Search

The immunologic effects of ampicillin polymers were studied in mice and rabbits. Polymerized ampicillin--in contrast to monomers--had a suppressive effect on the antibody response to penicilloylated proteins. Unfractionated ampicillin polymer, and high molecular weight fractions were found to have an immunogenic effect in mice, when tested by a haemolytic plaque assay. The immunogenic effect of variously sized polymers did not correspond to their antigenic activity as reflected by passive cutaneous anaphylaxis experiments.

This communication presents a sensitive and discriminative method for detection of protein impurities in penicillin preparations. Antibodies against various penicilloyl derivatives of high avidities and specificities raised in rabbits were coupled to microcrystalline cellulose. The amount of penicilloyl antigen present in a sample was calculated from the relative uptake of a radioiodinated penicilloylated albumin competing with the sample for binding to the antipenicilloyl immunosorbent. As little as 0.048 pmol/ml of penicilloylated human serum albumin could be detected. The accuracy of the determination was within +/- 23% (p less than 0.05). The pronounced specificities against the penicillin side chains demonstrated by the various immunosorbents were not displayed by the antibodies in passive cutaneous anaphylaxis experiments in guinea pigs. Furthermore, the immunosorbents showed the same pattern of specificity against monomeric penicillins as for penicilloylated proteins, but the former were considerably less efficiently recorded. The relatively small quantities of protein impurities in penicillin preparations, however, necessitated a separation from penicillin prior to analyses with the RIA. This was accomplished by fractionation on Sephadex G-50 fine, ginving a recovery of 80-90% of 0.1-2.5 ppm of penicilloylated protein.

Rabbit antibodies raised against protein conjugates of azidocillin (AzO), ampicillin (AmpO), benzylpenicillin (BPO) and crude 6-aminopenicillanic acid (6-APA) gave rise to two precipitation lines in immunoprecipitation tests against penicillin protein conjugates where the carrier was unrelated to the immunogens. This pattern verifies extensive cross-reactions between penicillins, which have previously been reported. The use of AzO-coated rabbit red cells in indirect haemagglutination (IHA) inhibition tests showed good reactivity (a) between the inhibitor AzO and antibodies raised against AzO, AmpO and BPO, (b) between the inhibitor AmpO and antibodies raised against AmpO and BPO, (c) between the inhibitor BPO and antibodies raised against BPO. When BPO-coated red cells were used, AzO, AmpO and BPO inhibited the IHA activity of antibodies against the BPO. Varying IHA results were obtained using differently conjugated rabbit red cells. Thus antibodies against AmpO gave higher titres when using AzO- than BPO-coated red cells while for antibodies against BPO the opposite was noted. The radiometric assay utilizing AzO-serum albumin as tracer as well as reference substance resulted in significantly higher values for antisera against AzO and AmpO compared with those for antisera against crude 6-APA. Antisera against BPO were not analysed in this respect. Furthermore, there was poor correlation between the results obtained with the indirect IHA technique and the radiometric immunoassay.

Antibody responses to penicilloyl were recorded in infants and children treated for about 1 week with intravenous ampicillin. In this single-blind, randomized study two commercially available preparations were compared, one of high-grade purity (less than 0.1 microgram antigen per gram) and the other slightly contaminated with high molecular weight proteinaceous material (1.5 microgram antigen per gram) according to a radioimmunoassay. The results showed no significant increase in the antibody titers in any of the patient groups. The immunogenic properties of high molecular weight proteinaceous impurities isolated from phenoxymethyl penicillin during manufacture were tested in rabbits given daily subcutaneous injections with 0.1-10 micrograms of the antigen over 10 day periods. When the 10-day period was repeated, the rabbits injected with more than 2 micrograms of antigen responded with both IgM/IgG and IgE antibodies. The clinical as well as the animal study indicates that antigen impurities in the preparations of the order of 2 micrograms per gram or less do not elicit a significant antibody response. However, the study in rabbits demonstrates that high molecular weight impurities can induce penicillin allergy if present in about 10-fold higher quantities than those usually found in the commercial penicillin preparations of today.

The effects of the different acyl side chains of azidocillin, ampicillin and benzylpenicillin on the immunogenic potency of penicilloylated antigens as well as on the specificity of the developed antibodies were investigated in CBA mice. The antigens used were penicilloylated bovine gammaglobulin (BGG), bovine serum albumin (BSA) and human serum albumin (HSA). Immunization was performed with injection of high doses of antigens together with adjuvant or by injection of minute amounts of antigen over periods of 10 days. IgE antibodies were recorded with PCA in rats and IgG antibodies were measured with a double antibody assay. The nature of the carrier as well as the number of epitopes was found to influence the development of antibodies irrespective of the immunization schedule used. The immunogenic activity of the PO-BSA antigens was related to the epitope density. The PO-BSA antigens were, in contrast to the PO-BGG antigens, weak immunogens in the CBA mice. The acyl side chains of the different penicillins influenced the induction and specificity of the IgE antibody responses obtained after daily treatment.

The ability of penicillins of varying lipophilicity to induce and elicit cellular allergic responses were analysed in guinea pigs. Epicutaneous application of the penicillins benzylpenicillin (Bp), cloxacillin (Clox) and Bacampicillin (Bamp) did not cause any unspecific skin irritation. Intracutaneously injected, however, Bamp, Clox and to a lesser extent Bp caused irritation at concentrations of 1.25%. Solutions of 0.12% of penicillins were inactive in this respect. Cellular allergic responses were induced with Bp, Bamp and Clox after repeated epicutaneous application. The magnitude of responses was related to the lipophilic properties of the penicillins, Bamp being superior. In the guinea pig maximization (GPM) test of Magnusson and Kligman employing intradermal injections of the penicillins with Freund's complete adjuvant, similar sensitizing abilities of the three penicillins were observed. The cellular allergic responses were elicited with Bp, Bamp, Clox and in addition ampicillin and the 1'-ethoxycarbonyloxyethyl ester of Bp. An extensive cross-reactivity between the penicillins was seen in Bp- and Bamp-sensitized animals, whereas the Clox-sensitized animals showed a specificity limited to Clox. Bamp was shown to possess a superior activity to elicit reactions, possibly due to its lipophilic properties together with an irritating effect exerted by the NH2 group.

BALB/C mice were immunized with conjugate of benzylpenicillin or ampicillin with Ascaris suum extract. The mice developed IgE antibodies to penicillin, which were found to react with commercially available penicillin preparations in the PCA system. Elimination of polymerized penicillin by Sephadex chromatography from tested preparations could not diminish their activities to elicit PCA. High pressure liquid chromatographic (HPLC) separation of PcG preparation yielded fractions unable to elicit PCA. On the other hand, all fractions from ABPC retained the activity even after the HPLC purification. A comparative estimation of cross-reactivity of IgE and IgE antipenicillin antibodies showed that IgE antibodies cross-reacted with a variety of beta-lactam antibiotics in a high degree.

Antibody responses to the penicilloyl moiety were recorded in patients treated intramuscularly with ampicillin for respiratory infections. For practical reasons this was performed at two clinics. At each clinic, two commercially available ampicillin preparations were used in a single-blind, randomised study. One of the preparations at each clinic was of high-grade purity and the other was slightly contaminated, according to a radioimmunoassay. There was the same heterogeneity in the patient material between the two clinics. The results showed a slight but significant (p less than 0.001) increase in IgM penicilloyl antibodies in the group of patients treated with the pure penicillin but not in those treated with the contaminated penicillin. When the patients in the two studies were analysed jointly, the increase in IgM antibodies was significantly greater (p less than 0.05) in the group treated with pure penicillin than in that treated with impure penicillin. The difference in antibody formation between the patient groups treated with pure or contamined ampicillin at each clinic was not significant, however. An immunological activity of the impurities suppressing the IgM antibody response is suggested.

Purified polymers have been isolated from 6-aminopenicillanic acid and from semi-synthetic penicillins and cephalosporins. All the polymers were shown to react with rabbit antibody of penicillins and cephalosporins. All the polymers were shown to react with rabbit antibody of penicilloyl specificity, as demonstrated by passive cutaneous anaphylaxis in guinea pigs, and the reactions were shown to be penicilloyl-specific by hapten inhibition experiments. The cephalosporin-derived polymers in addition reacted with rabbit antibodies raised to the corresponding cephalosporin conjugates of bovine gamma-globulin. Using direct skin tests, passive cutaneous anaphylaxis and an in vitro assay, no evidence was obtained that the polymers induced the formation of specific antibodies in baboons, guinea pigs and rabbits, but in baboons the induction of cell-mediated immunity, demonstrable by delayed skin test reactions, was shown.

Using a radioimmunoassay, it was shown that commercially available ampicillin preparations often contain penicilloylated high molecular weight impurities. These possess immunological activities and stimulate penicilloyl-specific antibody formation in mice treated according to a therapeutic schedule. Using purified and experimentally contaminated preparations it was also found that exposure of the animals to Escherichia coli and Bordetella pertussis bacteria could increase the antibody formation to small amounts of impurities. In addition, penicilloylated antigen could be recorded in serum from treated animals. The antigen formed by penicilloylation in vivo, however, was very weak and did not induce much antibody formation when injected together with Freund's adjuvant in mice or rabbits.

The present communication reveals a relationship between the epitope density of penicilloylated protein antigens and their antigenic activities in a radioimmunoassay (RIA), in passive cutaneous anaphylaxis (PCA) and in inducing antibody formation in mice. In the RIA and PCA a critical number of 2-4 penicilloyl residues per protein molecule was noted. At this level small changes in the number of substituents considerably influenced the antigenic activities. The molecular weight and the nature of the carrier proteins, myoglobin, bovine serum albumin (BSA) and dimeric BSA also affected the threshold concentration for efficient antigenic activity. The results with the RIA and PCA were significantly correlated to each other. Using penicilloylated BSA as immunizing antigen in mice it was found that an epitope number higher than 11 penicilloyl residues per protein molecule induced significant antibody formation after a single injection. Antigens with a lower degree of penicilloyl substitution were less immunogenic. An antigen carrying 0.6 penicilloyl residues per BSA molecule did not induce penicilloyl-specific antibodies even after three injections. The capacity of heavily penicilloylated proteins to induce and elicit penicillin allergy as revealed by the present results stresses the importance of limiting their presence in penicillin preparations.

35 individuals showing reactions to penicillin of anaphylactic shock, angioedema or urticaria were investigated. Their skin sensitivity was analysed using 16 different penicillin derivatives. In addition, the content of circulating reagins against the penicilloyl structure in the patient's sera were analysed using RAST. 17 of the patients had negative skin reactions and RAST results to all substances tested. The other 18 were skin test-positive to at least one derivative but showed markedly heterogeneous patterns of skin reactivity. 14 had positive reactions against penicilloyl structures accompanied by anti-penicilloyl reagins. Four patients showed doubtful reactions only to penicillin or penicilloate and/or penilloate. These patients also had very low levels of reagins against penicilloyl in their sera. Positive skin test results using monovalent penicillin derivatives such as penicillin, penicilloate, penilloate, penicilloyl amide, penicilloyl-formyl-lysine, penicillamine, which cannot form a multivalent antigen with penicillyol specificity, indicated formation of other derivatives of importance in penicillin allergy, e.g., penicillamine protein conjugates. Three patients showed skin reactions to ampicillin polymer and two to benzyl-penicillin polymer. The skin tests performed with the penicillin derivatives used do not seem to give more information on the sensitivity of the patients than does the RAST using penicilloyl structures.

The formation of homocytotropic antibodies (IgE) as determined by immunochemical characteristics against penicilloyl in rabbits was shown. The production of such antibodies to azidocilloyl-human serum albumin (AzO10-HSA) in alum was found optimal using 1 mg of antigen in the tested range of 0.01-8 mg. At the lowest dose (0.01 mg) only haemagglutinating but no IgE antibodies were formed. Immunization with azidocilloyl-bovine gamma-globulin (AzO9-BGG) resulted in a slower increase in antibody levels than was caused by AzO10-HSA. Univalent benzylpenicilloyl-epsilon-aminocaproate or ampicillin given together with the antigen upon immunization decreased the levels of penicllloyl specific IgE and haemagglutinating antibodies, but induced the formation of IgE antibodies against the carrier molecule. Further, the administration of penicilloyl specific IgG antibodies diminished formation of both IgE and haemagglutinating antibodies, but no antibodies specific for the carrier were formed. The usefulness of this animal model for the experimental study of penicillin allergy is discussed.