Low Ligand Density Research Articles

Apparent split tolerance of CD8+ T cells from beta 2-microglobulin-deficient (beta 2m-/-) mice to syngeneic beta 2m+/+ cells.

beta 2-microglobulin-deficient (beta 2m-/-) mice express reduced levels of MHC class I molecules and, consequently, have impaired positive selection of CD8+ T lymphocytes in the thymus. However, small numbers of CD8+ CTLs can be found in beta 2m-/- mice after immunization with allogeneic as well as syngeneic beta 2m+/+ tumor or spleen cells. It has been proposed, therefore, that because of the low ligand density in beta 2m-/- mice, negative selection does not remove cells capable of recognizing syngeneic MHC class I expressed at normal levels. We report here that beta 2m-/- CD8+ T cells are partially tolerant to syngeneic beta 2m+/+ cells. Despite the ability of beta 2m-/- mice to raise CD8+ CTLs against syngeneic beta 2m+/+ cells, these CD8+ cells do not proliferate and do not secrete IFN-gamma or IL-3/granulocyte-macrophage-CSF upon in vitro stimulation with syngeneic beta 2m+/+ cells. In contrast, all of these cellular responses are displayed by the beta 2m-/- CD8+ cells upon recognition of the allogeneic MHC class I. These in vitro findings of partial responsiveness to syngeneic and of full responsiveness to allogeneic MHC class I correlate well with the ability of beta 2m-/- mice to reject allogeneic, but not syngeneic, tumors in vivo. It appears, thus, that the significantly reduced levels of MHC class I molecules found in beta 2m-/- mice, although not capable of inducing deletion of all reactive clones, can induce deletion of high affinity clones and, therefore, maintain tolerance to self-MHC class I, even when expressed at much higher (beta 2m+/+) levels.

Kinetics of cell detachment: effect of ligand density.

Cell adhesion to substratum is often mediated by binding between cell surface receptors and substrate ligands. Substrates can be derivatized with different types and densities of ligands, but how substrate chemistry determines cellular function, such as adhesion strength, has not been demonstrated quantitatively. We employ a numerical methodology developed by Dembo and colleagues (9), who investigated membrane peeling under conditions of excess ligand density, to investigate the kinetics and strength of cell peeling from ligand coated surfaces for arbitrary ligand density. We show there are two asymptotic limits to peeling strength, as quantified by the critical tension: a high ligand density limit, where the critical tension is independent of ligand density and depends logarithmically on the receptor density; and a low ligand density limit, in which the critical tension depends logarithmically on the ligand density but is independent of receptor density. In between these limits, we numerically determine the critical tension. The critical tension is always a weak function of the dissociation constant between ligand and receptor. Furthermore, we show how the rate of peeling, for tensions above the critical tension, depends on ligand density and the mechanical properties of the receptor-ligand bonds. Interestingly, we illustrate when small increases in ligand density should alter cellular behavior, inducing a change to spreading onto a substrate from peeling up from a substrate. In total the predictions of this paper provide criteria for the design of ligand-coated substrate that provide for the proper adhesion strength and dynamics of detachment of cells from surfaces.

The CD8+ T cell repertoire in beta 2-microglobulin-deficient mice is biased towards reactivity against self-major histocompatibility class I.

Beta 2-Microglobulin-deficient (beta 2m -/-) mice are reported to lack cell surface expression of major histocompatibility complex (MHC) class I molecules, CD8+ T cells, and the ability to mount MHC class I-specific T cell responses. We have observed that beta 2m -/- mice possess CD8+ T cells that can be induced to perform strong allospecific cytotoxic responses against nonself-MHC class I by in vivo priming. We report that these beta 2m -/- cytotoxic T lymphocyte (CTL) differ from those induced in beta 2m-positive littermates in that they cross-react and kill cells expressing self-MHC class I at normal ligand density with beta 2m. beta 2m -/- CTL could even be induced in primary mixed lymphocyte culture by self-MHC class I expressing stimulator cells, whereas allogeneic stimulator cells failed to elicit a response under similar conditions. Cells with a reduced cell surface MHC class I expression were less sensitive, while syngeneic beta 2m -/- cells were resistant to the beta 2m -/- CTL. This antiself-MHC reactivity could not be induced when beta 2m -/- T cells matured in an environment with normal MHC class I expression in bone marrow chimeric mice. Antiself-MHC reactivity was also observed against human peptide loading-deficient cells expressing the appropriate murine class I molecules, suggesting that affinity to self-MHC class I may occur irrespective of peptide content. The results fit with a model where positive and negative selection of CD8+ T cells in beta 2m -/- mice is mediated by low levels of MHC class I free heavy chains. In this model, low ligand density on selecting cells leads to positive selection of rare T cells that bind to low levels of MHC class I free heavy chains, resulting in a very small peripheral CD8+ compartment. Due to low density of the selecting ligand, negative selection does not remove T cells recognizing beta 2m-positive cells expressing self-MHC class I at normal ligand density, which generates a T cell repertoire that would be autoreactive in a beta 2m-positive littermate. The first "MHC deficient" animals thus paradoxically provide a tool for direct demonstration and analysis of self MHC bias in the T cell repertoire.

T cell receptor antagonist peptides induce positive selection

We have used organ culture of fetal thymic lobes from T cell receptor (TCR) transgenic β 2 M(−/−) mice to study the role of peptides in positive selection. The TCR used was from a CD8 + T cell specific for ovalbumin 257–264 in the context of K b. Several peptides with the ability to induce positive selection were identified. These peptide-selected thymocytes have the same phenotype as mature CD8 + T cells and can respond to antigen. Those peptides with the ability to induce positive selection were all variants of the antigenic peptide and were identified as TCR antagonist peptides for this receptor. One peptide tested, E1, induced positive selection on the β 2 M(−/−) background but negative selection on the β 2 M(+/−) background. These results show that the process of positive selection is exquisitely peptide specific and sensitive to extremely low ligand density and support the notion that low efficacy ligands mediate positive selection.

A theoretical analysis for the effect of focal contact formation on cell-substrate attachment strength

For many cell types, growth, differentiation, and motility are dependent on receptor-mediated adhesion to ligand-coated surfaces. Focal contacts are strong, specialized, adhesive connections between cell and substrate in which receptors aggregate and connect extracellular ligand to intracellular cytoskeletal molecules. In this paper, we present a mathematical model to examine how focal contact formation affects cellular adhesive strength. To calculate adhesive strength with and without focal contacts, we use a one-dimensional tape peeling analysis to determine the critical tension necessary to peel the membrane. Receptor-ligand bonds are modeled as adhesive springs. In the absence of focal contacts, we derive analytic expressions for the critical tension at low and high ligand densities and show how membrane morphology affects adhesion. Then, focal contacts are modeled as cytoplasmic nucleation centers which bind adhesion receptors. The extent of adhesive strengthening upon focal contact formation depends on the elastic rigidity of the cytoskeletal connections, which determines the structural integrity of the focal contact itself. We consider two limits to this elasticity, very weak and rigid. Rigid cytoskeletal connections give much greater attachment strengths. The dependence of attachment strength on measurable model parameters is quite different in these two limits, which suggests focal contact structure might be deduced from properly performed adhesion experiments. Finally, we compare our model to the adhesive strengthening response reported for glioma cell adhesion to fibronectin (Lotz et al., 1989. J. Cell Biol. 109:1795-1805). Our model successfully predicts the observed detachment forces at 4 degrees C and yields values for the number of fibronectin receptors per glioma cell and the density of cytoskeletal connection molecules (talin) involved in receptor clusters which are consistent with measurements for other cell types. Comparison of the model with data at 37 degrees C suggests that while cytoskeletal cross-linking and clustering of fibronectin receptors significantly increases adhesion strength, specific glioma cell-substratum attachment sites possess little mechanical rigidity and detach through a peeling mechanism, consistent with the view that these sites of < or = 15 nm cell-substrate separation are precursors to fully formed, elastically rigid focal contacts.

Influence of Ligand Density on the Properties of Metal-Chelate Affinity Supports

A new procedure has been developed to immobilize iminodiacetic acid (IDA) onto the surface of silica supports, such as LiChrospher Si-1000 and 1.5 μm nonporous silica, for use in high-performance immobilized metal affinity chromatography (HPIMAC) of proteins. This IDA immobilization method has been achieved through the synthesis of a new silylation reagent, 1-(iminodiacetic acid di- tert-butylester)-3-glycidoxy-propyltrimethoxysilane (IDA-silane). Various modified silicas of different ligand densities have been prepared by using mixtures between 1 and 100% of the IDA-silane diluted with the corresponding 3-glycidoxy-propyltrimethoxysilane (GLYMO-silane). Frontal analysis was used with the IDA-Cu(II)-Concanavalin-A and IDA-Cu(II)-lysozyme systems to evaluate the capacity and the association constants for these HPIMAC sorbents. With these metal chelate sorbents the specific binding capacity per unit area increased continuously with the ligand density for the nonporous sorbents but reached a maximum at about 50% of the maximum ligand coverage for the porous sorbents. The association constant for the chelate-protein complex was highest for both concanavalin-A (Con-A) and lysozyme (HEWL) at the highest ligand density and decreased with lower ligand density. These observations have been evaluated in terms of the accessibility of histidine residues on the surface of the two test proteins and their ability to act as binding sites for the copper ions in the coordination complex. The experimental data indicate that both steric and conformational effects result in multiple classes of binding phenomena with Con-A and HEWL at high ligand concentrations. These experimental results provide a useful guideline for the design of silica-based sorbents for application in the HPIMAC of proteins.

Effects of stationary phase ligand density on high-performance ion-exchange chromatography of proteins

Cation-exchange matrices with ligand densities from 10 to 500 μmol/g were prepared by reaction of diglycolic anydride with diol-bonded silica. Lysozyme and cytochrome c were isocratically eluted from these columns under various conditions. The data was used to examine a retention model of proteins in which the slope ( Z number) of a plot of log k′ vs. log (1/sodium ion activity) was assumed to represent the number of points of binding between the protein and the matrix. As expected from the model, the Z number increased as the ligand density of the matrix increased. However, many qualitative and quantitative deviations from the model were also found, some of which may have been due to heterogeneity of the distribution of the ion-exchange sites on the matrix. An interesting observation was a change in elution order of lysozyme and cytochrome c as the ligand density changed; however, this may not have much practical benefit dual to the large band-broadening observed at low ligand densities.

Selective Activation of Th1‐ and Th2‐like Cells in vivo– Response to Human Collagen IV

Mice immunized with human collagen IV develop either antibody responses or T-cell proliferative responses as a function of the MHC genotype of the immunized mice. CD4+ T cells, similar to Th1 and Th2 cells, participate in these two types of responses, with CD4+ T-cell proliferative responses associating with IL-2 and IFN gamma release, and antibody production associating with CD4+ T-cell IL-4 and IL-5 release. Thus it would appear that the same antigen can induce responses consistent with either cell-mediated or humoral immunity depending on MHC class II genotype. In attempting to understand how MHC genotype controls the class of immunity observed several models are discussed. It was proposed, based on results obtained upon priming with human collagen IV, that the activation of Th1 and Th2 responses may be regulated at several levels (presentation by different APCs, presentation of different densities of the T-cell receptor ligand and presentation of different T-cell epitopes). With the identification of the peptide recognized by the CD4+ T cells, it was clear that the inability to induce Th1 responses in H-2b and the inability to induce Th2 responses in H-2s could not be accounted for by the failure to generate an immunodominant peptide during processing or the failure of the peptides to bind to the MHC class II molecules. Furthermore, the difference in the type of response generated could not be explained by the use of different peptides of the human collagen IV molecule in the two mouse strains, as a single peptide will induce both types of CD4+ T-cell response. However, it cannot be ruled out that the a2 peptide-class II interaction forms different T-cell ligands in the two strains either because the two class II MHC molecules are different or that the peptide is processed and reveals a different antigenic activity (Fox et al. 1988). Perhaps the most important finding from the peptide studies is that the lack of proliferative response in H-2b mice is not absolute, but can be overcome either by priming with high doses of the a2 peptide or by increasing the amount of peptide needed to elicit a recall response. It seems reasonable to speculate that changes in the dose required for priming Th1 or Th2 responses may reflect differences in the activation requirements of the two types of cells with Th1 cells requiring a high ligand density and Th2 cells a low ligand density.(ABSTRACT TRUNCATED AT 400 WORDS)

Studies on the interaction of a surface-bound ligand with a multi-valent high-molecular-weight ligate : The biotinycellulose—avidin system

Two types of biotinylated cellulose disks were examined: filter-paper disks to which biotin had been covalently attached directly to the paper surface (biotinylcellulose) and disks on which biotin was attached to polyacrylamide side-chains grafted onto the filter-paper surface (biotinylpolyacrylamide—cellulose). The amount of avidin taken up from solution by these disks was linearly related to the avidin input concentration and could be estimated by exposure to [ 14C]biotin. The avidin-binding capacity of the disks depended on the surface density of covalently attached ligand and exhibited hyperbolic, Langmuir-type behaviour for both types of disks. The [ 14C]biotin binding capacity of avidinylated disks, on the other hand, showed anomalous, biphasic behaviour: at higher ligand densities, a decrease in [ 14C]biotin binding was observed. The largest anomalies were obtained with biotinylpolyacrylamide—cellulose disks. Calculated ratios of bound [ 14C]biotin vs. amount of avidin tetramer (B/A 4) showed a similar biphasic behaviour. A constant value of B/A 4 = 3 was obtained at low ligand densities, whereas B/A 4 decreased monotonously with increasing ligand density and asymptotically approached B/A 4 = 1. The data could be explained by assuming that at high ligand densities tetrameric avidin interacts with more than one surface-bound biotin residue.

Protein immobilization on silica supports : A ligand density study

The immobilization of proteins on diol-bonded silica matrices containing carboxyl groups (spacer arms) was studied. It was found that the activated ester coupling method worked best with N,N′-dicyclohexylcarbodiimide as the condensing agent in the activation step. During protein coupling, the amount of protein immobilized was highest below pH 6. The optimum pore size of the silica was 300–1000 Å. The spacer arm ligand density was varied over as much as a 100-fold range and the effects on the total activities and specific activities of several proteins were studied. Two proteins exhibited up to two-fold increases in specific activity at low ligand densities. However, the total amounts of activity and protein immobilized decreased at low ligand densities.

Stationary phase contributions to retention in high-performance anion-exchange protein chromatography: ligand density and mixed mode effects

Several anion-exchange stationary phases (based on polyethyleneimine-coated silica) were synthesized so as to vary in ligand density and hydrophobicity. These materials were first examined for hemoglobin-binding capacity and then evaluated chromatographically. Protein binding, retention and resolution increased concomitantly with ligand density. Ferritin (molecular weight 440,000) could not be eluted from the more highly-charged surfaces, but was desorbed from a low ligand density support. The above parameters also varied with the hydrophobic character of the stationary phase. Retention and resolution increased as more hydrophobic moieties were addd. Data from a non-ionic hemoglobic-binding assay correlated reasonably well with anticipated matrix hydrophobicities. Possible explanations and applications of the observed phenomena are discussed.

Micropreparative protein purification by reversed-phase high-performance liquid chromatography

Micropreparative purification of the four subunits of phosphorylase kinase (molecular weights 16,680, 43,000, 113,000 and 132,000) by reversed-phase high-performance liquid chromatography has provided quantities sufficient for some of the first structural studies of these proteins. The best yield from a 25 × 0.46 cm I.D. column was obtained on a packing material (5μm C 18) bonded under conditions resulting in a relatively low ligand density; of a total 250 μg of protein applied, 76% was recovered. Low recoveries from 5–30-μg sample loads suggest partial irreversible adsorption. Retention of protein by the column after an initial micropreparative separation adversely affects resolution and recovery on subsequent separations. Incorporation of gradient wash steps between sample injections was necessary to maintain column performance and to prolong usable column lifetime. The use of a short column with a large diameter (6 × 1 cm) gave increased loading capacity above 400μg protein and enhanced recovery while maintaining good resolution. A less expensive reversed-phase support (17-μm C 18) provided adequate resolution for this separation.

Aromatic interaction chromatography : Acriflavin gel for separation of nucleotides, oligonucleotides and nucleic acids

Using aromatic ligands anchored on insoluble polysaccharide matrices, it is possible to separate many classes of synthetic and naturally occurring substances according to their aromaticity, hydrophobicity or electrostatic character. The strength of the interactions, mainly charge-transfer interactions which occur between ligand and solute, are functions of (i) the electron donor-acceptor properties of the ligand (type, number and electron-attracting or -releasing substituents on the aromatic ring), (ii) the coupling method, (iii) the degree of substitution and (iv) the adsorption and desorption conditions of the solute sample. Different applications of this type of chromatography to the separation of nucleotides, oligonucleotides and nucleic acids are presented. With acriflavin-dextran, methods have been developed for measuring cyclic AMP in cells pre-labelled with radioactive adenine and a rapid direct assay of 3′,5′-cyclic nucleotide phosphodiesterase activity. With highly substituted acriflavin-agarose it is possible to separate a series of oligonucleotides. The low acriflavin ligand density on agarose permits the separation of single-stranded from double-stranded nucleic acids.