Phenotypic and Functional Analysis of Activated B Cells ofAutoimmune NZB×NZW F1 Mice

Y.-L. YE*, J.-L. SUEN*, Y.-Y. CHEN* & B.-L. CHIANG*†

*Graduate Institute of Immunology and†The Graduate Institute of Clinical Medicine, College of Medicine, National Taiwan University, Taipei,Taiwan, China

(Received 17 June 1997; Accepted in revised form 18 November 1997)

Ye Y-L, Suen J-L, Chen Y-Y, Chiang B-L. Phenotypic and Functional Analysis of Activated B Cells ofAutoimmune NZB×NZW F1 Mice. Scand J Immunol 1998;47:122–126

Polyclonal B-cell activation is the central theme in the production of autoantibodies and possible activation ofautoreactive T cells in both human and murine lupus. The abnormal expansion of CD5þ B cells in murinelupus has been suggested, in particular, to be one of the most characteristic findings in these mice. Activated Bcells can be separated from the B cells of resting stage by the difference in cell density. The aim of this studywas to investigate the characteristics of different densities of the spleen cells separated by gradient density.Furthermore, the ability of anti-DNA antibody secretion in each percoll gradient fraction of B cells was alsoanalysed. The results showed: a higher percentage of CD5þ B cells, which corresponded to the activated B-cell population, in percoll gradient 1 and 2 fractions; that splenic B cells of NZB/W F1 mice had proliferativeresponse to interleukin (IL)-4 or IL-5 but not to IL-10 or interferon-g (IFN-g); and that B cells isolated bypercoll gradient produced anti-DNA antibody after stimulation with lipopolysaccharide (LPS) plus IL-5 andIFN-g, but not IL-4 and IL-10. These data suggest that B cells at different stages of activation expressdifferential characteristics and functions.

Dr Bor-Luen Chiang, Graduate Institute of Clinical Medicine, College of Medicine, National TaiwanUniversity No. 1, Chang-Teh Street, Taipei, Taiwan, Republic of China

INTRODUCTION

Murine lupus models such as NZB×NZW F1 (NZB/WF1),NZB.H-2bm12, NZB×SWR F1 (SNF1), MRL.lpr/lpr and BXSBmice have shed much insight into the understanding of thepathogenic mechanisms of lupus [1–3]. Among these murinelupus models, the natural course of NZB/W F1 mice is closer tohuman lupus than that of MRL.lpr/lpr and BXSB mice. Althoughthe exact mechanisms of murine lupus in NZB-derived mice arestill not well defined, multiple gene involvement has beensuggested by most researchers [4].

Polyclonal activation of B cells might be the critical back-ground abnormality in the pathogenesis of autoimmunity inNZB-derived mice [5–7]. The severity of polyclonal B-cellactivation might determine the production of autoantibodiesand the fate of the autoimmune mice. One of the major B-cellabnormalities in NZB/W F1 mice is the increased CD5þ (B-1a)B-cell frequency in both the spleen and the peritoneal cavity [8].These CD5þ B cells have been shown to be responsible forspontaneous secretion of IgM, which recognizes certain auto-antigens such as erythrocyte, thymocytes and single-stranded (ss)

DNA [9]. Although the relationship between CD5þ B cells andIgG anti-double-stranded (ds) DNA antibody production needs tobe further defined, it might play an important role in thepathogenesis of disease. A previous study has demonstrated T-helper cell dysregulation in autoimmune-prone NZB/W F1 micebut not in non-autoimmune C57BL/6 mice [10]. Cytokines, suchas IL-5 produced by T helper 2 (Th2) cells, are closely related tothe function and growth of autoreactive B cells and are respon-sible for autoantibody production [11, 12]. In addition, anti-interleukin (IL)-10 antibody-treated mice showed depletion ofCD5þ B cells but not conventional B cells, and delayed the onsetof disease in NZB/W F1 mice [13].

Activated cells can be separated from resting B cells usingdensity gradient centrifugation. Low-density B cells are largecells that proliferate to IL-2 and express cell-surface markersassociated with activation [14, 15]. Using this method, spleencells can be divided into five fractions, and phenotypes of eachcell fraction can be analysed by fluorescence-activated cell sorter(FACScan). Anti-DNA Ab secretion can also be studied byculturing each percoll fraction of B cells with lipopolysaccharide(LPS) plus different cytokines.

Scand. J. Immunol. 47, 122–126, 1998

122 q 1998 Blackwell Science Ltd

MATERIALS AND METHODS

Mice. All mice were maintained at the Animal Center, College ofMedicine, National Taiwan University. Two-month-old female NZB/WF1 mice were purchased from Charles River Japan (Tokyo, Japan). Theanimal room was on a 12-h-light and 12-h-dark cycle; constant tem-perature (258C6 28C) and humidity were maintained. Each experimentused 6–10 mice. Animal care and handling procedures conformed to theNIH Guide for the Care and Use of Laboratory Animals(NRC 1985).

Culture medium.The culture medium used for most of the experi-ments was RPMI 1640 medium supplemented with 10% fetal calfserum (FCS), 4 mM L-glutamine, 25 mM HEPES (pH 7.2), 5×10¹5

M 2-mercaptoethanol, 100 U/ml penicillin, 100mg/ml streptomycin and0.25 mg/ml amphotericin.

Reagents.Appropriate fluorescein isothiocyanate (FITC) or phycoer-ythrin (PE)-conjugated monoclonal antibodies such as anti-B220 (cloneRA3–6B2), anti-CD3-e (clone 145–2C11), anti-CD5 (Ly-1, clone 53–7.3), anti-CD23 (clone B3B4) anti-H-2Kd (clone AF3–12.1) and anti-I-Ad (clone AMS–32.1) (PharMingen, San Diego, CA) were used forphenotypic analysis of peritoneal exudate cells and spleen cells isolatedfrom NZB/W F1 mice. For anti-dsDNA antibody enzyme-linked immu-nosorbent assay (ELISA), horseradish peroxidase-conjugated goat anti-mouse IgG (g-chain specific; Jackson Inc., West Grove, PA) and anti-mouse IgM (m-chain specific; Jackson Inc.) were used as secondaryantibodies.

Cell separation.Spleens were removed and single cells were isolatedby lysing red blood cells with 0.144M ammonium chloride-0.17M Trisbuffer, then washing three times with Hank’s solution before use. B cellswere isolated by treating spleen cells with anti-thy 1.2 (HO-13–4;Cedarlane, Westbury, NY) plus complement (Cedarlane).

Activated cells could be separated from resting B cells using densitygradient centrifugation [12]. Low-density B cells are large cells and arewell activated. By contrast, high-density B cells are less activated.Sterile 1.5M NaCl (8 ml volume) containing 25 mM Hepes (pH 7.3)was added to 92 ml of Percoll (Pharmacia/LKB, Pistacataway, NJ).This stock solution of Percoll (92%) was then diluted to 40, 45, 50and 65% by the addition of RPMI 1640 medium containing 1 mg/mlbovine serum albumin (BSA), corresponding to calculated densities of1.060, 1.066, 1.073, 1.079 and 1.092 g/ml, respectively. Gradients wereformed by adding 2 ml of each fraction successively into 15 ml conicaltubes. Spleen cells or T-depleted cells were layered on top of the gradient(5–7×107) and the tubes were centrifuged for 25 min at 1000×g.Fractions 1–5 consisted of the 40%, 40/45%, 45/50%, 50/55% and55/65% interfaces to the bottom of the tube. The cells at each interface

were collected using a pipette, recovered by centrifugation, washedtwice and resuspended in medium. Different cell sizes of these fractionswere characterized by FACScan.

Proliferative response of different percoll gradient B cells by cyto-kines.To test proliferative activity of different percoll gradient B cells,1× 105 cells per well of percoll gradient B cells were added with LPS(Sigma Chemical Co., St. Louis, MO) (15mg/ml), IL-4 (6.25–200 U/ml),IL-5 (1.65–56 U/ml), IL-10 (0.78–25 U/ml) or interferon-g (IFN-g) (1–16 U/ml) (PharMingen) to 96-well round plates. After 3 days, 1mCi [3H]-thymidine per well was added to the culture and the cells were harvested18 h later.

FACScan analysis.Phenotypic analysis of spleen cells was carried outby FACScan (Becton Dickinson, Mountain View, CA). Aliquots of cells(2.5–5× 105) were suspended in 0.1 ml of PBS, 0.1% sodium azide andincubated at 48C for 30 min with predetermined optimal concentrationsof appropriate FITC- or PE-conjugated monoclonal antibodies(PharMingen). The cells were washed and resuspended in 0.5 ml ofPBS containing 0.1% sodium azide and subjected to FACScan analysis.A total of 5000 cells were counted and the frequency of each cell surfacemarker was determined using appropriate software (Becton Dickinson).The cells suspended in medium served as controls. Flow cytometry wasregularly calibrated with CaliBRITE beads (Becton Dickinson).

Anti-DNA antibody production by each fraction of B cells.To furtherinvestigate anti-DNA antibody production, each fraction of B cells wasstimulated with LPS (15mg/ml, Sigma) plus predetermined concentra-tions of various cytokines such as IL-4 (100 U/ml), IL-5 (3.85 U/ml), IL-10 (100 U/ml) or IFN-g (4 U/ml) (PharMingen). After 5 days of incuba-tion, the supernatant was collected and the anti-DNA antibody level wasassayed using ELISA. The ELISA method for anti-DNA antibodydetermination was modified as described previously [16]. Briefly,plates were initially coated with 100ml/well of 10mg/ml methylatedBSA (Sigma Chemical Co.). After overnight incubation at 48C, plateswere washed with PBS, then 100ml/well of 2.5mg/ml DNA was added.Native DNA antigens used in this experiment were prepared by phenol/chloroform extraction of calf thymus DNA (Sigma Chemical Co.). Forthe assessment of anti-ssDNA antibody, native DNA was denatured byboiling for 20 min and then incubated, with agitation, in ice for 20 min.Supernatants were added to the appropriate well and incubated at 378Cfor 50 min and at room temperature for 10 min. Then, the plates werewashed three times with PBS containing 0.05% Tween 20, followed bythe addition of peroxidase-conjugated rabbit anti-mouse IgG (Hþ l)antibodies or goat anti-mouse IgM (Jackson Inc.)-specific antibodies.After 2 h of incubation, adequate substrate (ABTS) was added toplates and 100ml of 5% SDS was used to stop the reaction. The results

Activated Autoreactive B cells and Autoantibody Production123

q1998 Blackwell Science Ltd,Scandinavian Journal of Immunology,47, 122–126

Table 1. Phenotypic analysis of different percoll fractions of spleen cells

Total cells P1† P2 P3 P4 P5

CD3þ 56.06 5.3* 17.76 5.4 21.66 2.6 31.66 8.7 58.26 10.2 71.86 4.9B220þ 33.46 5.7 64.46 15.4 61.26 1.7 57.26 9.6 24.26 7.8 10.56 5.7CD5þ/B220þ 5.86 1.2 11.16 5.0 13.26 1.9 6.36 1.7 2.46 0.5 1.96 1.2CD23¹/B220þ 22.66 16.7 50.26 8.8 45.36 2.3 28.06 10.3 16.56 8.7 9.06 4.8IAd 27.06 4.7 62.36 11.6 64.56 1.9 47.86 6.4 26.76 7.4 33.96 4.4Kd 98.06 0.6 95.66 1.7 97.76 1.8 98.36 0.4 98.56 0.5 95.76 1.9

* Mean 6 SD, n¼ 6.† P1, percoll fraction 1 with a density of 1.060; P2, percoll fraction 2 with a density of 1.066; P3, percoll fraction 3 with a density of 1.073; P4, percollfraction 4 with a density of 1.079; P5, percoll fraction 5 with a density of 1.092.

of anti-DNA antibody were presented as ELISA units (E U/ml) com-pared to monoclonal antibody (MoAb) 742H.7D with a concentration of746 0.5 ng/ml, specific for dsDNA, a gift from Dr M.E. Gershwin,University of California at Davis, Davis, CA.

Statistical analysis.Student’st-test andx2-test were used to analysethe data throughout the study. AP-value of<0.05 was consideredstatistically significant.

RESULTS

The percentage of CD5-positive B cells in each percoll gradientfraction

Phenotypic analysis of splenic cells from NZB/W F1 mice issummarized in Table 1. The data suggest that the use of FACScananalysis can denote a higher percentage of CD5þ B cells (B-1a) inpercoll gradient 1 and 2 fractions. This corresponded to activated

B-cell populations (Fig. 1). In addition, the frequency of B-1(CD23¹) cells was also higher in the cells of fraction 1 and 2.The data indicated that both B-1 (CD23¹) and B-1a (CD5þ)cells accounted for most of the activated B cells in autoimmuneNZB/W F1 mice.

Proliferative response of splenic non-T cells to differentcytokines

B cells of NZB/W F1 mice were isolated by treating spleen cellswith anti-thy 1.2 (HO-13–4) plus complement. After treatment,the percentage of T cells was< 5% by FACS analysis. Further-more, the cells were stimulated with different concentrations ofcytokines such as IL-4, IL-5, IL-10 and IFN-g or LPS as positivecontrol. The data suggested that splenic non-T cells had a betterproliferative response to IL-4 or IL-5 than to IL-10 or IFN-g

(Fig. 2).

In vitro anti-DNA antibody production after stimulation withmitogen or cytokines

After stimulation with LPS plus cytokine, the supernatants ofeach percoll gradient fraction B-cell culture were collected andassayed for the level of anti-DNA antibody. B cells stimulatedwith LPS plus IL-5 or IFN-g all produced high levels of anti-DNA antibody. (Fig. 3, A and B). By contrast, the level of anti-DNA antibody produced by B cells stimulated with LPS plus IL-4 or IL-10 was very low (Fig. 3, C and D). In general, B cells oflower density produced a higher level of anti-DNA antibody.

DISCUSSION

All murine lupus models such as NZB×NZW F1 (NZB/W F1),NZB.H-2bm12, NZB ×SWR F1 (SNF1), MRL.lpr/lpr and BXSBmice develop IgG anti-dsDNA antibody, a characteristic oflupus, and the mice die of uraemia in early life. Some genetic

124 Y.-L. Ye et al.

q1998 Blackwell Science Ltd,Scandinavian Journal of Immunology,47, 122–126

Pe

rce

nta

ge

75

Spleen

cells

P1 P2 P3 P4 P5

50

25

0

100B–1

B–1a

**

Fig. 1. The percentage of B-1 (CD23¹) and B-1a (CD5þ) B cells ineach fraction of B cells separated by percoll gradient. The datasuggested a higher percentage of B-1 and B-1a B cells in fractions 1and 2, which corresponded to an activated status of cells. The valuewas presented as mean6 SD, n¼ 6. *P<0.05.

SI

60

1

50

20

0

40

30

10

A

2 4 8 16 LPS

60

6.25

50

20

0

40

30

10

B

12.5 25 50 200 LPS100

SI

60

1.65

50

20

0

40

30

10

C

3.5 7 14 28 LPS

80

0.78

60

20

0

4030

10

D

1.56 3.13 6.25 25 LPS12.556(U/ml) (U/ml)

70

50

Fig. 2. Proliferative response of splenic non-Tcells to different cytokines: (A) IFN-g; (B)IL-4; (C) IL-5; and (D) IL–10. Splenic non-Tcells of NZB/W F1 mice had a proliferativeresponse to IL-4 or IL-5 but not to IFN-g orIL-10. The data were presented as mean6 SDfor three animals per time point from onerepresentative experiment out of threeperformed.

contributions like thelpr gene in MRL mice, and the Y chromo-some in BXSB mice have been well documented [17, 18]. Recentstudies on thelpr gene have further elucidated the nature ofsingle-gene effects on the development of systemic autoimmunediseases [19, 20]. Among these murine lupus models, the naturalcourse of NZB/W F1 mice is closer to human lupus than that ofMRL.lpr/lpr and BXSB mice. Although the exact mechanisms ofmurine lupus in NZB-derived mice are still not well defined,multiple gene involvement has been suggested by most research-ers [4]. Among these genes, polyclonal B-cell activation has beenregarded as one of the most important factors [21, 22].

One of the major B-cell abnormalities in NZB/W F1 mice isthe increased B1-a (CD5) B-cell frequency in both the spleen andthe peritoneal cavity [8]. These CD5þ B cells have been shown tobe responsible for spontaneous secretion of IgM, and the speci-ficity of these antibodies is directed to thymocytes and ssDNA[9]. Although the relationship between CD5þ B cells and IgGanti-dsDNA antibody production needs to be further defined, itmight play an important role in the pathogenesis of disease. Thedata here demonstrate a higher percentage of CD5þ B cells in thefractions with a low density of B cells, which corresponds toactivated cells. Previous data had suggested that both CD5þ andCD5¹ B cells of NZB/W F1 mice can produce IgG anti-dsDNAantibody in vitro and in vivo [23]. However, the data hereinfurther demonstrate a higher percentage of CD5þ B cells in thepercoll fractions of low-density B cells. Although either CD5þ orCD5¹ B cells can produce IgG anti-dsDNA antibody, CD5þ Bcells might represent a population of activated B cells.

The data suggest that splenic cells of NZB/W F1 miceproliferate to both IL-5 and IL-4 stimulation, but not to IL-10or IFN-g. IL-4 has been found to be a B-cell growth factor that

can induce B-cell switching to IgE or IgG1. Furthermore, it hasbeen well documented that certain cytokines, such as IL-5, canactivate B cells of murine lupus to proliferate and produceautoantibodies [11, 12]. Several studies have demonstrated thata higher level of IL-5 receptor (IL-5R) is expressed on CD5þ Bcells [24], which may account for both proliferative response andautoantibody production of CD5þ B cells stimulated by IL-5.The data here suggest that both IL-5 and IFN-g, in the presenceof B-cell mitogen, can stimulate autoreactive B cells of NZB/WF1 mice to produce anti-DNA antibody. IFN-g has been found toenhance proliferation ofStaphylococcus aureusCowan I (SAC)or anti-m-stimulated B cells when present in the early stage ofculture [25]. By contrast, IFN-g alone has been found to inhibitthe growth of B-1 cells [26]. All these findings agree with thedata that IFN-g can enhance the autoantibody production byLPS-stimulated B cells, but IFN-g alone cannot induce prolif-eration of activated B cells that are mainly B-1 cells. Severalpieces of evidence also suggested that there were higher levels ofIL-5 and IL-10 in autoimmune NZB/W F1 mice compared withnon-autoimmune mice [10]. IL-10 has been found to be anautocrine growth factor for B-1 cellsin vivo [13, 27]; however,the data presented here suggests no significant proliferativeresponse or autoantibody production of activated B cells stimu-lated with IL-10. Although more studies are needed, IL-10 mayindirectly mediate the growth of B-1 cells by other growth factorsor mechanisms. Furthermore, the production of Th2-relatedcytokines has been demonstrated in autoreactive T cells derivedfrom NZB/W F1 mice [28]. Th2-related cytokines, such as IL-5,produced by autoreactive T cells may play a critical role in theautoreactive B cell expansion and autoantibody production.

Hyperfunction of B cells may be the central background

Activated Autoreactive B cells and Autoantibody Production125

q1998 Blackwell Science Ltd,Scandinavian Journal of Immunology,47, 122–126

EU

P1

2

0

3

1

A4

P2 P3 P4 P5

ss IgMss IgGds IgMds IgG

P1

2

0

3

1

B4

P2 P3 P4 P5

EU

P1

2

0

3

1

C4

P2 P3 P4 P5 P1

2

0

3

1

D4

P2 P3 P4 P5

Percoll fractionsPercoll fractions

Fig. 3. Anti-DNA antibody production from Bcells stimulated with LPS plus differentcytokines: (A) IFN-g; (B) IL-4; (C) IL-5; and(D) IL-10. The data suggested that eitherIFN-g or IL-5 could stimulate splenic non-Tcells to produce anti-DNA antibodyin vitro.The data were presented as mean6 SD forthree animals per time point from onerepresentative experiment out of threeperformed.

abnormality in murine lupus models such as NZB/W F1 mice.The data here suggest that B-1 or B-1a cells may represent asubpopulation of B cells, which are more highly activatedin vivoand tend to produce more autoantibodies in autoimmune-proneNZB/W F1 mice.

ACKNOWLEDGMENT

This study was supported by grant no. NTUH 84118-A18 fromthe National Taiwan University Hospital of the Republic ofChina.

REFERENCES

1 Steinberg AD, Huston DP, Taurog JD, Cowdery JS, Raveche ES.The cellular and genetic basis of murine lupus. Immunol Rev1981;55:121–54.

2 Theofilopoulos AN, Dixon FJ. Etiopathogenesis of murine SLE.Immunol Rev 1981;55:179–216.

3 Theofilopoulos AN, Kofler R, Singer PA, Dixon FJ. Moleculargenetics of murine lupus models. Adv Immunol 1989;46:61–109.

4 Chused TM, McCoy KL, Lal RB, Brown EM, Baker PJ. Multigenicbasis of autoimmune disease in New Zealand mice. ConceptsImmunopathol 1987;4:129–43.

5 Kincade PW, Lee G, Fernandes G. Abnormalities in clonable Blymphocytes and myeloid progenitors in autoimmune NZB mice.Proc Natl Acad Sci USA. 1979;76:3464–8.

6 Kastner DL, Steinberg AD. Determinant of B cell hyperactivity inmurine lupus. Concepts Immunopathol 1988;6:22–88.

7 Moutsopoulos HM, Bohem-Truitt M, Kassan SS, Chused TM.Demonstration of activation of B lymphocytes in New ZealandBlack mice at birth by an immunoradiometric assay for murineIgM. J Immunol 1977;119:1639–44.

8 Herzenberg LA, Stall AM, Lalor PA, Moore WA, Parks DR,Herzenberg LA. The Ly-1 B cell lineage. Immunol Rev1986;93:81–102.

9 Kantor AB. The development and repertoire of B-1 cells (CD5 Bcells). Immunol Today 1991;12:389–91.

10 Lin L-C, Chen Y-C, Chou C-C, Hsieh K-H, Chiang B-L. T helpercell dysfunction in autoimmune prone NZB× NZW F1 mice: the roleof cytokines in systemic lupus erythematosus. Scand J Immunol1995;42:466–72.

11 Umland SP, Go NF, Cupp JE, Howard M. Responses of B cells fromautoimmune mice to IL–5. J Immunol 1988;142:1528–35.

12 Cawley D, Chiang B-L, Naiki M, Ansari A, Gershwin M. Compar-ison of the requirements for cognate T cell help for IgG anti-dsDNAantibody productionin vitro: T helper-derived lymphokines replaceT cell cloned lines for B cells from NZB. H-2bm12 but not B-2bm12

mice. J Immunol 1993;50:2467–77.

13 Ishida H, Hastings R, Kearney J, Howard M. Continuous anti-interleukin 10 antibody administration depletes mice of Ly-1 Bcells but not conventional B cells. J Exp Med 1992;75:1213–20.

14 Zubler RH, Lowenthal JW, Erard F, Hashimoto N, Devos R,Macdonald HR. Activated B cells express receptors for and prolif-erate in response to pure interleukin–2. J Exp Med 1984;160:1170–83.

15 Camp RL, Kraus TA, Birkeland ML, Pure E. High levels of CD44expression distinguish virgin from antigen-primed B cells. J ExpMed 1991;173:763–6.

16 Cawley D, Chiang B-L, Ansari A, Gershwin ME. Ionic bindingcharacteristics of monoclonal autoantibodies to DNA from NZB. H-2bm12 mice. Autoimmunity 1991;9:301–9.

17 Cohen I, Eisenberg RA.Lpr andgld: single gene models of systemicautoimmunity and lymphoproliferative disease. Ann Rev Immunol1991;9:243–69.

18 Merino R, Fosati L, Lacour M, Izui S. 1991. Selective autoantibodyproduction by Yaaþ B cells in autoimmune Yaaþ -Yaa¹ bonemarrow chimeric mice. J Exp Med 1991;174:1023–9.

19 Watanabe-Fukunaga R, Brannan CI, Copeland NG, Jenkins NA,Nagata S. Lymphoproliferation disorder in mice explained bydefects in Fas antigens that mediates apoptosis. 1991;356:314–7.

20 Watson ML, Rao JK, Gilkeson GSet al. Genetic analysis of MRL-lpr mice: relationship of the Fas apoptosis gene to disease manifesta-tions and renal disease modifying loci. J Exp Med 1992;176:1645–56.

21 Steinberg BJ, Smathers PA, Frederiksen K, Steinberg AD. Ability ofthe xid gene to prevent autoimmunity in (NZB× NZW) F1 miceduring the course of their natural history, after polyclonal stimula-tion, or following immunization with DNA. J Clin Invest1982;70:587–97.

22 Klinman DM. Polyclonal B cell activation in lupus-prone miceprecedes and predicts the development of autoimmune disease. JClin Invest 1990;86:1249–54.

23 Ye Y-L, Chuang Y-H, Chiang B-L.In vitro and in vivo functionalanalysis of CD5þ and CD5¹ B cells of autoimmune NZB×NZW F1mice. Clin Exp Immunol 1996;106:253–8.

24 Hitoshi Y, Yamaguchi N, Tominaga A, Takatsu K. Coexpression ofCD5 and IL-5 receptor on peritoneal B cells. Ann NY Acad Sci1992;651:261–3.

25 Li L, Young D, Wolf SF, Choi YS. Interleukin-12 stimulates B cellgrowth by inducing IFN-g. Cell Immunol 1996;168:133–40.

26 Velupillai P, Sypek J, Harn DA. Interleukin-12 and -10 and gammainterferon regulate polyclonal and ligand-specific expansion ofmurine B-1 cells. Infect Immun 1996;64:4557–60.

27 O’Garra A, Chang R, Go N, Hastings R, Haughton G, Howard M.Ly-1 B (B-1) cells are the main source of B cell-derived interleukin-10. Eur J Immunol 1992;22:711–7.

28 Chen Y-C, Ye Y-L, Chaing, B-L. Establishment and characterizationof cloned CD4¹ CD8¹ ab-TCR bearing autoreactive T cells fromautoimmune NZB/W F1 mice. Clin Exp Immunol 1997;108:52–7.

126 Y.-L. Ye et al.

q1998 Blackwell Science Ltd,Scandinavian Journal of Immunology,47, 122–126