Intracellular routing rather than cross-linking or rate of internalization determines the potency of immunotoxins directed against different epitopes of sIgD on murine B cells.

Several variables influence the potency of an immunotoxin (IT) prepared with a monoclonal antibody (mAb) and ricin A chain (IT-A). These include the affinity of the mAb, the nature and density of the target antigen (Ag), the epitope on the target Ag bound by the mAb, the type of cell target, and the rate of endocytosis and route of internalization of the bound IT-A. In a previous report, we demonstrated that anti-delta mAbs directed against epitopes which are putatively more proximal to the plasma membrane make more effective IT-As than those directed against epitopes that are putatively more distal from the plasma membrane. It is known that the latter mAbs cross-link sIgD less effectively than the former. Therefore, in the present study, we determined whether the differential cytotoxicity of IT-As directed against these epitopes is related to their ability to cross-link their specific surface antigen (sIgD). We further determined whether they were internalized at different rates by normal B cells. Our results show that neither cross-linking nor rate of internalization account for the different potencies of anti-Fc vs anti-Fd IT-As. However, when these IT-As were used in the presence of the lysosomotropic agent chloroquine, the less potent IT-A became 100-fold more potent and was as cytotoxic as the effective anti-Fc IT-A. Taken together with the results of other studies, these findings further support the hypothesis that the epitope specificity of a given mAb may be an important factor in determining the intracellular routing of an IT-A after internalization.     

Ricin A-chain containing immunotoxins directed against different epitopes on the CD2 molecule differ in their ability to kill normal and malignant T cells

Immunoconjugates consisting of mAb covalently coupled to plant or bacterial toxins or to cytotoxic drugs have provided novel experimental reagents for the treatment of malignancies and autoimmune diseases. In this report, we analyzed the efficacy of three ricin A chain-containing immunotoxins (IT-A) which recognize different epitopes on the CD2 molecule (E rosette receptor) on human T cells. Although all IT-A had similar binding avidities and A-chain activities, one (RFT11-A) was 100-1000-fold more effective in killing normal and malignant T cells than the others (35.1-A, 9.6-A). Immunoprecipitation experiments confirmed that all IT-A bound to the CD2 molecule. However, cross-blocking experiments, differential proteolysis with trypsin, and T cell co-activation experiments showed that the less effective IT-A, 35.1-A and 9.6-A, bound to an epitope far from the cell membrane (region I), whereas the more effective IT-A, RFT11-A bound to an epitope closer to the membrane (region II). Using cellular RIA and immunoelectron microscopy, it was shown that both RFT11-A and 35.1-A were rapidly internalized by T cells, but that their intracellular fates differed. The more toxic IT-A, RFT11-A, was retained for longer periods of time inside the cells and was more slowly degraded than the less effective IT-A, 35.1-A, which was rapidly transported to lysosomes, digested, and expelled. These results demonstrate that different IT-A targeting the same surface molecule can differ markedly in potency, and that the epitope recognized by an IT-A may have a significant impact on the ability of the IT-A to insert into cell membranes, translocate to the cytosol, and kill cells.     

Immunotoxins to human small-cell lung cancer.

Ricin A chain ITs directed against a variety of the common cell-surface antigens associated with SCLC exerted selective toxic effects on SCLC cell lines. The potency of the cytotoxic effects matched or exceeded that previously reported for ricin A chain ITs directed against identical or similar antigens on other types of carcinoma, suggesting that SCLC may be uniquely sensitive to this type of IT.     

Immunotoxins to human small-cell lung cancer

Ricin A chain ITs directed against a variety of the common cell-surface antigens associated with SCLC exerted selective toxic effects on SCLC cell lines. The potency of the cytotoxic effects matched or exceeded that previously reported for ricin A chain ITs directed against identical or similar antigens on other types of carcinoma, suggesting that SCLC may be uniquely sensitive to this type of IT.     

Potentiation of anti-carcinoembryonic antigen immunotoxin cytotoxicity by monoclonal antibodies reacting with co-expressed carcinoembryonic antigen epitopes

The initial step in ricin A-chain (RTA)-immunotoxin-mediated cell cytotoxicity involves binding to the target cell Ag through the antibody moiety. One of the factors influencing this is the affinity of the antibody component for the target cell Ag. Multiple epitopes on carcinoembryonic Ag have been mapped providing a range of mAb of known specificity. These have been used to show that the cytotoxicity of an immunotoxin containing RTA conjugated to an anti-carcinoembryonic Ag mAb (228-RTA) is potentiated by mAb recognizing different epitopes. The potentiating antibodies also increased the level of target cell binding of antibody 228. Cross-linking of cell bound antibody was not involved because monovalent fragments of a potentiating antibody were effective. The potentiating antibodies modified the binding affinity of 228 antibody increasing the t1/2 of antibody at the tumor cell surface. This increased the dwell time of cell bound antibody and using conjugates of 228 linked to albumin-tetramethylrhodamine it was shown to enhance conjugate endocytosis. These investigations indicate that enhanced antibody affinity leads to increased endocytosis of bound immunoconjugate and potentiates cytotoxicity.     

Generation of anti-idiotypic and anti-anti-idiotypic monoclonal antibodies in the same fusion. Support of Jerne's Network Theory

Upon immunization of mice with a mAb (290A-167) directed against an epitope of Lol p I (the major allergenic determinant of Lolium perenne), both anti-idiotypic (aId) mAb (Ab2) and anti-aId mAb (Ab3) were produced. The Ab2 displayed the following internal image properties of Lol p I: it can be affinity-purified on an immobilized Id column; its binding to the anti-Lol p I mAb (290A-167) is inhibited by Lol p I; it inhibits in a dose-response fashion the binding of the specific Id to Ag. It is recognized by anti-Lol p I antisera from different species such as mouse, human, and goat. The Ab3 which binds to Lol p I was also produced from the same fusion. This binding was inhibited significantly by aId mAb (Ab2), anti-Lol p I mAb (290A-167) and Lol p I. These data indicate that the two mAb with specificity for Lol p I (290A-167 and Ab3) share similar reactivity to the Ag and that aId mAb is the internal image of the epitope recognized by the Id. We showed also that the capacity of rabbit aId Ab directed against the 290A-167 Id to inhibit the binding of Ab1 and Ab3 to Ag was almost abolished by passage over a Ab3-coated Sepharose column. This would suggest that not only are the two mAb with reactivity to Lol p I (Ab1 and Ab3) directed against identical epitopes, but that they in fact shared identical idiotopes as well. The production of identical mAb upon immunization with either the Ag or the aId mAb supports that the conceptual framework proposed by Jerne finds its biologic application in the course of an immune response.     

Functional evidence for three distinct and independently inhibitable adhesion activities mediated by the human integrin VLA-4. Correlation with distinct alpha 4 epitopes.

The human integrin VLA (very late activation antigens)-4 (CD49d/CD29), the leukocyte receptor for both the CS-1 region of plasma fibronectin (Fn) and the vascular cell surface adhesion molecule-1 (VCAM-1), also mediates homotypic aggregation upon triggering with specific anti-VLA-4 monoclonal antibody (mAb). Epitope mapping of this integrin on the human B-cell line Ramos, performed with a wide panel of anti-VLA-4 mAb by both cross-competitive cell binding and protease sensitivity assays, revealed the existence of three topographically distinct epitopes on the alpha 4 chain, referred to as epitopes A-C. By testing this panel of anti-VLA-4 mAb for inhibition of cell binding to both a 38-kDa Fn fragment containing CS-1 and to VCAM-1, as well as for induction and inhibition of VLA-4 mediated homotypic cell adhesion, we have found overlapping but different functional properties associated with each epitope. Anti-alpha 4 mAb recognizing epitope B inhibited cell attachment to both Fn and VCAM-1, whereas mAb against epitope A did not block VCAM-1 binding and only partially inhibited binding to Fn. In contrast, mAb directed to epitope C did not affect cell adhesion to either of the two VLA-4 ligands. All mAb directed to site A, as well as a subgroup of mAb recognizing epitope B (called B2), were able to induce cell aggregation, but this effect was not exerted by mAb specific to site C and by a subgroup against epitope B (called B1). Moreover, although anti-epitope C and anti-epitope B1 mAb did not trigger aggregation, those mAb blocked aggregation induced by anti-epitope A or B2 mAb. In addition, anti-epitope A mAb blocked B2-induced aggregation, and conversely, anti-epitope B2 mAb blocked A-induced aggregation. Further evidence for multiple VLA-4 functions is that anti-Fn and anti-VCAM-1 antibodies inhibited binding to Fn or to VCAM-1, respectively, but did not affect VLA-4-mediated aggregation. In summary, we have demonstrated that there are at least three different VLA-4-mediated adhesion functions, we have defined three distinct VLA-4 epitopes, and we have correlated these epitopes with the different functions of VLA-4.     

Treatment of HIV tissue culture infection with monoclonal antibody-ricin A chain conjugates

mAb 907 is directed against the envelope protein of the HIV. The epitope recognized by this antibody is expressed in moderate density on the surface of tissue culture cells infected with the LAV/HTLV-IIIB strain of HIV. We have coupled antibody 907 to ricin A chain (RAC). The antibody-RAC conjugate inhibited protein synthesis and cell growth in HIV-infected cells. An irrelevant antibody conjugated to RAC had no effect. Most important, treatment of infected cells with the conjugate markedly inhibited the production of infectious virus, as measured by the production of viral foci on susceptible monolayer cells. Exposure of HIV-infected target cells to the conjugate for as short a period as 1 h resulted in cell death. Serum of AIDS patients inhibited, but did not completely suppress, the toxicity of the 907-RAC conjugate. A second antibody, designated BM-1, which recognizes a carbohydrate Ag on the surface of virally infected cells, was conjugated to RAC. The BM-1-RAC conjugate did not kill HIV-infected cells, highlighting the importance of the target Ag. Immunotoxins produced with antibodies that recognize Ag on the surface of HIV-infected cells may have utility in the therapy of AIDS.     

Marked difference in the in vivo antitumor efficacy between two immunotoxins targeted to different epitopes of common acute lymphoblastic leukemia antigen (CD10). Mechanisms involved in the differential activities of immunotoxins.

In the present study, two isotype-matching mAb, SN5d and SN5, which are directed toward two distinctively different epitopes of common acute lymphoblastic leukemia Ag (CD10) but show a very similar binding affinity to leukemia cells, were compared for their in vivo antitumor activity after conjugated to ricin A chain (RA). Our recently established nude mouse model carrying an ascitic tumor of NALM-6 human pre-B leukemia cells was used as the tumor model. A marked difference was observed in the in vivo antitumor efficacy between SN5d-RA and SN5-RA; SN5d-RA was much more effective than SN5-RA. Several experiments were carried out to gain information concerning the mechanisms involved in the different antitumor efficacy of the two immunotoxins. Although naked (unconjugated) mAb SN5d was much less effective than SN5d-RA conjugates in the in vivo tumor suppression, mAb SN5d was more effective than mAb SN5 in the in vivo tumor suppression. Additionally, marked differences were found between SN5d and SN5 in the induction of antigenic modulation and in the regulation of Ag biosynthesis and expression. Binding of SN5 to NALM-6 leukemia cells caused strong antigenic modulation (down-regulation of Ag expression) and strongly down-regulated Ag biosynthesis and cell surface expression of new Ag. In contrast, binding of SN5d to NALM-6 leukemia cells caused little modulation of overall cell surface expression of common acute lymphoblastic leukemia Ag; the decrease of old Ag by endocytosis after binding to mAb SN5d was compensated by newly exocytosed cell-surface expressed Ag. The present results appear to reveal a novel mechanism which regulates cytotoxic activities of antibodies and immunoconjugates.     

Evaluation of ricin A-chain immunotoxins directed against human T cells

We have synthesized four immunotoxins (ITs) by covalently coupling the A chain of ricin to murine monoclonal antibodies that recognize surface antigens on human T cells. Treatment of human peripheral blood lymphocytes with either 10.2-A, directed against the CD5 (Tp67) antigen, or 64.1-A, directed against the CD3 (Tp19) antigen, abolished protein synthesis in cells subsequently cultured with phytohemagglutinin (PHA). In contrast, two other ITs (9.6-A and 35.1-A), both directed against the CD2 (Tp50) antigen, had minimal effects on protein synthesis in PHA-stimulated cells. The binding of each IT to T cells was shown by immunofluorescence with fluorescein-conjugated goat anti-mouse immunoglobulin (FITC-GAMIg) and fluorescein-conjugated rabbit anti-ricin A-chain (FITC-RAR) antibodies. Activity of the ricin A chain in each IT was demonstrated by its ability to inhibit protein synthesis in a cell-free reticulocyte lysate assay. Ultrastructural immunoperoxidase analysis of IT internalization showed that ineffective and effective ITs were endocytosed at the same rate (50% of cells had labeled endosomes after 15 min). However, ineffective IT 35.1-A was more rapidly delivered to lysosomes (15-30 min) than effective ITs (10.2-A and 64.1-A) (greater than or equal to 30 min). The data support the hypothesis that there are several distinct pathways for internalization of ITs and that the ability of ricin A chain to reach and inactivate ribosomes may depend upon the specific membrane receptor involved in binding a given IT, its route of internalization, and the rate of entry of the IT into lysosomes.     

Identification of epitopes recognized by monoclonal antibodies directed against HTLV-I envelope surface glycoprotein using peptide phage display

Summary Phage peptide libraries constitute powerful tools for the mapping of epitopes recognized by monoclonal antibodies (mAbs). Using screening of phage displayed random peptide libraries we have characterized the binding epitopes of three mAbs directed against the surface envelope glycoprotein (gp46) of the human T-cell leukemia virus type I (HTLV-I). Two phage libraries, displaying random heptapeptides with or without flanking cysteine residues, were screened for binding to mAbs 7G5D8, DB4 and 4F5F6. The SSSSTPL consensus sequence isolated from constrained heptapeptide library defines the epitope recognized by DB4 mAb and corresponds to the exact region 249–252 of the virus sequence. The APPMLPH consensus sequence isolated from non constrained heptapeptide library defines the epitope recognized by 7G5D8 mAb and corresponds to the region 187–193 with a single amino acid substitution, methionine to leucine at position 190. The third consensus sequence LYWPHD isolated from constrained heptapeptide library defines the epitope recognized by 4F5F6 mAb. It corresponds to an epitope without direct equivalence with the virus sequence. The data presented here showed that 7G5D8 and DB4 mAbs are raised against linear epitopes while 4F5F6 mAb recognized a continoous topographic epitope.     

Cross-antigenicity between the major surface proteins (ospA and ospB) and other proteins of Borrelia burgdorferi

Two of the major surface Ag of Borrelia burgdorferi, the 31-kDa OspA and 34-kDa OspB proteins, are encoded by a 49-kb plasmid. In this study, mAb and monospecific polyclonal antibodies were used to define cross-antigenicity of the OspA and OspB protein to each other and to other lower molecular mass proteins by Western blot analysis. Two mAb studied, 105.5 and 184.1, were directed predominantly against the 31-kDa OspA protein. However, each also reacted with other minor bands, though with different specificities. Using V8 protease digestion and cleavage by cyanogen bromide, we demonstrated that each mAb reacted to the 31-kDa protein differently. Monospecific polyclonal rabbit and human antibodies directed against the 34-, 31-, 22-, and 20-kDa proteins were eluted from blots and used to further corroborate the cross-reactivity among these Ag. Rabbit antibodies to the 31- and 22-kDa Ag gave remarkably similar peptide maps after V8 protease digestion of the 31-kDa OspA protein, as did mAb 184.1, suggesting that this mAb recognized an immunodominant epitope common to the 22- and 31-kDa proteins. It seems likely therefore that the humoral immune response to Borrelia surface Ag may be due to a limited number of cross-reactive epitopes on distinct, but related, gene products.     

Protective efficacy of IgM monoclonal antibodies in experimental group B streptococcal infection is a function of antibody avidity

We have produced and characterized six mAb directed against group B streptococci (GBS). All antibodies are IgM. We have previously shown that some of these antibodies are highly protective in the treatment of experimental infections in neonatal rats, whereas others do not appear to have any protective efficacy. Using an ELISA, we demonstrate the specificity of both protective and nonprotective antibodies. Two antibodies, binding different epitopes, are directed against antigenic structures present on all GBS; two are specific for type III carbohydrate determinants; one binds to a protein Ag present on all type I and II GBS; and one appears to bind to type Ia GBS only. Quantitative absorption assays provide evidence that the difference between protective antibodies and nonprotective antibodies is the avidity that the antibody demonstrates for the epitope recognized on the surface of the bacteria; 10 to 15 times as much protective antibody binds to GBS as does nonprotective antibody. Direct binding experiments with radiolabeled antibody confirm this conclusion.     

Monoclonal antibodies directed against the galactose-binding lectin of Entamoeba histolytica enhance adherence.

The Entamoeba histolytica galactose-binding lectin is a surface glycoprotein composed of 170- and 35-kDa subunits. Inhibition of this lectin with galactose or anti-170 kDa subunit polyclonal antibody blocks amebic adherence to target cells and colonic mucin glycoproteins. We describe the properties of 10 mAb with specificity for the 170-kDa subunit. Based on competitive binding studies, six nonoverlapping antigenic determinants on the lectin were identified. The effect of the mAb on adherence of amebic trophozoites to both Chinese hamster ovary (CHO) cells and human colonic mucins was measured. Antilectin antibodies directed against epitopes 1 and 2 enhanced adherence, with the number of amebae having at least three adherent CHO cells increasing with the addition of epitope 1 mAb from 26 +/- 9 to 88 +/- 2% and the binding of colonic mucins increasing from 34 +/- 1 to 164 +/- 3 pg/10(5) amebae. Antibody-enhanced adherence remained 90 to 100% galactose inhibitable, occurred at 4 degrees C and was not Fc mediated. Univalent Fab fragments of epitope 1 mAb augmented mucin binding by 238% and CHO cell adherence by 338%. The binding of purified lectin to CHO cells was increased from 1.1 +/- 0.1 to 2.4 +/- 0.3 ng/10(3) CHO cells by mAb directed to epitope 1, demonstrating that enhanced adherence was due to direct activation of the lectin. mAb to epitope 3 bound to the lectin only upon its solubilization from the membrane and had no effect on adherence. Adherence to CHO cells and mucins was inhibited from 50 to 75% by mAb to epitopes 4 and 5; epitope 6 mAb inhibited amebic adherence to CHO cells but not mucins. The pooled sera from 10 patients with amebic liver abscess blocked the binding to the 170-kDa subunit of mAb directed to all six epitopes. Striking individual variations in the effects of immune sera on adherence were observed. Although the sera of all 44 South African patients with amebic liver abscess had high titer anti-lectin antibodies, 16 patients' sera significantly (more than 3 SEM) enhanced adherence whereas 25 patients' sera significantly inhibited adherence. Antilectin antibodies exert profound functional effects on the interaction of E. histolytica with target cells and human colonic mucins. Exploration of the clinical consequences of adherence-enhancing and inhibitory antibody responses may give insight into the role of antilectin antibodies in immunity to invasive amebiasis.     

Epitopes of the p70 and p80 (Ku) lupus autoantigens.

High titer autoantibodies to the Ku Ag, a DNA-protein complex containing 70- and approximately 80-kDa protein subunits (p70 and p80, respectively), are found in sera of certain patients with systemic lupus erythematosus and related disorders. Autoepitopes of the Ku Ag were identified and partially characterized by expressing fragments of the p70 and p80 cDNA as fusion proteins in bacteria. Systemic lupus erythematosus sera reacted on immunoblots with at least three epitopes of p70 (amino acids 560-609, 506-535, and 115-467), and three epitopes of p80 (amino acids 682-732, 558-681, and 1-374). These six antigenic regions had distinct amino acid sequences, and were also immunologically distinct, as determined by using immunoaffinity-purified auto-antibodies to particular epitopes. Detailed mapping of the strongly antigenic region near the C terminus of p70 revealed a complex conformational or discontinuous epitope, the antigenicity of which was abolished by deleting either amino acids 560-571 or 601-609. The C terminus of p80 may also contain a discontinuous or conformational epitope(s). Although only some sera reacted with p70 or p80 on immunoblots, all sera that immunoprecipitated the native Ku complex reacted with native Ku by ELISA, and inhibited the binding of mAb directed at epitopes of native Ku. Taken together, these studies indicate that anti-Ku autoantibodies target a diversity of independent epitopes located on p70, p80, and the intact Ku complex, and that a significant portion of the autoantibodies in most patients' sera is directed against conformational/discontinuous epitopes.     

Differential segmental flexibility and reach dictate the antigen binding mode of chimeric IgD and IgM: implications for the function of the B cell receptor

Mature, naive B cells coexpress IgD and IgM with identical binding sites. In this study, the binding properties of such IgM and IgD are compared to determine how size and shape may influence their ability to bind Ag and thus function as receptors. To dissect their intrinsic binding properties, recombinant IgM and IgD were produced in soluble form as monomers of the basic H(2)L(2) Ab architecture, each with two Ag binding sites. Since these sites are connected with a hinge region in IgD and structural Ig domains in IgM, the two molecules differ significantly in this region. The results show that IgD exhibited the larger angle and longer distance between its binding sites, as well as having the greater flexibility. Relative functional affinity was assessed on two antigenic surfaces with high or low epitope density, respectively. At high epitope density, IgM had a higher functional affinity for the Ag compared with IgD. The order was reversed at low epitope density due to a decrease in the functional affinity of IgM. Studies of binding kinetics showed similar association rates for both molecules. The dissociation rate, however, was slower for IgM at high epitope density and for IgD at low epitope density. Taken together, the results show that IgM and IgD with identical Ag binding regions have different Ag binding properties.     

The biologic activity of anti-T cell receptor V region monoclonal antibodies is determined by the epitope recognized

We have used 10 independently isolated mAb reactive with the Ag R on a cloned Th cell line to map three distinct epitopes and three subepitopes on the R. One of these epitopes is clearly on the V beta 8 region, as it is defined by the antibodies KJ-16 and F23.1, known to react with the V beta 8 family of variable regions, and a functional rearranged V beta 8 gene has been cloned from this cell line. Antibodies directed at a second epitope, believed to be on V alpha because it is unaffected by anti-V beta antibodies, are completely inhibited from binding by monoclonal anti-CD3 epsilon-chain antibodies. Because the cloned Th cell line used, D10.G4.1, responds to soluble monoclonal anti-TCR antibodies, it has been possible to compare the binding of anti-R antibodies with their ability to activate this cloned T cell line. We find that for antibodies all specific for the same or a closely related epitope, activation is proportional to binding, by using antibodies that differ by greater than 100-fold in avidity for the R. By contrast, antibodies directed at different epitopes on the R differ markedly in their ability to activate the D10.G4.1 cell line. We have tested whether these differences reflect differences in the orientation of cross-linking the TCR or possible conformational changes induced in the R by the antibodies, and our data support the latter hypothesis as an explanation for the differences in activation potency between antibodies.     

Epitope diversity of hepatitis B virus capsids: quasi-equivalent variations in spike epitopes and binding of different antibodies to the same epitope

To investigate the range of antigenic variation of HBV capsids, we have characterized the epitopes for two anti-capsid antibodies by cryo-electron microscopy and image reconstruction of Fab-labeled capsids to approximately 10A resolution followed by molecular modeling. Both antibodies engage residues on the protruding spikes but their epitopes and binding orientations differ. Steric interference effects limit maximum binding to approximately 50% average occupancy in each case. However, the occupancies of the two copies of a given epitope that are present on a single spike differ, reflecting subtle distinctions in structure and hence, binding affinity, arising from quasi-equivalence. The epitope for mAb88 is conformational but continuous, consisting of a loop-helix motif (residues 77-87) on one of the two polypeptide chains in the spike. In contrast, the epitope for mAb842, like most conformational epitopes, is discontinuous, consisting of a loop on one polypeptide chain (residues 74-78) combined with a loop-helix element (residues 78-83) on the other. The epitope of mAb842 is essentially identical with that previously mapped for mAb F11A4, although the binding orientations of the two monoclonal antibodies (mAbs) differ, as do their affinities measured by surface plasmon resonance. From the number of monoclonals (six) whose binding had to be characterized to give the first duplicate epitope, we estimate the total number of core antigen (cAg) epitopes to be of the order of 20. Given that different antibodies may share the same epitope, the potential number of distinct anti-cAg clones should be considerably higher. The observation that the large majority of cAg epitopes are conformational reflects the relative dimensions of a Fab (large) and the small size and close packing of the motifs that are exposed and accessible on the capsid surface.     

Induction of an anti-human type II collagen response by monoclonal anti-idiotypic antibody

Several distinct epitopes on human type II collagen were defined by using mAb. The presence of species-specific and species-nonspecific (common) epitopes was thus clarified. Anti-idiotypic mAb (Ab2) was developed against one of the antibodies (Ab1) reactive with species-specific epitopes. Thus Ab2 was demonstrated to recognize an idiotope expressed on the Ag-binding site (paratope) of Ab1, since the binding of Ab1 to human type II collagen was blocked by Ab2, and the binding of Ab2 to Ab1 was inhibited by soluble human type II collagen, but not by murine and bovine type II collagens. DBA/1 mice immunized with Ab2 coupled to keyhole limpet hemocyanin produced an antibody (Ab3) specifically reactive with human type II collagen. It was also demonstrated that Ab3 expressed an idiotype similar to that of Ab1. These findings indicate that anti-idiotypic antibody directed against mAb to human type II collagen mimics a species-specific epitope on human type II collagen. The anti-idiotypic antibody bearing internal image of type II collagen will open the way to isolation of the arthritogenic epitope on type II collagen.     

B cell responses to a peptide epitope. X. Epitope selection in a primary response is thermodynamically regulated

We examine the etiological basis of hierarchical immunodominance of B cell epitopes on a multideterminant Ag. A model T-dependent immunogen, containing a single immunodominant B cell epitope, was used. The primary IgM response to this peptide included Abs directed against diverse determinants presented by the peptide. Interestingly, affinity of individual monomeric IgM Abs segregated around epitope recognized and was independent of their clonal origins. Furthermore, affinity of Abs directed against the immunodominant epitope were markedly higher than that of the alternate specificities. These studies suggested that the affinity of an epitope-specific primary response, and variations therein, may be determined by the chemical composition of epitope. This inference was supported by thermodynamic analyses of monomer IgM binding to Ag, which revealed that this interaction occurs at the expense of unfavorable entropy changes. Permissible binding required compensation by net enthalpic changes. Finally, the correlation between chemical composition of an epitope, the resultant affinity of the early primary humoral response, and its eventual influence on relative immunogenicity could be experimentally verified. This was achieved by examining the effect of various amino-terminal substitutions on immunogenicity of a, hitherto cryptic, amino-terminal determinant. Such experiments permitted delineation of a hierarchy of individual amino acid residues based on their influence; which correlated well with calculated Gibbs-free energy changes that individual residue side chains were expected to contribute in a binding interaction. Thus, maturation of a T-dependent humoral response is initiated by a step that is under thermodynamic control.