Valency of CD3 binding and internalization of the CD3 cell-surface complex control T cell responses to second signals: distinction between effects on protein kinase C, cytoplasmic free calcium, and proliferation

We have studied the relationship of valency of CD3 stimulation and modulation of the CD3 receptor complex with biochemical and proliferative responses of T cells. Anti-CD3 Fab, as well as F(ab')2 and whole antibody caused rapid modulation of the CD3 antigen, whereas anti-CD3 conjugated to Sepharose did not. In the absence of monocytes, T cells stimulated with anti-CD3 Fab, F(ab')2, or F(ab')2-Sepharose showed differences in their ability to respond to second signals given by PMA, IL 1, IL 2, or antibodies to Tp67 and Tp44. None of the anti-CD3 signals alone caused resting T cells to produce IL 2, and only the Sepharose-bound anti-CD3 F(ab')2 caused T cells to express high levels of functional IL 2 receptors. Anti-CD3 F(ab')2-Sepharose-stimulated T cells produced IL 2 and proliferated in response to each of the second signals. Because anti-CD3-Sepharose did not cause modulation of the CD3 antigen, the ability of the Sepharose-bound antibody to induce T cells to express IL 2 receptors and to respond to individual second signals may be related to lack of modulation rather than valency of binding. Anti-CD3 Fab-stimulated T cells responded to PMA but required combinations of other second signals. T cells stimulated with unmodified anti-CD3 antibody or F(ab')2 fragments responded to PMA but did not respond to any other second signals alone or in combination. Stimulations that resulted in modulation (i.e., anti-CD3 whole antibody, anti-CD3 F(ab')2, or anti-CD3 Fab fragments) caused an increase in cytoplasmic calcium levels in resting T cells but blocked proliferation of T cells in response to mitogenic lectins or CD2 stimulation. Anti-CD3 F(ab')2 on Sepharose, however, did not block T cell proliferation. Whole bivalent anti-CD3 antibody or F(ab')2 fragments, but not monovalent Fab fragments, caused a rapid translation of protein kinase C activity from cytosol to membrane in the Jurkat T cell line. Because all of these modulate the receptor, these data indicate that the functional difference between monovalent and bivalent binding to CD3 is related to antibody valency and not to antigenic modulation. The use of Fab anti-CD3 stimulation that requires combinations of second signals for proliferation allowed an analysis of the functional relationships between IL 1, anti-Tp67, and anti-Tp44.     

In vivo cytotoxic efficacy of immunotoxins prepared from anti-CD5 antibody linked to ricin A-chain

Summary The antitumoral efficacy of various anti-CD5 immunotoxins, prepared with whole monoclonal antibody (mAb), F(ab)₂ or Fab fragment linked to native ricin A-chain (RTA) or partially deglycosylated ricin A-chain (dRTA), was examined in vivo in ascitic nude mice bearing a large burden of Ichikawa human tumour cells. We first demonstrated that after systemic administration of IgG-RTA or F(ab)₂-dRTA, the cytotoxic activity of immunotoxin molecules specifically bound to tumour cells was preserved. Secondly we showed, by using different immunotoxins with various targeting capacities, that their cytotoxic effect in vivo was related to the number of immunotoxin molecules bound per cell. However, even when antigen saturation was achieved after i.p. injection, the cytotoxic effect did not exceed 53% of the tumour burden. By contrast, when the immunotoxin was administered i.p. or i.v. with the enhancer monensin conjugated to human serum albumin and injected i.p., 90% of the tumour cells were killed. This potentiating effect was demonstrated even when the tumour localisation was as low as 5% of the saturation level. Such an effect could be completely prevented by addition of unconjugated monoclonal antibody, demonstrating the specificity of the immunotoxin-induced cytotoxicity in the presence of the enhancer. However this enhancement was demonstrated whatever the route of immunotoxin administration, i.p. or i.v., but was only observed when the enhancer was injected i.p. and not i.v.. These results emphasize the importance of optimizing the therapeutic course to improve the antitumoral efficacy of immunotoxins.     

The effect of antibody valency and lysosomotropic amines on the synergy between ricin A chain- and ricin B chain-containing immunotoxins

The cytotoxicity of A chain immunotoxins containing IgG or Fab fragments specific for the surface immunoglobulin of the Daudi cell line was assessed in the presence of B chain immunotoxins (IgG or Fab) or lysosomotropic amines, or both. The concentration required for 50% inhibition of protein synthesis (IC50) in Daudi cells was 1.3 X 10(-8) M for IgG-A and 5 X 10(-8) M for Fab-A. The toxicity of both A chain immunotoxins was enhanced twofold by ammonium chloride. In the presence of A chain immunotoxins and ammonium chloride, a maximum of 99 and 90% reduction of clonal precursors was obtained with IgG and Fab-A chain immunotoxins respectively. Immunotoxins containing ricin B chain and IgG or Fab fragments specific for the antibody portion of A chain immunotoxins were used as secondary "piggyback" immunotoxins to treat cells that were pretreated with A chain immunotoxins. Both B chain immunotoxins were nontoxic at 1 X 10(-6) M. When added to target cells pretreated with specific A chain immunotoxins, the IC50 of the A chain immunotoxins was decreased up to 16-fold in the absence of ammonium chloride. In contrast to the results obtained with A chain immunotoxins alone, ammonium chloride significantly increased the toxicity of the complete piggyback system, resulting in the killing of 99.999% or five logs of target cells in the clonal assay. This decreased the IC50 of A chain immunotoxins up to 116-fold when compared with A chain immunotoxin alone. This enhanced toxicity was independent of the valency of either immunotoxin.     

Purification of antibody Fab and F(ab')2 fragments using Gradiflow technology

The Gradiflow, a preparative electrophoresis instrument designed to separate molecules on the basis of their size and charge, was used to purify antibody Fab and F(ab')2 fragments. The method described is charge based, utilizing the difference in the pI between the antibody Fab/F(ab')2 fragments and antibody Fc fragments that occur after enzyme digestion of whole antibody molecules. This method of purification was successful across a range of monoclonal and polyclonal antibodies. In particular, F(ab')2 fragments were purified from a number of mouse monoclonal antibodies (both IgG1 and IgG2a isotypes) and Fab fragments were purified from egg yolk IgY polyclonal antibodies. This is a rapid purification method which has advantages over alternative methods that usually comprise ion exchange and gel filtration chromatography. This method may be applicable to most antibody digest preparations.     

Radiolocalisation of an anti-CEA monoclonal antibody (FO23C5) and its fragments in a colon carcinoma xenograft model

A new monoclonal antibody designated FO23C5 against a protein component of carcinoembryonic antigen (CEA) has been developed. A xenograft system of human colon cancer was used to compare the intact monoclonal IgG with its fragments (Fab')2 and Fab) and with an established anti-CEA antibody (MAb35) and the antibody AUA1 raised against the colon carcinoma cell line. We demonstrate that FO23C5 compares well with the existing anti-CEA antibody and with AUA1, and that F(ab')2 fragments perform best in achieving optimal tumour to normal tissue ratios compared with intact IgG and Fab fragment.     

Further studies on the biologic properties of guinea pig IgG1 antibodies. II. In vivo activation of C3 by anti-glomerular basement membrane antibodies.

Normal rats were injected with guinea pig anti-rat glomerular basement membrane antibodies of the IgG1 or IgG2 class or with their F (ab') 2 fragments, in order to study which antibody site triggers the alternate complement pathway in vivo. Both IgG classes were able to induce a heavy proteinuria and led to C3 deposition in the glomeruli in a pattern similar to their own distribution along the glomerular basement membrane, as shown by the immunofluorescence technique. The Fab(ab')2 fragment of IgG2 did not produce C3 binding or proteinuria. The F(ab')2 fragment of IgG1 was difficult to obtain devoid of Fc determinants. A F(ab')2 fragment of IgG1 still bearing Fc determinants led to C3 binding and proteinuria, whereas the true F(ab')2 fragment of IgG1 had none of these effects in two out of three animals.     

Triggering of the superoxide generation of macrophages by crosslinking of Fc gamma receptor

Crosslinking of monomeric IgG2 molecules bound to the Fc gamma receptors on the cell surface of guinea pig macrophages generated the triggering signal for the superoxide-generating system. A binding experiment indicated that macrophages have saturable binding sites for monomeric IgG2. Scatchard analysis of the binding data showed that macrophages have an average of 4 X 10(5) binding sites per cell and the association constant for the binding was 4.2 X 10(6) M-1. Binding of monomeric IgG2 to macrophages could be detected by subsequent reaction with the 125I-labeled F(ab')2 fragment of rabbit antibody specific for guinea pig Fab. Although binding of IgG2 monomer to Fc receptor did not stimulate superoxide release, further addition of the F(ab')2 fragment of anti-guinea pig Fab antibody did induce generation and release of superoxide, and the amount released was dependent on the dose of cell-bound IgG2. When macrophages were bound with a constant dose of IgG2 monomer in the first step, the superoxide release triggered by the addition of the F(ab')2 of anti-guinea pig Fab was dependent on the dose of the F(ab')2 fragment added. These results show that crosslinking of Fc receptors triggers the superoxide generation.     

Cytotoxicity of chimeric (human-murine) monoclonal antibody BR96 IgG, F(ab')2, and Fab' conjugated to Pseudomonas exotoxin.

We have made antigen-specific cytotoxic reagents by conjugating the chimeric antibody BR96 (chiBR96) to Pseudomonas exotoxin A (PE), as either native PE or a truncated form (LysPE40) devoid of the cell-recognition region (domain I). PE kills cells by ADP-ribosylation of elongation factor 2, thereby inhibiting protein synthesis. Chimeric BR96 immunotoxins were constructed by chemical conjugation of the toxin to Fab', F(ab')2, and intact IgG and purified by anion-exchange and gel-filtration chromatography. Chimeric BR96 [IgG and F(ab')2] immunotoxins were cytotoxic against tumor cell lines displaying the BR96 antigen, with EC50 values ranging from 0.1 to 110 pM. Immunotoxins constructed with chiBR96 Fab' were 50-100-fold less cytotoxic. Competition analysis showed that the immunotoxins were specifically active through their BR96 antigen-binding ability. The binding of chiBR96-PE and chiBR96-LysPE40 to antigen was equivalent to that of BR96 itself and these immunotoxins were found to internalize very rapidly, displaying 90% of their cytotoxicity within 1 h. Binding assays determined that chiBR96 F(ab')2-LysPE40 bound as well as chiBR96-LysPE40; however, chiBR96 Fab'-LysPE40 bound 20-fold less efficiently. The chiBR96 Fab'-LysPE40 internalized similarly to the F(ab')2 or the IgG immunotoxins. Therefore, the chiBR96 Fab'-LysPE40 immunotoxin is less cytotoxic toward target cells because of reduced antigen binding. This is may be due to the monovalent nature of chiBR96 Fab'-LysPE40. This study shows that the monoclonal antibody chiBR96-Pseudomonas exotoxin A immunotoxins can be effective at inhibiting protein synthesis in target cells.     

Activation of resting T lymphocytes by anti-CD3 (T3) antibodies in the absence of monocytes

The antigen receptor molecules on human T lymphocytes are noncovalently associated on the cell surface with the CD3 (T3) molecular complex. Perturbation of this complex with anti-CD3 monoclonal antibodies induces T cell activation. Previous studies have demonstrated that this process requires the participation of monocytes. In the present report, we demonstrate that purified, resting (G0 phase) T cells incubated with monoclonal anti-CD3 antibodies proliferate in response to purified interleukin 2 (IL 2), in a lymphokine dose-dependent fashion. Anti-CD3 antibody or IL 2 alone did not trigger cell division. The effect was specific for anti-CD3 antibodies because monoclonal antibodies reactive with other surface molecules (OKT4, OKT8, L368) were inactive. Furthermore, the same phenomenon was observed when anti-CD3 antibody Leu-4 (IgG1) was incubated with cells of individuals whose monocytes cannot process antibodies of the IgG1 subclass (Leu-4 nonresponders). In addition, both F(ab')2 and Fab fragments of anti-CD3 antibody OKT3 were also capable of rendering T cells receptive to the IL 2 growth signal. These data indicate that neither monocytes nor CD3 receptor cross-linking are required absolutely for resting T cell activation, provided that IL 2 is supplied exogenously. T lymphocytes treated with anti-CD3 antibodies proliferated in response to both purified mitogen-induced and recombinant IL 2. Antibodies to the IL 2 receptor (anti-Tac) inhibited the proliferation. Thus, the most likely mechanism for anti-CD3 antibody-mediated triggering is induction of IL 2 receptors.     

Structural effect of a recombinant monoclonal antibody on hinge region peptide bond hydrolysis

IgG hinge region peptide bonds are susceptible to degradation by hydrolysis. To study the effect of Fab and Fc on hinge region peptide bond hydrolysis, a recombinant humanized monoclonal IgG1 antibody, its F(ab')2 fragment, and a model peptide with amino acid sequence corresponding to the hinge region were incubated at 40 degrees C in formulation buffer including complete protease inhibitor and EDTA for 0, 2, 4, 6 and 8 weeks. Two major cleavage sites were identified in the hinge region of the intact recombinant humanized monoclonal antibody and its F(ab')2 fragment, but only one major cleavage site of the model peptide was identified. Hinge region peptide bond hydrolysis of the intact antibody and its F(ab')2 fragment degraded at comparable rates, while the model peptide degraded much faster. It was concluded that Fab region of the IgG, but not Fc portion had significant effect on preventing peptide bond cleavage by direct hydrolysis. Hydrolysis of hinge region peptide bonds was accelerated under both acidic and basic conditions.     

Antibodies to Tp67 and Tp44 augment and sustain proliferative responses of activated T cells

We have shown previously that binding of a monoclonal antibody (MAb) to Tp44 molecules increased the proliferation of anti-CD3-activated T cells by causing enhanced IL 2 receptor expression and IL 2 release. We now show that anti-CD5 (Tp67) antibodies have a similar effect under conditions in which monocytes are suboptimally activated or where monocytes are not present. The activity did not depend on antibody isotype or on the precise CD5 epitope recognized. Functional experiments indicated that both IL 2 production and IL 2 receptor expression were enhanced by antibody binding. Anti-Tp67 and anti-Tp44 appear to augment proliferation through distinct mechanisms, because both antibodies together had greater activity than either antibody alone. In neither system is the Fc portion of the antibody required, because F(ab')2 fragments had activity equivalent to that of the intact antibody and were effective at concentrations as low as 10 ng/ml. Fab fragments of anti-Tp67 were active, but Fab fragments of anti-Tp44 had no effect. Anti-Tp67 and anti-Tp44 were able to sustain continuous proliferation of anti-CD3-Sepharose-stimulated T cells for up to 2.5 wk without exogenous IL 2 or feeder cells. These experiments suggest that Tp67 and Tp44 are receptors that play a critical regulatory role in the control of T cell proliferation.     

Differences of T-cell activation by the anti-CD3 antibodies Leu4 and BMA030

Two anti-CD3 antibodies and their Fab/F(ab')2 fragments were compared with regard to their requirement for secondary signals and generations of intracellular messengers. The anti-CD3 antibody BMA030 was found to require monocyte contact to elicit T-cell mitogenesis. Cross-linking by plastic-bound goat anti-mouse antibodies (panning) failed to activate T cells, even in the presence of recombinant IL-1 or IL-2. In contrast, crosslinking of the anti-CD3 antibody Leu4 or Leu4 fragments was mitogenic in monocyte-free cultures. Measurements of intracellular Ca2+ ([Ca2+]i) and generation of inositol phosphates revealed that binding (+/- panning) of BMA030, Leu4, and their F(ab')2 fragments generated similar amounts of intracellular messengers and thus failed to explain the different responsiveness to passive crosslinking. Since the generation of these messengers was not necessarily followed by proliferation but was always observed when mitogenesis occurred, we conclude that the elevation of [Ca2+]i and the production of inositol phosphates are required but not sufficient to trigger mitogenesis.     

Roles of plasma proteins and surface negative charge of erythrocytes in erythrocyte aggregation

The effect of plasma proteins (and IgG fragments) and sialic acid content of erythrocytes on the aggregation of human erythrocytes was quantitatively examined by using a rheoscope combined with a television image analyser and a computer. (1) The velocity of erythrocyte aggregation by plasma proteins was increased with increasing in their molecular weight, i.e., IgG less than IgA less than fibrinogen less than IgM. F(ab')2. Fab and Fc could not induce the aggregation. (2) The aggregation induced by fibrinogen was accelerated by IgG and its peptic fragment, F(ab')2, but was unaffected by the plasmic fragments, Fab and Fc. The accelerating effect by IgG and F(ab')2 was inhibited by Fab and Fc. (3) The aggregation of erythrocytes was accelerated by decreasing the sialic acid content (due to the reduction of the electrostatic repulsive force among erythrocytes), and the effect of desialylation on the IgG-induced aggregation was greater than that of desialylation on the fibrinogen-induced aggregation. (4) The roles of plasma proteins and of sialic acid content of erythrocytes on the aggregation of erythrocytes were discussed.     

Influence of subunit-specific antibodies on the activity of the F0 complex of the ATP synthase of Escherichia coli. II. Effects of subunit c-specific polyclonal antibodies.

After incubation of F1-stripped everted membrane vesicles with antibodies against subunit c of the ATP synthase of Escherichia coli the proton translocation through the open F0 channel was blocked. Rebinding of F1 to those vesicles is affected by the antibody concentration used. In general, the use of F(ab')2 or Fab fragments prepared from anti-c antibodies gave similar results. However, using Fab fragments a higher amount of antigenic binding sites was necessary to block the F0 complex completely, whereas extremely low amounts of Fab fragments were necessary to inhibit the binding of F1. This can be explained by an antigenic determinant of subunit c, which is only accessible to the smaller Fab fragments with a molecular mass of approximately 50,000. Incubation of F1-containing everted membranes with anti-c antibodies showed that the binding of the antibodies resulted in a displacement of F1, while simultaneously the proton translocation through F0 has been blocked. Such a displacement can only be observed after incubation with IgG molecules or F(ab')2 fragments. Fab fragments were not able to displace the F1 part, indicating that the ability of antibodies and F(ab')2 fragments to produce cross-links is responsible for the loss of F1 from the membranes.     

A 44 kilodalton cell surface homodimer regulates interleukin 2 production by activated human T lymphocytes

We previously described a cell surface antigen, termed Tp44, detected by monoclonal antibody 9.3 on approximately 80% of mature human T lymphocytes. Analysis by SDS-polyacrylamide gel electrophoresis and isoelectric focusing demonstrated that this antigen consists of two identical 44 kilodalton glycopeptides that form a disulfide-linked homodimer. Competitive binding experiments showed that antibody 9.3 and an anti-CD3 antibody (64.1) recognize distinct antigenic determinants; furthermore, the binding of antibody 9.3 was unaffected by prior modulation of CD3. Thus, Tp44 has no detectable cell surface association with CD3. By itself, antibody 9.3 had no detectable effect on either IL 2 receptor expression or IL 2 release, and did not cause T cell proliferation even when monocytes were present and exogenous IL 2 was provided, indicating that binding of antibody 9.3 does not provide a primary signal for T cell activation. However, the proliferative responses of T lymphocytes activated by phytohemagglutinin, concanavalin A, or an anti-CD3 monoclonal antibody were strikingly enhanced in the presence of antibody 9.3, an effect associated with increased IL 2 receptor expression and increased IL 2 secretion. Antibody 9.3 enabled anti-CD3-Sepharose-activated T cells and anti-CD3 antibody-activated Jurkat cells to release IL 2 in the absence of monocytes. Fab fragments of antibody 9.3 had no effect on anti-CD3-induced IL 2 release by Jurkat cells, whereas F(ab')2 fragments had activity comparable to that of unmodified antibody, indicating that bivalent binding of Tp44 molecules is required for IL 2 secretion. Together, these results suggest that TP44 may function as a receptor for accessory signals in the activation of T cells.     

Improved binding of a bivalent single-chain immunotoxin results in increased efficacy for in vivo T-cell depletion.

Anti-CD3 immunotoxins exhibit considerable promise for the induction of transplantation tolerance in pre-clinical large animal models. Recently an anti-human anti-CD3epsilon single-chain immunotoxin based on truncated diphtheria toxin has been described that can be expressed in CHO cells that have been mutated to diphtheria toxin resistance. After the two toxin glycosylation sites were removed, the bioactivity of the expressed immunotoxin was nearly equal to that of the chemically conjugated immunotoxin. This immunotoxin, A-dmDT390-sFv, contains diphtheria toxin to residue 390 at the N-terminus followed by VL and VH domains of antibody UCHT1 linked by a (G(4)S)(3) spacer (sFv). Surprisingly, we now report that this immunotoxin is severely compromised in its binding affinity toward CD3(+) cells as compared with the intact parental UCHT1 antibody, the UCHT1 Fab fragment or the engineered UCHT1 sFv domain alone. Binding was increased 7-fold by adding an additional identical sFv domain to the immunotoxin generating a divalent construct, A-dmDT390-bisFv (G(4)S). In vitro potency increased 10-fold over the chemically conjugated immunotoxin, UCHT1-CRM9 and the monovalent A-dmDT390-sFv. The in vivo potency of the genetically engineered immunotoxins was assayed in the transgenic heterozygote mouse, tgepsilon 600, in which the T-cells express human CD3epsilon as well as murine CD3epsilon. T-cell depletion in the spleen and lymph node observed with the divalent construct was increased 9- and 34-fold, respectively, compared with the monovalent construct. The additional sFv domain appears partially to compensate for steric hindrance of immunotoxin binding due to the large N-terminal toxin domain.     

抗CD20嵌合抗体片段F(ab′)2突变体在大肠杆菌中的高效分泌表达

构建抗CD20嵌合抗体片段F(ab′)₂ 突变体 ,研究其在大肠杆菌中的高效表达及其表达产物的生物学活性。采用PCR法构建抗CD20嵌合抗体片段F(ab′)₂ 突变体 ,并用双脱氧终止法测定DNA序列 ;采用 19L发酵罐高密度发酵抗CD20嵌合抗体片段F(ab′)₂ 突变体 ,采用亲和色谱和分子筛色谱法纯化表达产物 ,并用SDS-PAGE和薄层激光扫描鉴定纯化产物 ;采用活细胞间接免疫荧光法测定纯化产物与靶细胞的结合活性 ;MTT法测定纯化产物对Raji细胞的生长抑制作用 ,并研究其作用机理。DNA序列测定结果表明 ,抗CD20嵌合抗体片段F(ab′)₂ 突变体已成功构建 ,表达可溶性产物的产量达 360mg L ,具有与Raji细胞 (CD20⁺)结合的活性 ,并抑制Raji细胞的生长 ,其作用机理为诱导Raji细胞凋亡。此突变体有望成为治疗非何杰金氏B细胞淋巴瘤的药物。     

On the fragmentation of monoclonal IgG1, IgG2a, and IgG2b from BALB/c mice

Methods for the production and purification of F(ab')2 fragments from BALB/c monoclonal IgG1, IgG2a, and IgG2b with pepsin and other proteases were examined. The overall susceptibility to degradation is IgG2b greater than IgG2a greater than IgG1. Stable F(ab')2 can be produced in good yield from IgG1 with pepsin at pH 3.5 to 4.0 and can be made directly by pepsin treatment of ascites fluids or cell culture supernatants containing IgG1. IgG2a is cleaved in two steps by pepsin, first to F(ab')2 and then to Fab'. With carefully chosen conditions, F(ab')2 can be obtained in acceptable yield. The primary cleavage for the IgG2a heavy chain appears to be on the COOH terminal side of the interheavy chain disulfides, and secondary cleavage is on the NH2-terminal side. For IgG2b the reverse is true, and F(ab')2 has not been obtained in useful amounts; however, the primary cleavage of IgG2b appears to be assymetric with respect to the two heavy chains, and Fab/c fragments that have one Fab plus Fc can be made. Digestion with elastase resulted in the best yield of Fab/c. This finding may provide a method for retaining cytotoxicity in monoclonal antibodies against cell surface antigens while eliminating their capacity to modulate. The cleavage patterns of the three classes of IgG are rationalized in terms of the structure of their hinge regions.     

Uptake of native and deglycosylated ricin A-chain immunotoxins by mouse liver parenchymal and non-parenchymal cells in vitro and in vivo

The therapeutic activity of ricin A-chain immunotoxins is undermined by their rapid clearance from the bloodstream of animals by the liver. This uptake has generally been attributed to recognition of the mannose-terminating oligosaccharides present on ricin A-chain by receptors present on the non-parenchymal (Kupffer and sinusoidal) cells of the liver. However, we demonstrate here that, in the mouse, the liver uptake of a ricin A-chain immunotoxin occurs in both parenchymal and non-parenchymal cells in equal amounts. This is in contrast to the situation in the rat, where uptake of the immunotoxin is predominantly by the non-parenchymal cells. Recognition of sugar residues on the A-chain portion of the immunotoxin plays an important role in the liver uptake by both cell types in both species. However it is not the only mechanism since, firstly, an immunotoxin containing ricin A-chain which had been effectively deglycosylated with metaperiodate and cyanoborohydride was still trapped to a significant extent by hepatic non-parenchymal cells after it was injected into mice. Secondly, deglycosylation, while eliminating uptake of the free A-chain by parenchymal and non-parenchymal cells in vitro, only reduced the uptake of an immunotoxin by either cell type by about half. Thirdly, the addition of excess D-mannose or L-fucose inhibited the uptake of free A-chain by mouse liver cell cultures by more than 80% but only inhibited the uptake of the native A-chain immunotoxin by about half and had little effect on the uptake of the deglycosylated ricin A-chain immunotoxin. Recognition of the antibody portion of the immunotoxin by liver cells seems improbable, since antibody alone or an antibody-bovine serum albumin conjugate were not taken up in appreciable amounts by the cultures. Possibly attachment of the A-chain to the antibody exposes sites on the A-chain that are recognised by liver cells in vitro and in vivo.