Isolation and characterization of high endothelial cell lines derived from mouse lymph nodes.

Two long-term cultured cell lines were established from BALB/c mouse axillary and cervical lymph nodes that exhibited a combination of functional, morphological, and phenotypic characteristics consistent only with high endothelial venule cells. Spleen lymphocytes selectively bound and migrated across the cell lines. On Matrigel, these cell lines formed tubules with lumens, a characteristic unique to endothelial cells. Morphologically the cells were 20-30 microns in diameter and exhibited contact inhibition. The cells were not myeloid in origin because they lacked sodium fluoride-inhibitable nonspecific esterase activity, myeloperoxidase activity, and F4/80 antigen. The cell line phenotypes were compared to high endothelial venule (HEV) cells in tissue sections. HEV cells in lymph node tissue sections expressed endoglin, PECAM-1, ICAM-1, VCAM-1, laminin, fibronectin, collagen IV, H2Kd, MECA 79, MECA 325, and vWF. The cell lines expressed endoglin, VCAM-1, fibronectin, and H2Kd. The cell line derived from cervical lymph nodes also expressed laminin and H2Dd. Neither cell line expressed collagen IV, IAd, ICAM-1, ICAM-2, dendritic cell antigen, or PECAM-1. They also did not express MECA antigens or intracellular vWF, consistent with reports of many cultured endothelial cells. To further substantiate cell ine identification, antiserum generated against the cell lines bound specifically to HEV cells in frozen lymph node tissue sections and to both of the lymph node-derived cell lines but not control cell lines. Thus, the lymph node derived-cell lines expressed molecules found on HEV cells in vivo and most importantly retained the functions of tubule formation, lymphocyte adhesion, and promotion of lymphocyte migration.

Effect of anti-alkaline phosphatase monoclonal antibody on B lymphocyte function.

Alkaline phosphatase (APase) is a glycosylphosphatidyl-inositol (GPI)-anchored protein appearing on the membranes of mitogen-stimulated B cells after progression into S phase of the cell cycle. Maximal APase expression occurs after peak proliferation and precedes maximal immunoglobulin (Ig) secretion. While APase is clearly an activation marker for mitogen-stimulated B cells, the physiologic role of APase in B cells has not been defined. Other GPI-anchored proteins have been assigned roles in transmembrane signaling since treatment with specific monoclonal antibodies (mAbs) can modulate and/or mimic the effect of mitogens or antigens. Thus, as an initial attempt to determine whether membrane APase (mAPase) plays a role in B cell activation, rat splenic B cells were treated with anti-APase specific mAb in the presence and absence of LPS plus dextran sulfate, known B cell mitogens. Anti-APase mAb alone did not induce proliferation or modulate mitogen-induced proliferation as measured by [3H]thymidine uptake and viable cell recoveries. However, the mAb augmented IgM secretion when used in a soluble form or cross-linked with anti-Ig. Both soluble and immobilized anti-APase mAb decreased the expression of APase activity by mitogen-stimulated B cells. Based upon these results we propose: (1) that transmembrane signaling may occur through mAPase as described for other GPI-anchored proteins such as Thy-1, CD55, CD59, CD24, CD73, Fc gamma III, Qa-2, Ly-6A/E and LFA-3, and (2) this signaling may be regulated by changes in protein phosphorylation caused by modulation of cellular phosphatases, specifically APase.

Expression of CD45 isoforms by Epstein-Barr virus-transformed human B lymphocytes.

CD45 is the most common protein tyrosine phosphatase (PTPase) in the membrane of white blood cells, serving as a potent regulator of lymphocyte activation and signal transduction. While the amino acid sequence of the intracellular domain of the molecule is conserved, that of the extracellular domain occurs in multiple isoforms, each of the result of alternative mRNA splicing. In T lymphocytes, the lowest relative molecular mass (Mr) form, CD45RO, is associated with acquisition of memory function, whereas the highest Mr isoform, CD45RA, occurs in "naive" T cells. Recently, B cells were also found to express CD45RO following in vitro activation. In order to more fully characterize the expression of CD45 on activated B cells, we have studied its appearance on Epstein-Barr virus-transformed (EBV-t) cells and have found heterogeneous expression of CD45RO and CD45RA. CD45RO expression was unstable with eventual loss by some EBV-t lines, and loss followed by reappearance in others. CD45RA and CD45RO varied independently whereas CD45 remained stable and high, suggesting a fluctuation in other CD45 isoforms. Immunostaining for CD45RB indicates that a probable 190-kDa isoform may be responsible for this observation. A similar bidirectional reversible shifting between CD45RA and CD45RO on T-cell lines has also been reported by Rothstein et al. In contrast to some reports on normal B cells, neither CD45RA nor CD45RO expression was associated with PCA-1 expression. Further evidence that these EBV-t lines may not correspond to a well-defined stage of B-cell differentiation is provided by the observation that a disproportionate loss of CD20 compared to CD19 was noted for several lines. The basis for the CD45 isoform switching, or any functional difference(s) in the expressed isoforms, is not yet known for human B cells.

Alkaline phosphatase on activated B cells characterization of the expression of alkaline phosphatase on activated B cells. Kinetics and membrane anchor.

Recently we reported that the expression of the enzyme alkaline phosphatase (APase) is a marker for B cell activation. Enzymatic activity was found only in activated B cells and not T cells. Using flow cytometry we showed that some of the APase was found on the cell membranes (mAPase) and by functional assays, some was spontaneously released into the tissue culture medium. In the present report the expression of mAPase on activated B lymphocytes is more fully characterized. Two mAb specific for rat APase were used to measure the kinetics of the membrane expression of mAPase. Within 48 h of activation, mAPase is detected by flow cytometry and increases coordinately with both the transferrin receptor and IL-2R. Maximal membrane expression of mAPase in terms of number of positive cells and mean fluorescent intensity, is detected by day 4 to 5 of culture. Using hydroxyurea and demecolcine to block cells at G1/S and G2/M, respectively, it appeared that the initial expression of mAPase occurred as cells progressed into S phase of the cell cycle. This was confirmed using two-color flow cytometric analysis with the Hoechst DNA stain 33342 and the FITC-labeled APase-specific mAb. Finally, using phosphatidylinositol-specific phospholipase C we were able to show that 60 to 80% of the mAPase is linked to the membrane via a glycosyl-phosphatidylinositol linkage. From this we have concluded that mAPase can be added to a growing list of glycoproteins that are anchored to the membrane by the glycosyl-phosphatidylinositol linkage and are expressed on differentiating B cells. This list now includes Thy-1, BLAST-1, Jlld, and mAPase.

Expression of the low Mr isoform of CD45 (CD45RO) in B-cell non-Hodgkin's lymphomas.

The CD45 antigen family consists of multiple molecular isoforms ranging from 180 to 220 relative molecular mass (Mr). The highest Mr isoforms are recognized by monoclonal antibodies (MoAbs) designated CD45RA, while those recognizing the low Mr isoform are designated CD45RO. About half of the T-cells in peripheral blood express CD45RA while the remainder express CD45RO. A switch from the high to the low Mr isoform of CD45 has been found in association with the process of T-cell stimulation and acquisition of "memory." B-cells normally express CD45RA, but not CD45RO. However, under stimulatory conditions, B-cells may be capable of undergoing an isoform switch and expressing CD45RO. The expression of this low Mr isoform of CD45 was investigated in lymphomas composed of monoclonal B-cells to determine if such a switch occurs in malignant B-cell populations. The vast majority (110/117 cases) of B-cell lymphomas expressed only CD45RA, while a very small number (7/117 cases) expressed CD45RO, but not CD45RA. There was no relationship between the CD45RO expression and the histologic subtype. The physiological significance of this unusual expression of CD45RO in a subpopulation of B-cell lymphomas is not clear. In that CD45RO, as defined by the MoAb UCHL 1, is typically used as a marker of T-cells in tissue sections, caution must be exercised in interpretation, since not all T-cells are reactive and some B-cell lymphomas are reactive.

Polyclonal activation of rat B cells. II. Dextran sulfate as a cofactor in mitogen-induced and antigen-induced differentiation of rat B lymphocytes.

Salmonella typhimurium mitogen (STM) is a polyclonal activator of rat B lymphocytes, triggering them to proliferate, but not differentiate, to antibody-secreting cells. When lymphokines in the form of a supernatant from Con A-stimulated splenocytes (CAS) are added to B cell cultures activated by STM, only a small number of cells are driven to differentiate. Only with the addition of a third signal provided by the polyanionic polysaccharide dextran sulfate (DXS) is significant rat B cell differentiation observed. In this study, we have shown that this requirement for DXS is not unique to the STM mitogen. LPS, Staphylococcus aureus Cowan I-fixed cells, and anti-Ig antibody all induced rat B cell proliferation with little differentiation, even in the presence of CAS. DXS was necessary to induce differentiation in all cultures costimulated with mitogen and CAS. The requirement for DXS for optimal B cell differentiation is also observed with other lymphokine preparations such as the supernatants from PMA-stimulated EL-4 cells and PHA-stimulated human T cells. Furthermore, this augmentative effect of DXS in rat B cell differentiation was not confined to polyclonal activation systems. Ag-specific IgG secretion was also increased when DXS was added to Ag and CAS costimulated cultures of B cells harvested from the draining lymph nodes of rats immunized with DNP-keyhole limpet hemocyanin. Within the polyclonal activation system, a method of staged additions of STM, DXS, and CAS to B cell cultures was used to investigate the role of DXS during B cell differentiation. Optimal differentiation occurred only when DXS was present in the B cell cultures in conjunction with CAS. The augmentation in differentiation seen with DXS did not appear to be due to the recruitment of an additional CAS-responsive B cell subset, because cycling, low density B cell blasts showed large increases in IgM secretion with subsequent exposure to DXS and CAS. These studies suggest tha DXS acts as a cofactor to various differentiation factors, augmenting polyclonal and Ag-specific rat B cell differentiation. The relevance of DXS to in vivo immune responses is discussed.

Late events in B cell activation. Expression of membrane alkaline phosphatase activity.

Alkaline phosphatase (APase) has been previously described as a membrane marker correlating with B cell proliferation after stimulation by selected B cell mitogens. We have found, however, that the appearance of B cell membrane APase correlates more closely with differentiation than with proliferation. This conclusion has been drawn from the following observations: 1) APase activity appears well after peak B cell thymidine uptake, 2) mitogens which stimulate only B cell proliferation (Salmonella typhimurium mitogen) fail to induce expression of the enzyme, and 3) when proliferation of mitogen-activated B cells is inhibited, APase activity is not suppressed and may even be augmented. In addition to membrane expression, APase is also spontaneously shed into the surrounding milieu, perhaps as a result of endogenous phospholipase activity. By using a group of well-characterized inhibitors, the APase activity was shown to belong to class I (similar to the bone/liver/kidney class). Because APase always appears in differentiating but not proliferating cells, we would propose that the enzyme appearance is a late marker of B cell activation, associated with cell progression to differentiation and consequent IgM synthesis.

Lymphokine-induction of memory B-cell differentiation: differential stimulation of large virgin and memory B-cell differentiation.

In order to compare and contrast the requirements of virgin and memory B cells for B-cell differentiation factors, a model system was developed in which low-density rat B cells isolated from 4-week primed antigen-draining lymph nodes were cultured in vitro. This large low-density cell population contained B cells which were 90% surface IgM positive and 60% IgD positive and showed moderately elevated Ia staining. When the cell population was stimulated with antigen plus lymphokines or lymphokines alone, antigen-specific IgG antibody was secreted; this was used as a measure of memory cell differentiation. When the cell population was stimulated with mitogen (lipopolysaccharide plus dextran sulfate) plus lymphokines, polyclonal IgG and IgM secretion was seen and was used as a measure of virgin B-cell differentiation. Using this system, we found that lymphokines contained in a Con A-induced rat spleen cell supernatant (CSN) were sufficient to drive both memory and virgin B-cell differentiation. In contrast, lymphokines contained in the supernatant from the murine T-cell hybridoma B151K12 (B151CFS) were able to induce large amounts of polyclonal IgM and IgG secretion but did not support memory B-cell differentiation. When recombinant human IL-2 was added to these cultures, it acted synergistically to augment virgin B-cell differentiation, but this combination of lymphokines was still not able to support memory B-cell differentiation. Furthermore, recombinant rat interferon-gamma and a commercial source of human BCGF, with or without IL-2, were unable to promote significant virgin or memory B-cell differentiation. These data support the hypothesis that memory B cells and virgin B cells differ in their lymphokine requirements for differentiation into antibody-secreting cells.

Lymphokine regulation of surface Ia expression on rat B cells.

Class II major histocompatibility antigens (Ia) play a major role in regulating T-B cell interactions; therefore, regulation of the amount of Ia on B cells may be an important point of control in the immune response. Mitogens in human and murine systems have been reported to increase Ia expression on B cells, and in the mouse the lymphokine BSF-1 (IL-4) markedly enhances Ia expression. This report describes studies of lymphokine and mitogen regulation of class II expression on rat B cells. Mitogens known to activate resting B cells, such as DxS/LPS, anti-IgM, STM, and Con A, induced increases in Ia expression. Highly purified murine IL-4 was found to have no Ia-enhancing activity on rat B cells, although the same preparation increased Ia expression eightfold on murine B cells. This confirms other recent reports that IL-4 is a species-specific lymphokine and will not cross even narrow phylogenetic barriers. Rat B cells were not refractory to lymphokine-induced enhancement of Ia expression, since lymphokine(s) contained in a Con A-induced supernatant enhanced Ia expression. Furthermore, murine IL-5-containing B151-CFS was able to markedly increase Ia expression on resting rat and mouse B cells. This activity was not lost after heat inactivation of B151-TRF2, indicating that B151-TRF1 (IL-5-like activity) was probably responsible for the increase in Ia expression. These results suggest Ia expression on rat B cells, like human and murine B cells, is an early activation event which is regulated by signals which act on resting B cells. Furthermore, while IL-4 is important in Ia regulation, it is not the only lymphokine involved, since the IL-4-free, B151-K12 supernatant was able to enhance Ia expression on resting rat and mouse B cells.

An alternative method of panning for rat B lymphocytes.

A new method of panning for B lymphocytes is described in which the ability of the sIg+ cells to adhere depends on the nature and concentration of nonspecific protein used rather than on the use of anti-immunoglobulin. Rat lymph node cells were suspended in 3% bovine serum albumin in Tris-buffered Hanks' and incubated in tissue culture flasks to allow adherence to the plastic. The recovered bed of adherent cells was shown by flow cytometry to be greater than 90% surface immunoglobulin positive and MHC class II positive while containing very few T cells. This adherent fraction was subsequently treated with anti-T cell antibody plus baby rabbit complement to produce a highly purified sIg+ cell population containing no detectable T cells. The sIg+ cells obtained by this panning procedure were functionally active in BCGF and BCDF assays. This method provides an easy and inexpensive alternative to conventional panning with anti-immunoglobulin and also eliminates the possibility of B cell activation by exposure to anti-immunoglobulin-coated surfaces.

Polyclonal activation of rat B cells. I. A single mitogenic signal can stimulate proliferation, but three signals are required for differentiation.

A water-soluble, proteinaceous preparation derived from the cell walls of Salmonella typhimurium Re mutants has recently been tested in our laboratory for its ability to act as a mitogen for rat lymphocytes. We have found this preparation (STM) to be a potent stimulator of B lymphocyte proliferation, as measured both by 3H-TdR incorporation and by cell cycle analysis performed with flow cytofluorometry. STM stimulates approximately 50% of rat B cells to enter cycle. Previous investigations by others have shown that at least two sets of signals are required for B cell differentiation; a) proliferation signals that may consist of both a stimulator of B cell conversion from G0 to G1 and growth factors, and b) differentiation signals that probably include at least two B cell differentiation factors (BCDF). When STM was tested in a differentiation system it did not drive purified B cells to differentiate to PFC, either alone or when supplemented with a supernatant from concanavalin A-stimulated spleen cells (CAS). However, when both CAS and dextran sulfate (DXS) were supplied to the STM-stimulated cells, a large number of PFC resulted. DXS does not act by stimulating an additional, CAS-responsive B cell subset, since it has only a marginal effect upon 3H-TdR uptake and does not increase the number of B cells in cycle when used together with STM. We postulate that the two agents may be acting sequentially: STM stimulates the B cells to proliferate, and DXS drives the proliferating cells to become responsive to CAS. This suggests that the signals for B cell differentiation must consist of at least three activities: a trigger to stimulate the cells to proliferate, a factor to drive the cells to a BCDF-responsive state, and a BCDF that can drive the cells to secrete antibody.

Polyclonal activation of primed rat B cells.

In recent years, murine and human virgin B lymphocytes have been used to examine the steps necessary for polyclonal activation. In these models mitogens are used in conjunction with lymphokines to determine which signals are responsible for regulating B-cell triggering, proliferation, and differentiation. While progress has been made in understanding these events as they occur in virgin B cells, very little evidence exists to suggest whether these models of activation also apply to the memory B-cell population. In this report we have described an antigen-specific, secondary in vitro immune response using cells isolated from lymph nodes draining the site of antigen injection. Unfractionated cells, B cells, and size-fractionated cells from dinitrophenyl-keyhole limpet hemocyanin (DNP-KLH)-primed rats were challenged in vitro with DNP-KLH, lipopolysaccharide plus dextran sulfate (LPS/DxS), and T-cell factors. We have consistently found, under all these conditions, that antigen challenge of primed cells results in the production of DNP-specific IgG antibody while stimulation with LPS/DxS plus T-cell factors results only in the polyclonal activation of virgin B cells; no antigen-specific IgG secretion is seen. This suggests that acquisition of memory status is associated with a loss in responsiveness to LPS/DxS-induced differentiation.

Lymphokine-like activity of 8-mercaptoguanosine: induction of T and B cell differentiation.

Lymphocyte proliferation and differentiation result from ordered cellular interactions governed by soluble products (lymphokines). Dissecting the individual steps in these processes has been difficult, due to a paucity of pure lymphokines. Recently, it was reported that the derivatized ribonucleoside 8-mercaptoguanosine (8MGuo) has both mitogenic and differentiative effects on murine B cells. In the present studies, we tested 8MGuo for its ability to stimulate both B and T cell responses. In contrast to the murine studies, 8MGuo does not stimulate rat B cells to proliferate and, when tested for B cell growth factor-like activity, no stimulation was observed. The addition of 8MGuo (0.5 to 1 mM final concentration) to mitogen-stimulated B cells led to a marked increase in IgM and a modest increase in IgG secretion. When mixed with conditioned medium, 8MGuo acted synergistically in stimulating secretion of both isotypes, arguing that 8MGuo has both B cell-differentiating factor-mu (BCDF-mu) and BCDF-gamma activity. 8MGuo had no IL 2-like activity when tested on a mouse IL 2-dependent cell line, and no IL 1-like activity on addition to mouse thymocytes with or without submitogenic doses of lectin. However, when added to cultures of murine allogeneic cells in which the stimulating cell populations had been heat-inactivated, 8MGuo induced the generation of specific allogeneic cytotoxic T lymphocytes. Together, these results suggest that a simple derivatized nucleoside can induce both T and B cell differentiation without concomitant proliferation, and thus represent a unique probe for studying events in lymphocyte differentiation.

Role of complement in the immune response.

Evidence has accumulated that shows that fragments of C3 are potent inhibitors of immune responses. A low-molecular-weight fragment of C3 and fragments possessing leukocyte-mobilizing activity have been shown to block both antigen- and mitogen-induced human T cell proliferation, and to block mixed lymphocyte culture responses and the generation of cytotoxic lymphocytes. The same fragments inhibit the development of secondary in vitro antibody responses of rat lymphocytes. C3b can be shown to inhibit the polyclonal activation of human lymphocytes by pokeweed mitogen, but it has no effect on T cell proliferation or on the generation of cytotoxic T cells. We now propose that different C3 fragments selectively act on various lymphocyte subsets and thus play a profound role in regulating both immune effector functions and the intensity of the immune response.

Complement inhibits immune responses: C3 preparations inhibit the generation of human cytotoxic T lymphocytes.

C3 fragments have been shown to inhibit mitogen- and antigen-induced human lymphocyte blastogenesis. In this study, C3 preparations and a small fragment of C3 (contained in a preparation called Fraction 2) were examined for their capacity to regulate cytotoxic T lymphocytes (CTL) and proliferative mixed lymphocyte responses (MLR). We found that both preparations inhibited the generation of allogeneic human CTL as well as MLR in a dose-related manner. In contrast, Fraction 1, which contained native C3 and C3b, did not inhibit the generation of CTL nor did it inhibit the MLR. The kinetics of inhibition of proliferation were divergent from the kinetics of inhibition of CTL generation; the active preparations inhibited proliferation significantly when added as late as day 5 of a 6-day culture, whereas no inhibition of CTL generation was seen when these preparations were added after day 3 of culture. Cultures in which C3 preparations caused complete inhibition of CTL generation had normal levels of the nonspecific, anomalous killer cell activity, as assayed on K 562 target cells. Furthermore, C3 preparations and Fraction 2 had no effect on the lytic function of differentiated CTL, on "spontaneous" natural killer cell activity or on interferon-induced augmentation of natural killer cell activity. These findings indicate that C3 fragments may play a negative role in the regulation of CTL responses.

The correlation between the activation state of B cells and their capacity for in vitro propagation of immunologic memory.

The B-cell population responsible for in vitro antigen-mediated proliferation and expansion of the memory B-cell population is a large activated blast. Such cells predominate early after antigen priming and can be regenerated by adjuvant (Bordetella pertussis) stimulation in vivo. Although these cells are proliferating in vivo, additional stimuli are needed for expansion of the memory population in vitro. These triggering requirements include specific antigen (DNP-OVA) and the assistance of adherent accessory cells. Although T cells are present in the culture, their role in the propagation of memory is not completely clear. Using the unrelated antigen, sheep erythrocytes, we have shown that "bystander" T-cell help can mediate differentiation of these memory B-cell blasts to AFC, but it cannot induce expansion of the memory-cell population. However, the fact that the TI-2 antigen DNP-Ficoll is a relatively ineffective inducer of memory-cell propagation (inducing an expanded response which is less than 10% of that induced by the T-cell-dependent antigen, DNP-OVA) suggests that T cells may be involved, possibly via production of B-cell growth factor. Thus, the minimal requirements for triggering the propagation of B-cell memory include (i) a blastogenic signal which can be mediated by adjuvant, (ii) specific antigen, and (iii) adherent accessory-cell help.

Inhibition of human lymphocyte blastogenesis by C3: the role of serum in the tissue culture medium.

Preparations of the third component of human complement (C3) inhibit human lymphocyte blastogenic response to mitogens and antigens when cultured in serum-free medium or in medium supplemented with 5% autologous serum (AS). In contrast, when the culture medium was supplemented with 5% foetal calf serum (FCS), C3 failed to inhibit responses to mitogen (concanavalin A) or to antigen (streptolysin O); some FCS lots allowed stimulation rather than inhibition of the lymphocyte responses. Moreover, when lymphocytes were cultured in serum containing equal amounts of FCS and AS, no inhibition was seen. Our findings may explain previous studies which suggest that C3 enhances or has no effect on lymphocyte responses.

Inhibition of secondary in vitro antibody responses by the third component of complement.

Purified human C3 was found to inhibit rat in vitro secondary antibody responses. Fifty percent inhibition of antibody-forming cell development occurred with C3 concentrations of 26 micrograms/ml. This decrease was not the result of a general toxicity or a shift in the antibody response kinetics. Using cell mixing experiments, we could not detect a C3-induced suppressor lymphocyte or macrophage. C3 was active when added to culture early (day 0 or 1 or during a 24-hr antigen prepulse) or late (day 3, 5, or 7)--the early addition being more suppressive. Regardless of the addition time, there was a characteristic 48- to 72-hr lag before the inhibitory effect was manifested. C3 could inhibit antibody-forming cell development after stimulation with the thymus-independent antigens, trinitrophenyl-Brucella abortus and dinitrophenyl-Ficoll, as well as the thymus-dependent antigens, dinitrophenyl-bovine gamma-globulin and chicken gamma-globulin suggesting that C3 was not selective for B memory cell subpopulations. Further characterization of our C3 preparation indicated that the majority of the suppressive activity resided in a small m.w. protein resembling the C3a fragment of C3. Human C3a preparations generated either by trypsin cleavage or zymosan activation of C3 were also tested in our antibody response system and were able to inhibit antibody-forming cell development. These data implicate C3 cleavage products as negative regulators of antibody formation.

Complement fragments suppress lymphocyte immune responses.

There has been controversy as to whether complement augments immune responses or inhibits them. In this article John Weiler and his colleagues discuss recent evidence that complement fragments can inhibit immune responses that depend upon cellular proliferation.

Role of T cells in the development of memory B cells. Quantitative and qualitative analysis.

The purpose of this investigation was to address the current controversy regarding the T-cell requirement for the generation of B-memory cells. We have circumvented the possible objection to previous experiments regarding residual T cells in T-deprived animals by examining memory cell generation in relation to the numbers of T cells participating in the immune response. Thymectomized and lethally-irradiated rats were reconstituted with foetal liver or a more mature stem cell source, neonatal liver. These animals were given graded doses of purified T cells one day before immunization with alum-precipitated DNP-BGG + Bordetella pertussis. Four weeks after priming, cell suspensions from experimental groups were adoptively transferred to carrier primed recipients and challenged with DNP-BGG in saline to assess memory cell development. The data demonstrate that in the absence of T cells only minimal memory development occurred. However, when T cells were present, the level of memory cell development increased with increasing numbers of T cells. By examining the relative affinity of the antibody produced in the primary and secondary responses, the increase in memory cell development in relation to increased numbers of T cells was shown to be due to the selective generation of high affinity memory B cells.