A differentiation pathway for B1 cells in adultbone marrowBrandt L. Esplina,b, Robert S. Welnera, Qingzhao Zhanga,c, Lisa A. Borghesid, and Paul W. Kincadea,1

aImmunobiology and Cancer Program, Oklahoma Medical Research Foundation, Oklahoma City, OK 73104; Departments of bMicrobiology and Immunologyand cPathology, University of Oklahoma Health Sciences Center, Oklahoma City, OK 73104; and dDepartment of Immunology, University of PittsburghSchool of Medicine, Pittsburgh, PA 15261

Edited by Leonard A. Herzenberg, Stanford University School of Medicine, Stanford, CA, and approved February 20, 2009(received for review November 14, 2008)

The recent description of a Lin�AA4.1�CD19�B220Lo/� B1-specifiedprogenitor (B1P) population in adult marrow adds support for theargument that these unique B cells arise from a distinct lineage.However, the origins of B1P were not investigated and theirdevelopmental relationships to conventional B2 cells remain un-clear. We now report that B1P development is IL-7R�-dependent,and negatively regulated by Bruton tyrosine kinase. Lymphoidcharacteristics of B1P were further studied with recombinationactivating gene (RAG)-1/GFP knock-in, RAG-1/Cre reporter, andVEX transgenic mice. Our results reveal that they are heteroge-neous with respect to lymphocyte affiliation. RAG-1� early lym-phoid progenitors and Lin�Sca-1�cKitLoIL-7R�� common lymphoidprogenitors from adult marrow efficiently generatedCD19�CD45R/B220Lo/� cells in vitro and in vivo. Moreover, earlylymphoid progenitors and common lymphoid progenitors pro-duced significant numbers of peritoneal CD11b�CD5� B1a andCD11b�CD5� B1b cells in vivo. Finally, 2-step transplantationexperiments established a differentiation pathway between con-ventional lymphoid progenitors, B1P, and mature B1 lymphocytes.Thus, our findings indicate that at least some B1P can be producedin adult bone marrow from primitive B2 progenitors, and suggesta developmental relationship between the major categories of Blymphocytes.

B1 lymphocytes � hematopoiesis � lymphopoiesis � B1 progenitors

Mature B lymphocytes are divisible into subsets that differwith respect to maturation, antigen experience, and func-

tional specialization. Two categories of B1 cells have attractedconsiderable attention, inasmuch as they are the first to appearduring fetal development (1), may represent a first line ofdefense against systemic bacterial infection in adults, and couldcontribute to autoimmunity and the formation of regulatory Bcells (2–4). The mature B1a subset that predominates in theperitoneal and pleural cavities of mice has been defined asIgMHiIgDLoCD11b�CD5�B220Lo/�, and an otherwise similarCD5� subset has been designated B1b (5). Considerable con-troversy has surrounded B1 cells, and particularly with respect totheir formation. For example, some findings suggest they arisevia a unique, and predominantly fetal, differentiation lineage (1,6, 7), whereas others indicate that the properties of B1 cells resultfrom the nature of signals delivered via their surface antigenreceptors (8, 9). Considering that particular Ig variable regiongenes are expressed by B1 cells and the corresponding antigenreceptors are poly-reactive (10), these are not mutually exclusiveconcepts. That is, developing B1 cells may be preferentiallyreceptive to low-level stimulation from self or environmentalantigens, leading to expression of distinctive surface markers.The recent discovery of a category of B1 restricted progenitors(i.e., B1Ps) in fetal and adult tissues might support the existenceof a separate lineage for these cells (11, 12). However, theirdevelopmental heritage has not been investigated. Of particularinterest is whether B1Ps are exclusively fetal derived, long-lived,and capable of self-renewal, or replenished by more primitivepopulations in the adult.

We now confirm that rare B1-biased progenitors are presentin adult bone marrow, but are unlikely to be homogeneous withrespect to lymphoid commitment and/or lineage progression. Amajor goal of our studies was to establish their developmentalrelationship to conventional lymphoid progenitors. B1Ps wegenerated in vivo from lymphoid progenitors efficiently pro-duced B1 cells upon secondary transplantation, suggesting ashared ancestry exists between B1 and B2 lymphocytes in adultlymphopoiesis.

ResultsDetailed Characterization of B1 Progenitors Suggests Hetero-geneity in Lineage Commitment and/or Lineage Progression.Lin�AA4.1�CD19�B220Lo/� cells in bone marrow were origi-nally described as progenitors with strong bias for producing B1rather than conventional B2 lymphocytes (11). However, B1Psare rare and flow cytometry did not reveal them to be a discretepopulation even when gating was stringently established withisotype-matched control antibodies. Therefore, we investigatedadditional parameters that would provide a background forfurther studies. Whereas B1Ps represent 0.1%–0.5% of nucle-ated marrow cells in BALB/c mice (11), they are consistently lessfrequent (0.05%–0.2%) in C57BL/6 mice (Fig. 1A). Consistentwith the fact that B lymphopoiesis is largely driven by IL-7 (13),B1Ps were reduced threefold in IL-7R��/� mice (Fig. 1). Brutontyrosine kinase (Btk)-KO mice have a severe deficiency inperipheral B1 cells (14, 15), and it is interesting that B1Pnumbers were elevated in Btk�/� marrow (Fig. 1). This isconsistent with previous reports that responsiveness of lymphoidprogenitors to IL-7 is abnormally high in related Xid�/� mice(16). Lymphopoiesis does not progress beyond the Pro-B stagein recombination activating gene (RAG)-1�/� or �-MT mice(17, 18), and we found nearly normal numbers of B1Ps in theirbone marrow.

Expression of RAG proteins and B-cell receptor (BCR)rearrangement represent important milestones in B lineagedifferentiation, and these parameters were used to gain addi-tional information about the nature of B1Ps. RAG-1/GFPknock-in mice permit detection of RAG-1 locus activity, impor-tant for identifying RAG-1� early lymphoid progenitors (ELPs)among the primitive lineage marker-negative, Sca-1�cKitHi

(LSK) subset (19). Our results demonstrated that approximately64% of Lin�Sca-1�cKitLoIL-7R�� common lymphoid progen-itors (CLPs) (20), 83% of CD19�B220�CD24�CD43�IgM�

Pro-B cells, and 84% of CD19�B220�CD43�IgM� Pre-B cells

Author contributions: B.L.E., R.S.W., and P.W.K. designed research; B.L.E. and Q.Z. per-formed research; L.A.B. contributed new reagents/analytic tools; B.L.E., R.S.W., Q.Z., L.A.B.,and P.W.K. analyzed data; and B.L.E., L.A.B., and P.W.K. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1To whom correspondence should be addressed. E-mail: kincade@omrf.ouhsc.edu.

This article contains supporting information online at www.pnas.org/cgi/content/full/0811632106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0811632106 PNAS � April 7, 2009 � vol. 106 � no. 14 � 5773–5778

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

June

2, 2

020

(21) were RAG-1/GFP� (Fig. 2A Top). Thus, it is informativethat less than half of the B1P expressed RAG-1/GFP. As inprevious studies, only 2% of IgMHiCD11b�CD5�/�B220Lo/� B1cells in the peritoneal cavity (B1, PC) were positive for thisreporter (22).

We constructed a new model by crossing RAG-1/Cre mice toanimals with tandem-dimer red fluorescent protein (tdRFP)knocked into the ROSA26 locus (23, 24). Thus, cells with alymphoid past are permanently marked, as was the case for 98%

of B1 PC cells (Fig. 2 A Middle). Nearly 60% of CLPs and 90%of Pro-B in bone marrow were RAG-1/tdRFP�, whereas ap-proximately 56% of B1Ps were labeled. It is noteworthy that theintensity of tdRFP� B1P was lower than that for mature B1 cells.As originally described for ROSA26 tdRFP mice (23), this tendsto be the case for mitotically active cells.

An Ig and T cell receptor (TCR) gene recombination substratesystem has been described that reports Ig/TCR rearrangementhistory by permanent expression of the VEX fluorochrome (25).In this model, CLP and Pro-B cells were shown to be approxi-mately 30% and 90% VEX-positive, respectively. We haveconfirmed those findings and now show that 48% of B1Ps hadrecombined the reporter substrate (Fig. 2 A Bottom). For com-parison, approximately 89% of CD19�IgMHiIgDLoCD11b� B1PC were labeled.

Lymphoid progenitors in bone marrow are highly sensitive tosex steroids, and we have previously used this parameter to assessrelatedness to other lineages (26–29). Therefore, we implantedtime-release estrogen pellets into male mice and evaluated them1 week later with respect to B1P (Fig. 2B). Whereas ELPsand CLPs were dramatically depleted, B1Ps were reduced byonly 31% (range, 26%–37% in 10 total mice). Numbersof CD19�B220�CD24�CD43�IgM� Pro-B andCD19�B220�CD43�IgM� pre-B cells were reduced by 46% and42%, respectively. Myeloid progenitors (Lin�Sca-1�cKitHi),spleen B cells, and B1 PC populations were unaffected. Wefound a similar modest decrease in B1P in pregnant mice (notshown), again indicating that only some are hormone-sensitive.

Most stem cells and ELPs spend considerable time in aquiescent, G0 state whereas later stages of B lymphopoiesisinvolve actively dividing Pro-B cells (30). In terminal stages, Bcells complete differentiation in the spleen without furtherexpansion. Therefore, we used intracellular Ki67 staining toassess the mitotic activity of B1Ps. There was a higher proportion

Fig. 1. KO models indicate dependence of B1P development on IL-7R�, andinhibition by Btk. (A) Frequencies � SEM of Lin�AA4.1�CD19�CD45R/B220Lo/�

B1Ps in marrow of various KO models. Plots are representative of 3 separateanalyses (3–5 per group) with at least 350,000 viable (i.e., propidium iodide-negative) gated events collected per plot. (B) Total numbers of B1P/femur �SEM were normalized to C57BL/6 WT (dashed line). All mutants were on aC57BL/6 background. CD45R/B220Lo/� gates were drawn according to isotypecontrols based on WT C57BL/6 profile. Student t test analysis was used to assessstatistical differences between means. (**, P � 0.01, ***, P � 0.001.)

Fig. 2. Ig gene rearrangement activity, estrogen sensitivity, and cell cycle status reveal heterogeneity within the B1P population. (A) RAG-1/GFP knock-in micewere used to detect RAG-1 locus activity (Top). RAG-1 expression history was studied with a (RAG-1/Cre � ROSA26 tdRFP) F1 model (Middle), and VEX transgenicmice (Bottom) were used to assess V(D)J recombinase history. (B) Estrogen pellets (0.1 mg) were implanted s.c. into 8- to 12-week-old male C57BL/6 mice for 1week. BM, Spl, and PC population cell numbers � SEM were normalized to control (dashed line). (C) Steady-state cell cycle status was analyzed in 2 separateexperiments (4 mice each) by intracellular Ki-67 staining in Pro-B, B1P, and mature spleen B cells. Average percentages � SEM of cycling (Ki67�) cells are shown.(D) C57BL/6 mice were administered an i.p. dose of 100 �g BrdU and maintained on BrdU-enriched water (8 mg/mL). Average BrdU incorporation � SEM wasevaluated by flow cytometry on days 2, 4, 7, and 10. Spl B, CD19�B220�IgM� spleen B cells. Shaded histograms in A and C depict WT and isotype-negative controls,respectively. Three experiments with 3 to 5 mice per group were performed unless otherwise stated.

5774 � www.pnas.org�cgi�doi�10.1073�pnas.0811632106 Esplin et al.

Dow

nloa

ded

by g

uest

on

June

2, 2

020

of Ki67-positive B1Ps (51%) than mature restingCD19�B220HiIgM� spleen B cells (11%), and likewise a con-siderably lower proportion than in Pro-B cells (80%; Fig. 2C). Asan additional assessment of proliferative activity, we measuredthe incorporation of BrdU into the DNA of these populations invivo over a 10-d period (Fig. 2D). B1Ps incorporated BrdU at amuch slower rate than cycling Pro-B cells, but more rapidly thanresting splenic B cells. Consistent with previous reports (31),more than 75% of Pro-B cells had incorporated BrdU by day 4.This contrasted with 35% and 8% labeling of B1P and splenic Bcells, respectively. After 10 d of continual BrdU administration,nearly all Pro-B cells, but still only 63% of B1Ps and 21% ofsplenic B-cells, had incorporated BrdU.

In summary, most cells in the previously definedLin�AA4.1�CD19�B220Lo/� subset have characteristics associ-ated with previously described lymphoid progenitors. Theirrequirement for IL-7R� signaling, and negative regulation byBtk indicate that they have initiated development in the Blymphocyte pathway. However, their numbers are minimallyaffected by mutations related to Ig gene recombination orpre-BCR expression. Only 35% of B1Ps have an active RAG-1locus, 56% have a history of RAG-1 expression, and 48% haveevidence of Ig gene recombination. A subset was depleted byestrogen treatment, approximately one third were in active cellcycle, and approximately two thirds incorporated BrdU in a 10-dperiod. Although B1Ps are not homogeneous in any of theserespects, these results may indicate a continuum of maturationand lymphoid lineage progression.

Adult Lymphoid Progenitors Are a Source of B1P. The earliesthematopoietic origin of bone marrow B1Ps had not been pre-viously investigated, and initial studies concluded that B1 cellsmay arise only during fetal life (6, 7). Therefore, we sorted aseries of well characterized adult marrow progenitors and as-sessed their potential for generating B1Ps. These includedhematopoietic stem cell-enriched LSK, RAG-1�/GFP� ELPs,CLPs, Pro-B cells, and myeloid progenitors (MP). An additionalAA4.1�CD19�CD24�CD21�IgM� transitional-type 1 (T1)fraction of spleen B cells was included in some experiments asa negative control. We used a previously described trans-wellculture system in which conditions were modified by addition ofIL-7 and removal of IL-3 and IL-6 to improve viability of theearliest progenitors (11, 32, 33). Measurements were made after7 or 12 d of culture. Unlike LSK, ELPs, and CLPs (Fig. 3A),myeloid-restricted progenitors and Pro-B and spleen T1 cells hadno ability to make B1P cells in these conditions. Observation ofserial cultures revealed that B1Ps were made more rapidly fromCLPs than RAG-1� ELPs, but ELPs eventually gave slightlyhigher yields [supporting information (SI) Fig. S1].

Some cells with the potential to make B1Ps might not havebeen detected under these artificial conditions. Therefore, thesame progenitor populations were transferred to sublethallyirradiated congenic Ly5.1 recipients before flow cytometryanalysis of bone marrow. As in the trans-well culture system,LSKs, ELPs, and CLPs were sources of B1Ps (Fig. 3B). Of these,the highest yields were observed with CLPs and ELPs intemporal fashion 18 and 21 d after transplantation, respectively.Pro-B, myeloid progenitors, and spleen T1 cells never producedappreciable numbers of B1Ps during the 28-d reconstitutionperiod (Fig. 3B, and not shown). Note that, in contrast to CLPs,LSK-derived B1Ps were not convincingly detected before day 28,yet they were consistently found in all 48 recipients in 3experimental series. In a single experiment, yields of B1P werehigher where LSK recipients were examined on day 35 (notshown).

These findings indicate how cells with theLin�AA4.1�CD19�B220Lo/� staining properties of B1P can bemade from hematopoietic cells in adult bone marrow. We

detected no evidence of affiliation with myeloid cells, and theywere not made from relatively differentiated Pro-B or spleen T1cells. Rather, they arose efficiently in vitro and in vivo from CLPsand then ELPs.

Lymphoid Progenitors in Adult Bone Marrow Can Generate B1 Lym-phocytes. As discussed earlier, it has been controversial whetherall B1 cells arise from a fetal wave of lymphopoiesis or also canbe made from adult progenitors (1, 6, 7, 34). To address this,adult marrow progenitors were sorted to high purity fromC57BL/6 (CD45.2�) mice and transplanted into sublethallyirradiated Ly5.1 (CD45.1�) congenic mice. Four weeks later, weanalyzed the peritoneal cavity of recipient mice by flow cytom-etry to determine the relative yields of B1 and conventional B2cells (Fig. 4A Top and 4B). Here, CD19� B1 and B2 cells werediscriminated as IgMHiIgDLo and IgMLoIgDHi, respectively. Asmay be expected, essentially no B1 cells were produced whenmyeloid progenitors or Pro-B or T1 cells were transplanted (notshown). Consistent with their original description (11), B1Pswere highly B1-biased (Fig. 4A Top Right). In contrast, morebalanced B1-to-B2 ratios were observed with PC cells recoveredfrom mice transplanted with ELPs or CLPs. When we analyzeddonor B1 subsets, we found that CD19�IgMHiCD11b�CD5�

B1a cells were slightly more numerous than their CD5� B1bcohorts when B1Ps were transplanted (Fig. 4A Lower). This ratiowas similar to that observed for peritoneal B1 subsets in normalmice (35, 36). It is perhaps noteworthy that the B1 cells gener-ated from B1Ps were slightly more homogeneous with respect toCD5 densities than those generated from ELPs or CLPs (Fig. 4ALower). In more limited experiments, we transplanted fetal liverlymphoid progenitors and found them to be more efficient inproducing B1 cells than adult progenitors (not shown).

Consistent with published results (11), B1Ps were poor sourcesof either CD23HiCD21Lo follicular or CD23LoCD21Hi marginal-zone B2 lymphocytes resident in the spleen (Fig. 4C). In fact,

Fig. 3. Conventional lymphocyte progenitors are a source of B1Ps. (A)Twelve-day trans-well cultures were initiated with LSK, ELP, CLP, Pro-B, MP,and T1 cells to assess potential for generating CD19�B220Lo/� progeny. Plotsrepresentative of 3 separate experiments, with average percentage of viableAA4.1� cells shown. (B) Adult marrow CD45.2� progenitor populations (103 to104) were transplanted into sublethally irradiated (650 R) Ly5.1 (CD45.1�)recipients. Marrow B1P production was analyzed 14 to 28 d later. Averageyield/input � SEM of 3 independent experiments with 4 mice per group pertime point are shown.

Esplin et al. PNAS � April 7, 2009 � vol. 106 � no. 14 � 5775

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

June

2, 2

020

B1Ps primarily generated CD23LoCD21Lo B cells, a fractioncontaining immature, transitional, and B1 lymphocytes. In con-trast, conventional lymphoid progenitors produced the wholespectrum of splenic B cells. Utilization of VH11 Ig genes is aproperty unique to B1 lymphocytes (37–39), and we found thatdonor-type CD45.2�CD19� cells recovered from the PC of CLPrecipients included lymphocytes with VH11 sequences (see Ma-terials and Methods and Fig. S2 A). Importantly, VH11 sequenceswere not detected in CD45.1�CD19�B220�CD43�IgM� hostB2 cells simultaneously harvested from the spleen.

Finally, 2-step transplantation experiments were conducted toexplore precursor-product relationships between primitive lym-phoid progenitors and B1P and B1 cells. CD45.2� CLPs or ELPswere transferred into sublethally irradiated Ly5.1 (CD45.1�)recipients. Donor-derived B1Ps were isolated from marrow ofthe CLP or ELP transplanted mice 18 and 21 d later, respectively,before secondary transplantation into Ly5.1 hosts. Four weekslater, PC and spleen cells were recovered to assess B1 potential(Fig. 4D and not shown). We found that ELP- and CLP-derivedB1Ps exhibited B1 bias similar to freshly isolated B1Ps, andpreferentially generated IgMHiIgDLo B1 cells (compare to Fig.4A Top Right). Recipient spleens also contained few CD23� B2cells (not shown). Furthermore, we detected B1 cell distinctiveVH11 sequences in donor-derived CD19� cells recovered fromspleens and PC of these mice (see Materials and Methods and Fig.S2B).

Thus, previously characterized lymphoid progenitors in adultbone marrow can produce B1 cells when transplanted to suble-thally irradiated adult recipients. Unlike B1Ps, ELPs and CLPshave the additional potential to generate substantial numbers ofB2 cells. A complete differentiation sequence was indicated bythe successful transfer of ELPs or CLPs, recovery of newlyformed B1Ps from bone marrow, and generation of B1 cells insecondary recipients. We conclude that B1Ps represent anintermediate stage between primitive lymphoid progenitors andB1 cells.

DiscussionB1 cells have long been considered to represent specializedcomponents of the immune system, and identification of B1-

biased progenitors was informative with respect to their origins.The goal of our study was to determine the hematopoietic sourceof B1P and assess their relationship to better characterizedmultipotent progenitors. We found that numbers of B1Ps weregreatly reduced in the absence of IL-7� receptors, and expandedin mutant Btk�/� mice. Although these characteristics help toestablish their B lymphoid affiliation, other findings suggest thatthis is a heterogeneous population, as almost half had no historyof RAG-1 expression or Ig gene recombination. B1Ps wereslowly generated from ELPs, and more rapidly from CLPs, butnot detectable in cultures initiated with Pro-B cells. Thesefindings suggest a differentiation sequence that was supported byresults from single and double transplantation experiments. Thatis, ELPs or CLPs generate B1Ps, which in turn are heavily biased,but not completely restricted, to produce B1 cells. Although thisis a plausible series of events, the findings do not exclude thepossibility that some adult B1 cells arise during fetal life or thatothers derive from the differentiation of mature B2 cells.

Although B1Ps respond to thymic stromal lymphopoietin, IL-7is a more potent stimulus for their expansion in culture (32).Furthermore, a 40-fold reduction in B1Ps was reported inIL-7-deficient mice (40). Although we found that numbers ofB1Ps were reduced in IL-7R��/� mice, they were not altogetherabsent, and approximately 30% of B1Ps are IL-7R�� (ref. 11).Btk is required for late stages of B lineage differentiation inhumans and mice, but IL-7 responsiveness of marrow progeni-tors is exaggerated in Btk�/� mice (16). This is consistent withour result that B1Ps were much more abundant in those animals.Therefore, most, but not necessarily all, CD19�CD45R/B220Lo/�

cells are IL-7R�-dependent lymphoid progenitors that are neg-atively regulated by Btk.

Expression of RAG proteins and BCR rearrangement repre-sent important milestones in B lymphopoiesis, so 3 reportersystems were used to further characterize B1Ps. Again, we foundevidence of heterogeneity, even when care was taken to excludecells with significant amounts of CD45R/B220. Rare CD19�

marrow cells with no history of RAG-1 expression or Ig substraterecombination comprise roughly half of B1Ps, and we have nofurther information concerning their fate. These determinations

Fig. 4. Conventional lymphocyte precursors represent a developmental link between B1 and B2 cells in adult lymphopoiesis. (A–C) CD45.2� donor ELPs, CLPs,and B1Ps (103 to 104) were transplanted into CD45.1� recipients and analyzed for B-cell potential in PC (A) and spleens (C) 4 weeks post-transplantation. Averageyield/input � SEM of donor CD19�IgMHiIgDLoCD11b� B1 PC, and CD19�IgMLoIgDHiCD11b� B2 cells, respectively, were quantified in B. (D) Two-step transplan-tations in which CLP- or ELP-derived B1Ps were serially transplanted, and B1 potential in the PC was assessed 4 weeks later. Average percentages of CD19� donorIgMHiIgDLo B1 and IgMLoIgDHi B2 cells in recipient PC are shown. Pre-gating of viable donor cells is indicated (Left). All transplants involved sublethal irradiation(650 R). Data are representative of 3 experiments with 3 to 5 recipients per group.

5776 � www.pnas.org�cgi�doi�10.1073�pnas.0811632106 Esplin et al.

Dow

nloa

ded

by g

uest

on

June

2, 2

020

were made with previously described RAG-1/GFP and VEXmice, as well as a new (RAG-1/Cre � ROSA26 tdRFP) F1model. A previous study revealed that at least some B1Ps haveundergone D-J, but not V-DJ, recombination (11). The fre-quency of B1Ps with sufficient RAG-1/RAG-2 to trigger sub-strate recombination in the reporter mice was similar to those inthe CLP subset, yet much less than Pro-B cells. This is consistentwith our inability to generate them from Pro-B cells, a subsetconsidered to have largely undergone V-DJ rearrangements(41–43).

Estrogen sensitivity has been a useful criterion for identifyingand characterizing lymphopoietic cells in bone marrow (26–29).Whereas most ELPs and CLPs disappeared from marrow ofhormone treated mice within 1 week, there was only a 30%reduction in numbers of B1Ps. This could again reflect theirheterogeneity, but it also represents a distinctive property. Notethat marrow precursors for plasmacytoid dendritic cells are alsorelatively hormone resistant, and decline only after extendedtreatment (28, 44).

Another important issue relates to the mitotic activity of B1Ps,and our experiments suggest it is less than for the somewhat moremature Pro-B cells. Although not directly compared to B1Ps inthis study, most primitive ELPs and pro-lymphocytes/CLPs arequiescent (30, 45). We conclude that the proliferative status ofB1Ps must be similar to conventional lymphopoietic progenitorsand sufficient to sustain the small B1 compartment. Moreover,B1Ps were present in 13-month-old mice (not shown), suggestingthey must be slowly replaced throughout life.

There is a gradual loss of options with lineage progression, andwe asked which cells could generate B1Ps in culture or upontransplantation. The results were unambiguous, and B1Ps didnot arise from myeloid progenitors or cells thought to beirrevocably committed to the conventional B2 lineage. In con-trast, they were produced from CLPs, and more slowly from LSKor the RAG-1� ELPs included in that primitive subset. Impor-tantly, experimentally produced B1Ps were heavily biased toproduce B1 lymphocytes. We estimate that individual ELPs orCLPs generated 250 or 150 CD19� cells, respectively, in singlefemurs of transplanted mice. This translates to nearly 4,000marrow lymphocytes per progenitor/animal, and almost all ofthem would represent conventional B2 cells. In contrast, B1Psproduced only an average of 15 peritoneal B1 cells within 4weeks of transplantation. Although these represent low yields, itshould be remembered that B1 cells are normally rare, and littleexpansion of B1Ps might be needed to sustain them.

Our experiments demonstrate an early developmental linkagebetween the major categories of B lymphocytes, and show thatthe potential to make B1 cells is retained at least to the CLPstage. Indeed, B1Ps can be found in the marrow of aged mice.However, many of the B1 cells in adult tissues might have beengenerated during fetal life, and still others might representtrans-differentiation of conventional B2 cells during the courseof immune responses. It also seems possible that the progeny ofB1Ps initially resemble B2 cells and require antigen-dependentinstruction or selection for an ultimate B1 fate. Further study isneeded to address these important questions and to learn ifparticular culture conditions favor B1 versus B2 restriction inlymphoid progenitors.

Materials and MethodsMice. C57BL/6 (CD45.2 alloantigen; Jackson Laboratory), B6-SJL/Ly5.1 (CD45.1alloantigen; Jackson Laboratory), and B6-RAG-1/GFP knock-in mice were bredand maintained in the Laboratory Animal Resource Center at the OklahomaMedical Research Foundation (Oklahoma City, OK). RAG-1�/�, IL7R��/�, and�-MT mice were purchased from the Jackson Laboratory. VEX and Btk�/� micehave been described previously (25, 46). A new model was created by crossingRAG-1/Cre mice to animals with tdRFP knocked into the ROSA26 locus (23, 24).These animals will be described in more detail elsewhere. All mice were 8 to

14 weeks old, and male and female mice were used with equal probability,save in the estrogen experiments (as described later).

Isolation of Cell Populations and Flow Cytometry. Cell manipulations wereperformed in Hanks’ balanced salt solution with 5% FCS. Marrow cells wereisolated from the long bones of donor mice, and erythrocytes were lysed bybriefly re-suspending in NH4Cl� hypotonic solution. PC cells were acquired bylavage with 10 mL Hanks’ balanced salt solution plus 5% FCS. To isolateprogenitor populations for culture and transplantation, BM cells were en-riched by negative selection by labeling marrow with Ly6G�C/Gr-1 (RB6–8C5),CD11b/Mac-1 (M1/70), TER-119, CD3 (17A2), CD8 (53–6.7), CD19 (1D3), andB220 (14.8), and then immuno-magnetically depleted with the BioMag goatanti–rat IgG system (Qiagen). All cells were treated with Fc-receptor block(2.4G2) before fluorescent staining and sorting. After staining marrow withbiotin-anti-lineage markers (Gr-1, CD11b, CD19, CD45R/B220, Ter119, CD3�,CD8�, and NK1.1), APC-anti–cKit (2B8) and PE-Cy5-anti-Sca-1 (D7; eBio-science), PE-anti-IL-7R� (A7R34; eBioscience), and secondary streptavidin PE-Texas Red (Caltag Laboratories), Lin� populations were sorted using either aMoFlo (Dako Cytomation) or FACSAria cytometer (BD Biosciences) into specificpopulations. LSK cells were obtained as Lin�ckitHiScaI�RAG-1/GFP�, whereasELPs were sorted as Lin�ckitHiSca-1�RAG-1/GFP�. CLPs were sorted asLin�ckitLoSca-1�IL-7R��. Myeloid progenitors were sorted as Lin�Sca-1�ckitHi,and Pro-B cells were sorted as CD19�B220�CD24�CD43�IgM�. Anti-CD19 andCD45R/B220 Abs were excluded from lineage depletions when isolating B1Progenitors (11). Whole spleen samples were immuno-magnetically depletedfor T cells (CD3, CD8) and myeloid cells (Gr-1, CD11b) before isolation ofAA4.1�CD19�CD24�CD21�IgM� T1 cells. Dead cells were excluded by pro-pidium iodide staining (Molecular Probes). Purification of each subset wasachieved by double sorting, and confirmed by post-sort analysis. Cells har-vested from killed recipient mice were analyzed to determine the phenotypesof resident or donor cells. Abs included biotinylated CD3 (145–2C11), CD8(53–6.7), CD11b (M170), TER-119, Gr-1 (RB6–8C5), IgM (R6–60.2), and NK1.1(PK136); CD5 (53–7.3) APC, IgD (11–26c.2a) PE, CD23 (2G8) APC (SouthernBiotech), CD21 (7G6) PE, IgM (R6–60.2) PerCpCy5.5, CD19 (1D3) APC-Cy7, CD93(AA4.1) APC, B220 (RA3–6B2) PE-Cy5, and CD45.2 (104) FITC. Flow cytometrywas performed on a BD LSRII (BD Biosciences), and FlowJo software (Treestar)was used for data analysis. All Abs came from BD PharMingen unless otherwisestated.

Intravenous Progenitor Transfers. Recipient mice were given 650 Rads ofradiation from a 137Cs source (Mark I �-irradiator; J.L. Shepard and Associates).Mice were anesthetized with isoflurane (Isosol; Vedco), and sorted popula-tions were infused intravenously by retro-orbital injection. Host mice were 8-to 12-week-old B6-SJL/Ly5.1 (CD45.1 alloantigen); donor cells were from 8- to14-week-old C57BL/6 or B6-RAG-1/GFP mice expressing the CD45.2 alloanti-gen. PC, spleen, and BM cells from one femur were collected separately fromrecipient mice, and CD45.2� donor-derived cells were assayed by flow cytom-etry, as described earlier.

Estrogen Treatment In Vivo. Time-release pellets of 17�-estradiol, 0.1 mg/pellet, 21-d release (Innovative Research of America) were implanted s.c. into8- to 12-week-old male C57BL/6 mice with a 10-gauge precision trocar. After1 week, mice were killed and PC, spleen, and BM cells were analyzed by flowcytometry.

Cell Cycle Kinetics Analysis. Single-cell suspensions of C57BL/6 marrow andspleen were fixed and permeabilized using the BD Cytofix/Cytoperm kit (BDBiosciences), and stained intracellularly with FITC anti-Ki-67 (B56) or MOPC-21isotype control. Mice were given an i.p. injection of 100 �g/100 �L BrdU in PBSsolution (BD Biosciences), with continual administration of BrdU-enricheddrinking water (0.8 mg/mL; Sigma-Aldrich) thereafter for the duration of theexperiment. At defined time points, marrow from a single femur was collectedfrom 4 mice and analyzed by intracellular staining with FITC anti-BrdU (FITCBrdU flow kit; BD Biosciences).

Trans-well Cultures. Sorted progenitors were cultured in a trans-well systempreviously described (11) for 7 to 12 days, with OP9 stromal cells provided byJ.C. Zuniga-Pflucker (Toronto, ON, Canada) plated to sub-confluence in theupper chamber. The RPMI 1640 contained 10% FCS, 20 ng/mL Flt3-L, 20 ng/mLstem cell factor, 10 ng/mL IL-7, 10 ng/mL recombinant human thymic stromallymphopoietin, 1 mM L-glutamine, 100 U/mL penicillin, and 100 �g/mL strep-tomycin. Cells were fed by replacing half the volume with fresh media andcytokines every 4 days.

Esplin et al. PNAS � April 7, 2009 � vol. 106 � no. 14 � 5777

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

June

2, 2

020

Ig Gene Repertoire Analysis. For all samples, cells were resuspended in TRIzolReagent (Invitrogen). RNA was extracted using standard protocol. RT-PCR wasperformed using OneStep RT-PCR Kit (Qiagen). For positive and negativecontrols, forward primers included MH3, MH4, MH5, MH6, MH7, MH11,MH12–1, and MH12–2. DS-MurCmu-Int was used as reverse primer. The PCRproduct was cloned into TA vector from TOPO cloning kit (Invitrogen).

White colonies were picked for mini-prep extraction and sequencing. Forsamples, forward primers were M-VH2, M-VH3, M-VH4, M-VH5, M-VH6, M-VH8, M-VH10, M-VH13, M-VH11(B), M-VH12(B)-1, and M-VH12(B)-2. JH-1 andJH-2 were used as reverse primers. PCR primers were digested with restrictionenzyme Age1 and SalI (New England Biolabs). The digestion product wascloned into AB vector. Colonies were subsequently chosen for sequencing.

Primer Sequences: MH3, CAGGTTACTCTGAAAGWG; MH4, GAGGTCCARCT-GCAACAR; MH5, CAGGTCCAACTVCAGCAR; MH6, GAGGTGAASSTGGTGGAA;MH7, GATGTGAACTTGGAAGTG; MH11, GAAGTGCAGCTGTTGGAG; MH12–1,CAGATGCAGCTTCAGGAG; MH12–2, CAGATTCAGCTTAAGGAG; DS-MurCmu-Int, AGG AGA CGA GGG GGA AGA C; M-VH2, CTGCAACCGGTGTACATTCACAGGTCCAACTGCAGCAG; M-VH3, CTGCAACCGGTGTACATTCACAGGTC-CAGCTGCAGCAA; M-VH4, CTGCAACCGGTGTACATTCA GAGGTCCAGCTG-CAACAA; M-VH5, CTGCAACCGGTGTACATTCA GAGGTCCAGCTGCAACAG; M-

VH6, CTGCAACCGGTGTACATTCA CAGGTGCAGCTGAAGCAG; M-VH8,CTGCAACCGGTGTACATTCAGAAGTGAAGCTGGTGGAG; M-VH10, CTGCAAC-CGGTGTACATTCACATGTCCAACTGCAGCAG; M-VH13, CTGCAACCGGTGTA-CATTCACAGATCCAGTTGGTGCAG; M-VH11(B), CTGCAACCGGTGTACATTCA-GAAGTGCAGCTGTTGGAG; M-VH12(B)-1, CTGCAACCGGTGTACATTCACAGATGCAGCTTCAGGAG; M-VH12(B)-2, CTGCAACCGGTGTACATTCA CA-GATTCAGCTTAAGGAG; JH-1, TGCGAAGTCGACGCTGAGGAAACGGT-GACCGTG; JH-2, TGCGAAGTCGACGCTGAGGAGACTGTGAGAGTG.

Statistics. Prism V3.02 software (GraphPad) was used for statistical analysis.Intergroup comparisons were performed with the unpaired Student t test. Pvalues were 2-tailed and considered significant if less than 0.05.

ACKNOWLEDGMENTS. We thank Drs. Carol Webb and Mark Lang for scientificconsultation, Tara Khamphanthala and Karla Garrett for technical assistance,Jacob Bass and Dr. Diana Hamilton for cell sorting, and Shelli Wasson foreditorial assistance. This work was supported by National Institutes of HealthGrants AI020069, AI058162, AI069024 (to P.W.K.), AR054529 (to L.A.B.), andF30AG031646 (to B.L.E.). P.W.K. holds the William H. and Rita Bell EndowedChair in Biomedical Research.

1. Hayakawa K, Hardy RR, Herzenberg LA (1985) Progenitors for Ly-1 B cells are distinctfrom progenitors for other B cells. J Exp Med 161:1554–1568.

2. Hayakawa K, Hardy RR, Parks DR, Herzenberg LA (1983) The ‘‘Ly-1 B’’ cell subpopula-tion in normal, immunodefective, and autoimmune mice. J Exp Med 157:202.

3. Hayakawa K, et al. (1984) Ly-1 B cells: functionally distinct lymphocytes that secrete IgMautoantibodies. Proc Natl Acad Sci USA 81:2494–2498.

4. Yanaba K, et al. (2008) A regulatory B cell subset with a unique CD1dhiCD5� pheno-type controls T cell-dependent inflammatory responses. Immunity 28:639–650.

5. Stall AM, Adams S, Herzenberg LA, Kantor AB (1992) Characteristics and developmentof the murine B-1b (Ly-1 B sister) cell population. Ann N Y Acad Sci 651:33–43.

6. Hardy RR, Hayakawa K (1991) A developmental switch in B lymphopoiesis. Proc NatlAcad Sci USA 88:11550–11554.

7. De Andres B, et al. (2002) The first 3 days of B-cell development in the mouse embryo.Blood 100:4074–4081.

8. Wortis HH, et al. (1995) B-cell activation by crosslinking of surface IgM or ligation ofCD40 involves alternative signal pathways and results in different B-cell phenotypes.Proc Natl Acad Sci USA 92:3348–3352.

9. Haughton G, Arnold LW, Whitmore AC, Clarke SH (1993) B-1 cells are made, not born.Immunol Today 14:84–87.

10. Kantor AB, Merrill CE, Herzenberg LA, Hillson JL (1997) An unbiased analysis ofV(H)-D-J(H) sequences from B-1a, B-1b, and conventional B cells. J Immunol 158:1175–1186.

11. Montecino-Rodriguez E, Leathers H, Dorshkind K (2006) Identification of a B-1 Bcell-specified progenitor. Nat Immunol 7:293–301.

12. Tung JW, et al. (2006) Phenotypically distinct B cell development pathways map to thethree B cell lineages in the mouse. Proc Natl Acad Sci USA 103:6293–6298.

13. Izui S, et al. (1981) Low-calorie diet selectively reduces expression of retroviral envelopeglycoprotein gp70 in sera of NZB x NZW F1 hybrid mice. J Exp Med 154:1116–1124.

14. Scher I (1982) The CBA/N mouse strain: An experimental model illustrating the influ-ence of the X-chromosome on immunity. Adv Immunol 33:1–71.

15. Hardy RR, Hayakawa K, Parks DR, Herzenberg LA (1984) Murine B cell differentiationlineages. J Exp Med 159:1169–1191.

16. Narendran A, et al. (1993) B cell developmental defects in X-linked immunodeficiency.Int Immunol 5:139–144.

17. Mombaerts P, et al. (1992) RAG-1-deficient mice have no mature B and T lymphocytes.Cell 68:869–877.

18. Kitamura D, Roes J, Kuhn R, Rajewsky K (1991) A B cell-deficient mouse by targeteddisruption of the membrane exon of the immunoglobulin � chain gene. Nature350:423–426.

19. Igarashi H, et al. (2002) Transcription from the RAG1 locus marks the earliest lympho-cyte progenitors in bone marrow. Immunity 17:117–130.

20. Kondo M, Weissman IL, Akashi K (1997) Identification of clonogenic common lymphoidprogenitors in mouse bone marrow. Cell 91:661–672.

21. Hardy RR, et al. (1991) Resolution and characterization of pro-B and pre-pro-B cellstages in normal mouse bone marrow. J Exp Med 173:1213–1225.

22. Kuwata N, et al. (1999) Cutting edge: Absence of expression of RAG1 in peritoneal B-1cells detected by knocking into RAG1 locus with green fluorescent protein gene.J Immunol 163:6355–6359.

23. Luche H, et al. (2007) Faithful activation of an extra-bright red fluorescent protein in‘‘knock-in’’ Cre-reporter mice ideally suited for lineage tracing studies. Eur J Immunol37:43–53.

24. Soriano P (1999) Generalized lacZ expression with the ROSA26 Cre reporter strain. NatGenet 21:70–71.

25. Borghesi L, et al. (2004) B lineage-specific regulation of V(D)J recombinase activity isestablished in common lymphoid progenitors. J Exp Med 199:491–502.

26. Medina KL, Strasser A, Kincade PW (2000) Estrogen influences the differentiation,proliferation and survival of early B lineage precursors. Blood 95:2059–2067.

27. Medina KL, et al. (2001) Identification of very early lymphoid precursors in bonemarrow and their regulation by estrogen. Nat Immunol 2:718–724.

28. Pelayo R, et al. (2005) Derivation of two categories of plasmacytoid dendritic cells inmurine bone marrow. Blood 105:4407–4415.

29. Welner RS, et al. (2007) Interferon-producing killer dendritic cells arise via a uniquedifferentiation pathway from primitive c-KitHi CD62L� lymphoid progenitors. Blood109:4825–4931.

30. Pelayo R, et al. (2006) Cell cycle quiescence of early lymphoid progenitors in adult bonemarrow. Stem Cells 24:2703–2713.

31. Cancro MP, et al. (2001) xid mice reveal the interplay of homeostasis and Bruton’styrosine kinase-mediated selection at multiple stages of B cell development. IntImmunol 13:1501–1514.

32. Montecino-Rodriguez E, Dorshkind K (2006) Stromal cell-dependent growth of B-1 Bcell progenitors in the absence of direct contact. Nat Protoc 1:1140–1144.

33. Maeda K, et al. (2005) IL-6 blocks a discrete early step in lymphopoiesis. Blood106:879–885.

34. Marshall AJ, Wu GE, Paige GJ (1996) Frequency of VH81x usage during B cell develop-ment: initial decline in usage is independent of Ig heavy chain cell surface expression.J Immunol 156:2077–2084.

35. Kantor AB, Herzenberg LA (1993) Origin of murine B cell lineages. Annu Rev Immunol11:501–538.

36. Tung JW, et al. (2004) Identification of B-cell subsets: an exposition of 11-color (Hi-D)FACS methods. Methods Mol Biol 271:37–58.

37. Reininger L, et al. (1987) Novel V genes encode virtually identical variable regions of sixmurine monoclonal anti-bromelain-treated red blood cell autoantibodies. J Immunol138:316–323.

38. Hardy RR, et al. (1989) A single VH gene is utilized predominantly in anti-BrMRBChybridomas derived from purified Ly-1 B cells. Definition of the VH11 family. J Immunol142:3643–3651.

39. Carmack CE, Shinton SA, Hayakawa K, Hardy RR (1990) Rearrangement and selectionof VH11 in the Ly-1 B cell lineage. J Exp Med 172:371–374.

40. Jensen CT, et al. (2008) FLT3 ligand and not TSLP is the key regulator of IL-7-independent B-1 and B-2 B lymphopoiesis. Blood 112:2297–2304.

41. Corcoran AE, Riddell A, Krooshoop D, Venkitaraman AR (1998) Impaired immunoglob-ulin gene rearrangement in mice lacking the IL-7receptor. Nature 391:904–907.

42. Ehlich A, Martin V, Muller W, Rajewsky K (1994) Analysis of the B-cell progenitorcompartment at the level of single cells. Curr Biol 4:573–583.

43. Loffert D, Ehlich A, Muller W, Rajewsky K (1996) Surrogate light chain expression isrequired to establish immunoglobulin heavy chain allelic exclusion during early B celldevelopment. Immunity 4:133–144.

44. Harman BC, et al. (2006) Mouse plasmacytoid dendritic cells derive exclusively fromestrogen-resistant myeloid progenitors. Blood 108:878–885.

45. Passegue E, et al. (2005) Global analysis of proliferation and cell cycle gene expressionin the regulation of hematopoietic stem and progenitor cell fates. J Exp Med 202:1599–1611.

46. Khan WN, et al. (1995) Defective B cell development and function in Btk-deficient mice.Immunity 3:283–299.

5778 � www.pnas.org�cgi�doi�10.1073�pnas.0811632106 Esplin et al.

Dow

nloa

ded

by g

uest

on

June

2, 2

020