Thioether-bonded constructs of Fab'gamma and Fc gamma modules utilizing differential reduction of interchain disulfide bonds.

We describe a two-stage preparation of chemically engineered Ab constructs, employing as modules Fab'gamma from mAb or rAb, and Fc from human normal IgG1. A multivalent, optionally multispecific F(ab')(n) core is formed in stage one, and one or more Fc modules added in stage two. Examples include bispecific Fab(2)Fc(2) (for simplicity, primes and Greek letters are omitted from names of final constructs) and trivalent Fab(3)Fc(2), which are designed to kill neoplastic cells. An essential element in the construction is the availability of the Fab' in two reduced forms, Fab'(-sulfhydryl (SH))(5) and Fab'-SH. The first is obtained by full reduction of the interchain disulfide bonds (SS) in the F(ab')(2) fragment of IgG. Fab'-SH is obtained by disulfide-interchange reactions on Fab'(-SH)(5), whereby the gamma-light SS is reconstituted, an unusual intrachain SS forms in the gamma-chain hinge, and one hinge SH remains. F(ab')(2) and F(ab')(3) cores are built using partially reduced modules, being given intermodular thioether links that resist reduction. These cores are then fully reduced, making available SH groups for addition of the Fcgamma modules. In the final constructs, all intermodular links embody tandem thioether bonds arising at hinge-region cysteines. Cytotoxic activities of representative constructs, and some enhancements deriving from multiple modules, are assessed. In guinea pigs, catabolism of Fab(2)Fc(2) yielded a t(1/2) similar to that of human IgG1, although the serum Fab(2)Fc(2) revealed some proteolytic breakdown not shown by the IgG1. Immunotherapy of a guinea-pig leukemia confirmed the ability of these constructs to kill target cells in vivo.

Conjugation of human Fc gamma in closed-hinge or open-hinge configuration to Fab'gamma and analogous ligands.

We describe a method for linking human normal Fc gamma1, via stable thioether bonds emerging from its hinge, to any molecule expressing a free sulfhydryl (SH) group. The Fc hinge may be closed by a disulfide (SS) bond or left open. Preparation begins with reduction of the Fc hinge to release four SH groups from its two parallel inter-gamma SS bonds. When the Fc is required in normal closed-hinge configuration, one SH group is alkylated with N-ethylmaleimide under limiting conditions, and one of the inter-gamma SS bonds is reconstituted by SS interchange. The residual SH group, to be used for linking, is left as a 4-dithiopyridyl group suitable for storage. When the Fc is required for conjugation the 4-dithiopyridyl is replaced by a metastable maleimidyl group, which reacts rapidly with SH on the partner molecule to form a tandem thioether link. If the partner is Ab Fab'gamma, linking to cysteines in the Fab'gamma hinge yields derivatives such as FabFc and FabFc2. Chimeric FabFc Abs (mouse Fab'gamma/human Fc gamma1) invoked cellular cytotoxicity in vitro, using human cell lines as targets and human lymphocytes as effectors, whether the Fc hinge was open or closed. The same Abs could kill the same targets by activating human complement, but only when the Fc hinge was closed. Both effector functions were enhanced by the presence of a second Fc in FabFc2. This method of Fc addition can be used to predict the performance of recombinant chimeric Abs and to provide novel molecular geometries.

Tumour cell killing using chemically engineered antibody constructs specific for tumour cells and the complement inhibitor CD59.

Immunotherapy using MoAbs is inefficient due to limited activation of human effectors by mouse antibodies and multiple protective mechanisms available to host cells against autologous complement. We have used chemically engineered antibody constructs and human complement in vitro to specifically target and kill neoplastic B lymphoid cells (Raji). Fab'gamma Fc gamma2 chimaeric antibody (specific for human CD37) was used to activate the classical pathway of human complement on Raji cells, whilst CD59 was neutralized using one of two different bispecific F(ab'gamma)2 antibody constructs which contained both cell-targeting (anti-CD19 or anti-CD38) and CD59-neutralizing moieties. When either bispecific construct was used to neutralize CD59, 15-25% of cells were lysed. If CD55 was also neutralized using specific antibody, Raji cells were efficiently killed (70% lysis). When added to a mixture of target (Raji) and bystander (K562) cells, one bispecific antibody (anti-CD38 x anti-CD59) could be specifically delivered to Raji, avoiding significant uptake on CD59-expressing bystander cells (K562). The second bispecific antibody (anti-CD19 x anti-CD59) bound equally well to either cell type. Cell-specific targeting was dependent upon combination of a low-affinity anti-CD59 Fab'gamma with a high-affinity anti-tumour cell Fab'gamma. When Raji and K562 cells were mixed and incubated with a combination of the engineered constructs and anti-CD55 antibodies, Raji cell lysis (30-40%) was observed in the absence of K562 killing. We propose that combinations of these constructs may be of use for treatments such as ex vivo purging of autologous bone marrow or in vivo targeting of tumour cells.

Linkage of silver to antibodies through 2-imino thiolane.

Earlier studies described the linkage of silver to antibodies using SH groups generated by the reduction of the SS groups using ascorbic acid (1) analogous to the Thakur and DeFulvio technique for linking technetium to antibodies. This work describes the linkage of silver to IgG after introducing SH groups by coupling the IgG to 2-imino thiolane. The protein was dissolved in sodium acetate buffer pH 4.5 containing 1 mM EDTA by dialysis/gel chromatography in a concentration of 20 mg/mL. 2-Imino thiolane dissolved in Tris-HCl acetate buffer, pH 8.2, 0.2M was added to give a final dilution of 0.2 mM 2-imino thiolane. The excess of 2-imino thiolane was removed by dialysis or G-25 Sephadex gel chromatography and then the protein was reacted with silver nitrate 0.1 mM. The unreacted SH groups were blocked by adding iodoacetamide to a concentration of 5 mM. The nonprotein reagents again were removed by dialysis or gel chromatography. The thiol groups were titrated using 1.5 mM 2 2-Py-SS-Py prior to and after addition of silver. It was observed that depending on the concentration of silver, 50-80% of the SH groups were coupled to silver. Higher concentrations of silver led to insoluble precipitates and should be avoided.

Mechanisms in removal of tumor by antibody.

We report two preliminary trials of antibody treatment of B-cell lymphoma. Advanced lymphoma was treated with chimeric FabFc2, in which mouse Fab' gamma is linked to two human Fc gamma 1 fragments so as to recruit natural effectors to tumor targets. Terminal lymphoma was treated with bispecific antibody (BsAb) which recruits the ribosome-inactivating protein saporin. These different mechanisms led to interesting differences in patterns of tumor clearance. Eight patients were treated with chimeric antibody of two specificities, each at 12 mg/kg: anti-CD37, plus either anti-CD38 or anti-CD19 according to tumor phenotype. On completion of the 3-wk treatment, residual plasma antibody had a half-life exceeding 10 d. Tumor cells in blood disappeared rapidly. However significant reductions in solid masses occurred in only three patients, becoming apparent 3-4 wk after beginning treatment and then continuing slowly. Five patients were treated with preformed immune complexes of saporin and F(ab' gamma)2 BsAb. Although doses of saporin reached 10 mg weekly, contact with the tumor can only have been fleeting: plasma antibody was undetectable (< 0.5 micrograms/mL) 48 h after infusion, whereas the saporin disappeared even faster and was undetectable (< 4 ng/mL) at 24 h. Tumor cells disappeared from the blood more slowly than occurred with chimeric antibody. In contrast shrinkage of extravascular tumor was more rapid, and occurred in all patients, but proved less durable.

Menadione increases hepatic tight-junctional permeability. Its effect can be decreased by butylated hydroxytoluene and verapamil.

Infusion of menadione at two different doses [2.7 mg and 5.5 mg in 100 microliters of dimethyl sulphoxide (DMSO)] into perfused rat livers for 30 min caused no or a 6-fold increase respectively in junctional permeability to horseradish peroxidase as compared with controls receiving 100 microliters of DMSO alone. The total glutathione (GSH) contents in these livers measured at the end of the experiments were 115% and 53%, compared with the controls. The free-radical scavenger butylated hydroxytoluene (BHT) (final concn. 5 microM) protected against the GSH depletion caused by the higher dose of menadione and partially decreased the menadione-induced increase in junctional permeability. Verapamil, a Ca2(+)-channel blocker which was added into the perfusion medium (final concn. 40 microM) 10 min before the infusion of 5.5 mg of menadione, completely abolished the effect of menadione on junctional permeability. Menadione exposure therefore increases tight-junctional permeability in the liver; this may involve a depletion of GSH and a subsequent increase in intracellular Ca2+.

Oestradiol 17 beta-glucuronide increases tight-junctional permeability in rat liver.

By using rat liver perfusion under one-pass conditions with a single pulse of horseradish peroxidase (HRP), the biliary output of HRP was used as an indicator of paracellular permeability change caused by the cholestatic compound oestradiol 17 beta-glucuronide (E17G). Since E17G reduced bile flow, we have also used, during the assessment of junctional permeability after E17G treatment, the choleretic compound taurodehydrocholate to enhance bile flow back to control levels. At both low and restored bile flow rates, the acute administration of E17G (3.4 mumol) increased the HRP peak height, thereby indicating that one of the hepatotoxic actions of E17G is to increase the permeability of hepatic tight junctions. The action of E17G in affecting bile acid secretion and biliary volume are also explored.

Retrograde intrabiliary injection of amphipathic materials causes phospholipid secretion into bile. Taurocholate causes phosphatidylcholine secretion, 3-[(3-cholamidopropyl)dimethylammonio]-propane-1-sulphonate (CHAPS) causes mixed phospholipid secretion.

The control of biliary phospholipid and cholesterol secretions by bile acid was studied by using the technique of retrograde intrabiliary injection. Taurocholate (TC), a moderately hydrophobic bile acid, taurodehydrocholate (TDHC), a hydrophilic non-micelle-forming bile acid, and 3-[(3-cholamidopropyl)-dimethylammonio]propane-1-sulphonate (CHAPS), a detergent, were individually administered by retrograde intrabiliary injection (RII) into the biliary tree, and bile acids, phospholipids and cholesterol subsequently appearing in the bile were measured. TC (1.3 mumol; 45 microliters) injected retrogradely provoked a 3.5-fold increase in biliary phospholipid output for 40 min, as compared with the saline control. Injection of 2.7 mumol of TC (90 microliters) caused a 7.5-fold increase in phospholipid output, which reached a peak at 12 min after RII, and phospholipid output continued for 40 min. Cholesterol output was also elicited under these conditions, showing both dose-dependency and extended secretion. Injection of 1.8 mumol of TDHC caused very little increase in either biliary phospholipid or cholesterol. Injection of 0.9 mumol of CHAPS (45 microliters) provoked a single substantial peak of phospholipid output in the 3 min bile sample. T.l.c. analysis of the phospholipid extracts of the bile collected after each compound showed, for TC, a single compound which co-migrated with the phosphatidylcholine standard, whereas for CHAPS substantial amounts of other phospholipids were present.

The calcium ionophore A23187 increases the tight-junctional permeability in rat liver.

The effect of an increase in intracellular Ca2+ concentration on tight-junctional permeability in rat liver was studied by using the calcium ionophore A23187. Infusion of 100 microliters of dimethyl sulphoxide containing various amounts of A23187 over 30 min into isolated perfused livers was followed by a pulse of horseradish peroxidase (HRP) under single-pass conditions. The first biliary HRP peak, a measure of junctional permeability, was increased 4-fold with 100 micrograms of A23187. There were, however, no significant effects on bile flow or on aspartate aminotransferase leakage as compared with the control at this dosage, and thus the increase in junctional permeability was occurring without evidence of appreciable cholestatic or hepatocellular damage. Higher dosages of A23187, however, caused not only an increase in HRP peak height but also changes in bile flow and increases in aminotransferase leakage, indicating more extensive effects at these higher dosages. A second peak of HRP secretion, occurring 20-25 min after the HRP pulse, was also elevated approx. 3.5-fold; this may indicate that pinocytosis and transcellular movement of HRP are also increased under these conditions.

1-Naphthylisothiocyanate-induced permeability of hepatic tight junctions to proteins.

We have studied the early action of 1-naphthylisothiocyanate (ANIT) in relation to its effect on the permeability barrier formed by hepatic tight junctions. Materials having different Mr values [inulin (5000), horseradish peroxidase (HRP) (40,000), ovalbumin (also 40,000) and pig gamma-globulin (IgG) (160,000)] were individually pulsed, within 1 min, into perfused rat livers operating under single-pass conditions. In untreated rats, a small peak of HRP and ovalbumin and a comparatively larger peak of inulin were observed in the bile at 7 min. In rats treated with ANIT, with increasing duration of ANIT treatment the inulin peak increased proportionally, whereas the HRP and ovalbumin peaks remained unchanged until after 10 h of ANIT exposure; gamma-globulin was not detected in the 7 min bile sample until after 14 h of ANIT treatment. Bile flow in all rats remained approximately the same until after 14 h of ANIT pretreatment, when substantial bile-flow reduction was observed. Phenobarbitone pretreatment increased the effect of ANIT and massively elevated the first HRP peak; it also shortened the time (to 4 h) at which the increase in permeability to this protein was observed. In contrast, the first HRP peak was virtually abolished in rats that had received the mixed-function-oxidase inhibitor SKF 525A. These experiments suggest that (i) ANIT progressively increased the permeability of the junctional barrier before the reduction in bile flow, (ii) the ANIT-increased permeability change seems to be inversely dependent upon the Mr of the infused proteins, and (iii) metabolites of ANIT were involved in the development of the junctional permeability change.

Transcytosis and paracellular movements of horseradish peroxidase across liver parenchymal tissue from blood to bile. Effects of alpha-naphthylisothiocyanate and colchicine.

The pathways for the entry of horseradish peroxidase (HRP) into bile have been investigated using the isolated perfused rat liver operating under one-pass conditions. Following a 1 min one-pass infusion of HRP, two peaks of HRP activity were noted in the bile. The first, at 5-7 min post-infusion, correlated with the biliary secretion of the [3H]methoxyinulin which was infused simultaneously with the HRP. The second peak of HRP activity occurred at 20-25 min, and correlated with the biliary secretion of 125I-IgA, which was also infused simultaneously with the HRP. If the isolated livers were perfused with a medium containing 2.5 microM-colchicine, the biliary secretion of IgA and the second secretion peak of HRP were inhibited by 60%. If rats were pretreated for 12h with alpha-naphthylisothiocyanate (25mg/100g body wt.) prior to liver isolation, the biliary secretion of [3H]methoxyinulin and the first secretion peak of HRP were increased. Taken together, these results suggest that HRP enters the bile via two routes. The faster route, which was increased by alpha-naphthylisothiocyanate and correlated with [3H]methoxyinulin entry into bile, was probably paracellular, involving diffusion across tight junctions. The slower route, which was inhibited by colchicine and correlated with the secretion of IgA, was probably due to transcytosis, possibly within IgA and other transport vesicles.

Immunohistochemical localization of choline acetyltransferase in rabbit forebrain.

uinea pig antiserum specific to the purified bovine choline acetyltransferase was used to demonstrate the localization of this enzyme in rabbit forebrain by the peroxidase-antiperoxidase immunohistochemical method. Choline acetyltransferase was localized in olfactory bulb, olfactory tract, olfactory tubercle, piriform cortex, septum, diagonal band, basal ganglia, thalamus, hypothalamus, subthalamus, habenula, cerebral cortex, hippocampal region, corpus callosum, internal capsule, fornix, longitudinal striae and other areas. The findings reflect the distribution of cholinergic axons and, possibly, their terminals. These observations correlate well with biochemical determinations of choline acetyltransferase and with previously proposed cholinergic pathways.

Localization of choline acetyltransferase at neuromuscular junctions.

Choline acetyltransferase (ChAc) was localized at the neuromuscular junctions of rabbit diaphragm with immunohistochemical staining, which produced a brown color, in combination with histochemical staining for cholinesterase, which produced a blue color, in the same section. The innervations of nerve terminals at the neuromuscular junctions were comparable when a silver impregnation was substituted for the immunohistochemical staining of ChAc.

Immunohistochemical localization of choline acetyltransferase in the human cerebellum.

Guinea pig antiserum specific to purified bovine choline acetyltransferase was found to cross-react with human enzyme. The peroxidase-antiperoxidase immunohistochemical method was then used to demonstrate the localization of choline acetyltransferase in formalin-fixed and paraffin-embedded human cerebellum from normal as well as from Huntington's disease brains. Choline acetyltransferase was localized exclusively in the mossy fibers and the glomeruli of the cerebellar folia. These immunohistochemical findings reveal the distribution of cholinergic axons and their terminals. The results are not only similar to our previous studies using the same method on the localization of choline acetyltransferase in rabbit cerebellum, but also demonstrate that some mossy fibers are cholinergic as suggested by others.

Immunohistochemical localization of choline acetyltransferase in rabbit spinal cord and cerebellum.

Guinea pig antiserum specific for purified bovine choline acetyltransferase has been shown to cross-react with rabbit enzyme. We used the peroxidase-antiperoxidase immunohistochemical method to demonstrate the localization of choline acetyltransferase in formalin-fixed and paraffin-embedded sections of rabbit spinal cord and cerebellum. In the spinal cord, in agreement with our and others' previous results using immunofluorescent techniques, choline acetyltransferase was found in the cell bodies of the ventral horn motor neurons. In the cerebellum, choline acetyltransferase was localized exclusively in the mossy fibers and the glomeruli of the cerebellar folia. The immunohistochemical findings in the cerebellum reveal the morphological detail of cholinergic axons and their terminals. The results are consistent with published biochemical data on the cerebellar distribution of choline acetyltransferase.

The Light-harvesting Chlorophyll a/b-Protein Complex of Chlamydomonas reinhardii.

The molecular organization of chlorophyll in Chlamydomonas reinhardii has been shown to be essentially similar to that in higher plants. Some 50% of the chlorophyll in Chlamydomonas reinhardii chloroplast membranes has been shown to be located in a chlorophyll a/b-protein complex. The complex was isolated in a homogeneous form by hydroxylapatite chromatography of sodium dodecyl sulfate extracts of the chloroplast membranes. Its absorption spectrum exhibits two maxima in the red region at 670 and 652 nm due to the presence of equimolar quantities of chlorophylls a and b in the complex. Preparations of the chlorophyll-protein also contain some of each of the carotenoids observed in the intact chloroplast membrane, but not in the same proportions. The native complex (S value = 2.3S) exhibits a molecular weight of 28,000 +/- 2,000 on calibrated sodium dodecyl sulfate-polyacrylamide gel electrophoresis. However, on the basis of its amino acid composition and other data a more probable molecular weight of about 35,000 was calculated. Each 35,000 dalton unit contains three chlorophyll a and three chlorophyll b molecules, and on the average one carotenoid molecule conjugated with probably a single polypeptide of 29,000 daltons. Comparison of spectral and biochemical characteristics demonstrates that this algal chlorophyll-protein is homologous to the previously described major light-harvesting chlorophyll a/b-protein of higher plants. It is anticipated that the Chlamydomonas complex functions solely in a light-harvesting capacity in analogy to the function determined for the higher plant component.