A Developmental Switch in B Lymphopoiesis (B-Cell Development/Bone Marrow/CD51 B Cells/Fetal Liver/Hematopoiesis) R

Home , Lymphopoiesis

Proc. Nati. Acad. Sci. USA Vol. 88, pp. 11550-11554, December 1991 Immunology A developmental switch in B lymphopoiesis (B-cell development/bone marrow/CD51 B cells/fetal liver/hematopoiesis) R. R. HARDY* AND K. HAYAKAWA Institute for Cancer Research, Fox Chase Cancer Center, 7701 Burholme Avenue, Philadelphia, PA 19111 Communicated by Robert P. Perry, September 30, 1991 (receivedfor review August 6, 1991)

ABSTRACT B and T lymphocytes are generated from until after the generation of an IgM' cell. To clarify these hematopoietic stem cells during both fetal and adult life. A issues, our strategy has been first to identify B-cell progen- critical unresolved issue is whether the differentiation path- itors in fetal liver and adult bone marrow and then to ways in lymphopoiesis are the same in fetal and adult animals investigate whether they show a predisposition to generating or whether they differ, similar to the hemoglobin switch in phenotypically distinct types of B cells. erythropoiesis. We report here that a developmental switch We recently found (15) that immature B-lineage cells in occurs in B lymphopoiesis. We isolated "pro-B" cells (i.e., cells mouse bone marrow, defined as cells bearing the B-cell- that have initiated, but not completed, heavy-chain gene rear- restricted form of the common leukocyte antigen (CD45R, rangement) from fetal and adult sources and investigated their B220) (16) but lacking surface IgM expression, could be B-cell progeny generated both in vitro and in vivo. Most of the fractionated further based on expression of sialophorin cells from fetal liver, but few from adult bone marrow, (CD43). CD43 is a glycoprotein present on peripheral T, but expressed CD5. Further, fetal pro-B cells failed to generate not B, cells (17). However, the earliest B-lineage cells in bone cells expressing high levels of IgD in severe combined immu- marrow do express CD43 and so are resolvable from later nodeficiency mice, whereas adult pro-B cells gave rise to CD5- stage pre-B and B cells as a B220'CD43' fraction, consti- B cells bearing IgD at levels comparable to the bulk of cells in tuting 3% of bone marrow. Cells in this fraction express the spleen of adult mice. Thus, all committed B progenitors in variable levels of another surface molecule, the heat stable fetal liver of day 16 gestation mice give rise to phenotypically antigen (HSA) (18). Among them, HSA' cells in bone mar- distinct progeny when compared to cells at a comparable row proliferate rapidly in Whitlock-Witte culture on a sup- differentiation stage in the bone marrow of adult animals. We port stromal layer where they give rise to IgM' cells. conclude that the cohort of B-lineage progenitors in early fetal The status of the immunoglobulin heavy-chain gene loci of development is committed to a differentiation pathway distinct cells in this fraction was determined by use of a PCR assay from that seen in the adult. (15). We amplified DNA segments that are known to be deleted upon rearrangement, either diversity (D) tojoining (J) Developmentally regulated differences in the cells that pre- or variable (V) to DJ. This allowed us to classify the dominate in the hematopoietic system during fetal versus B220+CD43+HSA+ fraction in bone marrow as "pro-B": it adult life have been observed in both the erythroid and contained cells with extensive D-J, but not V-DJ, rearrange- lymphoid lineages. Erythrocytes produced during the fetal ments. We have now resolved a comparable fraction in fetal stage express the "i" cell surface antigen and possess a liver and compared the progeny of these two pro-B popula- distinctive fetal hemoglobin (1-5). T cells bearing the vS tions both in vitro and in vivo. Intriguingly, the B cells T-cell receptor class are abundant in the fetal stage, but are generated by these developmentally distinct precursors show rare in the adult (6-9). Phenotypic and functional differences striking phenotypic differences: the progeny of bone marrow in B cells present in early and adult stages of ontogeny have pro-B resemble the bulk of B cells found in adult mice, been documented (10-12). Most B cells in neonatal mice whereas those of the fetal pro-B resemble a small subpop- express only low levels of IgD, unlike the bulk of B cells in ulation of B cells found in the adult, many of which bear the the adult, which bear high levels (12). Furthermore, CD5 CD5 molecule. expression is more frequent on B cells found early in ontogeny (13). MATERIALS AND METHODS It has been unclear whether these differences in B-cell phenotype reflect a stable feature of cells generated at Animals. Fetal liver was obtained from timed matings of different times in ontogeny or instead whether most B cells BALB/cAnN mice. Bone marrow was obtained from 3- to in the neonate are simply intermediates that have not yet 4-month-old BALB/cAnN female mice. Two- to four-month- matured into the adult type. Our previous cell transfer old C.B-17 scid female mice [severe combined immunodefi- experiments showed that liver from newborn mice was much ciency (SCID) mice] were used for cell transfer recipients. All more effective in reconstituting CD5+ B cells when compared mice were bred in our animal facility. to bone marrow from adult mice (14), suggesting that CD5+ Determination of Immunoglobulin Gene Rearrangement in B cells might arise from distinct precursors present in fetal Pro-B Cell Fractions by PCR. Single-cell suspensions of bone liver but absent from adult bone marrow. However, since the marrow or fetal liver were stained simultaneously with flu- IgDhigh B-cell population (IgD"+ B) was also generated in orescein-labeled anti-CD43 (S7), phycoerythrin-labeled anti- these newborn liver transfers, it was not established whether B220 (RA3-6B2), allophycocyanin-labeled anti-Thy-1.2 both types of differentiation were occurring simultaneously (30H12), and biotin-labeled anti-HSA (30F1); the biotin re- in the liver. Furthermore, the stage at which commitment to agent staining was revealed by a second-step incubation with a particular B-cell phenotype (such as CD5+) has not been defined. Conceivably, this commitment might not take place Abbreviations: FACS, fluorescence-activated cell sorter; HSA, heat stable antigen; SCID, severe combined immunodeficiency; PC, phosphatidylcholine; V, variable; D, diversity; J,joining; PerC, PerC The publication costs of this article were defrayed in part by page charge cells, cells washed out ofthe peritoneal cavity; BrMRBC, bromelain- payment. This article must therefore be hereby marked "advertisement" treated mouse erythrocytes. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 11550 Downloaded by guest on September 27, 2021 Immunology: Hardy and Hayakawa Proc. Nadl. Acad. Sci. USA 88 (1991) 11551 Texas Red-avidin as described (15). Sort gates are drawn on clonal proliferation and the effects of differential cell sur- the plots in Fig. 1, which present contours enclosing equal vival, thereby providing a more uniform progeny. Pre-B cells probabilities of cells (5%). Fig. 1 Inset shows the HSA (B220'CD43-; ref. 15) were absent from the bone marrow of distribution, which is also gated on an intermediate level of recipients >2 months after cell transfer (data not shown). expression; any anti-Thy-i-stained cells (data not shown) These animals differed from those reconstituted with unsep- were excluded. Sorted cells represent 1% offetal liver and 2% arated or hemopoietic stem cell-enriched fractions where of bone marrow. Reanalysis of sorted fractions showed continuous generation of B- and T-lineage cells (and other purities in excess of95%. DNA was prepared from 1-2 x 10- cell types) from precursors has been suggested (20). Conse- cells, and regions 5' of DFL16.1, 5' of heavy-chain joining quently, we find that using pro-B cells requires 100-1000 region 1 (JH1), and within the actin gene were amplified by times more cells (104-10W cells) to obtain significant levels using oligonucleotide primers described previously (15). (>5%) of peripheral B-cell generation as compared to that One-fifth of the total sample was used for a PCR reaction. seen with stem cell-enriched fractions. Analysis with allo- Ten-microliter aliquots (one-tenth ofthe reaction) were with- type-specific anti-IgM reagents (anti-IgMa, RS3.1; anti-IgMb, drawn after 18 and 22 cycles, size fractionated on 1.5% AF6-78) showed that all B cells were ofdonor BALB/c origin agarose gels, blotted, and then probed and scanned to reveal (data not shown). levels of DNA produced. Probes used were generated by Analysis ofAntigen Binding by Flow Cytometry. Phosphati- cloning the PCR products into the Sma I site of pBSM13- dylcholine (PC)-containing vesicles incorporating the fluo- (Stratagene), which permitted generation of high specific rescent dye Texas Red were produced following a published activity RNA transcripts. DNA from adult liver served as a procedure (32). Cells washed out of the peritoneal cavity nonrearranged standard (for determination of percent reten- from SCID mice reconstituted with pro-B cells from either tion of the germ line). DNA from two Abelson murine fetal liver or adult bone marrow were incubated with the leukemia virus-transformed lines (1-8 with VDJ/DJ and 3-1 fluorescent vesicles, together with fluorescein-labeled anti- with VDJ/VDJ) was used to demonstrate the specificity of IgM and phycoerythrin-labeled anti-CD5. After washing, the assay. The reliability of this method for determining the cells were analyzed by fluorescence-activated cell sorting rearrangement status of immunoglobulin genes was evalu- (FACS). Cells falling within a "Iymphoid gate" (excluding ated in a previous publication from this laboratory (15). small debris and granular cells) were then analyzed for Pro-B Stromal Cultures. Pro-B cells (B220CD43+HSA+) vesicle binding. PC vesicle binding to B cells is specific, since were sorted onto preestablished layers of the FLST2 line as vesicles generated in a similarfashion, but lacking PC, are not described (15). Medium (RPMI 1640 supplemented with 50 bound (data not shown). juM 2-mercaptoethanol and 5% fetal calf serum) was replen- ished at 4-day intervals. Cells recovered from the stromal layer after 10-14 days were stained with fluorescein-labeled RESULTS anti-IgM (331.12) and allophycocyanin-labeled anti- Definition ofPro-B Cells in Fetal Liver. B220+ cells are first CD5(53-7) plus propidium iodide (1 Ag/ml) to exclude dead detected at significant levels (1-2%6) in the fetal liver of day cells and then analyzed on the cell sorter. 15-16 gestation mice (Fig. 1). These B220+ cells all express Pro-B Cell Transfers. Cells (105) in the pro-B fractions from CD43 at levels similar to that seen in bone marrow, whereas day 16 fetal liver or adult bone marrow of BALB/c mice pre-B or B cells are very rare (

A Day 16 FL 3 Month BM B

'6,s

CD 10 coC\l 5 of D 0 C\Jcom 1 Actin AWOL- 5 Of JH .1 1 10 100 .1 1 10 100 CD43 >

FiG. 1. Isolation ofpro-B cell fractions from day 16 fetal liver (FL) and adult bone marrow (BM). (A) Multiparameter FACS analysis ofthese tissues resolves a set ofcells with a comparable phenotype, B220+CD43+. (Inset) Histogram of HSA distribution within this population. Sorting gates for pro-B sorting are drawn on the contour plots as is the intermediate HSA' level (15). (B) Autoradiograph of PCR analysis of deletion of DNA segments between V and D or D and J. Although the 5' region of the D band in this particular sample of adult liver (germ line) DNA appears to be less intense than those in fetal liver and bone marrow pro-B samples, this was not routinely observed in other experiments (see Table 1). HO, a germ line heavy-chain gene. Downloaded by guest on September 27, 2021 11552 Immunology: Hardy and Hayakawa Proc. Natl. Acad Sci. USA 88 (1991) Table 1. D-J, but not V-DJ, rearrangement in FL Pro B BM Pro B B220'CD43'HSA+ cell fraction -CD5or Relative retention 6c 75 control| 75- of germ-line DNA

DNA sample 5' region of D 5' region of J i; 50- 50- FL pro B 1.00 + 0.02 0.45 + 0.02 ( 25 BM pro B 1.00 ± 0.02 0.19 ± 0.03 a: 25 1-8 (VDJ/DJ) 0.53 ± 0.05 0.02 ± 0.01 ± ± 0.01 3-1 (VDJ/VDJ) 0.04 0.01 0.01 0.1 1 10 100 0.1 1 10 100 The intensities of PCR-amplified DNA bands in samples sorted Fluorescence intensity from fetal liver (FL), bone marrow (BM), and two Abelson murine leukemia virus lines (1-8 and 3-1) are expressed as ratios of the FIG. 2. Expression of CD5 on IgMI cells generated after a corresponding bands in adult liver samples. Values represent means 2-week culture offetal liver (FL) and adult bone marrow (BM) pro-B + SE for 14 determinations (from five separate PCR amplifications). cells. Histograms are gated for IgM' cells (15-30% of total cells). Control staining is with an allophycocyanin-labeled conjugate of an This is demonstrated by PCR analysis quantitating deletion of irrelevant specificity. DNA segments between V and D or between D and J. A representative autoradiograph (2-h exposure) shows disap- marrow-derived B cells do not. This result shows that fetal pearance of the 5' JH band, with retention of 5' D band. pro-B cells differ from adult pro-B cells in their potential to Numerical values shown in Table 1 are the average and generate distinct phenotypes of B cells in short-term culture. standard error from 14 determinations (five PCR amplifica- Differentiative Potenthl ofPro-B Cels in Vivo. Cell transfer tions). Since cells in this fraction have initiated, but not into irradiated immunodeficient SCID mice (23) provides an completed, heavy-chain gene rearrangement, we classify environment for lymphoid differentiation, one that permits them as pro B. generation of mature cells (ref. 24, unpublished results). We DNA isolated from this fraction from day 16 gestation liver therefore transferred limiting numbers of pro-B fractions shows somewhat less D-J rearrangement than thatfrom adult from fetal and adult BALB/c mice into SCID mice and then bone marrow. However, expression of BP-1/6C3 (data not examined the phenotype of lymphocytes generated in the shown), a cell surface molecule that appears after initiation of periphery 3 weeks after cell transfer (Fig. 3). Since distinct D-J rearrangement, suggests that this day 16 fraction corre- subpopulations of B cells are identifiable in spleen and sponds to the earlier (BP1/6C3-) pro-B cell stage in bone peritoneal cavity (19), we focused our analysis on these marrow (15). This interpretation is further supported by the locations. FACS analysis of spleen and peritoneal cavity fact that we detect less rearrangement (and fewer BP1/6C3' (PerC) cells revealed B-cell progenitor activity: IgM' B cells cells) on day 15 and more rearrangement (together with a could be readily found in the recipients, whereas T cells higher proportion of BP1/6C3' cells) on day 17. Therefore, (CD5+IgM-) were undetectable (<1%) in spleen and perito- these B220'CD43' cells in fetal liver appear to progress as a neal cavity (Fig. 3 A, B, E, and F), and the thymus was not cohort through B-lineage differentiation. reconstituted (data not shown). These transferred pro-B cells Differentiative Potential of Pro-B Cells in Vito. We have differentiate in the recipients and do not establish a self- compared the differentiative potential of this pro-B cell replenishing pro-B population in spleen or bone marrow. fraction from fetal liver with that of pro-B cells in the bone Similar numbers of B cells were recovered in the two types marrow ofadult animals. First we employed a defined in vitro of transfers from both locations (see Table 2). All recipients system. Early B-lineage cells can undergo limited B-cell receiving day 16 fetal pro-B cells generated CD5' B cells in differentiation when cultured on certain cloned adherent both spleen and peritoneal cavity (Fig. 3 A and B). Further, stromal cell lines (21, 22). B220'CD43+HSA' (pro-B) cells most B cells in these recipients express low levels ofIgD (Fig. from both adult and fetal sources generate IgM' B cells at 3 C and D). In contrast, CD5' B cells were not generated similar frequencies during short-term culture on a stromal using adult pro-B cells (Fig. 3 E and F), and most B cells in layer. However, as shown in Fig. 2, the majority of fetal- these animals expressed high levels of IgD (Fig. 3 G and H). derived B cells express CD5, whereas the majority of bone Thus, as is clear in Fig. 4, two distinctive B-cell profiles FL Pro B Transfer Adult BM Pro B Transfer Spleen PerC Spleen PerC 7 F

LO FIG. 3. B cells repopulated in SCID mice 3 weeks after transfer 0 of day 16 fetal liver (FL) or adult bone marrow (BM) pro-B frac- W - ~n i tions show distinct IgM, IgD, and CD5 phenotypes. Cell suspensions prepared from spleen (Spleen) or cells washed out ofthe 10c 1 D-- peritoneal cavity (PerC cells) were analyzed by flow cytometry. C) 1 _} Boxes show the gates used to de- 1 43i ) !~~~~~ fine CD51 B cells in CD5/IgM ( analyses or IgD" B cells in IgD/ 1 _ IgM analyses. Representative .1 data from numerous analyses (see IgM --- -- ::s Fig. 4 for distribution) are shown. Downloaded by guest on September 27, 2021 Immunology: Hardy and Hayakawa Proc. Natl. Acad. Sci. USA 88 (1991) 11553 Table 2. Reciprocal repopulation of CD5+IgDVow and CD5-IgDEhi" B cells Phenotype of reconstituted B cells Transferred cells Tissue CD5+ CD5- IgD++ IgD+ Fetal liver pro B Spleen 4.9 (0.6) 6.3 (0.8) 1.5 (0.2) 7.3 (0.9) PerC 45.7 (2.2) 27.1 (1.5) 9.3 (0.6) 52.5 (2.4) Bone marrow pro B Spleen 0.4 (0.1) 5.9 (0.9) 5.9 (0.9) 1.8 (0.3) PerC 3.5 (1.2) 52.7 (2.7) 43.9 (2.7) 13.3 (1.2) Percentages of B cells (SE in parentheses) reconstituted in SCID mice 3 weeks after pro-B transfers. Number of independent determinations: 28 for fetal liver and 15 for bone marrow transfer using 105 pro-B cells. The percentages of cells reported within a lymphoid gate (excluding erythrocytes and granular cells) are as follows: for fetal liver transfer, 48.6% (±1.1%) and 27.0%o (±1.3%) of total cells fall within a lymphoid gate for spleen and peritoneal cells, respectively (mean and SE); for bone marrow transfers, the corresponding values are 47.0%o (±1.2%) and 18.1% (±1.4%). Cell recovery was typically 1-2 x 107 for spleen cells and 1-3 x 106 for PerC cells in both kinds of transfer. Cells were analyzed as in Fig. 3. emerged in both sites: CD5+IgDlow from the fetal pro-B cell sulting in the accumulation of B cells with these specificities. transfer and CD5-IgDhJgh from the adult pro-B cell transfer. A striking example ofthis is the high frequency ofB cells with Data from analyses of mice long after cell transfer (2-5 reactivity to a determinant(s) on mouse erythrocytes re- months) yielded essentially the same conclusion: no IgD"i& vealed by treatment with the proteolytic enzyme bromelain B-cell generation by fetal pro-B and no CD5' B-cell gener- (BrMRBC) (26-30). These BrMRBC-specific B cells can also ation by adult pro-B transfer (data not shown). In summary, be recognized by their specific ability to bind vesicles con- our data demonstrate that the differences in the two types of taining PC (28, 31), presumably one component of the anti- B cells generated from these two B-progenitor sources are genic determinant. As shown in Fig. 5, when analyzed 2 due to characteristics inherent in each B-cell progenitor. This months after transfer, significant numbers (10%) of CD5+ B in situ generation of distinct B progenitors appears to be cells in the fetal liver pro-B recipient mice bind PC vesicles determined by critical developmental timing, since we have (as was found with normal adult PerC cells). In contrast, the found that pro-B cells in the liver later than day 17 ofgestation CD5-IgD`+ B cells reconstituted from adult bone marrow already include B-cell progenitors that can generate IgDhi" B pro-B cells do not. In accord with these data, a high level of cells (data not shown). anti-BrMRBC antibody secretion (and anti-thymocyte au- An Autoantibody Specificity Characteristic of CD5+ B Cells toantibody, another specificity characteristic ofCD5+ B; ref. Arises from Fetal Liver Pro-B Cells. We have investigated 25) was obtained with cultures of PerC cells from fetal liver whether the B cells repopulated in SCID mice by pro-B cells pro-B recipients (data not shown). These results indicate that show functional similarities to the B cells present in situ. The pro-B cells isolated from fetal liver generate a B-cell popu- CD5+ B-cell population in adult mice shows enrichment of lation in SCID recipients possessing characteristic biases in certain autoreactive specificities (25-29), probably due to antibody specificities seen with CD5+ B cells. positive selection by auto-(or environmental-) antigens re- DISCUSSION Differences in the immune response correlated with ontogeny CD5eB/Total B IgD B/Total B have been observed in previous studies (11). Responsiveness 10..8.0 O Fetal Pro B to a variety ofantigens is more limited at early developmental A AdultProB times, and the induction of immunological tolerance is more easily obtained in the fetal through neonatal period (10, 33). In studies of B lymphocytes, functional differences in young 0.8- are of COo animals usually ascribed to the relative "immaturity" At 00 A A BM Pro-B FL Pro-B 0.6 0 0 -.i c0 LO QC ck 10 MA= o 8 0.4

~~~00 0 em PC 0.2- binding FIG. 5. Demonstration of a high frequency of anti-BrMRBC (PC binding) B cells in the CD5' B-cell fraction generated from fetal liver pro-B cells. Peritoneal washout cells from fetal liver (FL) or bone marrow (BM) pro-B repopulated SCID mice were incubated with Spl PerC Spi PerC fluorescein-labeled anti-IgM, phycoerythrin-labeled, anti-CDS and Texas Red fluorochrome-loaded PC vesicles (31, 32) and then FIG. 4. Scatter plot of data from repopulated SCID mice dem- analyzed*~ by FACS. This procedure has been shown to reveal onstrating distinctive phenotypes generated from fetal and adult antigen-specific binding via the immunoglobulin molecule (32). Ten pro-B cells. Ratios of CD5+ B cells out of total B cells and IgD++ percent of B cells generated in fetal liver pro-B repopulated individ- cells out of total B cells are derived from the type of analyses shown uals bound PC (most also expressed CD5) compared to <1% of B in Fig. 3 and summarized in Table 2. Data from 28 fetal pro-B and 15 cells in bone marrow pro-B-repopulated individuals. Figures are adult pro-B reconstituted individuals are shown (some points over- gated for IgM+ B cells. Representative data from four analyses are lap). Spl, spleen. shown. Downloaded by guest on September 27, 2021 11554 Immunology: Hardy and Hayakawa Proc. Natl. Acad. Sci. USA 88 (1991) the cells present in the lymphoid organs, with the implicit for Cancer Research, Fox Chase Cancer Center) and Drs. J. Erikson assumption that such cells have the potential to mature. and A. Caton (Wistar Institute) fora critical reading ofthis manuscript. However, the experiments described above demonstrate that This work was supported by grants from the National Institutes of fetal and adult pro-B cells show a reciprocal ability in gener- Health (CA-06927, RR-05539, AI-26782, and CA-37252), the American ating CD5' and IgDh'O B cells. Thus, we suggest that func- Cancer Society (IM-529), and the Pew Charitable Trust (86-5043HE) tional distinctions previously documented can be ascribed to and by an appropriation from the Commonwealth of Pennsylvania. physiological differences inherent in each type of B cell. 1. Hakomori, S. (1981) Semin. Hematol. 18, 39-61. In both mouse and humans, the 67-kDa glycoprotein 2. Papayannopoulou, T., Nakamoto, B., Manna, M., Lucarelli, G. & known as CD5 is largely restricted to T cells. In adult mice, Stamatoyannopoulos, G. (1986) Blood 67, 99-104. most B cells in the peripheral lymphoid organs lack CD5 and 3. Zanjani, E. D., Lim, G., McGlave, P. B., Clapp, J. F., Mann, L. I., bear high levels of IgD (CD5-IgD" 'h) (34), whereas CD5' B Norwood, T. H. & Stamatoyannopoulos, G. (1982) Nature (Lon- cells are found at readily detectable levels only in restricted don) 295, 244-246. anatomical sites-for example, the peritoneal cavity (19). 4. Wood, W. G., Bunch, C., Kelly, S., Gunn, Y. & Breckon, G. (1985) Nature (London) 313, 320-323. These CD5' B cells are also characterized by low levels of 5. Enver, T., Raich, N., Ebens, A. J., Papayannopoulou, T., Costan- IgD and exhibit functional differences from CD5-IgDO B tini, F. & Stamatoyannopoulos, G. (1990) Nature (London) 344, cells, which has led us to consider them a distinct B-cell 309-313. subset in mice, originally termed Ly-1 B (35). Our current 6. Havran, W. L. & Allison, J. P. (1988) Nature (London) 335, 443-445. findings demonstrate that CD5 expression on B cells arises 7. Ikuta, K., Kina, T., MacNeil, I., Uchida, N., Peault, B., Chien, frequently in the course offetal pro-B cell differentiation and Y. H. & Weissman, I. L. (1990) Cell 62, 863-874. that all progeny of B-lineage differentiation in early ontogeny 8. Ito, K., Bonneville, M., Takagaki, Y., Nakanishi, N., Kanagawa, O., Krecko, E. G. & Tonegawa, S. (1989) Proc. Natl. Acad. Sci. possess a distinctive surface phenotype. Thus, it appears USA 86, 631-635. likely that CD5 expression serves to mark a population of 9. Pardon, D. M., Fowlkes, B. J., Bluestone, J. A., Kruisbeek, A., fetal-derived B cells in a similar fashion as the y chain of Maloy, W. L., Coligan, J. E. & Schwartz, R. H. (1987) Nature hemoglobin marks a fetal-derived erythrocyte (4). (London) 326, 79-81. We found here that a characteristic specificity of CD5' B 10. Nossal, G. J. V. (1983) Annu. Rev. Immunol. 1, 33-62. cells, to BrMRBC (PC), is enriched in SCID mice repopulated 11. Klinman, N. R. & Press, J. L. (1975) Transplant. Rev. 24, 41-83. 12. Hardy, R. R., Hayakawa, K., Parks, D. R. & Herzenberg, L. A. with pro-B cells from fetal liver but not adult bone marrow. (1983) Nature (London) 306, 270-272. Thus, a functional CD5' B population is generated selec- 13. Hardy, R. R. & Hayakawa, K. (1986) Immunol. Rev. 93, 53-79. tively from fetal pro-B cells, suggesting that B cells generated 14. Hayakawa, K., Hardy, R. R., Herzenberg, L. A. & Herzenberg, early in ontogeny can persist through life as a part ofthe adult L. A. (1985) J. Exp. Med. 161, 1554-1568. immune system. However, the process whereby the CD5' B 15. Hardy, R. R., Carmack, C. E., Shinton, S. A., Kemp, J. D. & Hayakawa, K. (1991) J. Exp. Med. 173, 1213-1225. repertoire ofadult mice is formed from the primary repertoire 16. Coffman, R. L. & Weissman, I. L. (1981) Nature (London) 289, of the fetal and early postnatal periods is not known. We 681-683. speculate that, in the absence of continuing replenishment 17. Gulley, M. L., Ogata, L. C., Thorson, J. A., Dailey, M. 0. & from unrearranged precursors, the adult CD5' B-cell popu- Kemp, J. D. (1988) J. Immunol. 140, 3751-3757. lation consists predominantly of long-lived self-renewing 18. Pearse, M., Gallagher, P., Wilson, A., Wu, L., Fisicaro, N., Miller, cells and that recruitment into this pool requires antigenic J. F., Scollay, R. & Shortman, K. (1988) Proc. Natl. Acad. Sci. USA 85, 6082-6086. selection. However, this remains to be proven. Further, 19. Hayakawa, K., Hardy, R. R. & Herzenberg, L. A. (1986) Eur. J. whether the distinctive specificities enriched in the CD5' Immunol. 16, 450-456. population (such as anti-BrMRBC) are deleted from the bone 20. Mulier-Sieburg, C. E., Whitlock, C. A. & Weissman, I. L. (1986) marrow-derived B-cell population, or simply not expanded, Cell 44, 653-662. remains to be determined. 21. Denis, K., Treiman, L. J., St. Claire, J. 1. & Witte, 0. N. (1984) J. B differ the be Exp. Med. 160, 1087-1101. Why would cells that from adult type 22. Hardy, R. R., Kishimoto, T. & Hayakawa, K. (1987) Eur. J. generated fetally? One possibility would be to allow for the Immunol. 17, 1769-1774. generation of B cells with reactivities that might otherwise be 23. Bosma, G. C., Custer, R. P. & Bosma, M. J. (1983) Nature (Lon- eliminated. We have noted previously that certain (likely don) 301, 527-530. beneficial) reactivities to self and environmental determi- 24. Dorshkind, K. (1989) Curr. Top. Microbiol. Immunol. 152,169-172. nants are uniquely enriched in the CD5+ B-cell population 25. Hayakawa, K., Carmack, C. E., Hyman, R. & Hardy, R. R. (1990) J. Exp. Med. 172, 869-878. (25-27). Exposure to antigen during the differentiation of the 26. Carmack, C. E., Shinton, S. A., Hayakawa, K. & Hardy, R. R. adult B-cell population probably leads to functional inacti- (1990) J. Exp. Med. 172, 371-374. vation ("anergy") or deletion as suggested by experiments 27. Hardy, R. R., Carmack, C. E., Shinton, S. A., Riblet, R. J. & with transgenic mice (36, 37). In a young animal, the B-cell Hayakawa, K. (1989) J. Immunol. 142, 3643-3651. population is in the process of being generated and so 28. Mercolino, T. J., Locke, A. L., Afshari, A., Sasser, D., Travis, infection at this time fail to induce a If W. W., Arnold, L. W. & Haughton, G. (1989) J. Exp. Med. 169, might response. 1869-1877. instead, germ-line-encoded specificities to common patho- 29. Pennel, C. A., Mercolino,T. J., Grdina, T. A., Arnold, L. W., Haugh- gens lead to positive selection in fetal-derived B cells, then ton, G. & Clarke, S. H. (1989) Eur. J. Immunol. 19, 1289-1295. such cells would serve as an early protective population, an 30. Hayakawa, K., Hardy, R. R., Honda, M., Herzenberg, L. A., inherited immune system. A difference in positive selection Steinberg, A. D. & Herzenberg, L. A. (1984) Proc. Natl. Acad. Sci. for germ-line-encoded self-reactivities between CD5+ and USA 81, 2494-2498. B cells be due to 31. Mercolino, T. J., Arnold, L. W., Hawkins, L. A. & Haughton, G. CD5-IgDhiO would likely physiological (1988) J. Exp. Med. 168, 687-698. distinctions between fetal and adult B cells. Animals repop- 32. Mercotino, T. J., Arnold, L. W. & Haughton, G. (1986) J. Exp. ulated using pro-B cells derived from different stages of Med. 163, 155-165. ontogeny will provide an experimental system for asking such 33. Weigle, W. 0. (1973) Adv. Immunol. 16, 61-122. questions. Finally, a comparison of the functional properties 34. Hayakawa, K., Hardy, R. R., Parks, D. R. & Herzenberg, L. A. and genes expressed in pro-B cells at these two developmen- (1983) J. Exp. Med. 157, 202-218. 35. Hayakawa, K. & Hardy, R. R. (1988)Annu. Rev. Immunol. 6,197-218. tal stages will help to determine the mechanism for generating 36. Goodnow, C. C., Crosbie, J., Adelstein, S., Lavoie, T. B., Smith, such distinctive progeny. G. S. J., Brink, R. A., Pritchard, B. H., Wotherspoon, J. S., Lo- blay, R. H., Raphael, K., Trent, R. & Basten, A. (1988) Nature We thank Ms. Susan A. Shinton for excellent technical help in this (London) 334, 676-682. work. We thank Drs. M. Bosma, B. Perry, and M. Weigert (Institute 37. Nemazee, D. A. & Burki, K. (1989) Nature (London) 337, 562-566. Downloaded by guest on September 27, 2021