An Essential Role for BAFF in the Normal Development of B Cells Through a BCMA-Independent Pathway Barbara Schiemann, Jennifer L. Gommerman, Kalpit Vora, Teresa G. Cachero, Svetlana Shulga-Morskaya, Max Dobles, Erica Frew, Martin L. Scott

Supplementary Material

Supplemental Figure 1. Generation of BAFF^-/- and BCMA^-/- mice. (A) BAFF and (B) BCMA knockout strategies, with diagrams of wild type genomic loci, knockout constructs, and targeted loci are shown, with predicted sizes of normal and mutant alleles. Dark, thick vertical lines represent exons; blue horizontal lines represent intronic DNA. Thin, horizontal black lines show the lengths of the wild type (wt) and knockout (KO) alleles identified by digestion of ES cell DNAs with EcoRV, detected by hybridization with probes derived from the indicated regions. Southern blots (marked ES) identifying heterozygous ES cells are shown for each locus. The panel marked PCR shows the products of multiplex PCR reactions that span the initiating ATGs of each wild type locus and the tailless human CD2 reporter gene introduced as part of the knockouts (1). These reactions can simultaneously detect wild type and knockout alleles. Northern blots (marked RNA) were done using coding sequence probes from the deleted BAFF (A) or BCMA (B) alleles. mRNA sizes are 1.8 kb (BAFF) and 0.9 kb (BCMA). RNA was purified from splenocytes and mice of the following genotypes: 1=wild type, 2=BCMA knockout, 3=BAFF knockout. For the BAFF knockout, a 1.8 kb Pst1 fragment encoding amino acids 1-43 of BAFF was inserted into pNNO3. The resulting plasmid, pSAB316, was targeted by ET cloning (2). The PCR product employed included 55 nt from the sequence just upstream of the initiating ATG, adding in frame DNA sequence that encoded a tailless human CD2 reporter (1), a neomycin resistance cassette, and part of the pNNO3 polylinker. Subsequent cloning steps and the sequence of the construct are available on request. For the BCMA knockout, an 8.9 kb HindIII fragment containing the entire coding region for mouse BCMA was subcloned into pBSK(+). This fragment was targeted by ET cloning (2) using a PCR product that included 64 nucleotides upstream of the initiating ATG, the same hCD2 and neomycin resistance cassette described above, an additional EcoRV site just downstream of the neo marker, and 60 additional nucleotides of downstream genomic homology. A second subclone of the locus, a 5.5 kb EcoR1 genomic fragment containing the initiating ATG in pET17b, was then digested with HindIII. Into it was ligated the HindIII fragment from the previously prepared, hCD2-containing plasmid. This step yielded a knockout construct (pBJSK4) that contained approximately 2 kb more 5' sequence homology than the original HindIII fragment. The complete sequences of the HindIII portion of the targeting construct and the probe are available on request. Electroporation of D3 ES cells with linearized targeting constructs, G418 selection, and identification of homologous recombinant clones was carried out by standard methods. Homologous recombinant clones containing each disrupted allele were subcloned, karyotyped, and injected into C57Bl/6 blastocysts. Chimera breeding to C57Bl/6 mice yielded heterozygous mice from two BAFF and two BCMA clones. Crosses of these animals yielded knockout BAFF ^-/- or BCMA ^-/- mice that were outwardly normal and survived at least 6-8 months of age without unusual morbidity. Heterozygote matings yielded homozygous-null mice at the expected Mendelian ratios (0.25, n> 100 for each). At necropsy, all major organs including thymus, spleen, and lymph node were present although average spleen weights of BAFF knockout animals were significantly reduced (wild type spleen 89 mg ± 16 mg; BAFF KO 46 mg ± 11 mg, p<0.05; BCMA KO 93 mg ± 13 mg).

Medium version | Full size version

---

Supplemental Figure 2. Immunohistochemistry of wild type and BAFF knockout spleens. The upper row of photos shows wild-type (wt) tissue sections, the lower row shows BAFF^-/- (BAFF KO) tissue sections. Antigens detected are indicated below each column of photos. (A) Spleen dendritic cells were stained for CD11c (brown), T cells for CD4 and CD8 (blue). (B) Spleen antigen presenting cells were stained for MAdCAM-1 (brown, ring-shaped structures) and B cells for B220 (blue). (C) Spleen follicular dendritic cells were stained for FDC-M1 (blue). Staining was carried out on spleen sections obtained 6 days after i.p. immunization with 100 g of NP-KLH. Primary antibodies from Pharmingen were used at a 1:50 dilution. In panel A, they included anti-CD4-FITC (RM4-5), anti-CD8a-FITC (53-6.7), anti-CD8b.2-FITC (53-5.8), and anti-CD11c-biotin (HL3) ; these were detected with anti-FITC-AP (Roche 1 426 338) and streptavidin-HRP (Jackson Immunoresearch, 016-030-084) as described in the references for the main text. In panel B, the primary antibodies were anti-CD45R/B220-FITC (RA3-6B2) and anti-MAdCAM-1-biotin (MECA 89); the detecting reagents and conditions were the same as in A. In panel C, the primary antibody was FDC-M1 culture supernatant kindly provided by J. Tew; it was detected with biotin-coupled mouse F(ab')2 anti-rat Ig as described in the references for the main text.

Medium version | Full size version

---

Supplemental Figure 3. FACS analysis. Cells from 7 week old mice wild type (marked wt), BAFF ^-/- (BAFF KO), or BCMA ^-/- (BCMA KO) mice were stained for the antigens indicated in each panel, then analyzed after electronic scatter gating. The numbers within each panel are percentages of gated events. (A) Reduced splenic B cells (B220⁺; rectangular region) from BAFF^-/-, but not from BCMA^-/- mice, are shown in comparison to wild type cells. (B) Reduced mature splenic B cells (B220⁺, IgD⁺, IgM^b; upper left rectangular region) from BAFF^-/-, but not from BCMA^-/- mice, are shown in comparison to wild type cells. Immature B cells (B220⁺, IgD^b, IgM^hi; right rectangular region) are similar among all three samples. (C) Normal bone marrow pro-B cells (B220+, CD43⁺; right rectangular region) and immature cells (B220⁺, CD43^-; left rectangular region) from wt, BAFF^-/-, and BCMA^-/- mice are shown. The immature population defined by this stain also normally contains a small proportion of mature recirculating B cells (B220^hi, IgD^hi, IgM^lo) cells that are absent from BAFF knockout marrow (Supplemental Fig. 3D). (D) Reduced bone marrow mature recirculating B cells (B220⁺, IgD⁺, IgM^b; rectangle) from BAFF^-/-, but not from BCMA^-/- mice, are shown in comparison to wild type cells. The plots shown are gated for both scatter and B220-positivity. The percentages given are derived from the number of scatter gated cells.

Medium version | Full size version

---

The staining reagents used were obtained from Pharmingen and included the following: anti-B220-PE and anti-B220-CyC (RA3-6B2), anti-IgM-biotin (R6-60.2), anti-CD1d-PE (1B1), anti-CD3-PE (145-2C11), anti-CD4-FITC and anti-CD4-biotin (GK 1.5), anti-CD8-FITC (53-6.7), anti-CD11b-FITC (M1/70), anti CD11c-PE and anti CD11c-FITC (HL3), anti-CD21-FITC (7G6), anti-CD23-biotin and anti-CD23-FITC (B3B4), anti-CD24-FITC (M1/69), anti-CD43-PE (S7), anti-Ly76-PE (TER 119), anti-GR-1-biotin (RB6-8C5), anti-TCR-APC (H57-597), anti-CD5-CyC (53-7.3), anti CD16/32 (2.4G2), and streptavidin-APC. Anti-IgD-PE (11-26) was obtained from Southern Biotechnology Associates. FACS staining was carried out on red-cell depleted (ammonium-chloride) samples from marrow, spleen, and thymus per protocols described elsewhere (www.pharmingen.com).

---

Supplemental Figure 4. Resting serum immunoglobulin (Ig) levels. Concentrations of the indicated Ig isotypes were determined by ELISAs of sera from unimmunized 6-9 week old wild type (wt), BAFF^± (het), and BAFF^-/- (KO) mice. Each symbol represents a single animal of the genotype indicated on the X-axis. The Y-axis of each plot indicates Ig concentrations determined by parallel dilution of standards obtained from Southern Biotechnology Associates. Horizontal bars within each panel indicate the means for each genotype.

Medium version | Full size version

---

Supplemental Figure 5. A model for BAFF signaling. This figure, showing possible stages at which BAFF signaling may occur, is modified from Martin and Kearney's review (3). It seems clear that BAFF is necessary for progression to or survival of T2 transitional B cells, but not B1 cells. If the B1 subset develops as a genetically determined separate lineage, then BAFF independence may be an important functional attribute of these cells. If they are selected from existing early B2 cells, our data show that this likely occurs at the T1 stage or earlier. Based on the A/WySnJ phenotype, BAFF-R is likely required for T2 B cell survival. However, BAFF signaling through TACI alone is not sufficient to maintain MZ cells in these mice. Therefore, the impairment of T-independent antibody responses in TACI-deficient mice (4) may therefore involve either an additional ligand or a different function of BAFF that is mediated through TACI. BCMA, which is preferentially expressed late in B cell ontogeny, may still act as a BAFF receptor or perhaps through another mechanism. The role of APRIL may be a more general one involving proliferation at different stages of development or in sites other than the spleen.

Medium version | Full size version

---

Supplemental Table 1. B and T cells in the spleens of wild type, BAFF-/-, and BCMA-/- mice.
animal	genotype	Spleen cell number*	%B cells	B cell number*	%T cells	T cell number*
134.3f	wt	13.0	57.9	7.5	36.1	4.7
134.2f	wt	16.5	53.1	8.8	38.9	6.4
144.1f	wt	11.4	54.2	6.2	41.3	4.7
144.2f	wt	8.7	49.0	4.3	47.9	4.2
	Mean**	12.4 (3.3)	53.6 (3.7)	6.7 (1.9)	41.1 (5.0)	5.0 (1.0)
108.1f	BAFF	4.4	22.3	1.0	54.5	2.4
108.4f	BAFF	3.3	15.8	0.5	59.4	2.0
108.3f	BAFF	3.6	26.9	1.0	63.8	2.3
108.5f	BAFF	7.5	22.1	1.7	67.0	5.0
	Mean**	4.7 (1.9)	21.8 (4.6)	1.0 (0.5)	61.2 (5.4)	2.9 (1.4)
144.4f	BCMA	10.3	70.2	7.2	26.5	2.7
144.5f	BCMA	8.5	60.4	5.1	36.5	3.1
143.1f	BCMA	10.8	55.6	6.0	30.1	3.3
143.4f	BCMA	12.8	48.8	6.2	42.4	5.4
	Mean**	10.6 (1.8)	58.8 (9.0)	6.2 (0.9)	33.9 (7.0)	3.6 (1.2)
*10-7
**Mean (standard deviation)

---

References

1. I. Riviere, M. J. Sunshine, D. R. Littman, Immunity 9, 217 (1998).

2. Y. Zhang, J. P. Muyrers, G. Testa, A. F. Stewart, Nat Biotechnol 18, 1314 (2000).

3. F. Martin, J. F. Kearney, Curr Opin Immunol 13, 195 (2001).

4. G. von Bulow, J. M. van Deursen, R. J. Bram, Immunity 14, 573 (2001).