A human lymphoid cell line with receptors for both sheep red blood cells and complement

The human lymphoid cell line MOLT 4, from a patient with acute lymphocytic leukemia, was initially considered to be derived from T lymphocytes, on the basis of rosette formation with sheep erythrocytes (E). This cell line has now also been found to form rosettes with sheep erythrocytes sensitized with rabbit antibody and mouse complement (EAC). Evidence is presented that the formation of both E and EAC rosettes is due to two separate receptors on the MOLT cells: (a) EAC rosettes were formed more rapidly and were more stable than E rosettes; (b) preincubation of MOLT with an EAC membrane preparation inhibited resetting with EAC and not with E; (c) MOLT formed rosettes with EAC prepared from trypsinized E, but did not bind to trypsin-treated E alone. The implications of this finding, in regard to the derivation of this cell line, are discussed.     

Expression of specific binding sites on Candida with functional and antigenic characteristics of human complement receptors

Receptors for C3 degradation fragments (CR1, CR2, and CR3) are present on many human cells including phagocytes and lymphoid cells and may be critical in the attachment of invading microorganisms. In these studies Candida were found to mimic the human CR by binding erythrocytes coated with specific human C3 fragments. Yeast forms of Candida species were adhered to glass slides and were allowed to germinate. Sheep erythrocytes (E) were coated with IgM (EA) and human complement components to prepare EA, EAC14, EAC3b, EAC3bi, and EAC3d. These test cells were then examined for adherence to the organism. Antibodies to human CR1, CR2, and CR3 were used to evaluate their potential for blocking adherence of the test erythrocytes to Candida. Fluorescein-labeled antibodies to human complement receptors were also used to characterize the binding sites. EAC3bi and EAC3d, but not E, EA, or EAC14, bound extensively to the germ tubes and pseudohyphae of Candida albicans and C. stellatoidea. EAC3b bound infrequently. Other Candida species, generally considered less pathogenic, bound significantly fewer specific test erythrocytes than C. albicans. Monoclonal antibodies to human CR1 and CR3 (3D9, 1B4, C511, 2B6, anti-B2, Mo1, and anti-Mac-1), in general, did not block adherence of test erythrocytes. Blocking of adherence of EAC3bi and EAC3d test erythrocytes coated with small quantities of C3 fragments occurred with high concentrations of monoclonal (anti-CR2) HB-5 and polyclonal (anti-CR2) anti-GP 140. Immunofluorescence studies demonstrated binding of Mo-1 to the germinated forms of the organism, whereas binding of the other antibodies was not seen. These studies suggest a surface constituent on the organism similar to CR on human cells. Additional studies are necessary to further define the molecular nature of the binding site. The ability of organisms to mimic human CR may be more generalized than previously known and may serve as a mechanism for modification of the inflammatory and immune response.     

Binding of fluid-phase complement components C3 and C3b to human lymphocytes.

It is known that a population of B-lymphocytes has receptors for the third component of complement, C3, and that these lymphocytes may be identified by their ability to form rosettes with sheep erythrocytes coated with covalently bound fragments of complement component C3. Human tonsil lymphocytes, enriched for B-cells, form rosettes with sheep erythrocytes coated with antibody and complement components C1, C4b and C3b (EAC143b cells). Fluid-phase C3 will inhibit rosette formation between EAC143b and human tonsil lymphocytes over the same concentration range as fluid-phase C3b. C3 is not cleaved to C3b during incubation with lymphocytes or with lymphocytes and EAC143b cells. Fluid-phase 125I-labelled C3 and 125I-labelled C3b bind to lymphocytes in a specific manner. The characteristics of binding of both radioiodinated C3 and radioiodinated C3b are very similar, but the binding oc C3 is again not a result of cleavage to C3b. Salicylhydroxamic acid does not inhibit binding of 125I-labelled C3 to tonsil lymphocytes at concentrations that completely inhibit binding of 125I-labelled C3 to EAC142 cells via the nascent binding site of C3b. It is concluded that C3 and C3b share a common feature involved in binding to lymphocytes bearing receptors for the third component of complement.     

Complement inhibitor(s) released by leukocytes. I. Pretreatment of sheep erythrocytes with supernatants of mouse spleen and thymus cells inhibit whole complement activity and C2 utilization.

Sheep erythrocytes pretreated with supernatants of mouse spleen or thymus cells become resistant to lysis by guinea pig complement. The inhibitory activity (IA) reduces the utilization of C2 by EAC14. Because IA binds to the surface of sheep erythrocytes and does not inhibit C1 irreversibly, it is probably a hitherto undescribed inhibitor of complement.     

Fc receptors on human T lymphocytes. II. Cytotoxic capabilities of human T gamma, T mu, B, and L cells.

Subpopulations of human peripheral blood lymphocytes were prepared by rosetting techniques employing neuraminidase-treated sheep erythrocytes (SRBC_n), sheep erythrocytes coated with IgM and murine complement (EAC′), and bovine erythrocytes coated with IgG and IgM. The isolated subpopulations were tested in assays of natural cytotoxicity (NC), antibody-dependent cellular cytotoxicity (ADCC), and mitogen-induced cellular cytotoxicity (MICC). B cells (SRBC_n?, EAC′+) did not mediate cytotoxicity. L cells (SRBC_n?, EAC′?) mediated NC and ADCC but not MICC. T cells (SRBC_n+) mediated NC, ADCC, and MICC. Separation of T cells into Fc-IgG (Tγ) and Fc-IgM (Tμ) subsets revealed that Tγ cells mediated NC, ADCC, and MICC while Tμ cells mediated only MICC. Thus MICC but not NC or ADCC was solely T-cell mediated. Tγ and L cells were functionally distinguishable in that Tγ cells but not L cells mediated MICC. Tγ cells and Tμ cells differed with regard to NC and ADCC effector function while both subsets mediated MICC.     

Complement inhibitor(s) released by leukocytes. III. Evidence for a "new" C1 inhibitor in the supernatants of short-term cultures of mouse spleen and thymus cells.

Mouse spleen or thymus cells in short-term culture release a factor, designated S, that binds to sheep erythrocytes (E). Supernatant-treated sheep erythrocytes (SE) are capable of fixing and transferring the activated first component of guinea pig complement. SEC1, however is not capable of initiating hemolysis by the rest of the complement components. SE is capable of binding but not activating native C1; native C1 bound to SE seems irreversibly inhibited. Evidence is presented that S may be the same factor as the previously described inhibitor released by mouse spleen or thymus cells that inhibits the utilization of C2 by EAC14.     

Morphological differences in the interactions between human mononuclear cells and coated or uncoated sheep red blood cells

Summary Enriched preparations of E, EA and EAC rosettes formed by human peripheral blood mononuclear cells were freeze-etched and examined electron-microscopically. In E rosettes only lymphocytes were involved, whereas in EA and EAC rosettes lymphocytes and mononuclear phagocytes participated as rosette-forming cells. In EA and EAC rosettes, cytoplasmic extensions of the rosette forming cell were seen to penetrate the sheep red blood cell, whereas E rosettes showed a broad zone of adherence without penetration. None of the three types of rosettes showed an interspace between the membranes. Unlike E rosettes, EA and EAC rosettes showed polarity in the adherence of sheep red blood cells. These observations made by freeze-etch electron microscopy indicate distinct morphological differences between rosettes formed with coated or uncoated erythrocytes.The authors wish to thank Prof. Dr. A. Cats, Dr. P.C.J. van Breda Vriesman and Dr. J.C.H. de Man for helpful discussion and criticism; the assistance of Miss R. Kleinjan and Mrs. A.C. Scheurkogel-van Efferen is gratefully acknowledged. This work was supported in part by a grant of the Praeventiefonds     

Characterization of the C3 receptor induced by herpes simplex virus type 1 infection of human epidermal, endothelial, and A431 cells

Herpes simplex virus type 1 (HSV-1) infection induces the appearance of viral analogues of human Fc IgG and C3 receptors on the surface of human cells. The virally induced C3 receptor(s) has been broadly defined as a C3b receptor, but its ligand binding characteristics have not been rigorously defined. In this study, human epidermal cells, A431 cells, and human umbilical vein endothelial cells infected with HSV-1 demonstrated rosetting with sheep erythrocytes (E) coated with IgG (E-IgG) or the complement components C3b (EAC3b) or iC3b (EAC3bi), but not with E-IgM, C4 (EAC14), C3d (EAC3d), or E alone. Rosetting was markedly enhanced by pretreatment of HSV-1-infected cells with neuraminidase. Unlike human C3 receptors, the HSV-1-induced C3 receptor was found to be trypsin resistant. To determine whether HSV-1 induced CR1-like receptors or CR3-like receptors, infected cells were pretreated with EDTA, which is known to inhibit native CR3 function. EDTA failed to prevent rosetting with EAC3bi. Furthermore, blocking studies using monoclonal antibodies against CR1 and CR3 revealed that the anti-CR1 antibody 5C11 consistently blocked EAC3b and EAC3bi rosetting with HSV-1-infected cells in a dose dependent manner, but monoclonal antibodies against CR3 did not. This study indicates that the HSV-1-induced C3 receptor is an analogue of CR1.     

Membrane fluidity change in erythrocytes induced by complement system.

The structural change in erythrocyte membranes induced by antibody and complement was studied using phospholipid spin-labels. Sheep erythrocytes were labeled with phosphatidylcholine spin-label and various intermediate cells (erythrocyte-antibody complex (EA), EA bound with complement components from C1 to C7 (EAC1-7), EAC1-8, and EAC1-9) were prepared. Electron spin resonance spectra of EA, EAC1-7, and EAC1-8 were very similar to that of the erythrocytes, while that of EAC1-9 was markedly different. The overall splitting value for the lysed EAC1-9 (53 G) was much smaller than that for the erythrocytes (57 G), indicating a marked fluidization around the phosphatidylcholine label. The unlysed EAC1-9 membranes contained a limited fraction of the fluidized area. When EA was reacted with complement in the presence of 36% bovine serum albumin, the membranes were fluidized similarly to the lysed EAC1-9, although the hemolysis was largely blocked. The membranes of unlysed EAC1-9 prepared in isotonic (ethylenedinitrilo)tetraacetic acid were also fluidized, but to somewhat smaller extent. The role of C9 in the modification of erythrocyte membranes was also demonstrated using Mg2+ ghosts, which were prepared by hypotonic hemolysis in the presence of Mg2+. The membranes of Mg2+ ghost of EAC1-7 were markedly fluidized when bound with C8 and C9, but not affected by binding of C8 only. The component C8 was found to give a latent effect on the membranes that caused irreversible fluidization upon osmotic shock. The terminal component thus creates a fluidized area in the erythrocyte membranes through which small ions and molecules may diffuse more easily and the resulting osmotic unbalance may finally cause hemolysis.     

The modifications of the final stages of the complement reaction by alkali metal cations.

We studied the effects of alkali metal cations on the terminal stages of complement lysis of human and sheep HK erythrocytes. Sensitized erythrocytes (EA) were reacted with limited amounts of complement for 1 hr at 37 degrees C in buffer containing 147 mM NaCl (Na buffer), which resulted in 10-40% lysis. The unlysed cells were washed with Na buffer at 0-2 degrees C and incubated for 1 hr at 37 degrees C in buffers containing 147 mM of the various alkali metal cations. Although additional lysis (25 to 65%) occurred with K, Rb, or Cs buffer, only minor degrees developed with Na or Li buffer, only minor degrees developed with Na or Li buffer. Intermediate levels occurred with 100 mM of the divalent alkali cations. Halogen ions and SCN-(147 MM), Ca++ (0.15mM), and Mg++ (0.5 mM) did not alter the effect of the alkali metal cations. Lysis occurring in K+, Rb+ or Cs+ proceeded without lag, was temperature dependent with an optimum of 43 degrees C, and had a pH optimum of 6.5. Lysis in K and Na buffers was unaffected by 10(-3) to 10(-5) M ouabain. Experiments with mixtures of cations indicated that Na+ had a mild inhibitory effect that could be totally overcome by K+, partially by Rb+, and not at all by Cs+. Li+ had a strong inhibitory effect, 6 X 10(-5) M causing 50% inhibition in buffers containing 147 mM K+, Rb+, or Cs+. By using intermediate complexes of EA and purified complement components we demonstrated that K+ enhances the lytic action of C8 on EAC1-7 as well as that of C9 on EAC1-8. It was known that Li+ facilitates lysis when acting on the entire complement reaction. We found that Li+ enhanced the lytic action of C8 on EAC1-7, with a kinetic that differed from that of the K+ effect. In addition, Li+ inhibited the enhancing effect of K+ upon lysis of EAC1-8 by C9. This occurred at concentration of Li+ similar to those which inhibited the additional lysis by K+, Rb+, and Cs+ of cells that were pretreated in Na buffer with the entire complement sequence. We propose that the major effects of alkali metal cations on complement lysis are due to their interaction with C8 and/or membrane constitutes.     

Complement-mediated hemolytic activity of succinylated concanavalin A: preparation and activity of cell intermediates

Succinylated concanavalin A (SCon A) lyses sheep erythrocytes (E) in the presence of complement, whereas the native tetravalent lectin, Con A, is inactive. We have studied the ability of E-SCon A (ES) to interact with early acting guinea pig (gp) or human (hu) complement components (C1, C2, C4) and found that cell intermediates ESC1, ESC4, ESC14, and ESC142 can be generated that are analogous to intermediates conventionally prepared with E and rabbit IgM (pentameric) anti-Forssman antibody. Titration of gp or hu C1, C4, and C2, and quantification of the number of activated C1 molecules bound to ESC4 by the C1 fixation and transfer test showed in each case that an average of one effective site per cell was sufficient to cause cell lysis. Determination of tmax for optimal formation of ESC142 sites depended on the species combination of components used to make the intermediates, and the decay of ESC142 and EAC142 sites or sites generated with ESC4, EAC4, and trypsin-activated C2 were similar. The sugar alpha-D-methylglucopyranoside (alpha-MGP) inhibited binding of SCon A to E and eluted the lectin from ES, whereas galactose was nearly inactive, consistent with lectin sugar-binding selectivity. In contrast, both sugars were ineffective in eluting SCon A or C4hu from ESC4hu, indicating that C4hu blocked the interaction between lectin and alpha-MGP, perhaps by steric hindrance. SCon A is a divalent functional analogue of IgM anti-Forssman antibody that may be a uniquely suited reagent specific for cell membrane glycoconjugates for studying the mechanism of binding and activation of complement components.     

Studies of complement receptors on cytotoxic effector cells in human peripheral blood

The question of whether cells bearing complement receptors (CR) mediate cytotoxicity in vitro against allogeneic Chang liver cell targets was investigated by assessing peripheral blood mononuclear cells (PBMC) from normal humans for cell surface characteristics and cytotoxic capacity before and after depletion of CR+ cells capable of forming rosettes with sheep erythrocytes coated with 19S antibody and mouse complement (EAC) and depletion of Fc receptor-bearing cells capable of forming rosettes with human O+ erythrocytes coated with Ripley antibody (EA-Ripley). PBMC depleted of CR+ cells by density centrifugation contained markedly reduced proportions of phagocytes and sIg + cells and increased proportions of both sIg ?, FcR+ cells as well as cells forming rosettes with sheep erythrocytes (E). PBMC depleted of CR+ cells mediated cytotoxicity to an extent equal to or greater than that mediated by unfractionated PBMC in assays of spontaneous cell-mediated cytotoxicity (SCMC), antibody-dependent cellular cytotoxicity (ADCC), and mitogen-induced cellular cytotoxicity (MICC). Cells harvested from the EAC-rosette enriched pellet mediated cytotoxicity 5- to 10-fold less than unfractionated PBMC; however, the cytotoxic activity of the pellet could not be attributed to CR + effector cells since similar cytotoxic activity was present in cell pellets obtained by density centrifugation of PBMC which had been incubated with E coated with 19S antibody or E alone. PBMC depleted of EA-Ripley rosette-forming cells contained decreased proportions of sIg?, FcR+ cells and increased proportions of CR+ cells; PBMC so depleted contained virtually no SCMC and ADCC effector cell activity. These findings indicate that at least the majority of effector cells which mediate SCMC, ADCC, and MICC do not bear CR.     

The computer analysis of volume distribution curves. Demonstration of two erythrocyte populations of different size in the young guinea pig and analysis of the mechanism of immune lysis of cells by antibody and complement.

Guinea pig, rat and sheep erythrocytes were sized electrically using the hydrodynamic focusing technique. The experimental curves were approximated with a computer by linear and logarithmic normal distributions. Rat and guinea pig erythrocytes from adult animals were best approximated by one linear normal distribution. Two populations (I, II) of erythrocytes with different mean volume could be demonstrated in young guinea pigs by this analysis. Population I erythrocytes are small, have a lower electrophoretic mobility and are mainly present at birth. They are gradually replaced by the larger population II erythrocytes. Both types of erythrocytes are probably the result of separate differentiation pathways. The analysis of erythrocyte volume distribution curves during immune lysis by antibody and complement shows that intact and ghost erythrocytes are measured by electrical sizing. No volume changes were observed up to the EAC1-8 intermediate. After the addition of C9, a C9 dose-dependent part of the erythrocytes swell permanently to spheroids. The spheroid transformation is a temperature-dependent, all or nothing reaction which is independent of protein osmotic forces from the interior of the cell.     

Alternative complement pathway activation by C4b deposited during classical pathway activation

Sheep erythrocytes (E) sensitized with anti-E antibody (A) were reacted with guinea pig C1 (C1gp) and human C4 (C4hu) or guinea pig C4 (C4gp) to prepare EAC1, 4b. Treatment of the EAC1, 4b with a buffer containing EDTA removes C1rgp and C1sgp, resulting in the formation of EAC4b. EAC4b prepared in this way were found to be lysed by human or guinea pig serum in a gelatin Veronal-buffered saline containing 2 mM MgCl2 and 8 mM EGTA (Mg-EGTA-GVB). In the hemolytic sensitivity of EAC4bhu, essentially no difference was noted whether IgG or IgM antibodies were used for preparation of EAC4bhu. The extent of the hemolysis of EAC4bhu was dependent on the dose of C4bhu. Because EAC4bhu were lysed even by C2-deficient human serum, C3 convertase of the classical complement pathway would not be involved in the hemolysis of EAC4bhu. Furthermore, the reactivity of EAC4bhu with serum in Mg-EGTA-GVB remained even after treatment of the intermediate cells with 1 mM PMSF, indicating that any remaining C1gp was not responsible for the hemolysis. Therefore, the hemolysis of EAC4b by sera in Mg-EGTA-GVB was considered to be mediated via activation of the alternative complement pathway (ACP). Pretreatment of EAC4bhu with anti-C4hu antibody or C4-binding protein suppressed the hemolysis of EAC4bhu via the ACP activation. Furthermore, EAC4bhu were more sensitive to hemolysis by the reaction with a mixture of C3, B, D, and H followed by rat serum in EDTA-GVB than EAC1qgp were. These results indicate that C4b molecules on the cell membrane participate in the activation of ACP.     

Mouse skin graft prolongation with donor strain bone marrow and anti-lymphocyte serum: surface markers of the active bone marrow cells

The survival of C3H/HeJ skin grafts on B6AF1 mice treated with anti-lymphocyte serum (ALS) can be significantly prolonged by the injection of the host with C3H/HeJ bone marrow. Although the prolongation is apparently due at least in part to the ultimate presence in the host of specific suppressor cells of donor origin, little is known about the nature of the cells in the marrow inoculum that are responsible for this effect. The present investigation was undertaken to characterize surface markers of the active bone marrow cells. Marker-positive populations were either depleted and enriched by panning techniques or depleted by killing with specific antibody and complement, and then were assayed for their ability to prolong graft survival. Cells that were adherent to anti-Ia-coated plates extended graft survival only slightly better than did treatment with ALS alone, whereas nonadherent (Ia-depleted) cells, as well as cells treated with anti Ia and complement, retained good prolonging activity. Similarly, panning on anti-immunoglobulin (Ig)-coated plates produced an active, Ig+-depleted population and an inactive adherent population, and killing of Thy-1+ cells with antibody and complement did not compromise the ability of the bone marrow inoculum to prolong graft survival. Complement receptor-positive (EAC+) and Fc gamma receptor-positive cells (EA+) were separated by panning on plates coated with sheep erythrocytes/antibody/complement and erythrocytes/7S antibody respectively. Adherent, EAC+-enriched cells were only slightly active, whereas the nonadherent, EAC-depleted population was fully active in graft prolongation. However, both Fc gamma R+ (EA+)-enriched and depleted populations were active, with the enriched fraction producing significantly better prolongation than the depleted population. Thus, the bone marrow cells that can prolong skin graft survival in ALS-treated mice appear to be Ia-, Thy-1+, largely complement receptor negative, and Ig-, but are largely positive for Fc gamma receptors.     

E and EAC rosettes in man and nonhuman primates and the effect of thymosin in vitro.

E (erythrocyte) and EAC (erythrocytes coated with antibody and complement) rosettes were quantitated in baboons, cebus monkeys and cotton-topped marmosets. There was poor correlation between E rosette levels and other parameter of T-lymphocyte function. In nonhuman primates, E rosettes were increased in the presence of thymosin fraction V, while human E rosettes were not affected.     

Immunoglobulin, complement, and histocompatibility antigen studies in keloid patients.

The increased collagen synthesis and deposition, which is characteristic of keloids, may be related to an immune response initiated by wounding. Therefore, we examined various systemic and localized immune parameters in keloid patients to establish if such factors are related to keloid pathogenesis. To determine if there is a systemic immune response, we compared the serum levels of IgG and IgM in keloid patients to those in a closely matched population. In addition, we measured complement levels (Clq, C3, and C4) and receptors for sheep (E), mouse erythrocytes (MRBC), and complement (EAC) on blood lymphocytes. All of these were in the normal range in the keloid patients. However, the extractable IgG from keloid tissue was significantly increased (compared to normal skin and normal scar controls), suggesting a localized immune response. To determine whether keloid formation is associated with a specific histocompatibility locus, human lymphocyte antigen (HLA) profiles of 45 keloid patients were analyzed; no significant differences in the incidence of HLA-A and B antigens were found (compared to 200 controls). These studies suggest that there is a localized immune response involved in keloid pathogenesis, one which is not related to either the HLA-A or B histocompatibility loci.     

Separation of human lymphocyte subpopulations. I. Transformation characteristics of rosette-forming cells

Lymphocytes were isolated from human peripheral blood by carbonyl-iron treatment and Ficoll-Isopaque centrifugation. The lymphocytes were allowed to form rosettes with sheep erythrocytes, either uncoated (E) or coated with antibody and complement (EAC).In 32 experiments E rosettes were separated from free lymphocytes on a Ficoll density gradient. Thus, depleted (non-E) and enriched (E) fractions were obtained. It was found that in the original suspension 24 ± 7.2% of the lymphocytes formed rosettes with EAC and 56 ± 8% with E. In fraction non-E these values were 56 ± 11.4 and 22 ± 8.9%, respectively; in fraction E 8 ± 3.8 and 79 ± 8.8%. Moreover, the percentages of Ig-bearing cells among the fractions were found to follow closely those of CRL.In a series of lymphocyte culture experiments these fractions were compared with the original suspension and a control suspension (rosetted, not separated), as well as with a recombined population (non-E + E). It was found that fraction non-E showed an increased response to PHA and PWM, and an enhanced MLC stimulatory capacity, whereas fraction E was decreased in these respects. Moreover, fraction E displayed significantly reduced spontaneous DNA synthesis.It is concluded that the responses to PHA or PWM, as well as the capacity to stimulate allogeneic cells, are not solely dependent on the cells forming rosettes with E.     

Resialylation of sialidase-treated sheep and human erythrocytes by Trypanosoma cruzi trans-sialidase: restoration of complement resistance of desialylated sheep erythrocytes.

Trypanosoma cruzi trans-sialidase (TS) is a recently described enzyme which transfers alpha(2-3)-linked sialic acid from host-derived sialylated glycoconjugates to parasite surface molecules [Schenkman et al. (1991) Cell, 65, 1117]. We report here on the ability of TS to transfer sialic acid from donor sialyl-alpha(2-3)lactose to sialidase-treated sheep and human erythrocytes. Up to approximately 50% resialylation of both desialylated red cells could be attained. Resialylation of desialylated sheep erythrocytes restores their resistance to lysis by human complement. This ascribes a possible biological role for T. cruzi TS and demonstrates directly that sialic acid is solely responsible for preventing alternative pathway activation of human complement by sheep erythrocytes.     

E rosette formation at 37°:A property of mitogen-stimulated human peripheral blood lymphocytes

Peripheral blood lymphocytes from healthy humans formed stable E rosettes with sheep erythrocytes (SRBC) at 37°C after culture with phytohemagglutinin or the divalent cation ionophore A23187. Cells manifesting this phenomenon exhibited “blast” morphology, appeared by 16 hr of culture, increased dramatically in percentage and absolute number by 62 hr, and persisted in large numbers for the duration of culture (182 hr). Unstimulated lymphocytes formed rosettes at 4°C but not at 37°C. Increased “stickiness” due to surface-bound lectin mitogen was not the cause of rosette formation at 37°C.Formation of E rosettes at 37°C has previously been considered a property of lymphocytes less differentiated than the circulating T cell (e.g., thymocytes, leukemic lymphoblasts). The present findings indicate that this property can be “reexpressed” during blastogenesis in culture.This observation also demonstrates technical problems associated with the use of SRBC to quantitate lymphocytes with complement receptors (B cells) by the EAC rosette assay in culture. False positives resulted from 37°C E rosette formation, but this was overcome by replacing the SRBC with guinea pig erythrocytes in the EAC assay.