Fas (CD95)-independent regulation of immune responses by antigen-specific CD4 CD8+ T cells

Antigen-activated T cells of the CD4⁺CD8^– and the CD4^–CD8⁺phenotype are susceptible to antigen receptor-stimulated celldeath. This form of apoptotic cell death has been shown to bedependent on the expression of the Fas (CD95) antigen and canoccur via an autocrine mechanism involving the concomitant up-regulationof Fas and its ligand on activated T cells. Mutations in genesencoding Fas (lpr) and the Fas ligand (gld) contribute to thedevelopment of an autoimmune syndrome similar to systemic lupuserythematosus in mice. These observations led to the suggestionthat the Fas signaling pathway is an important regulator ofimmune responses in vivo. Here we evaluated the importance ofthe Fas pathway in regulating immune responses by male antigen-specificCD4^–CD8⁺ T cells. We found that the in vivo eliminationof male antigen-activated cells was independent of Fas expressionby these cells. However, the elimination of these activatedcells was inhibited by the transgenic expression of Bcl-2, aprotein that inhibits multiple forms of apoptotic cell death.The transgenic Bcl-2 protein also inhibited the death of maleantigen-activated cells following IL-2 deprivation. Cell deathresulting from IL-2 deprivation occurred efficiently in maleantigen-activated Fas^- cells. We propose that the rapid deletionof male antigen-activated Fas^– cells in vivo is due tolimiting amounts of IL-2 that are available in the microenvironmentof the activated cells at the peak of the response.     

gld/gld mice are unable to express a functional ligand for Fas

Mice homozygous for either the lpr or gld genes develop phenotypically identical autoimmune disorders. The gene responsible for the pathology in lpr/lpr mice encodes the Fas antigen, a protein associated with the induction of programmed cell death. To determine if the defect associated with gld represents a mutation in the ligand for Fas, we have assessed the ability of lymphoid cells from homozygous gld/gld mice to lyse target cells in a Fas-dependent manner. Using an antagonistic antibody to Fas, we demonstrate that activated T cells from normal and lpr mice are capable of inducing Fas-mediated lysis of tumor target cells. In contrast, activated T cells from gld/gld mice fail to induce lysis of tumor targets, although cells from gld mice are able to lyse specific allogeneic targets following mixed lymphocyte culture. In addition, activated T cells from gld/gld homozygous animals are not capable of binding to a Fas.Fc fusion protein at high levels, whereas activated T cells from normal and lpr/lpr animals bind Fas.Fc efficiently. These data indicate that mice homozygous for gld are unable to express a functional ligand for Fas.     

Critical role of dendritic cells in determining the Th1/Th2 balance upon Leishmania major infection

The onset of T_h1 immunity is in part regulated by genetic background.To elucidate the cell type carrying critical factors determiningthe T_h1 response, we employed Rag-2^–/– mice on Leishmaniamajor-susceptible BALB/c and -resistant B10.D2 backgrounds.By using bone marrow (BM) chimeras generated by the transplantationof B10.D2 BM cells into BALB/c-Rag-2^–/– mice, andvice versa, it was shown that hematopoietic cells carry factorsdetermining the disease outcome and T_h1 response against L.major infection. B10.D2-Rag-2^–/– mice reconstitutedwith BALB/c CD4⁺ T cells exhibited a T_h1 response and controlledL. major infection. Wild-type BALB/c mice inoculated with L.major-parasitized B10.D2-Rag-2^–/– splenocytes alsoexhibited a T_h1 response and a mild disease outcome, whereassuch a T_h1 response was not induced when CD11c⁺ dendritic cells(DCs) were depleted from parasitized B10.D2-Rag-2^–/–splenocytes. T_h1 response was reconstituted by the additionof L. major-parasitized B10.D2 DCs but not L. major-parasitizedBALB/c DCs to DC-depleted parasitized B10.D2-Rag-2^–/–splenocytes. These results indicate that DCs determine the outcomeof the disease upon L. major infection.     

NK1.1+ T cells from IL-7-deficient mice have a normal distribution and selection but exhibit impaired cytokine production

Particular subsets of T cells expressing the NK1.1 antigen havebeen proposed to play an immune regulatory role by their fastand strong production of cytokines, in particular IL-4. We soughtto determine factors driving the functional differentiationof NK1.1⁺ T cells. Since NK1.1⁺ T cells are exquisitely sensitiveto IL-7 stimulation, we analyzed the development, selectionand IL-4 production of NK1.1⁺ T cells in IL-7 deficient mice(IL-7–/–mice). Besides a sharp reduction of allT cell subsets, NK1.1⁺ T cells develop at normal relative frequenciesin IL-7–/–;mice. They also undergo a normal selectionprocess, as revealed by the biased V_ß TCR repertoireidentical to the one in IL-7+/+ mice. However, NK1.1₊ T cellsfrom IL-7+/+ mice were found to be impaired in IL-4 and IFN-production in in vitro and in vivo models. In addition, IL-7was able to restore IL-4 production by NK1.1⁺ thymocytes fromIL-7–/– mice. Finally, IL-7 but not IL-4 was ableto maintain and increase IL-4 production by NK1.1⁺ thymocytesfrom normal mice. These data suggest that the functional maturationof NK1.1⁺ T cells requires a cytokine-driven differentiationprocess, in which IL-7 plays a major role.     

Characterization of the non-functional Fas ligand of gld mice

Mice homozygous for either the gld or Ipr mutation develop autoimmunediseases and progressive lymphadenopathy. The Ipr mutation Ischaracterized by the absence of unctional Fas, whereas gld miceexhibit an inactive FasL due to a point mutation proximal tothe extracellular C-terminus. The structural repercussions ofthis amino acid substitution remain unknown. Here we reportthat FasL Is expressed at similar levels on the surface of activatedT lymphocytes from gld and wild-type mice. Using a polyclonalanti-FasL antibody, Indistinguishable amounts of a 40 kDa proteinare detected In both gld and wild-type splenocytes. The molecularmodel of FasL, based on the known structure of TNF-, predictsthat the PheLeu gld mutation is located at the protomer interfacewhich Is close to the FasR Interaction site. We conclude thatthe gld mutation allows normal FasL biosynthesis, surface expressionand ollgomerlzatlon, but induces structural alterations to theFas binding region leading to the phenotypic changes observed.     

Proliferation and apoptosis of B220+ CD4 CD8 TCR{alpha}{beta}intermediate T cells in the liver of normal adult mice: implication for Ipr pathogenesis

Small numbers of T cells have been isolated from the normalmouse liver and many of these are of the CD4^–CD8^–TCRß⁺phenotype. Larger numbers of such cells are present in the liversof mice homozygous for the Ipr mutation and the liver has beenproposed to be the site of an extrathymlc T cell developmentpathway that is expanded in Ipr/lpr mice. Using a modified separationprocedure that increases the liver T cell yield, we have beenable to characterize a subset of CD4^–CD8^–TCRß^intermediateT cells that express the B220 epltope of the CD45 molecule,and resemble in this and many other ways the accumulating Tcells in Ipr lymph nodes. These cells are an actively dividingpopulation and even in healthy, unmanipulated mice a large proportionof them are undergoing apoptosis. We propose the model thatthe normal liver is a major site for T cell destruction andthat the Ipr defect results in failure of this process withleakage of B220⁺CD4^–CD8^–TCRß⁺ cells fromthe liver to peripheral lymphoid tissues, particularly lymphnodes.     

Emergence of CD52 , glycosyiphosphatidylinositol-anchor deficient lymphocytes in rheumatoid arthritis patients following Campath-1H treatment

CD52 is a glycosylphosphatidyl-inositol (GPI)-llnked glycoproteinexpressed at high levels on normal T and B lymphocytes and atlower levels on monocytes, while being absent on granulocytesand bone marrow stem cell precursors. The emergence of CD52^– lymphocytes of both T and B cell lineages was observedIn three out of 25 rheumatoid arthritis patients treated withthe humanized antibody Campath-1H in phase II clinical trial.Whereas the majority of CD52^– B cells had disappearedfrom the peripheral blood by 3 months post-treatment, both CD52^–CD4⁺ and CD8⁺ T cells persisted in the circulation for at least20 months. In the two patients that were tested, the GPI-anchoredsurface molecules CD55 and CD59 were also absent on the CD52^–cells, although expression of other cell surface transmembraneproteins (CD3, CD4 and CD2) was unaffected. The CD52^–cells maintained a stable phenotype in vitro despite repeatedre-stimulation in culture. Both CD52^– and CD52⁺ clones,established from one of the patients, responded to a similarextent to several T cell mitogens, as assessed by proliferation,suggesting that a general defect in expression of GPI-llnkedmolecules does not impair T cell activation. These data showthat an immune attack against a GPI-anchored surface moleculecan result in the selection of a GPI-anchor-deficient cell population.Despite the persistence of CD52^– T cells in the peripheralblood, no adverse reactions associated with the presence ofthese cells were noted in any of the patients; in fact theyresponded with longer remission times after Campath-1H treatmentthan the other patients in the trial. Received 16 May 1995, accepted 27 November 1995.     

Co-stimulation via CD28 induces activation of a refractory subset of MRL-Ipr/Ipr T lymphocytes

Peripheral lymphoid tissues of Ipr mice contain a large proportionof TCRß/CD3⁺CD4^–CD8^– T cells that lacksurface CD2 and express the B cell isoform of CD45, B220. Thissubset of T cells does not proliferate or produce IL-2 in responseto mitogenic signals or TCR–CD3 ligation. At the sametime, these abnormal T cells display several characteristicsof an activated phenotype. Collectively, these properties ofIpr CD4^–CD8^– T cells have functional parallels withanergic T cells. A critical co-stimulatory molecule implicatedin the prevention of or recovery from anergy is CD28, whichbinds the ligand BB1/B7 on certain accessory cells. Ipr CD4^–CD8^–T cells express normal levels of CD28 which is capable of transducinga strong proliferative signal to these cells in co-stimulationwith mitogens. However, proliferation of Ipr CD4^–CD8^–T cells in response to CD28 co-stimulation does not reach thelevels observed in normal T cells stimulated under similar conditions.Stimulation with anti-CD28 mAb in conjunction with phorbol myristateacetate and lonomycin promotes cell cycling in the CD2^–subset of CD4^–CD8^– T cells, and results in a slightinduction of CD2 levels during the course of the culture period.However, the majority of cells obtained at the end of the cultureperiod remain TCRß+ CD4^–CD8^–, CD2^low/–and B220^high, similar to freshly isolated CD4^–CD8^–Ipr T cells. In contrast, if IL-2 is included in the cultures,a strong shift toward a CD2⁺ phenotype is observed by a majorityof the Ipr T cells. Upon repeat stimulation, these Ipr CD4^–CD8^–T cells can now proliferate in an IL-2-dependent manner whenstimulated with only anti-CD3 mAb or mitogens, in the absenceof exogenous IL-2 or anti-CD28 mAb. These data show that thehyporesponsiveness of Ipr CD4^–CD8^– T cells doesnot result from a lack of CD28 expression, that it is not afixed state, and that it can be reversed by the induction ofcell cycling in the presence of IL-2. These observations extendthe parallels between Ipr CD4^–CD8^– T cells and anergicT cells.     

Contribution of Fas ligand to cardiac allograft rejection

Effector mechanisms for allograft injury remain unclear. Inthe present study, we verified the contribution of Fas and Fasligand (FasL) to cardiac allograft rejection by utilizing theFas-deficlent lpr or FasL-deficient gid mice as the donor orrecipient. Cardiac myocytes prepared from normal mice, but notthose from lpr mice, constitutively expressed Fas and were susceptibleto FasL-mediated lysis. Survival of cardiac allografts was substantiallyprolonged when gld or lpr mice were used as the recipient. Incontrast, cardiac allografts from ipr mice were normally rejectedwithout a delay. Histological examination of the grafts in thegld or lpr recipients demonstrated a lesser cellular infiltrationand much milder myocyte damage. Proliferative response and cytotoxicT lymphocyte induction against the donor-type alloantigens werenot impaired in the gld or lpr recipients. These results indicatea substaintial contribution of FasL to cardiac allograft rejection,independent of Fas in the grafts. This raises a possibilitythat FasL may be more generally involved in tissue damage associatedwith various diseases than expected from the expression of Fasin the target organs.     

Reversal of T cell unresponsiveness by augmentation of antigen presenting cell function

We have previously described epltopes of the 18 kDa proteinof Mycobacterium leprae which stimulate T and B cell responsesin different strains of mice. A series of overlapping 20-merpeptides that span the 18 kDa protein were used as immunogensto examine T and B cell recognition of different epltopes. Strain-specificvariation in the epltopes which induce the strongest responseswas affected by genes linked to the H-2 complex and the T cellresponses revealed by re-challenge with antigen were at leastpartially controlled by factors other than T cell specificity.We have examined the responses to one such antigen, peptlde1–20, which contains strongly immunogenlc epltopes forT and B cells. T cells from draining lymph nodes of peptlde1–20 immunized B10.BR, but not BALB/c mice, proliferatedIn vitro In response to rechallenge with peptlde 1–20or whole protein. Immunization with the same peptlde also inducedspecific antibody only in B10.BR mice. However, Immunizationof BALB/c mice results in ‘silent’ priming of Tcells since these can be induced to respond In vitro to thisantigen when cultured with activated macrophages as antigenpresenting cells (APC). The failure of APC from mBALB/c miceprimed with peptlde 1–20 to stimulate CD4⁺ proliferationwhen re-challenged In vitro and the failure to elicit antibodyresponses to peptlde 1–20 are presumably due to the samedefect in antigen-presenting cell function, since presentationof peptlde 1–20 by activated macrophages is sufficientto restore both responses. The failure of APC to stimulate responsesto this antigen in this model may be generally applicable toother cases of apparent non-responsiveness, and may have importantimplications for the understanding of T cell activation requirements.     

Suppression and Reversal ofgldDisease by Parabiosis with Normal Mice

The disruption of the Fas receptor or Fas ligand by thelprorgldmutations, respectively, results in severe autoimmune and lymphoproliferative disease due to the failure of Fas-mediated deletion of self-reactive lymphocytes. Recently, we have shown in mixed chimeras thatgld-induced autoimmunity could be corrected by normal bone marrow, in particular by normal T cells. In contrast,lpr-mediated autoimmunity could not be influenced by normal bone marrow-derived cells. In the present report, we have studied the role of normal lymphocytes in suppressing or reversinggld-induced autoimmunity by parabiosis with normal mice. Our results show a suppression of lymphadenopathy, fewer CD4⁻CD8⁻T cells, and lower levels of autoantibody production ingldmice parabiosed with normal mice at 4–6 weeks of age. Thegldmice parabiosed with normal mice at 4 months of age also exhibited a substantial reduction of both total and CD4⁻CD8⁻T cells in the periphery 2 months after surgery. However, they showed little reduction of autoantibodies compared togldmice parabiosed withgldmice. In contrast, olderlprmice did not exhibit any reduction in lymphadenopathy or autoantibody production after parabiosis with normal mice. The prevention or reversal of lymphadenopathy in parabiosedgldmice suggests that ongoing Fas-mediated deletion in the periphery may play an important role in maintaining self-tolerance. The relative irreversibility of autoantibody synthesis in older parabiosedgldmice suggests that autoantibody-producing B cells or their committed precursors are long lived and do not express functional Fas receptor.     

Spontaneous and antibody directed cytotoxicity of double-negative T cells from autoimmune mice

MRL/Mp-lpr/lpr (MRL/lpr) mice develop a syndrome similar tosystemic lupus erythematosus in humans. This strain of miceis characterized by the progressive accumulation of CD4^–CD8^–(double-negative; DN) T cells which express increased levelsof cell adhesion molecules such as CD44 and heat stable antigen(HSA). The DN T cells exhibited a higher level of spontaneouscytolytic activity and contained a higher level of serine esteraseas compared with T cells of MRL/Mp-+/+ (MRL/+) mice. We alsofound that mAbs against CD44, Mei-14, CD45R, and HSA could augmentthe cytolytic activity of DN T cells of MRL/lpr mice. Antibody-mediatedaugmentation of cytolytic activity of DN T cells was due toconjugate formation in which the Fc portion of mAb bound tothe Fc receptor on target cells and the Fab portion of mAb boundto corresponding cell surface antigens on DN T cells. The antibody-mediatedaugmentation of cytolytic activity was not detected in T cellsof MRL/+ mice and lymphokine activated killer (LAK) cells ofC57BL/6 mice. In contrast, anti-CD3 mAbs could augment the cytolyticactivity of DN T cells, T cells as well as LAK cells. mAbs againstLFA-1 and VLA-4 failed to augment the cytolytic activity ofthree different effector cells. It should be noted that antl-CD3mAb-mediated cytolytic activity of DN T cells was substantiallyreduced by anti-LFA-1 mAb. However, CD44, Mel-14, CD45R as wellas HSA-mediated cytolytic activity of DN T cells was not inhibitedby anti-LFA-1 mAb. The cell-cell and cell-matrix interactionsthrough cell adhesion molecules might augment the antigen non-specificcytolytic activity of DN T cells in vivo.     

Pervasive and stochastic changes in the TCR repertoire of regulatory T-cell-deficient mice

We hypothesize that regulatory T-cell (Treg)-deficient strainshave an altered TCR repertoire in part due to the expansionof autoimmune repertoire by self-antigen. We compared the Vβfamily expression profile between B6 and Treg-lacking B6.Cg-Foxp3^sf/Y(B6.sf) mice using fluorescent anti-Vβ mAbs and observedno changes. However, while the spectratypes of 20 Vβ familiesamong B6 mice were highly similar, the Vβ family spectratypesof B6.sf mice were remarkably different from B6 mice and fromeach other. Significant spectratype changes in many Vβfamilies were also observed in Treg-deficient IL-2 knockout(KO) and IL-2R KO mice. Such changes were not observed withanti-CD3 mAb-treated B6 mice or B6 CD4⁺CD25^– T cells.TCR transgenic (OT-II.sf) mice displayed dramatic reductionof clonotypic TCR with concomitant increase in T cells bearingnon-transgenic Vβ and V families, including T cells withdual receptors expressing reduced levels of transgenic V andendogenous V. Collectively, the data demonstrate that Treg deficiencyallows polyclonal expansion of T cells in a stochastic manner,resulting in widespread changes in the TCR repertoire.     

Induction of CD5 antigen on human CD5 B cells by stimulation with Staphylococcus aureus Cowan strain I

We have examined whether the CD5 phenotype could be inducedon human B cell surfaces by the polycional B cell stimulator,Staphyiococcus aureus Cowan strain I (SAC). Fresh tonsillarB cells were prepared by Percoll density gradient from E^–cells. The proportion of CD5⁺ B cells In the 50/60% and 60/70%interface high-density fractions varied between 1.2 and 10.2%depending on the tonsil preparations when they were placed onthe in vitro culture 12–60 h prior to flow cytometrlcanalysis. The expression of CD5 antigen obviously increasedin the presence of SAC (1:10⁵ v/v). The percentage of CD5⁺ Bcells varied from tonsil to tonsil, from 25.1 to 65.9% in aseries of experiments. The CD5⁺ B cells were found both amongCD23⁺CD25⁺CD71⁺ and CD23^–CD25^–CD71^– B cells.The level of CD5 expression was related to the cell size eniargement.The addition of anti-CD5 antibody in the culture blocked theCD5 induction by SAC without interfering with the expressionof other activation markers. A time-course study showed thatCD5 antigen appeared to be induced on the cell surface duringthe G₀ to G₁ phase transition in the cell cycle. When CD5⁺ andCD5^– B cells were separated by magnetic isolation, theCD5^– B cells showed DNA synthesis to the stimulation bySAC and expressed CD5 antigen on their cell surface. These resultssuggest that human CD5^– B cells can express the CD5 phenotypeby stimulationwith the polyclonal B cell stimulator, SAC.     

Induction of phosphocholine-specific antibodies in X-linked immune deficient mice: in vivo protection against a Streptococcus pneumoniae challenge

X-linked Immune deficient (XID) mice are susceptible to infectionwith Streptococcus pneumoniae because they fail to mount animmune response to the Immunodomlnant phosphochollne (PC) epltopeon the bacterial cell wall. It is difficult to induce PC-speclflcantibodies in XID mice because PC-specific B cells expressingthe T15-, M167- and M603 Idiotype (Id), which provide protectionagainst S. pneumoniae, are deleted in these mice via an antigen-specific,receptor-mediated process. In addition, the standard PC hapten,p-dlazophenylphosphochollne (DPPC), induces high affinity phenylphosphochollne(PPC)-speciflc antibodies in XID mice, which are not protectiveagainst S. pneumoniae. We have used a novel PC hapten, p-nitrophenyl-6-(O-phospho-choline)hydroxyhexanoate(EPC), to induce PC-specific antibodies in XID mice. The immuneresponse to EPC-keyhole limpet hemacyanln (KLH) is dominatedby IgGi, V_H1⁺, US-Id^–, PC-inhlbitable antibodies. A smallIgM antl-PC response having a consistent T15-ld⁺ component isalso induced in XID mice, whereas normal mice produce a largeIgM response dominated by T15-ld⁺ antibodies. The immune responseto EPC-KLH remains predominantly PC-lnhlbltable even after multipleimmunizations, while the response to DPPC–KLH becomesdominated by PPC-speclflc antibodies. C.CBA/N mice immunizedtwice with EPC–KLH are protected against 10⁴ S. pneumoniaewhile as few as 10 bacteria are 100% lethal for the unlmmunlzedcontrols. The ability of EPC-protein to induce a long-lived,PC-speclflc response should make this hapten a potential TDvaccine candidate for S. pneumoniae     

Frequent deletion of the transgene in T cell receptor {beta} chain transgenic mice

TCR V_ß8.1 transgenic mice were generated using a genomicTCR V_ß gene construct under the control of its promoterand enhancer. Among three lines of transgenic mice, one lineexpressed the transgenic TCR on only 70% of peripheral T cells,while the other two lines expressed it on almost all matureT cells. T cells which lacked expression of the transgenic TCRß chain expressed endogenous TCR ß chains.The molecular basis underlying the lack of transgene expressionin T cells of this line of transgenic mice was investigated.The transgenic TCR^– cells were isolated by two methods.First, Thy-1⁺ V_ß8.1/8.2^– cells were purifiedfrom peripheral T cells using cell sorting. Second, transgenicTCR^– T cell clones were established. In both cases, Southernblotting indicated that V_ß8.1^– T cells had deletedthe transgenic TCR gene. Thus, deletion of the transgenic TCRcan occur in a high proportion of T cells, which allows rearrangementand expression of endogenous TCR ß chains.     

A specific intercellular pathway of apoptotic cell death is defective in the mature peripheral T cells of autoimmune lpr and gld mice

Homozygosity for either of the unlinked murine autosomal recessive mutations lpr or gld leads to autoimmunity characterized by peripheral accumulation of CD4^?/CD8^? “double-negative” T cells, autoantibodies and various forms of tissue pathology. Recently, the gene affected by lpr was identified as fas, whose product acts as a trigger for programmed cell death or apoptosis. Data reported here indicate that the Fas receptor and its ligand, the wild-type form of the gld gene product, are essential for antigen-stimulated peripheral T cell apoptosis. Furthermore, the wild-type gld gene product is a non-cell-autonomous protein that is produced by activated T cells. Apoptotic elimination of antigen-receptor-triggered peripheral T cells appears to be abnormal in lpr and gld mice, and this deficiency causes peripheral T cells to accumulate resulting in lymphadenopathy. These findings support the importance of apoptotic regulation of lymphocyte persistence after antigen encounter in vivo.     

A novel form of self tolerance dictated in the thymus of transgenic mice with autoreactive TCR {alpha} and {beta} chain genes

Transgenic (TG) mice with TCR and ß chain genes froma CD4-dependent auto-l-A^k reactive T cell clone were generated.H-2^k TG mice had a large number of thymic and splenic CD4 Tcells expressing the autoreactive TCR without manifestationof autolmmunlty. The cells were not anergic, as they could respondto autologous antigen presenting cells and antl-TCR antibodiesin vitro to proliferate and to produce interleuklns. Variousdegrees of down-regulation of CD2 and CD44 was observed in TGmice, Indicating the presence of a defective co-stlmulatoryprocess in TG T cells. These features indicate that the selftolerance in autoreactive TCR TG mice is due not to clonal deletionand anergy but to a novel mechanism where T cells cannot sufficientlyrespond to normally existing self ligand in vivo. That suchan in vivo unresponsiveness of autoreactive T cells is dictatedin the thymus during CD4 T cell differentiation as an atypicalform of positive selection of autoreactive T cells was suggestedby the abnormal surface expression of CD69 and HSA.     

ER-MP12 antigen, a new cell surface marker on mouse bone marrow cells with thymus-repopulating ability: II. Thymus-homing ability and phenotypic characterization of ER-MP12-positive bone marrow cells

In the accompanying paper we showed that six distinct subsetsof bone marrow (BM) cells can be identified using the mAb ER-MP12and ER-MP20 in two-colour immunofluorescence analysis. Uponintrathymic transfer into sublethally irradiated mice thymus-repopulatingability was restricted to ER-MP20^– BM cells expressingeither high or intermediate levels of the ER-MP12 antigen (1–2%and –30% of BM nucleated cells respectively). The highestfrequency of thymus-repopulating cells was found in the minorsubset of ER-MP12⁺⁺20^– BM cells. In the present studywe demonstrate that upon intravenous transfer, thymus-homingand-repopulating BM cells are exclusively confined to the ER-MP12⁺⁺20^–and ER-MP12⁺20^– subpopulations, the highest frequencybeing detected among ER-MP12⁺⁺20^– BM cells. Analysis ofthe peripheral blood leucocytes of reconstituted mice showedthat not only prothymocytes but also progenitorcells of theB cell lineage as well as the myelold lineage were present withinboth subsets. Three-colour flow cytometric analysis revealedthat ER-MP12⁺⁺20^– BM cells in particular were phenotyplcallyheterogeneous with respect to the expression of the cell surfacemarkers Thy-1, Sca-1, CD44, B220 and c-kit. Taken together ourdata demonstrate that ER-MP12 positively identifies BM cellswith the ability to home to and repopulate the thymus. The phenotypicheterogeneity displayed by the ER-MP12⁺⁺20^– BM subset,containing the highest frequency of thymus-homing and-repopulatingcells, provides a basis for further separation of prothymocyteactivity from other haematopoietic activities in the BM of themouse.