE 1. Eff
cts of neonatal administration of R36A on the subsequent antibody response to PC. Mice were injected with 2 X10 R36A at the ages indicated on the abscissa, rested, and challenged with 2 xl08 R36A at 7 weeks of age.Immune sera were quantitated as described in Methods.Data indicate levels of anti-PC antibodies (solid bars), T15 VH expressing antibodies (cross-hatched bars), and T15 antibodies (stippled bars) in the sera of control and treated mice.

Relationship of Antigen Dose to Its Priming Capacity In mice pretreated with R36A at 2 or 4 days after birth, the concentration of T15-reactive IgM anti- bodies in the serum after challenge with R36A as adults was comparable to that of control mice.How- ever, T15-positive antibodies constituted only a small portion of the total anti-PC antibodies (Fig. 1).

These results suggest that lack of T15 dominance in these mice did not reflect suppressive effects of the antigen on T15 / B cells.To test this hypothesis, the following experiment was conducted.Groups of 2-day-old mice were injected with 2 xl08 R36A (the optimal immunogenic dose in adult mice), or tenfold dilutions thereof, with 2 X10 R36A being the lowest dose tested.Following a rest period of 6 weeks, the mice were challenged with 2 xl08 R36A.

As shown in Fig. 2, mice pretreated with 2 xl08 R36A failed to mount a serum-antibody response following challenge with the same dose of antigen as adults, suggesting that this dose had rendered available PC-specific B cells functionally "inactive.Neonatal exposure to 2 X10 and 2 X10 R36A resulted in priming of the non-T15 component in the adult response to PC. 2 xl05 R36A yielded a small increase over contr)l anti-PC levels, however, 2 xl04 or lower doses of R36A failed to alter the PCspecific immune response.It is of note that in all mice primed with 2 X10 or lower doses of R36A, the level of T15 / IgM antibodies was compar

le to that in the group of control mice challenged
at tlqe same age and that none of the doses tested resulted in either the induction or suppression of T15 anti- bodies.Thus, the resulting expansion of the T15- component in the adult response to PC in neonatally treated mice did not appear to occur at the expense of T15 / B cells.Furthermore, the results of this and the previous experiment showed that priming of non-T15 antibodies in normal mice results only at appropriate times during development.


Characteristics of the Adult Antibody Response in

Mice treated neonatally with PC Hybridomas were generated from normal control mice and from mice treated with 2 X10 R36A 2 days after birth after challenge with 2 xl0 pneumococci at 7 weeks of age to examine in detail the molecular characteristics of T15-antibodies generated in the latter group.Monoclonal antibodies produced by these hybridomas were analyzed for reactivity with PC-BSA and with a panel of anti-V H and anti- idiotypic antibodies (described in Cerny et al., 1982;    Effects of neonatal administration of different doses Of R36A on the subsequent antibody response to PC. Mice were injected 2 days after birth with doses of R36A indicated on the abscissa, rested, and challenged with 2 xl08 R36A at 7 weeks of age.Immune sera were quantitated as described in Methods.Data indicate levels of anti-PC antibodies (solid bars), T15 VH-expressing anti- bodies (cross-hatched bars), and T15 antibodies (stippled bars) in the sera of control and treated mice.

seven monoclonal antibodies derived from normal untreated mice react with PC-BSA, TC-54, AB1-2, GB4-10, and MAID5-4 (Table 2), as does.the proto- type T15 / antibody BH8 previously generated in our laboratory (Pollok et al., 1984).However, only one of nine monoclonal antibodies derived from mice neonatally injected with R36A showed PC binding associated with the characteristic T15-idiotype profile, The other eight hybridoma antibodies reacted with PC-BSA and the VH S107-speci ic antibody TC54 but not with the T15-specific antiidiotype antibodies and were c tegorized as T15-.Dot blot analyses of total cytoplasmic RNA using V. $107 as well as VK22-, VK8-and VK 24-specific probes suggested that while all of the hybridomas generated from normal untreated mice expressed productively rearranged VH $107 as well as V K 22 as does the prototype hybridoma BH8, only the singular T15-idiotype positive hybridoma from the R36A-treated mice utilized this combination of VH and VK genes.RNA from the remaining eight hybri- domas from the latter group hybridized with the V H $107 probe consistent with the reactivity of their antibody products with TC54.However, six of these hybridoma expressed V K 24-encoded light chains characteristic of M167-idiotype-.bearingantibodies and one hybridoma expressed light chains encoded by VK 8 genes characteristic of M603 antibodies.The light-chain gene used by the remaining hybridoma could not be determined.

Failure to Induce Protective Immunity in Sera of Mice Neonatally Exposed to R36A

Anti-PC antibodies expressing the T15 rather than non-T15 idiotypes have been previously shown to be much more protective against virulent S. pneumoniae (Briles et al., 1982).We therefore sought to test whether the anti-PC antibodies generated in adult mice that had been primed as neonates with small doses of R36A would., protect against infection by the virulent strain WU2.Sera .werecollected and pooled from mice that had been injected with R36A 2 days after

at 7 weeks
of age and that expressed the lowest content of T15 antibodies, or from control mice, following challenge as adults as described (Fig. 1) and were transferred into groups of recipient CBA/NxDBA/2 F1. male XID mice.Sera were diluted in saline to contain approxi- mately 30/g of IgM anti-PC antibody in 0.2 ml and injected intraperitoneally in recipient X!D mice.One hour later, these mice were injected intravenously with 100 live S. pneumoniae (WU2) and mortality was measured over a period of 7 days.

As shown in Table 3, all mice in the control group injected with WU2 alone or with a dilution of serum from normal unimmunized mice comparable to that of the immune serum pool died with 3 days.Five of seven mice that received immune sera from normal adult mice challenged within R36A survived the length of the experiment.However, all four mice that received sera from neonatally immunized mice that had been treated with R36A died within the first 3 days.These results show that neonatal (premature) exposure to R36A does not induce protec- tive immunity even though high titers of IgM anti- PC antibodies are produced in such mice following exposure to R36A as adults.bDifferent from none/R36A at p <.0.001.

Different from none/R36A at p < 0.06.


DISCUSSION

The phenomenon of B-cell clonal dominance in immune responses to several exogenous antigens has been repeatedly observed in inbred strains of mice.The mechanisms by which dominant idiotypes are selected are not well understood.Three hypotheses have emerged to explain this phenomenon.

The first, derived from analysis of the PC response suggested a role for environmental antigen in selec- tive expansion of antigen-specific B-cell clones (Etlinger et al, 1982).Later, Klinman and Stone  (1983) observed that the T15 idiotype was dominant in bone-marrow-derived surface immunoglobulin negative pre-B cells capable of differentiating into antibody-secreting foci in organ cultures.They therefore proposed that dominance of the T15 idio- type was regulated at the level of DNA rearrangements and was the result of a biased association of VH $107 with V K 22 in PC-specific antibodies (Klinman and Stone 1983).Finally, based on ou studies on the functional characteristics of mono- donal antiidiotypic antibodies derived from perinatal B-cell hybridomas (Vakil and Kearney, 1986; Vakil et   al., 1986), we proposed that dominant expression of the T15 idiotype .resultsfrom selective expansion of B cells expressing this idiotype by autoantiidiotypic B cells during ontogeny, as had been suggested previously by K6hler (1979).

It has been demonstrated that-in mice lethally irradiated and reconstituted with normal adult bone marrow, the T15 idiotype fails to dominate the antibody response to PC (Kaplan et al., 1978;  Augustin et al., 1977; Wemhoff and Quintans, 1987)   despite the fact that bone-marrow-transplanted mice are exposed to an environment similar to normal conventionally reared mice.Furthermore, molecular mechanisms operating at the level of DNA re- arrangements in the donor hemopoietic tissue in marrow-transplant recipients are not likely to be dis- tinct from those operating in normal mice.Thus, the first two hypotheses stated before are unable to explain the loss of T15 dominance in lethally irradiated bone-marrow-transplanted mice.It appears therefore that although exposure to antigen may aid in the expansion of dominant idiotypes, it is not the limiting factor in the establishment of dominant B-cell clones.

W have previously reported the stimulatory effects of the anti-T15 antibody BD2 and the antianti-T15 antibody DB3, products of perinatal B-cell hybridomas (Vakil and Kearney, 1986).Exper iments described in the current report establish that the antigen PC does not play a significant role in the initial expansion of T15 / B cells during ontogeny.We have assessed (i) the functional status and idio- typic characteristics of PC-responsive precursors in neonatal mice and-(ii) the effects of exposure of newborn mice to smll amounts of antigen on sub- sequent immune response to PC as adults.

The frequencies of T15 / B cells as a function of age, directly assayed by immunofluorescent staining of IgM B cells by anti-T15 antibodies AB1-2 and GB4-10, have been reported (Bast et al., 1984; Pollok   and Kearney, 1984).Here we demonstrate that pre- cursors of PC-responsive B cells are present in the spleens of 4-dayoold BALB/c mice and can develop into antibody-forming foci in fragment cultures.However, the responding B cells do not express the T15 idiotype.PC-responsive T15 / B cells begin to appear around 6 days after birth when they con- stitute about 29% of the PC-reactive precursors and they increase in numbers nearing adult proportions of T15 idiotype by day 10.S milar results have been reported by Fung and K6hler (1980) and Teale and  Kearney (1986).

To permit detection of minor changes in T15 expression as a function of age, we pooled donor spleen cells only from mice born within 18 hours of each other.The trend of increase in percent T15 / foci between days 4 and 42 calculated using a rank test is statistically significant with p<0.001.Sigal and Klinman have previously conducted similar experiments using PC-Hemocyanin as the stimu- lating antigen (Siga!et al., 1977), and reported that PC-responsive precursors were detected at a fre- quency of 0.87 per 106 cells in spleens of 4-5-day- old (BALB/c xC57BL/6)F1 mice and that these lacked the T15 idiotype.However, when spleen cells pooled from 6-8-day-old, 8-10-day-old, or adult mice were analyzed in similar assays, T15 / B cells were found to dominate in the population of PC-responsive B cells.Because of the rapid emergence of the T15 idiotype, the detection of non-T15 PC-specific precursors only at very low frequencies in spleens of 3- or 4-day-old donor mice precluded further analysis.Our results confirm that T15-B cells are functional before T15 / B cells during ontogeny, however, the latter are selectively expanded and rapidly out- number the former, leading to the dominance of th T15 idiotype later in life.

The relative appearance of functional T.15-and T15 / PC-reactive B cells during ontogeny is also re- flected by their response to antigen priming during the first two weeks of life.Antigen exposure up to 4 days after birth resulted in the priming of T15- antibodies, whereas exposure at 6 days and beyond primed both T15-and T15 / B cells.The content of T15 idiotype in the subsequent response to PC re- flects closely the proportion of T15 / B cells within the PC-reactive B-cell pool at the time of priming.These observations suggest that there is selective and rapid expansion of T15 / B cells in neonatal mice between 6 and 10 days of life.To determine whether antigen plays a role in the events that lead to the dominant expression of the T15 idiotype, we analyzed the priming capacity of various doses of R36A in 2-day-old mice.Three findings were of note.Firstly, the optimal adult dose of 2 xl0 bac- teria appeared to induce tolerance to PC when injected into 2-day-old mice.Secondly, in mice pre- treated with doses of R36A lower than the optimal adult dose, there was no obvious impairment of the T15 / component of the anti-PC response at 7 weeks of age.Thirdly, it is apparent that doses of R36A smaller than 2 xl0 fail to produce any effect that is readily evident in a subsequent immune response to PC.

Although we have not conducted experiments to determine whether the PC-reactive cells have been physically deleted in mice neonatally injected with 2 xl0 R36A impairment of the anti-PC response is obvious.Considering that T15-B cells may be the only PC-responsive B cells present at this age, it is not clear why this treatment also affects the T15 response since B cells expressing this idiotype do not become functional until about 6 days after birth.However, it is probable that with high doses of R36A, antigen may persist in the neonatal mice at a high concentration long enough to inactivate emerging T15 / B cells.Alternately, the functional elimination of T15-B cells may interfere with idio- typic interactions, which we propose are responsible for the development of functionally competent T15 / B cells (Vakil and Kearney, 1991).Injection of 2 X10 or lower doses of R36A into neonatal mice does not alter T15 expression, suggesting that T15 / B cells may be absent in 2-4-day-old mice, or if present, they are functionally incompetent and unable to respond to the antigen.The failure of neonatal injec- tion of small doses of R36A (< 2 xl05 organisms) to alter the levels of anti-PC antibodies or T15 ex- pression is consistent with the observation that the T15 idiotype is invariably dominant in mice raised under germfree or conventional environments (Sigal  et al., 1977; Etlinger and Heusser, 1986), despite the ubiquitous nature of PC antigen,in the latter (Potter,, 1977).Furthermore, in lethally irradiated, bone- marrow-transplanted mice, the absence of T15 dio- type suggests that small amounts of environmental PC are not sufficient for the selective expansion of T15 / B cells.

We have examined the characteristics of the PC- reactive component of the immune response that is primed by early exposure to R36A.Hybridomas generated from mice neonatally exposed to R36A showed that the majority of hybridizing B cells produced anti-PC antibodies that utilize the V H $107 heavy-chain gene, however, it is in combination with VK24 rather than V K 22. Therefore, these anti- bodies did ot arise from T15 / B cells by somatic mutation, nor do they belong to the group-II or group III anti-PC antibodies described by Chang et   al. (1984).These non-T15 antibodies also failed to confer prOtective immunity in recipient XID mice against virulent S. pneumoniae.We have further reported in a separate study that mice primed with R36A 2 days after birth also fail to respond to a structurally distinct but idiotypically linked antigen c1,3 dextran (Vakil and Kearney, 1991).Therefore, our results suggest that premature antigen priming may lead to dominance by clones secreting anti- bodies that do not protect against infection, and as well, distort immune responses to unrelated antigens.
hese considerations may be relevant to present attempts to develop and use immunogenic polysaccharide vaccines in human infants.

We have previously reported the generation of autoantiidiotypic B cell hybridomas from un-of either T15 idiotype or of the connected J558 idio- manipulated fetal and newborn mice.Passive type during this time also interfered with the transfer of the purified IgM anti-T15 antibody BD2 development of endogenous T15 / B cells, pre- into newborn mice resulted in an enhancement of sumably by blocking the interaction between anti- the immune response to PC in these mice as adults.T15 and T15 B cells (Vakil and Kearney, 1991).These

In fact, treatment with BD2 on days 5 and 7 induced observations suggest that autoantiidiotypic B cells maximal enhancement of response to PC with domi-are obligatory components of idiotypic cascades nant expression of T15 idiotype; treatment on day 10 responsible for the establishment of idiotype domi- induced suboptimal enhancement; and treatments nance.beyond day 10 produced minimal or no enhance- To ether, these observations suggest that domi- ment (Vakil and Kearney, 1986).Thus, the age nance of the T15 idiotype and, therefore, the window during which antibody BD2 induces maxi-development of protective immunity to PC- mat enhancement of T15 coincides with the age at expressing pathogens is not directed by antigen.

which spleen cells from donor mice normally Rather, antigen administration to adults results in acquire the capacity to generate T15 / foci in fragthe selective expansion of T15 / B cells initially ment cultures.It also corresponds with the developselected and expanded by endogenous antiidiotypic mental stage when T15 / B cells can be primed by antibodies during development.

R36A to give an enhanced response to PC later in This phenomenon is not restricted to the anti-PC life.The studies presented here support.ourhypothresponse.Similar experiments have been conducted esis that idiotypic dominance of T15 results from the by injection of suboptimal doses of the selective expansion of particular B cells and that T-independent antigen oH,3 dextran into newborn endogenous BD2-1ike antiidiotypic B cells and/or mice to investigate the phenomenon of idiotype their antibody products may be involved in the dominance of the J558.When such treatments were establishment of T15/B cells in young unimmunized followed by antigen challenge at 6 to 7 weeks of mice.age, we likewise observed a skewing of the immune The mechanisms involved in the proposed B-cell response to an overproduction of antibodies lacking interactions are not obvious.We have observed that the normally dominant J558 idiotype (Vakil and  development of anti-T15 B cells precedes that of Kearney, 1991).Therefore, the dominant expression T15 / B cells (Pollok and  Kearney, 1984).Similarly, of one or a few idiotypes commonly observed in anti-T15 antibodies that were proposed to regulate immune responses to bacterial antigens does not the expression of T15 idiotype in vivo have been result from environmental priming during developobserved in serum and cultures of spleen cells from ment.Rather, we suggest that it may be the result 4-7-day-old mice (Strayer and K6hler, 1976).Per-of regulatory B-cell circuits within the immune haps, during ontogeny, antiidiotypic B cells delive