Antigenic peptide binding to MHC class II molecules at increased peptide concentrations.

Affinity-purified major histocompatibility complex (MHC) class II molecules are known to bind antigenic peptides in vitro. The percentage of MHC class II molecules occupied with such peptides is usually very low and varies significantly depending upon the sequence and size of a given antigenic peptide. The present study describes a method by which complete saturation of affinity-purified MHC class II with antigenic peptide can be achieved by simply incubating purified MHC class II molecules at neutral pH in the presence of several 100-fold molar excess of antigenic peptide. Complexes of human HLA-DR2 and a peptide analog from human myelin basic protein MBP (83-102)Y83 were selected for this study. The on-rate kinetic results showed saturation of MHC class II occupancy at 300-500-fold molar excess peptide concentrations. The specificity of the MBP (83-102)Y83 peptide binding to HLA-DR2 at higher peptide concentration was demonstrated by incubating an equivalent amount of another epitope from myelin basic protein [MBP (1-14) peptide] as well as by competitive binding assays. The quantitation of bound peptide was carried out using biotinylated-MBP (83-102)Y83 peptide which showed 100-125% occupancy of HLA-DR2 with a recovery of 100%. The presence of a single peptide entity in purified complexes was confirmed by reverse-phase narrowbore HPLC analysis of the acid extracted supernatant and by mass spectrometry analysis. Two-dimensional gel electrophoresis (IEF/SDS) of purified HLA-DR2 and DR2.MBP (83-102)Y83 complexes showed the absence of various endogenous polypeptides in 100% loaded complexes. These results demonstrate that higher peptide concentrations can be useful in generating MHC class II-peptide complexes of defined composition. Such complexes of MHC class II occupied with a single peptide may have significant clinical relevance for antigen-specific therapy of various autoimmune diseases and may provide better understanding of MHC-peptide-TCR interactions.

Role of tumor necrosis factor species in the sequence-dependent effects of interleukin-2--tumor necrosis factor immunotherapy in mice: implications for preclinical cytokine testing.

We previously demonstrated that recombinant human tumor necrosis factor (hTNF) is synergistic with human interleukin-2 (hIL-2) for in vivo regression of murine tumors. In mice, the timing of cytokine administration is critical in achieving synergy. Because hTNF exhibits negligible binding to the type II murine TNF receptor (TNF-R), we questioned whether murine TNF (mTNF) would have therapeutic benefits, scheduling requirements, and toxic effects similar to those of the hIL-2-hTNF combination. To evaluate the biological effects of TNF-R types I and II interaction, we directly compared the effects of mTNF and hTNF in combination with hIL-2 on in vivo tumor regression and in vitro activation of murine splenocytes. Our results demonstrate for the first time that (a) the cytokine combination hTNF-hIL-2 is consistently more efficacious than mTNF-hIL-2 in in vivo murine immunotherapy models; (b) the in vivo antitumor effects of hTNF-hIL-2 and not mTNF-hIL-2 are critically dependent upon cytokine scheduling; and (c) in vitro culture of splenocytes with mTNF-hIL-2 enhances cellular proliferation, Lyt 2, and TNF-RI expression compared with hTNF-hIL-2. Collectively, these studies extend the previous findings of species-specific mTNF-R responses and reveal that optimal scheduling and efficacy of TNF-hIL-2 combination therapy in murine tumor models is critically dependent upon the TNF species employed.

Human monoclonal antibodies to glycolipid A that exhibit complement species-specific effector functions.

Two human IgM monoclonal anti-glycolipid A antibodies (MAbs) were evaluated for their abilities to bind to various endotoxins and pathogenic gram-negative bacteria and to activate complement pathways, thereby accomplishing bactericidal and opsonic effector functions. Both MAbs cross-reacted with glycolipid A mutant lipopolysaccharides from rough colony-forming gram-negative bacteria and with selected endotoxins from smooth colony-forming bacteria. However, MAb 10235 bound to all clinical isolates of Escherichia coli and Klebsiella pneumoniae tested but only very weakly to Pseudomonas aeruginosa, whereas MAb 10058 bound to all three genera. Several strains of serum-insensitive organisms were selected for evaluation of antigen-specific, complement-mediated effector functions for the two MAbs. Assessment of bactericidal and opsonic activities showed that neither MAb was able to activate complement from nonprimate species (mouse, rat, rabbit, guinea pig, or sheep). However, both MAbs were highly effective in using primate sources of serum complement to mediate these effector functions.

Effect of production method on the systemic clearance rate of a human monoclonal antibody in the rat.

Pharmacologic studies of human immunoglobulins (IgM) in non-primate animal models, whether directed toward efficacy or toxicity, rely on pharmacokinetic parameters to achieve optimal doses and schedules. In rodents, human IgMs have an effective circulatory half-life of 11 h and a plasma clearance rate of 0.044 ml/min/kg. Studies of a new group of human monoclonal antibodies (hMAb) specific for Gram-negative bacteria and endotoxin revealed an IgM molecule, hMAb-10058, which, when purified from tissue culture medium, exhibited a suprisingly short circulatory lifetime in rodents. Investigations into possible explanations for this short circulatory half-life resulted in the development of a simple and efficient method for producing hMAbs in the immunodeficient NIH-3 mouse (bg x nu x XID). This method of production of hMAb-10058 had dramatic effects on its half-life. Whereas hMAb-10058 produced in serum-free, defined medium had a clearance rate of 14.4 ml/min/kg and an effective half-life of 0.12 h, the same hMAb-10058 raised in mouse ascites had a decreased clearance rate of 0.092 ml/min/kg and an increased effective half-life of 12 h. This 100-fold enhancement of the hMAb's half-life was not affected by the purification process. Some potential molecular structures involved in the circulatory half-life of this hMAb are discussed.

Contrast-enhanced magnetic resonance imaging of tumor-bearing mice treated with human recombinant tumor necrosis factor alpha.

Pharmacological effects of recombinant human tumor necrosis factor alpha (TNF) were studied in a mouse fibrosarcoma model using magnetic resonance imaging enhanced with a macromolecular contrast agent, albumin(gadolinium-diethylenetriamine pentaacetic acid)35. TNF was administered i.v. in a dose of 150 micrograms/kg, 60 to 80 min prior to imaging. Contrast-enhanced and nonenhanced magnetic resonance images of TNF-treated (n = 10) and untreated (n = 8) Meth A fibrosarcomas were obtained at 2.0 Tesla using T1-weighted spin-echo pulse sequences. Serial images spanning an interval of 60 to 120 min after TNF administration showed that the TNF-treated tumors enhanced significantly more overall than did untreated tumors (43% versus 31%). The most marked differential tumor enhancement was observed in the tumor rim (59% versus 40%). Nontumorous tissue, including muscle and brain, revealed no significant enhancement differences between TNF-treated animals and controls. The observed tumor enhancement corresponded strongly with Evans blue staining; the TNF-treated tumors stained deep blue, while untreated tumors and normal tissues observed did not stain. The different enhancement and Evans blue staining patterns between TNF-treated tumors and untreated tumors are attributed to TNF-induced changes in tumor capillary integrity. The data indicate that TNF effects on tumors include an increased capillary permeability for macromolecules at early times after administration. The ability to detect changes in capillary permeability in vivo using contrast-enhanced magnetic resonance imaging may prove to be clinically useful to monitor tumor response to TNF.

Schedule dependency of the antitumor activity and toxicity of polyethylene glycol-modified interleukin 2 in murine tumor models.

Modification of recombinant human interleukin 2 (rhIL-2) with monomethoxy polyethylene glycol has been shown to alter its pharmacokinetic properties. Therefore, we investigated the pharmacological parameters of schedule and dose in order to assess the impact on the in vivo antitumor activity of this modification. The antitumor efficacy, as well as the toxicity, of polyethylene glycol-interleukin 2 (PEG-IL-2) was compared to that of rhIL-2 in three transplantable syngeneic murine tumor models, Meth A fibrosarcoma, B16 melanoma, and Pan-02 pancreatic carcinoma. At equitoxic dose levels, the antitumor activity of PEG-IL-2 was far superior to that of rhIL-2 in all three tumor models. This efficacy of PEG-IL-2 was dose dependent and was greatest on a Q7D x 2 schedule in Meth A and B16. When the same total doses were further divided and delivered on any of several alternative schedules, either the efficacy was reduced or the toxicity of the treatments was increased. In Pan-02, a rhIL-2-resistant tumor, PEG-IL-2 treatment on either the Q7D x 2, Q4D x 3, or Q3D x 4 schedule resulted in approximately a 200% increase in lifespan; however, the toxicity of the treatment increased as the interval between doses was shortened. Simulations of the pharmacokinetic profiles of these various regimens suggested that the toxicity of PEG-IL-2 and rhIL-2 was related to the minimum plasma concentration that was obtained and the time interval between peak levels. The efficacy of the treatment was associated with the interleukin 2 plasma peak height, since a dose response was observed; however, peak plasma concentration did not appear to be the only parameter which determined efficacy. We hypothesize that this observed schedule dependence is also affected by the kinetics of the host's biological response to rhIL-2.

Contrast-enhanced MRI of tumors. Comparison of Gd-DTPA and a macromolecular agent.

The study aim was to define potential differences and advantages in magnetic resonance (MR) patterns of tumoral contrast enhancement using either a small molecular, extracellular fluid contrast enhancer [Gd-DTPA] or a macromolecular agent [albumin-(Gd-DTPA)20], designed for primary intravascular biodistribution. MR images of 25 mice with implanted fibrosarcomas were obtained before and repeatedly for up to 120 minutes after injection of either Gd-DTPA [0.2 mmol/kg, n = 11] or albumin-(Gd-DTPA) [0.0029 mmol/kg, n = 14]. Histologically, this hypovascular tumor contained zones of viable tissue and non-viable, necrotic tissue. Using either type of contrast media, the viable portions enhanced strongly, up to 152% and the necrotic portions enhanced poorly, less than 31%. However, the time-course of enhancement differed between contrast agents. Gd-DTPA tended to provide maximal enhancement soon after administration with no significant changes over two hours. Enhancement from albumin-(Gd-DTPA) was weak initially, corresponding to tumor hypovascularity, but over two hours the signal of the viable tumor zones progressively increased in intensity. This gradual tumoral accumulation of the macromolecular agent within the tumor was considered to reflect abnormal capillary permeability, associated with neovascularity. Thus, the increasing intensity within the neoplastic tissues over time, reflecting abnormal capillary permeability for macromolecules, may serve as a useful, albeit indirect, marker of neoplasia.

The role of oxidant injury in tumor cell sensitivity to recombinant human tumor necrosis factor in vivo. Implications for mechanisms of action.

The intracellular glutathione levels of two human tumor lines and seven murine tumor lines were determined in order to investigate the role of oxidant injury in tumor cell sensitivity to human rTNF (rhTNF). Correlations were found between high intracellular glutathione levels and in vivo tumor resistance to rhTNF, and on the other hand, low glutathione levels and rhTNF sensitivity. The transplantable murine fibrosarcoma, Meth A, a TNF-sensitive line in vivo, was less sensitive to rhTNF and host toxicity was reduced when the hosts were pretreated with uric acid, a major reactive oxygen scavenger in humans and certain other primates. Conversely, pretreatment of the tumor-bearing hosts with DL-buthionine-(S,R)-sulfoximine, an inhibitor of GSH biosynthesis, resulted in an increased sensitivity of Meth A to rhTNF. This effect was not limited to tumor-bearing mice, as rats pretreated with diethyl maleate, a compound which irreversibly binds glutathione, were more sensitive to rhTNF toxicity than control rats. On the other hand, pretreatment with N-acetyl cysteine, an oxidant scavenger, reduced the toxicity of rhTNF treatment in rats. The data are consistent with the hypothesis that tumor cell sensitivity to rhTNF in vivo is dependent on its capacity to buffer oxidative attack. In addition, host toxicity is also related to the production of reactive oxygen species. Activated effector cells such as granulocytes and macrophages are hypothesized to produce most of this damage by their respiratory burst and oxidant release, although the direct action of rhTNF may also contribute to oxidative injury in vivo.

Sequence dependence of administration of human recombinant tumor necrosis factor and interleukin-2 in murine tumor therapy.

Simultaneous administration of recombinant human tumor necrosis factor (rhTNF) and interleukin-2 (rhIL-2) has been shown to block tumor take in murine models. We investigated the effects of sequence and schedule of administration as a function of tumor burden with two tumor models (B16 and Meth A). rhTNF followed by rhIL-2 had extraordinary antitumor efficacy, but rhIL-2 followed by rhTNF was much less effective. Sequential rhTNF/rhIL-2 therapy resulted in complete tumor regression, whereas simultaneous therapy resulted in complete tumor regression, whereas simultaneous therapy resulted in only reduced growth rate. Experiments with genetically immunodeficient mice suggested that T cell factors may be required for synergistic antitumor activity.

Synergistic effects of combination therapy with human recombinant interleukin-2 and tumor necrosis factor in murine tumor models.

Human recombinant interleukin 2 (IL-2) and tumor necrosis factor (TNF) were evaluated individually and in combination for their antitumor efficacy in vivo, using five s.c. murine tumors: L1210 leukemia, P815 mastocytoma, B16 melanoma, EL-4 lymphoma, and the methylcholanthrene-induced sarcoma, Meth A. While only the s.c. methylcholanthrene-induced tumor exhibited regression and/or cures in response to immunomodulatory therapy with either agent alone, the simultaneous administration of a maximally tolerated dose of TNF and IL-2 given daily from within 1 day (B16 melanoma), 3 days (L1210 leukemia and P815 mastocytoma) or 5 and 10 days (EL-4 lymphoma and methylcholanthrene-induced sarcoma) after tumor cell implant resulted in no tumor takes (growth). The TNF dose was apparently rate limiting in that reduction of the amount of TNF in the combination by 50% resulted in the loss of curative effects, while IL-2 doses could be reduced by 90% (depending upon tumor type) and still result in an efficacious combination. The synergy seen in combination IL-2 and TNF therapy appeared to be dependent upon tumor burden, but somewhat independent of tumor location. For example, no tumors were seen in the artificial pulmonary metastasis model of the B16 melanoma, and the percentage of extension of median lifetime (test versus control) greater than 150% was seen in the i.p. B16 melanoma, as well as several other i.p. models of the five tumor types. On the other hand, no significant extension of lifetime (greater than 150%) was seen with either lymphokine alone when administered i.p. at maximally tolerated dose for any of the five tumors tested here. Results are discussed in relation to potential immune modulatory events which may be occurring during combination treatments.

Antitumor activity of an immunotoxin in a nude mouse model of human ovarian cancer.

An immunotoxin composed of an antibody to the human transferrin receptor (454A12) and ricin A chain (RTA) was shown to inhibit the growth of NIH:OVCAR-3 tumors in a nude mouse model of human ovarian cancer. Inhibition of tumor growth by 454A12-RTA was related to the dose administered. The antitumor activity of the immunotoxin was blocked by coinjection of excess antibody with immunotoxin. An immunotoxin made using 454A12 and recombinant ricin A chain (rRTA) had an activity similar to that made with native RTA. The administration of 10 micrograms or greater of the immunotoxin 454A12-RTA/rRTA had significant antitumor activity. The injection of 30 micrograms of an irrelevant immunotoxin, MOPC21-RTA, or 30 to 500 micrograms of the 454A12 antibody had no antitumor activity.

Fc:Fc interactions revealed by spin-labeled IgG heterosaccharides in model immune complexes.

Dynamic properties of spin-labelled heterosaccharides in the Fc-region of murine monoclonal antihapten immunoglobulin G were studied in model immune complexes (IC) as a function of the IC size. Model IC dimers, trimers and oligomers were formed using bivalent photoaffinity antigens. The ESR spectrum exhibits two components. The rotational correlation time of the less-immobilized species is shorter than 10(-10) sec, and that of the more-immobilized component is in the order to 10(-9) approximately 10(-8) sec depending on the IC size. Fraction of the more-immobilized spin labels increases, and the mobility of this component decreases with increase in IC size (i.e., mobility: monomers approximately equal to dimers greater than trimers much greater than immune-complex precipitates). These data strongly suggest the existence of Fc:Fc interactions in IC, and provide the basis for a model in which such interactions underlie the initial mechanism by which the information of antigen binding to Fab region is transferred into organized Fc:Fc association structure for IgG effector activities.

Evolution of antibody structure and effector functions: comparative hemolytic activities of monomeric and tetrameric IgM from rainbow trout, Salmo gairdnerii.

Monomeric and tetrameric IgM anti-haptin antibodies isolated from the sera of rainbow trout (S. gairdnerii) by immunoaffinity chromatography were compared both immunochemically and with regard to their functional abilities to lyse haptenated trout erythrocytes in the presence of trout complement. The two populations had similar binding affinities for hapten and apparently identical L chains, but differed in H chain peptide maps and immunoreactivity with rabbit anti-trout H chain serum. These differences could not be attributed to J-chain. The abilities of the two antibody subpopulations to activate C to lyse haptenated trout erythrocytes also differed dramatically. Such functional differences are not simply explained by the greater avidity of the tetrameric form since preliminary studies show that the monomeric form of trout IgM activates C via an alternative pathway mechanism while the tetrameric form activates both classical and alternative pathway mechanisms. Results suggest divergent evolution of antibody structures involved in the familiar effector functions (C activation, transport, etc.).

Human aglycosyl-IgG exhibits increased hydrophobicity. Binding/fluorescence studies with 8-anilinonaphthalene-1-sulfonic acid (ANS).

Human monoclonal, aglycosyl-IgG produced in vitro in the presence of tunicamycin, was compared with its native and acid pH-altered counterparts for their respective abilities to bind the fluorescent hydrophobicity probe, 8-anilinonaphthalene sulfonate. A novel technique based on continuous-flow dynamic dialysis (Sparrow et al., 1982, Anal. Biochem. 123:255-264) allowed binding studies under non-equilibrium conditions. While the native IgG conformation exhibits two, weak ANS binding sites (ca. 10(3) l/mol), aglycosyl-IgG has one weak and one moderate affinity (least squares average Ka = 2 X 10(4) l/mol) site, and the acid conformer binds yet another two ANS molecules with moderate affinity (4 X 10(4) l/mol). Increases in affinity and in the number of sites correlate roughly with increased relative percent fluorescence by conventional fluorimetry. The fluorescence lifetime of ANS bound to altered IgGs is about 10% longer (T2 = 15 nsec by time-resolved fluorimetry) than that for native IgG. All populations also exhibit a rapid decay component (T1 = 3 nsec) analogous to that seen for ANS in 50% aqueous dioxane. Results are discussed in relation to structural role(s) for IgG-linked heterosaccharides.

Complement C1q binding affects spin-labeled heterosaccharides of rabbit antibodies in immune but not artificial immunoglobulin G aggregates.

IgG anti-hapten antibodies were purified from the sera of rabbits homozygous for allotypic determinants d11 and d12 in the constant region of the heavy chain. Correlative with this determinant is the absence (d11) or presence (d12) of an oligosaccharide chain just below the hinge region of the IgG molecule. Both d11 and d12 molecules contain a complex heterosaccharide chain located near the carboxyl terminus of the second constant region domain. The two populations of IgG antibodies were thus selectively labeled with the spin probe Tempamine in their second constant region domains by reductive amination primarily of terminal N-acetylneuraminic acid residues. Chemical and enzymatic cleavages showed about 80% of the attached spin labels were N-acetylneuraminic acid-associated. Analysis of probe adducts by ESR spectrometry showed the presence of slower and faster moving subcomponents. Formation of immune complexes by antigen induces slight but significant restrictions of spin label mobility for both d11 and d12 IgG molecules. This restriction is qualitatively different from that seen in glutaraldehyde-, carbodiimide-, or ethanol-induced aggregates of the same IgG antibodies. The addition of purified complement C1 subcomponent C1q to immune aggregates resulted in marked immobilization of spin labels, the rotational correlation time of which was 30-40 mu s for both d11 and d12 molecules (evaluated by saturation transfer spectroscopy). A similar spin probe immobilizing effect is not seen when C1q binds to chemically aggregated IgG antibodies (which also do not activate C1). A novel model is proposed in which C1q is hypothesized to juxtapose Fc moieties in a discrete fashion required for subsequent C1 activation processes mediated by immune complexes.

Effects of hyperthermia on plasma glycoprotein catabolism by the isolated perfused rat liver.

Clearance and degradation of the glycoprotein, asialofetuin (AF), by the isolated perfused rat liver at supranormal temperatures were investigated. The half-life for disappearance of AF was similar at 37, 41, and 42 degrees C, P greater than 0.05. There was a significant difference between the amount of hydrolysis of AF at 37, 41, and 42 degrees C, P less than 0.05. This indicates that there was significant retardation of lysosomal proteolysis or receptor endocytosis by the hepatocyte at elevated temperatures.

Induction in rainbow trout of an acute phase (C-reactive) protein by chemicals of environmental concern.

1. Rainbow trout, Salmo gairdneri, produce elevated amounts of a serum acute phase (C-reactive) protein (CRP) when administered a variety of chemicals of environmental importance. 2. Compounds administered in doses which induce the cytochrome(s) P450 catalytic enzymes in trout hepatic microsomes also induce serum CRP. 3. However, an interferon-inducing virus does not induce CRP. Interferon induction by the virus is not significantly inhibited by chemicals which induce trout cytochrome(s) P450. 4. Simultaneous administration of chemicals and virus or virus alone results in depression of P450 protein production and only minor induction of CRP. 5. Thus, as with mammals, a reciprocating relationship appears to exist between the hemeprotein monooxygenase and immune systems of this freshwater teleost, and C-reactive protein appears to fit the reciprocating scheme closer to the cytochromes P450 response.

Acute phase (C-reactive) protein-like macromolecules from rainbow trout (Salmo gairdneri).

A protein which reacts with the Cx-polysaccharide of Streptococcus pneumoniae and is inhibited by phosphorylcholine was isolated from the serum of rainbow trout by affinity chromatography. The protein, which exists in monomeric and oligomeric forms in non-immune trout serum, is very similar with regard to specificity and size to the Cx-reactive protein from rabbits. A semi-quantitative analytical method for evaluating bacterial agglutination with an electronic particle counter and size distribution analyzer was developed to compare natural and acute serum levels of trout and rabbit Cx-reactive proteins. Results indicate that the poikilotherm has much higher concentrations in normal serum. The trout serum protein can also be rapidly induced to yet higher levels by both chemical and physical stress. The implications for such a protein in the teleost's natural defense system and overall homeostasis are discussed.