Mutations Affecting Either Generation or Survival of Cells Influence the Pool Size of Mature B Cells

Immunity, Vol. 10, 619–628, May, 1999, Copyright 1999 by Cell Press

Mutations Affecting Either Generation orSurvival of Cells Influence the Pool Sizeof Mature B Cells

the periarteriolar lymphocyte sheath (PALS), where theybecome part of the B cell–rich follicular areas (MacLen-nan and Gray, 1986; Lortan et al., 1987). Immature Bcells in BM and spleen can be distinguished from theirmature counterparts by their short lifespan (3–4 days

Antonius G. Rolink,*‖ Thomas Brocker,†Horst Bluethmann,‡ Marie H. Kosco-Vilbois,§Jan Andersson,* and Fritz Melchers**Basel Institute for ImmunologyGrenzacherstrasse 487

versus 15–20 weeks), their lower B220 and IgD, andCH-4005 Baseltheir higher HSA and IgM expression levels (Forster andSwitzerlandRajewsky, 1990; Hardy et al., 1991; Allman et al., 1992,†Max Planck Institute for Immunobiology1993). Recently, we have described the specificity of aStubeweg 51mAb, called 493, which recognizes a protein with anD-79108 Freiburgapproximate m.w. of 130–140 kDa (Rolink et al., 1998).GermanyThe expression of this protein (pB130-140) within the‡Department of Roche GeneticsB cell lineage, as determined by FACS, is restricted toHoffmann-La Roche Ltd.pro-, pre-, and immature B cells, while mature B cellsCH-4070 Baseldo not express the protein in detectable amounts onSwitzerlandthe cell surface. This allows us to distinguish, in a simpler§Serono Pharmaceutical Research Instituteway than previously, immature from mature B cells14 Chemin des Aulxfound in bone marrow and spleen (Rolink et al., 1998).CH-1228 Plan-les-OuatesUsing this mAb in combination with BrdU-labeling kinet-Switzerlandics, we confirmed the results published by others show-ing that only a small proportion (10%–20%) of immatureB cells generated in the bone marrow enter the matureSummarylong-lived peripheral B cell pool. Moreover, our previousstudy indicated that the major loss during the differentia-The mature B cell compartment of MHC class II–defi-tion from immature to mature B cells occurs in the BMcient B6 I-Aa2/2 and the btk-defective CBA/N mouseand/or in the homing from the BM to the spleen, whilestrain is 4- to 5-fold smaller than in wild-type B6 mice.the majority, if not all, of the immature splenic B cellsThe defect in B6 I-Aa2/2 mice is intrinsic to B cellsdo enter the mature compartment (Rolink et al., 1998).and due to a 4- to 5-fold reduced lifespan, which how-

Why and how only such a small proportion of imma-ever can be normalized by an I-Ead transgene, but onlyture B cells is selected from the BM is largely unknown.when expressed early during B cell development. TheThe specificity of the expressed IgM molecule mightreduced number of mature B cells in the btk-defectiveplay a role (Gu et al., 1992). Evidence for a signalingCBA/N mouse is due to a 4- to 5-fold lower numberfunction through the expressed surface IgM comes from

of immature splenic B cells entering the mature com-mice expressing only a truncated form of Iga (mb-1)

partment. The combined defects of reduced lifespan (Torres et al., 1996) who have a block in B cell develop-and impaired generation in double mutant mice result ment. Particularly interesting is a similar defect at thein a severe deficiency in the mature B cell pool. same stage of B cell development seen in invariant (iI)

chain–deficient mice (Shachar and Flavell, 1996), withan even more severe block in double mutant mice lack-Introductioning both invariant chain and DM functions (Kenty et al.,1998). However, it remains to be elucidated why defects

B cell development in the bone marrow (BM) can bein invariant chain or DM chain synthesis affect B cell

divided into different stages, based on the rearrange- development. Moreover, the btk mutant mouse CBA/Nment status of the IgH and IgL chain loci (Ehlich et al., and the btk-deficient mouse both show a major defect1994; Ten Boekel et al., 1995) and the expression of in the transition from immature to mature B cells. Thisintracellular and surface bound markers (Rolink et al., defect is even more pronounced in CBA/N/CD402/2 and1994). The first cells during this developmental process btk2/2/CD402/2 double mutant mice (Oka et al., 1996;expressing IgM at their surface are the immature B cells Khan et al., 1997).(Rolink et al., 1994). Experiments by Osmond and his Here, using the novel mAb 493 in combination withcolleagues have shown that adult mice produce about BrdU labeling, we analyze MHC class II and btk mutant2 3 107 of these immature B cells per day (Osmond, mice for their capacity to generate immature and mature1991). Around 10%–20% of the immature B cells made B cells. Our results indicate that MHC class II mutantin the bone marrow migrate to the spleen (Allman et al., mice as well as btk mutant mice produce similar num-1993; Rolink et al., 1998) and enter through the terminal bers of immature B cells in the bone marrow, indistin-branches of central arterioles, thus arriving in the mar- guishable from wild-type control animals. Likewise, theginal zone blood sinusoids (MacLennan and Chan, 1993) efficiency by which both types of immature B cells homefrom which they then penetrate into the outer zone of to the spleen is equal in mutant mice compared to wild-

type controls. Moreover, the efficiency of immaturesplenic B cells to differentiate into mature B cells in‖ To whom correspondence should be addressed (e-mail: antibody@

bii.ch). MHC class II mutant mice appears not to be affected.

Immunity620

Figure 1. Defective B Cell Maturation inI-Aa2/2 Mice

(A) B220/pB130-140 expression on bone mar-row and spleen of adult B6, B6 I-Ab2/2, andB6 I-Aa2/2 mice. Bone marrow and spleencells were labeled with mAb 493 and anti-B220 mAb and were analyzed by flow cytom-etry. Numbers in quadrants represent per-centage of total gated lymphoid cells. Theabsolute numbers of spleen cells were 6.6 6

1.6 3 107 (B6), 5.9 6 1.2 3 107 (B6 I-Ab2/2),and 3.4 6 0.9 3 107 (B6 I-Aa2/2 mice).(B) BrdU-labeled B2201 spleen cells. AdultB6, B6 I-Ab2/2, and B6 I-Aa2/2 mice weregiven BrdU for 4 days in their drinking water.Splenic cells were analyzed for BrdU incorpo-ration. Histograms represent BrdU labeling ofB2201 gated spleen cells. Numbers representpercentage of BrdU-positive B2201 spleencells.

However, the efficiency of immature splenic B cells to al., 1993) that B cell development in B6 I-Ab2/2 is unal-tered. In marked contrast, the B220/pB130-140 expres-differentiate into mature B cells in btk mutant mice ap-

pears to be 4- to 5-fold reduced resulting in a smaller sion in bone marrow and spleen of the B6 I-Aa2/2 mice isabnormal. Thus, in the bone marrow of adult B6 I-Aa2/2splenic mature B cell compartment. The half-life of these

mature B cells appears unaffected. In contrast, the half- animals, the B220high pB130-1402 population, represent-ing the recirculating mature B cells (Rolink et al., 1998),life of mature B cells in MHC class II mutant mice is

reduced 4- to 5-fold, also resulting in a smaller mature is almost absent (Figure 1A). However, B cell develop-ment up to and including immature B cells in the boneperipheral B cell compartment compared to wild-type

mice. Moreover, the two mutations act in concert as the marrow of these mice is normal based on the numberof c-kit1 and CD251 pro- and pre-B cells (data notdouble mutant btk2/2/MHC Aa2/2 mice have a severely

reduced, almost undetectable, number of mature B cells shown). In the spleen of B6 I-Aa2/2, about 50% of theB2201 cells coexpress pB130-140 (Figure 1A). The num-in the spleen.ber of immature B cells in the spleen of these mice iscomparable to the number found in wild-type B6 andResultsB6 I-Ab2/2 mice, while the number of mature B cells is4- to 5-fold lower (Table 1). Thus, B6 I-Aa2/2 mice areB6 I-Aa2/2 but Not B6 I-Ab2/2 Mice Have a Defect

in the Generation of Mature B Cells severely impaired in the generation and/or maintenanceof mature B cells. This conclusion was further supportedMAb 493 was used to analyze the immature and mature

B cell compartments in bone marrow and spleen of by analyzing the BrdU incorporation of splenic B cellsshown in Figure 1B. Thus, about 15% of the B2201MHC class II–deficient mouse strains B6 I-Ab2/2 and B6

I-Aa2/2, which lack I-Ab or I-Aa expression, respectively. spleen cells in B6 and B6 I-Ab2/2 mice are BrdU positiveafter 4 days of BrdU feeding, while about 50% of theAs shown in Figure 1A, the B220/pB130-140 expression

in the bone marrow and the spleen of B6 I-Ab2/2 mice B2201 spleen cells in B6 I-Aa2/2 mice are labeled byBrdU during the same time. Thus, based on both theis comparable to that observed in normal B6 mice. More-

over, the numbers of immature and mature splenic B staining with mAb 493 and the BrdU labeling, we con-clude that B6 I-Aa2/2 mice export normal numbers ofcells are practically the same (Table 1). This confirms

previous findings (Cosgrove et al., 1991; Markowitz et immature B cells from the marrow that home at normal

Regulation of the Mature B Cell Pool Size621

MHC Class II Expression Early during B CellTable 1. Absolute Numbers (3106) of Immature and MatureDevelopment Is Required for the FormationSplenic B Cells in Wild-type B6 versus Several Mutant Miceof the Long-Lived Mature B Cell Compartment

Number of Number ofTo determine whether the I-Aa2/2-dependent defectImmaturea Maturea

could be repaired by the expression of MHC class II,Mouse Strain B Cells (6SD) B Cells (1SD)we crossed two different Ead transgenes (herein called

B6 8.90 6 0.65 38.90 6 2.20 EAD-107 and EAD-J) back onto B6 I-Aa2/2 mice. In bothB6 I-Ab2/2 9.00 6 1.10 32.50 6 1.90

cases, B6 I-Aa2/2 mice are capable of expressing theB6 I-Aa2/2 6.10 6 0.75 8.80 1 1.72transgenic I-Ead in combination with the endogenousB6 I-Aa2/2 EAD-107 5.00 6 0.81 7.72 1 1.53

B6 I-Aa2/2 EAD-J 8.32 6 0.55 32.20 6 3.51 I-Ebb on their cell surface.CD3e2/2 9.77 6 1.06 26.52 6 4.12 As shown in Figure 2, the B6 I-Aa2/2 mice comple-CD42/2 9.30 6 0.62 24.20 6 2.20 mented with the EAD-107 transgene have a B220/LAG-32/2 7.87 6 0.42 29.37 6 3.47 pB130-140 expression pattern in bone marrow andCBA/N 7.1 6 1.82 7.75 6 2.57

spleen indistinguishable from that observed in the non-B6 I-Aa2/2 CBA/N 3.52 6 0.73 0.87 6 0.33transgenic, I-Aa-defective littermates (compare Figures

a Immature B cells are defined as B2201/pB130-1401 while mature 1A and 2A). Moreover, the number of immature andB cells are B2201/pB130-1402. Numbers represent the mean 6 SD

mature B cells in the spleens of these mice is practicallyof at least five individual mice analyzed at 10–15 weeks of age.identical, indicating that the EAD-107 transgene cannotValues in bold represent significant differences compared to wild-rescue the B cell defect of B6 I-Aa2/2 mice (Table 1).type control mice.However, B6 I-Aa2/2 mice that carried the EAD-J trans-gene show a normalized B220/pB130-140 expression in

rates to the spleen. However, they are either impaired bone marrow and spleen (Figure 2A). Also, the numbersin the transition of immature to mature B cells within of immature and mature B cells in spleens of these micethe spleen or have a mature B cell compartment that are comparable to those found in B6 mice (Table 1).has a shorter half-life time than that of mature B cells Thus, the EAD-J transgene can rescue the B cell defectfrom wild-type B6 animals. in B6 I-Aa2/2 mice while the EAD-107 transgene cannot.

To measure half-lives of B cells, mice were given BrdU The BrdU labeling studies summarized in Table 2 sup-for 14 days in their drinking water. As shown in Table port this conclusion and show that the half-life of mature2, over 90% of the immature splenic B cells and about B cells in B6 I-Aa2/2 mice is increased to that of wild-15% of the mature splenic B cells in B6 and B6 I-Ab2/2 type B cells with the introduction of the EAD-J transgene,mice are BrdU positive after this period of labeling. Thus, while the lifespan of mature B cells in B6 I-Aa2/2 micemature B cells in these two strains have a half-life of carrying the EAD-107 transgene remains decreased rel-about 15–20 weeks. This conclusion is further supported ative to wild-type (Table 2).by the finding that 2 weeks after withdrawal of the BrdU Since the EAD-107 transgene cannot correct the de-only a minor fraction (2%–3%) of BrdU-positive mature velopmental defect in B6 I-Aa2/2 mice whereas theB cells disappeared (Table 2). EAD-J transgene can, the obvious question is—what is

In B6 I-Aa2/2 mice, over 90% of the immature splenic the difference between these two transgenic strains ofB cell compartment is also BrdU positive after 14 days mice? Although harboring identical Ead coding se-of labeling. However, in marked contrast to B6 and B6 quences, the construct used to generate the EAD-107I-Ab2/2 mice, about 50% of the mature splenic B cells transgenic mice consists of the Ead gene with only 1.4in B6 I-Aa2/2 mice are BrdU positive after 14 days of kb of its 59-noncoding region. For the EAD-J trans-labeling (Table 2). This, together with the finding that 2 genics, on the other hand, a 14 kb fragment was usedweeks after withdrawal of the BrdU only 25% of the containing the Ead gene flanked by approximately 9 kbmature splenic B cells are still BrdU positive, indicates of 59-noncoding region. This suggested that the expres-that the half-life of mature B cells in B6 I-Aa2/2 mice is sion of these two transgenes might be different. To testreduced to about 4 weeks. Thus, the 4- to 5-fold de- this, cell surface I-E expression in bone marrow andcrease in number (Table 1) appears to be due to a de- spleen of B6 I-Aa2/2 mice harboring either the EAD-creased lifespan from 15–20 weeks to 4 weeks of mature 107 or the EAD-J transgene was analyzed. As shown in

Figure 2B, I-E expression is hardly, if at all, detectableB cells in B6 I-Aa2/2 mice.

Table 2. Percentage of BrdU Positive Immature and Mature Splenic B Cells

After 2 Weeks Continuous LabelingAfter 2 Weeks Continuous Labeling and 2 Weeks Chase

Immature Mature Immature MatureStrain of Mice B Cellsa B Cellsa B Cellsa B Cellsa

B6 95.0 6 4.1 18.3 6 2.1 3.1 6 2.5 14.7 6 2.7B6 I-Ab2/2 92.3 6 6.2 17.0 6 2.3 4.1 6 0.6 14.9 6 3.1B6 I-Aa2/2 91.0 6 3.1 53.2 6 3.7 2.0 6 0.3 24.2 6 1.8B6 I-Aa2/2 EAD-107 93.2 6 0.7 49.5 6 1.6 1.8 6 0.4 23.4 6 0.7B6 I-Aa2/2 EAD-J 95.1 6 2.3 22.3 6 2.9 2.0 6 1.5 21.0 6 2.6CBA/N 93.6 6 2.8 21.0 6 3.0 1.5 6 0.8 18.7 6 2.5

a Immature B cells are defined as B2201/pB130-1401 while mature B cells are B2201/pB130-1402. The percentages represent the mean 6

SD of at least three individual mice analyzed. Values in bold represent significant differences compared to wild-type control mice.

Immunity622

Figure 2. Complementation of the B Cell De-fect in I-Aa2/2 Mice by an I-Ea Transgene

(A) B220/pB130-140 expression on bone mar-row and spleen cells of adult B6 I-Aa2/2 micecomplemented with the EAD-107 or theEAD-J transgene. Spleen and bone marrowcells were labeled with mAb 493 (anti-pB130-140) and an anti-B220 mAb and were ana-lyzed by flow cytometry.(B) I-E expression on bone marrow andspleen cells of B6 I-Aa2/2 mice comple-mented with the EAD-107 or the EAD-J trans-gene. Spleen and bone marrow cells werelabeled with the I-E-specific mAb 14-4-4S(Pharmingen) and an anti-B220 mAb and thenanalyzed by flow cytometry. Numbers inquadrants indicate percentage of total gatedlymphoid cells. Total number of spleen cellswere 3.3 6 0.6 3 107 (EAD-107) and 6.0 6

0.7 3 107 (EAD-J).

in the bone marrow of EAD-107 transgenic mice. On the cell precursors is required for mature B cells to becomelong lived.other hand, I-E is strongly expressed on the majority of

B2201 B cell precursors in EAD-J transgenic mice. Inthe spleen of EAD-J transgenic mice, all B cells express The Mature B Cell Maintenance Defect in B6 I-Aa2/2

Mice Is Intrinsic to the B Cellshigh levels of the I-E transgene. In the EAD-107 mice,only a fraction of the B cells have detectable I-E on their To test whether the mature B cell defect in B6 I-Aa2/2

mice is B cell intrinsic or due to the absence of Aasurface. Double labeling analysis with an anti-I-E mAband mAb 493 or anti-CD23, respectively, on spleen cells expression in or on other cells, BM chimeric mice were

made. B6 mice were lethally irradiated and injected ei-of the B6 I-Aa2/2 EAD-107 transgenic mice revealed thatthe majority of the immature pB130-1401 B cells do not ther with 5 3 106 B6 Ly5.1 or with 5 3 106 B6 I-Aa2/2

(Ly5.2) or with a mixture consisting of 2.5 3 106 BM cellsexpress I-E at all, or only at very low levels, while themajority of the mature CD231 B cells express the I-E of each. As shown in Table 3, B6 Ly5.1 chimeras have

an immature, as well as a mature, splenic B cell compart-transgene (data not shown).We conclude from these experiments that a difference ment similar in size to wild-type B6 mice, while in the B6

I-Aa2/2 chimeras these are similar to mutant B6 I-Aa2/2in transgene expression appears to be the most likelyexplanation for why the EAD-J transgene can, and the mice, i.e., with a reduced size of the mature B cell com-

partment. Thus, MHC class II expression on nonhemato-EAD-107 cannot, correct the B cell defect in B6 I-Aa2/2

mice. Moreover, these findings suggest that I-Aa or I-Ea poetic cells of the B6 host does not rescue the B celldefect of the B6 I-Aa2/2 mouse.expression on immature or possibly even on earlier B

Regulation of the Mature B Cell Pool Size623

Table 3. Absolute Numbers (3106) of Immature and Mature Splenic B Cells in Lethally Irradiated B6 Mice Reconstituted with BoneMarrow Cells from Various Mouse Strains

Number of Immature B Cells De- Number of Mature B Cells Derivedrived froma froma

Bone Marrow Cell Origin B6 Ly5.1 B6 I-Aa2/2 B6 Ly5.1 B6 I-Aa2/2

B6 Ly5.1 7.2 6 0.9 — 25.7 6 3.9 —B6 I-Aa2/2 — 6.2 6 0.7 — 8.3 6 2.4B6 Ly5.1 and B6 I-Aa2/2 4.2 6 0.6 3.9 6 1.2 22.7 6 2.9 5.4 6 2.1

a Spleen of chimeric mice were analyzed by FACS after triple labeling with Cy5-conjugated anti-B220 (clone RA3-6B2, Pharmingen), biotinylatedmAb493 (anti-pB130-140), and FITC-conjugated anti-Ly5.1 (clone A20, Pharmingen) or FITC-conjugated anti-Ly5.2 (clone 104, Pharmingen).The biotinylated mAb was revealed with PE-conjugated streptavidin (Southern Biotechnology). Immature B cells are defined as B2201/pB130-1401 while mature B cells are B2201/pB130-1402. The percentages represent the mean 6 SD of at least three individual mice analyzed.Values in bold represent significant differences compared to wild-type control mice.

The immature bone marrow and splenic B cell com- A Defective btk Gene Dramatically Reducesthe Efficiency of Immature Splenic B Cellspartments of the B6 mice reconstituted with the mixture

containing 2.5 3 106 normal B6 Ly5.1 and 2.5 3 106 B6 to Differentiate into Mature B CellsPrevious studies have shown that mice carrying a defec-I-Aa2/2 bone marrow cells consisted of about 50% B6

Ly5.1 and 50% B6 I-Aa2/2-derived cells (data not shown; tive btk gene have a severely reduced mature B cellcompartment (Hardy et al., 1983; Khan et al., 1995, 1997;Table 3). Thus, even under competitive conditions, B6

I-Aa2/2 B cell development is normal up to and including Oka et al., 1996; Klaus et al., 1997; see also Table 1).However, as also shown previously, B cell developmentthe immature splenic B cell compartment. However, the

B6 Ly5.1-derived mature splenic B cell compartment is up to and including immature B cells in the bone marrowis not affected (Oka et al., 1996). Moreover, as shownabout 4-fold the size of that derived from the B6 I-Aa2/2

bone marrow in these mixed chimeras. In conclusion, in Table 1, the number of immature splenic B cells isnormal, while the number of mature B cells in the spleenthe B cell maturation defect of the B6 I-Aa2/2 mouse

cannot be rescued by MHC class II–expressing hemato- is 4- to 5-fold reduced in btk mutant CBA/N mice. Todetermine whether this reduction is due to a shorterpoetic cells and therefore is most likely due to an intrinsic

B cell defect. half-life time of the mature B cells or to a decreasedefficiency of immature splenic B cells to differentiateSince the Bcl-2 protein is a critical determinant of

lymphocyte lifespan, we tested whether there are mea- into the mature compartment, BrdU labeling and chaseexperiments were performed. Similar to all other mousesurable differences in the level of Bcl-2 expression in

splenic B cells from B6 I-Aa2/2 compared to wild-type strains tested, over 90% of the immature splenic B cellsin CBA/N mice are BrdU positive after 14 days of labelingB6 mice. Using the novel mAb specific for mouse Bcl-2,

we detected no differences in the FACS profiles from (Table 2). Similar to B6 and B6 I-Ab2/2 mice, about 20%of the mature B cells in CBA/N mice become BrdU posi-permeabilized mature splenic B cells of mutant and wild-

type B6 mice (data not shown). tive after 14 days of labeling (Table 2). Moreover, 2 weeksafter withdrawal of BrdU (chase), 18% of the mature Bcell compartment in CBA/N mice remain BrdU positive,indicating that the half-life of mature B cells in theseMolecules Known to Interact with MHC Class II

Are Not Involved in the B Cell Maturation mice is comparable to wild-type B6 mice. Thus, the 4- to5-fold reduction in size of the mature B cell compartmentDefect of B6 I-Aa2/2 Mice

The TCR-CD3 complex as well as CD4 are essential is most likely due to an impaired transition from imma-ture to mature B cells in the spleen and not, like in themolecules on helper T cells involved in interactions with

MHC class II molecules on antigen-presenting cells, i.e., B6 I-Aa2/2 mice, due to a reduced half-life of mature Bcells.also on B cells. Also, the LAG-3 molecule has been

shown to be able to interact with MHC class II molecules(Huard et al., 1995). To test whether these interactions Severe Reduction in the Number of Mature B Cellsmight play a role in this Aa-dependent B cell maturation, in B6 I-Aa2/2/CBA/N btk-DefectiveB cell development was analyzed in CD32/2, CD42/2, Double Mutant Miceand Lag-32/2 mice. Since the origin of the defect in the mature B cell com-

The results of these analyses presented in Table 1 partment in B6 I-Aa2/2 mice and CBA/N mice was differ-show that no obvious defects in the transition from im- ent, we tested whether these two defects could actmature to mature B cells are detectable in the spleen synergistically. Therefore, B6 I-Aa2/2/CBA/N btk-defec-of these mice. Additionally, like in MHC class II2/2 mice, tive double mutant mice were bred and analyzed.no defects at other points of B cell development in bone As shown in Table 1 and Figure 3, B6 I-Aa2/2 btk-marrow are detectable also in these mutant mice (data defective double mutant mice have about half the num-not shown). Therefore, normal B cell development ap- ber of immature splenic B cells as their single mutantpears to be independent of the interactions between counterparts. However, mature B cells are practicallyMHC class II molecules and the TCR-CD3 complex absent in these mice—not only in the spleen (Figure 3)

but also in other lymphoid organs, such as lymph nodesand/or CD4 or LAG-3.

Immunity624

be true for defects in DM or invariant chain synthesis(Shachar and Flavell, 1996; Kenty et al., 1998). However,in all likelihood, this cannot be the reason for the short-ened lifespan of a mature B cell, since B cell maturationand lifespan is normal in B6 I-Ab2/2 mice. If MHC classII expression is required, how do B6 I-Ab2/2 mice man-age? Ruberti et al. described heterodimers of AbdEad

that play an important role in the immune responseagainst a peptide from sperm whale myoglobin (Rubertiet al., 1993). Several laboratories have also shown thatAbEa heterodimers can be formed by in vitro transfec-tion experiments (Anderson and David, 1989; Andersonet al., 1989; Matsunaga et al., 1990; Mineta et al., 1990).Hence, the B6 I-Ab2/2 mice might be able to form hetero-Figure 3. Severe B Cell Maturation Defect in B6 I-Aa2/2/CBA/N Dou-dimers between Aa and Eb, and the expression of theseble Mutant Miceheterodimers might be the reason for the normal B cellB220/pB130-140 expression on spleen cells from adult CBA/N anddevelopment. However, it must be noted that the exis-B6 I-Aa2/2/CBA/N double mutant mice. Spleen cells were labeled

with mAb 493 and an anti-B220 mAb and were analyzed by flow tence of AaEb heterodimers has never been docu-cytometry. Numbers in quadrants indicate percentage of total gated mented experimentally, and by analyzing cells from thelymphoid cells. B6 I-Ab2/2 mice we have also not found evidence for

the existence of AaEb heterodimers either. Alternatively,it might be hypothesized from the experiments pre-and bone marrow (data not shown). The low percentagesented here that Aa alone can associate with other mole-of splenic B2201 cells that do not express the 493 markercules to form a complex involved in B cell maturationare also CD192 (data not shown) and therefore probablyand maintenance either directly, i.e., in contact with ado not belong to the B cell lineage. Thus, the defectsligand in the environment of bone marrow, or indirectly,of the B6 I-Aa2/2 and btk-defective CBA/N mouse actby providing structures for such a contact.synergistically, resulting in the complete absence of a

Shachar and Flavell (1996) some time ago reportedmature B cell compartment. Immunohistochemical anal-that mice lacking the invariant chain (Ii) associated withysis of spleen cryosections showed the presence of BMHC class II molecules have an impaired B cell matura-cell follicles in the CBA/N and B6 I-Aa2/2 single mutanttion. Based on the findings that B cell development inmice (Figure 4). In marked contrast, the B6 I-Aa2/2/btk-B6 I-Ab2/2 mice is normal and that irradiated wild-typedefective double mutant mice have no proper B cellmice reconstituted with bone marrow from Ii2/2 micefollicles. Instead, their B cells are scattered looselystill showed the defect, the authors concluded that thearound the T cell areas and throughout the red pulp.block in the Ii2/2 mice is an intrinsic feature of B cellsthat is independent of MHC class II expression. MoreDiscussionrecently, Kenty et al. (1998) analyzing mice lacking thenonconventional MHC class II product DM also ob-

Recently, we have described a mAb, called 493, thatserved a B cell maturation defect, and interestingly,

enabled us to distinguish immature from mature B cells found an even more severe defect in Ii2/2DM2/2 double(Rolink et al., 1998). Here, we have used this novel mAb mutant mice. Thus, these two studies also, like oursto analyze B cell development in mice deficient in MHC presented herein, point at an important role for MHCclass II expression. class II molecules in B cell maturation.

Our analyses confirm those of others (Cosgrove et al., Like B6 I-Ab2/2 mice, mice lacking the MHC class II1991; Grusby et al., 1991; Markowitz et al., 1993) that B transactivator (CIITA) have an impaired MHC class IIcell development in B6 I-Ab2/2 mice is unaffected. Thus, expression. They do not show an obvious B cell matura-the sizes of the precursor, immature, and mature B cell tion defect (Chang et al., 1996). However, the Ii and DMcompartments as well as the lifespan of these cells in genes are still expressed in CIITA2/2 mice, although atB6 I-Ab2/2 mice are indistinguishable from those found about 10-fold reduced steady state levels. Moreover,in wild-type B6 mice. In B6 I-Aa2/2 mice, the size and very low but significant amounts of I-Aa transcripts wereturnover of the precursor and immature B cell compart- also detectable in these mice (Chang et al., 1996). Thus,ments are also comparable to wild-type B6 mice. How- the possibility that CIITA2/2 mice produce low amountsever, we find a 4- to 5-fold reduced number of mature of MHC class II protein, especially I-Aa, might explainB cells in B6 I-Aa2/2 mice. BrdU labeling and chase the absence of an obvious B cell maturation defect inexperiments indicate that a severe reduction in the life- these mice.span of mature B cells in B6 I-Aa2/2 mice (4 weeks In this paper, we show that an Ead transgene thatversus 15–20 weeks in B6 and B6 I-Ab2/2 mice) is the gives rise to MHC class II expression on precursor anddirect cause of this smaller peripheral compartment. immature B cells can rescue B cell maturation in B6

The obvious question arising from our analyses is why I-Aa2/2 mice, while an Ead transgene that is expressedthe lifespan of a mature B cell is shortened when MHC only on mature B cells cannot. This makes us favor theclass II I-Aa chains are not synthesized early in B cell idea that MHC class II expression on precursors and/ordevelopment. One possible reason could be a conges- immature B cells is required for optimal B cell maturation.tion of the endoplasmic reticulum with improperly as- Indeed, the early expression of the EAD-J transgene on

precursor B cells in the bone marrow of B6 I-Aa2/2 micesembled MHC class II molecules, and this would also

Regulation of the Mature B Cell Pool Size625

Figure 4. B6 I-Aa2/2/CBA/N Double MutantMice Lack B Cell Follicles

Cryosections from (A) B6 I-Aa2/2, (B) CBA/N,and (C). B6 I-Aa2/2/CBA/N mice were incu-bated with Texas red–conjugated goat anti-mouse IgM (red) and rat anti-Thy 1 (T24) fol-lowed by mouse anti-rat IgG conjugated toFITC (green). While the single mutant miceshow development of B cell zones (red)around the T cell containing PALS (green),these structures are absent or drastically re-duced in the double mutant (C).

(Figure 2) resembles that of wild-type MHC class II ex- tion-defective B6 I-Aa2/2 mice carrying the EAD-107transgene (Brocker et al., 1997). We, therefore, favor thepression found in normal mice (Tarlinton, 1993).

Positive selection of CD4 T cells—a process depen- idea that MHC class II expression on precursors and/orimmature B cells, and not on another cell type, is re-dent on the expression of MHC class II molecules on

thymic epithelial cells—is normal in the B cell matura- quired for optimal B cell maturation. This conclusion is

Immunity626

strongly supported by the finding that the B6 I-Aa2/2 B differentiate into the mature compartment. Since in bcl-2cell maturation defect is still prominent in the environ- transgenic CBA/N mice a complete normalization of thement of an irradiated wild-type B6 mouse reconstituted mature B cell compartment with respect to numbers iswith either B6 I-Aa2/2 bone marrow alone or with a observed (Woodland et al., 1996), the half-life time ofmixture of wild-type and B6 I-Aa2/2 bone marrow. the immature splenic B cells in these mice might form

The results of this paper indicate that MHC class II the basis for this defect. Thus, by virtue of the novelexpression, or at least I-Aa chain expression, is not mAb 493, we have now defined two different mutantneeded for the generation of normal numbers of imma- mice, B6 I-Aa2/2 and CBA/N, which both have a severelyture B cells in bone marrow (expressing MHC class II reduced mature B cell compartment. The two mutantsmolecules) nor for their transit to mature B cells. How- differ however in that B6 I-Aa2/2 mice show normal inputever, it is mandatory for mature B cells to obtain a long but decreased maintenance of mature B cells, whereaslifespan in the periphery. The mechanism by which MHC CBA/N mice display decreased input but normal mainte-class II participates in this B cell maturation process is nance of the mature B cell pool. Consequently, btk-unknown. Either MHC class II expression by itself on defective B6 I-Aa2/2 double mutant mice are devoid ofimmature B cells without interaction with a ligand is mature B cells. In conclusion, the size of the mature Bsufficient and necessary, or a binding of a ligand, maybe cell pool in spleen is controlled by reactions affectingon other cells, is needed. The three molecules known the rate of production or the lifespan of the cells. Whento interact with MHC class II, namely the TCR-CD3 com- mutations that reduce either of the two are combined,plex, CD4, and LAG-3, do not appear to be involved a severe mature B cell deficiency results that can be(Table 1). Thus, another, yet unknown, molecule capable seen in many peripheral lymphoid organs.of interacting with MHC class II might be involved, andthis molecule might either be expressed on immature B Experimental Procedurescells or on cooperating cells that are not the classical

MiceCD41 helper T cells.C57BL/6 (herein called B6) mice and the btk mutant CBA/N miceMembers of the Bcl-2 family of protooncogenes playwere purchased from Biological Research Laboratories (Fullinsdorf,a major role in regulating cellular responses to apoptoticSwitzerland). CD3e2/2 (Malissen et al., 1995) and B6 I-Ab2/2 (Cos-

signals, and the Bcl-2 protein is a critical determinant grove et al., 1991) mice were originally obtained from the CNRSof lymphocyte lifespan (Veis et al., 1993; Newton and animal facility (CNRS/CDTA, Orleans, France). Mice deficient forStrasser, 1998). A possible reason for decreased life- LAG-3 (Lag-32/2; Miyazaki et al., 1996) were kindly given to us by

Dr. Petter Hoglund, IGBMC, Universite Louis Pasteur, Strasbourg,span of mature B cells in MHC class II I-Aa2/2 miceFrance. B6 I-Aa2/2 (Kontgen et al., 1993) mice were originally madecould be decreased levels of Bcl-2 protein expressionat Basel Institute for Immunology, while the B6 Ead transgenic 107.1(Lu et al., 1999). However, after cytoplasmic immunoflu-line (herein called EAD-107) (Widera et al., 1987) was kindly given

orescent staining and FACS analyses we could not de- to us by Dr. R. Flavell, Yale University. The B6 Ead transgenic linetect any differences in the level of Bcl-2 expression (herein called EAD-J) originally described by Yamamura et al. (1985)between either immature or mature splenic B cells of was given to us by Dr. S. Kashiwamura, Osaka University, Osaka,

Japan. CD42/2 (Rahemtulla et al., 1991) and B6 Ly5.1 mice weremutant or wild-type mice.purchased from Jackson Laboratories. All mutant and transgenicMaintenance and survival of both naive and memorymice as well as the intercrosses between these were bred underT cells critically depend on interactions between TCRpathogen-free conditions at the Basel Institute for Immunology.and its MHC ligand (Tanchot et al., 1997). Mature B

cell maintenance and survival is dependent upon BcR Production of BM Chimeric Miceexpression (Lam et al., 1997). Also, the maintenance of B6 mice were irradiated with 950 rad and 4–6 hr later injected withB cell memory requires presence of antigen, indicating 5 3 106 BM cells. Chimeric mice were analyzed by FACS after 10

weeks of reconstitution.a role for BCR ligand interaction in this process (Mac-Lennan and Gray, 1986). However, since the defect de-

BrdU Labeling of Cellsscribed herein occurs after the expression of Ig andBrdU labeling of cells in vivo and determination of BrdU-positivetherefore, most likely, does not influence the B cell re-cells was performed as previously described (Rolink et al., 1998).

ceptor specificities, it is unlikely that the decreased life-span is due to a BCR ligand interaction. Antibodies, Flow Cytometric Analysis,

While the effects of defects in the normal maintenance and Cell SortingCell surface labeling was performed as described in Rolink et al.of longevity of mature B cells are visible in MHC class(1996, 1998), and the labeled cells were analyzed by FACScan orII I-Aa2/2 mice, defects in the rate of production of ma-FACS Calibur (Becton Dickinson). Determination of bcl-2 content ofture B cells can be seen in btk-defective CBA/N mice.splenic B cells was done using the FITC-conjugated Bcl-2 AntibodyThis mouse strain shows a wide array of immune de-Reagent Set (Pharmingen). For sorting of labeled cells, the MoFlo

fects. These include a failure to make antibody re- (Cytomation) or the FACStar PLUS (Becton Dickinson) was used.sponses to type 2 T-independent antigens (Scher,1982b) and the absence of B-1 cells (Scher, 1982a; ImmunohistologyHardy et al., 1983). Moreover, while B cell development Spleens from nonimmune CBA/N, B6 I-Aa2/2, and B6 I-Aa2/2/CBA/N

mice were embedded in Tissue-Tek (Sakura Finetek U. S. A.) andin bone marrow is little affected (Khan et al., 1995; Okafrozen on dry ice. Seven millimeter thick cryosections were fixed inet al., 1996) and immature B cell numbers in spleenacetone for 10 min and sequentially incubated in rat anti-mouseare relatively normal, CBA/N mice have previously beenThy-1 (clone T24), mouse anti-rat IgG (H 1 L chain-specific) F(ab9)2analyzed and found to have a severely reduced mature conjugated to FITC (Jackson Immunoresearch Labs) and Texas red–

B cell compartment (Oka et al., 1996; Klaus et al., 1997). conjugated goat anti-mouse IgM (Southern Biotechnology Associ-Here, we indicate that this defect is due to the decreased ates). After washing in PBS and mounting, the sections were photo-

graphed using a Zeiss Axiophot epifluorescence microscope.frequency by which CBA/N immature splenic B cells

Regulation of the Mature B Cell Pool Size627

Acknowledgments al. (1995). Defective B cell development and function in Btk-deficientmice. Immunity 3, 283–299.

The Basel Institute for Immunology was founded and is supported Khan, W.N., Nilsson, A., Mizoguchi, E., Castigli, E., Forsell, J., Bhan,by F. Hoffmann-La Roche Ltd., Basel, Switzerland. We thank Andrea A.K., Geha, R., Sideras, P., and Alt, F.W. (1997). Impaired B cellGroenewegen, Mireille Riedinger, and Nadja Straube for excellent maturation in mice lacking Bruton’s tyrosine kinase (Btk) and CD40.technical assistance and Werner Metzger and Ernst Wagner for Int. Immunol. 9, 395–405.excellent animal facilities. We also thank Rhodri Ceredig, Klaus Kar- Klaus, G.G., Holman, M., Johnson-Leger, C., Elgueta-Karstegl, C.,jalainen, Christoph Schaniel, and Raul Torres for the critical reading and Atkins, C. (1997). A re-evaluation of the effects of X-linkedof our manuscript. immunodeficiency (xid) mutation on B cell differentiation and func-

tion in the mouse. Eur. J. Immunol. 27, 2749–2756.

Kontgen, F., Suss, G., Stewart, C., Steinmetz, M., and Bluethmann,Received February 15, 1999; revised April 15, 1999.H. (1993). Targeted disruption of the MHC class II Aa gene inC57BL/6 mice. Int. Immunol. 5, 957–964.

References Lam, K.P., Kuhn, R., and Rajewsky, K. (1997). In vivo ablation ofsurface immunoglobulin on mature B cells by inducible gene tar-

Allman, D.M., Ferguson, S.E., and Cancro, M.P. (1992). Peripheral geting results in rapid cell death. Cell 90, 1073–1083.B cell maturation I. Immature peripheral B cells in adults are heat-

Lortan, J.E., Roobottom, C.A., Oldfield, S., and MacLennan, I.C.stable antigenhi and exhibit unique signaling characteristics. J. Im-(1987). Newly produced virgin B cells migrate to secondary lymphoidmunol. 149, 2533–2540.organs but their capacity to enter follicles is restricted. Eur. J. Immu-

Allman, D.M., Ferguson, S.E., Lentz, V.M., and Cancro, M.P. (1993). nol. 17, 1311–1316.Peripheral B cell maturation. II. Heat-stable antigenhi splenic B cells

Lu, L., Chaudhury, P., and Osmond, D.G. (1999). Regulation of cellare an immature developmental intermediate in the production ofsurvival during B lymphopoiesis: apoptosis and bcl- 2/Bax contentlong lived marrow-derived B cells. J. Immunol. 151, 4431–4444.of precursor B cells in bone marrow of mice with altered expression

Anderson, G.D., and David, C.S. (1989). In vivo expression and func- of IL-7 and recombinase-activating gene-2. J. Immunol. 162, 1931–tion of hybrid Ia dimers (E alpha A beta) in recombinant and 1940.transgenic mice. J. Exp. Med. 170, 1003–1008.

MacLennan, I.C.M., and Gray, D. (1986). Antigen-driven selection ofAnderson, G.D., Banerjee, S., and David, C.S. (1989). MHC class II virgin and memory B cells. Immunol. Rev. 91, 61–85.A alpha and E alpha molecules determine the clonal deletion of V

MacLennan, I., and Chan, E. (1993). The dynamic relationship be-beta 61 T cells. Studies with recombinant and transgenic mice. J.

tween B-cell populations in adults. Immunol. Today 14, 29–34.Immunol. 143, 3757–3761.

Malissen, M., Gillet, A., Ardouin, L., Bouvier, G., Trucy, J., Ferrier,Brocker, T., Riedinger, M., and Karjalainen, K. (1997). Targeted ex- P., Vivier, E., and Malissen, B. (1995). Altered T cell development inpression of major histocompatibility complex (MHC) class II mole- mice with a targeted mutation of the CD3-epsilon gene. EMBO J.cules demonstrates that dendritic cells can induce negative but not 14, 4641–4653.positive selection of thymocytes in vivo. J. Exp. Med. 185, 541–550.

Markowitz, J.S., Rogers, P.R., Grusby, M.J., Parker, D.C., andChang, C.H., Guerder, S., Hong, S.C., van Ewijk, W., and Flavell, Glimcher, L.H. (1993). B lymphocyte development and activationR.A. (1996). Mice lacking the MHC class II transactivator (CIITA) independent of MHC class II expression. J. Immunol. 150, 1223–show tissue-specific impairment of MHC class II expression. Immu- 1233.nity 4, 167–178.

Matsunaga, M., Seki, K., Mineta, T., and Kimoto, M. (1990). Antigen-Cosgrove, D., Gray, D., Dierich, A., Kaufman, J., Lemeur, M., Benoist, reactive T cell clones restricted by mixed isotype A beta d/E alphaC., and Mathis, D. (1991). Mice lacking MHC class II molecules. Cell d class II molecules. J. Exp. Med. 171, 577–582.66, 1051–1066.

Mineta, T., Seki, K., Matsunaga, M., and Kimoto, M. (1990). ExistenceEhlich, A., Matin, V., Muller, W., and Rajewsky, K. (1994). Analysis of mixed isotype A beta E alpha class II molecules in Ed alpha gene-of the B-cell progenitor compartment at the level of single cells. introduced C57BL/6 transgenic mice. Immunology 69, 385–390.Curr. Biol. 4, 573–583.

Miyazaki, T., Dierich, A., Benoist, C., and Mathis, D. (1996). Indepen-Forster, I., and Rajewsky, K. (1990). The bulk of the peripheral B-cell dent modes of natural killing distinguished in mice lacking Lag3.pool in mice is stable and not rapidly renewed from the bone marrow. Science 272, 405–408.Proc. Natl. Acad. Sci. USA 87, 4781–4784.

Newton, K., and Strasser, A. (1998). The Bcl-2 family and cell deathGrusby, M.J., Johnson, R.S., Papaioannou, V.E., and Glimcher, L.H. regulation. Curr. Opin. Genet. Dev. 8, 68–75.(1991). Depletion of CD41 T cells in major histocompatibility complex

Oka, Y., Rolink, A.G., Andersson, J., Kananaka, M., Uchida, J., Yasui,class II-deficient mice. Science 253, 1417–1420. T., Kishimoto, T., Kikutani, H., and Melchers, F. (1996). ProfoundGu, H., Forster, I., and Rajewsky, K. (1992). Study of murine B-cell reduction of mature B cell numbers, reactivities and serum immuno-development through analysis of immunoglobulin variable region globulin levels in mice which simultaneously carry the XID and CD40-genes. Ann. N.Y. Acad. Sci. 651, 304–310. deficiency genes. Int. Immunol. 8, 1675–1685.Hardy, R.R., Hayakawa, K., Parks, D.R., and Herzenberg, L.A. (1983). Osmond, D.G. (1991). Proliferation kinetics and the lifespan of BDemonstration of B-cell maturation in X-linked immunodeficient cells in central and peripheral lymphoid organs. Curr. Opin. Immunol.mice by simultaneous three-colour immunofluorescence. Nature 3, 179–185.306, 270–272. Rahemtulla, A., Fung-Leung, W.P., Schilham, M.W., Kundig, T.M.,Hardy, R.R., Carmack, C.E., Shinton, S.A., Kemp, J.D., and Haya- Sambhara, S.R., Narendran, A., Arabian, A., Wakeham, A., Paige,kawa, K. (1991). Resolution and characterization of proB and pre- C.J., Zinkernagel, R.M., et al. (1991). Normal development and func-pro B cell stages in normal mouse bone marrow. J. Exp. Med. 173, tion of CD81 cells but markedly decreased helper cell activity in1213–1225. mice lacking CD4. Nature 353, 180–184.Huard, B., Prigent, P., Tournier, M., Bruniquel, D., and Triebel, F. Rolink, A., Grawunder, U., Winkler, T.H., Karasuyama, H., and Melch-(1995). CD4/major histocompatibility complex class II interaction ers, F. (1994). IL-2 receptor a chain (CD25,TAC) expression definesanalyzed with CD4- and lymphocyte activation gene-3 (LAG-3)-Ig a crucial stage in preB cell development. Int. Immunol. 6, 1257–1264.fusion proteins. Eur. J. Immunol. 25, 2718–2721. Rolink, A.G., Melchers, F., and Andersson, J. (1996). The SCID butKenty, G., Martin, W.D., Van Kaer, L., and Bikoff, E.K. (1998). MHC not the RAG-2 gene product is required for Sm-Se-heavy chain classclass II expression in double mutant mice lacking invariant chain switching. Immunity 5, 433–443.and DM functions. J. Immunol. 160, 606–614. Rolink, A.G., Melchers, F., and Andersson, J. (1998). Characteriza-Khan, W.N., Alt, F.W., Gerstein, R.M., Malynn, B.A., Larsson, I., Rath- tion of immature B cells by a novel monoclonal antibody, by turnover

and by mitogen reactivity. Eur. J. Immunol. 28, 3738–3748.bun, G., Davidson, L., Muller, S., Kantor, A.B., Herzenberg, L.A., et

Immunity628

Ruberti, G., Paragas, V., Kim, D., and Fathman, C.G. (1993). Selectionfor amino acid sequence and J beta element usage in the beta chainof DBA/2V beta b- and DBA/2V beta a-derived myoglobin-specificT cell clones. J. Immunol. 151, 6185–6194.

Scher, I. (1982a). CBA/N immune defective mice: evidence for thefailure of a B cell subpopulation to be expressed. Immunol. Rev.64, 117–136.

Scher, I. (1982b). The CBA/N mouse strain: an experimental modelillustrating the influence of the X-chromosome on immunity. Adv.Immunol. 33, 1–71.

Shachar, I., and Flavell, R.A. (1996). Requirement for invariant chainin B cell maturation and function. Science 274, 106–108.

Tanchot, C., Lemonnier, F.A., Perarnau, B., Freitas, A.A., and Rocha,B. (1997). Differential requirements for survival and proliferation ofCD8 naive or memory T cells. Science 276, 2057–2062.

Tarlinton, D. (1993). Direct demonstration of MHC class II surfaceexpression on murine pre-B cells. Int. Immunol. 5, 1629–1635.

Ten Boekel, E., Melchers, F., and Rolink, A. (1995). The status of Igloci rearrangements in single cells from different stages of B celldevelopment. Int. Immunol. 7, 1013–1019.

Torres, R.M., Flaswinkel, H., Reth, M., and Rajewsky, K. (1996).Aberrant B cell development and immune response in mice with acompromised BCR complex. Science 272, 1804–1808.

Veis, D.J., Sorenson, C.M., Shutter, J.R., and Korsmeyer, S.J. (1993).Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis,polycystic kidneys, and hypopigmented hair. Cell 75, 229–240.

Widera, G., Burkly, L.C., Pinkert, C.A., Bottger, E.C., Cowing, C.,Palmiter, R.D., Brinster, R.L., and Flavell, R.A. (1987). Transgenicmice selectively lacking MHC class II (I-E) antigen expression on Bcells: an in vivo approach to investigate Ia gene function. Cell 51,175–187.

Woodland, R.T., Schmidt, M.R., Korsmeyer, S.J., and Gravel, K.A.(1996). Regulation of B cell survival in xid mice by the proto-onco-gene bcl-2. J Immunol 156, 2143–2154.

Yamamura, K., Kikutani, H., Folsom, V., Clayton, L.K., Kimoto, M.,Akira, S., Kashiwamura, S., Tonegawa, S., and Kishimoto, T. (1985).Functional expression of a microinjected Ed alpha gene in C57BL/6transgenic mice. Nature 316, 67–69.