- Email: [email protected]
Immediate hypersensitivity to penicillins. Identification of a new antigenic determinant
Accepted Manuscript Title: Immediate hypersensitivity to penicillins. Identification of a new antigenic determinant Authors: Sonia Matas, Marta Broto, Merc`e Corominas, Ramon Lleonart, Ruth Babington, M.-Pilar Marco, Roger Galve PII: DOI: Reference:
S0731-7085(17)31594-7 http://dx.doi.org/10.1016/j.jpba.2017.08.024 PBA 11470
To appear in:
Journal of Pharmaceutical and Biomedical Analysis
Received date: Revised date: Accepted date:
20-6-2017 10-8-2017 17-8-2017
Please cite this article as: Sonia Matas, Marta Broto, Merc`e Corominas, Ramon Lleonart, Ruth Babington, M.-Pilar Marco, Roger Galve, Immediate hypersensitivity to penicillins.Identification of a new antigenic determinant, Journal of Pharmaceutical and Biomedical Analysishttp://dx.doi.org/10.1016/j.jpba.2017.08.024 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Immediate hypersensitivity to penicillins. Identification of a new antigenic determinant. Sonia Matas1,2, Marta Broto1,2, Mercè Corominas3, Ramon Lleonart3, Ruth Babington2,1, M.-Pilar Marco1,2, Roger Galve1,2,# 1
Nanobiotechnology for Diagnostics (Nb4D), Department of Chemical and Biomolecular
Nanotechnology, Institute for Advanced Chemistry of Catalonia (IQAC) of the Spanish Council for Scientific Research (CSIC). 2CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Jordi Girona 18-26, 08034 Barcelona, Spain. 3
Allergology-Internal Medicine, Bellvitge University Hospital, IDIBELL, Feixa Llarga
sn, 08907 L’Hospitalet de Llobregat, Barcelona. CORRESPONDING AUTHOR #
To whom correspondence should be addressed. Phone: 934006100 (Ext. 5117). Fax:
932045904. E-mail: [email protected]
1
Graphical Abstract
Highlights:
A strategy based on the use of peptide probes, small labeled and chemically active peptides, has been used for identification of new allergenic epitopes. A new antigenic determinant of penicillins has been isolated and identified using these peptides probes. A magneto-ELISA has been developed for the detection of specific IgEs from sera of penicillin allergic patients.
Abstract The study of adverse drug reactions (ADRs) constitutes a challenge in the area of Medicine. Drugs generate a large number of the total registered hypersensitivity reactions, where penicillins are responsible for more than half of them. In vitro tests in the market are not efficient enough since they lack in sensitivity and specificity. This is the reason why in vivo tests are carried out, with the subsequent danger to the patient’s life. It is essential to discover new β-lactam antigenic determinants to develop more 2
effective detection systems and thus, obtain better explanations of the allergic mechanisms related to these drugs. We propose a strategy based on the use of “peptide probes”, small labeled and chemical active peptides which have been structurally modified for reacting with the β-lactam moiety at different conditions. The probes also contain a biotin group for application in an immunoassay format. Three different amoxicillin adducts have been obtained, purified and characterized by HPLC-MS and NMR techniques. These results have helped us to elucidate and propose a new antigenic determinant for β-lactams, named the “penamidyl” epitope. All the adducts have been validated and evaluated with sera from different penicillin allergic patients by means of a Magneto-ELISA, immunochemical technique that has allowed us to detect specific IgEs in a very high percentage of the serum samples. An immunoassay has been developed, validated and applied as a diagnostic tool for the detection of specific IgEs in the sera of penicillin allergic patients using a new antigenic determinant.
Abbreviations SI (supporting information), NBD-Cl (4-chloro-7-nitrobenzofurazan), NBD-OH (4hydroxy-7-nitrobenzofurazan), dimethylaminopyridine),
DIC
DIEA
(N,N’-diisopropylcarbodiimide), (N,N-diisopropylethylamine),
DMAP HATU
[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium
(4(1-
3-oxid
hexafluorophosphate), DKP (diketopiperazine), Fmoc (fluorenylmethoxycarbonyl group), tBu (tert-butyl group), DNP (2,4-dinitrophenyl group), TIS (triisopropylsilane), TFA (trifluoroacetic acid), r.t. (room temperature), STV-HRP (streptavidin linked to horseradish peroxidase), MP (magnetic particles), TMB (3,3’,5,5’-tetramethylbenzidine), PVP (polyvinylpyrrolidone), PP (peptide probe), MW (molecular weight), AX (amoxicillin), AP (ampicillin), PG (penicillin G), PV (penicillin V), MDM (minor determinant mixture), PPL (penicilloyl polylysine), HSA (human serum albumin), BSA (bovine serum albumin), PLys (PLL or PL, polylysine), IgE (immunoglobulin E), IgG (immunoglobulin
G),
RP-HPLC
(reverse-phase
high-performance
liquid
chromatography), UPLC-MS (ultra-performance liquid chromatography tandem massspectrometry), NMR (nuclear magnetic resonance).
3
Keywords: β-Lactam Hypersensitivity. Amoxicillin. Antigenic Determinant. Adverse Drug Reactions (ADRs). PenamidylIntroduction Since Alexander Fleming discovered penicillin in 1928, antibiotics have been widely used to treat humans against bacterial infections. The presence of these drugs has been linked to many undesirable consequences such as bacterial resistance [1] and adverse drug reactions (ADRs) [2]. Many of ADRs are predictive reactions, but in some cases drugs can also induce, through immunological mechanisms, hypersensitivity reactions with symptoms that can lead to severe reactions and anaphylaxis. According to the modified classification of Gell-Coombs [3], there are mainly 7 types of hypersensitivity reactions based on different immunological mechanisms and involving the participation of immunoglobulins, complement activation and T cells. Allergy is often associated with a type I hypersensitivity reaction, which is mediated by IgE antibodies produced against an external agent such as pollen, mites, food or drugs that acts as antigen [4]. The most widely used antibiotics are β-lactams due to their broad antibacterial spectrum and low toxicity. These include penicillins, cephalosporins, carbapenems and monobactams, and represent 65% of the world antibiotic market [5, 6]. Allergy to lactams represents about 30% of the total hypersensitivity drug reactions [7, 8], and penicillin and its derivatives are responsible for 75% of death from anaphylaxis [9]. Allergic reactions to -lactam drugs can be classified according to their temporal profile in immediate, accelerated and delayed, which is related with the clinical manifestations and the pathogenic mechanisms [10]. Most of the immediate and accelerated reactions are mediated by IgEs, with clinical manifestations of immediate hypersensitivity [11]. Unlike other allergens, most drugs are low molecular weight compounds incapable of eliciting an immune response. Although the mechanisms involved are complex, for several drugs the hypersensitivity phenomenon has been based mainly on the "hypothesis of the hapten", in which the drug itself or some metabolites covalently binds with an endogenous molecule, usually a blood protein (i.e. HSA, human serum albumin), forming an antigen capable of initiating an immunological reaction [12]. The sites for the binding are not well known and much research has been done to determine the chemical reactions of the haptens. The nature of the drug and the reaction mechanisms influence the outcome of the haptenized protein [13-15]. We can find several examples in the literature of the 4
behavior of this hapten reactivity. For instance, the use of peptide probes has been frequently applied in the study of allergic contact dermatitis (ACD) [16, 17]. In this context, Troutman et al. [18] described the utility of lysine as a nucleophile for screening around 70 contact allergens. The studies done for the identification of antibiotic allergens, and particularly for β-lactams, have been based on similar strategies, either working with whole proteins (HSA, BSA or PLys [19]) or with selected amino acids [20] or dendrimers [21]. Few antigenic determinants have been described until now (Figure 1A), but only the major one (penicilloyl) has been postulated to be the real precursor for this type of allergies. Concerning the specific example of the amoxicillin, one of the most worldwide used β-lactams, only the amoxicilloyl derivative has been isolated. An article recently published by Garzón et al. [22] describes the identification of a possible adduct between amoxicillin and the lysine 190 of the HSA. The diagnostic of penicillin allergy represents in many cases a serious challenge. This is because the potential risk of in vivo methods and the lack or low performance of in vitro tests. Initially, when a patient reports a clinical picture of hypersensitivity, a comprehensive study of the clinical history is essential to judge if the patient has had a reaction related to the drug. The study aims to determine whether the reaction produced is due to a selective allergic reaction against this drug or due to another idiosyncratic adverse reaction [23, 24]. Basically, skin tests and exposure controlled tests using increasing amounts of drug doses are performed [25, 26]. Penicillin skin testing is the most reliable way to evaluate patients for IgE-mediated penicillin allergy. Skin testing is conducted using benzylpenicilloyl polylysine (PPL, major determinant), penicillin G and some penicillin analogues (minor determinants, Figure 1B). Testing with both the major and minor determinants can be identified up to 97% of patients with an immediate hypersensitivity to these drugs. However, in vivo tests are invasive techniques with a certain risk for the patient. Thus, several in vitro diagnostic tests have been developed for the diagnosis of drug allergies, being the immunochemical techniques the most used ones. These methods are based on the detection of the specific IgEs using coated antigenic determinants. Examples of immunoassays are the RAST (radioallergosorbent assay), ELISA (enzyme-linked immunosorbent assay) or FIA (fluorescent immunoassay) [27, 28]. Other in vitro techniques based on cellular studies are the LTT (lymphocyte transformation test), the CAST (cellular allergen stimulation test) and the BAT (basophil activation test) [29]. However, nowadays the use of in vitro diagnostic tests is far from 5
representing an effective and reliable alternative to the in vivo assays. The low sensitivity and the lack of specificity of in vitro tests are due to several causes. Likely, the low concentration of specific IgEs in the serum could be one of the reasons. However, the main cause is probably related to the antigenic determinants used in these techniques. All of the developed immunoassays, including the commercial ones (i.e. 3gAllergy test from Siemens, Immuno CAP test from Thermo-Fisher), are based exclusively on the detection of specific IgEs by means of the penicilloyl determinants. These companies provide "specific tests" for penicillin G, penicillin V, ampicillin and amoxicillin, showing poor sensitivities and low true positives, so currently their values are only complementary to the in vivo tests [30, 31]. Concerning cell tests, they are not well standardized, and their use is exclusively reduced to individual laboratories that apply their own methodologies. The low sensitivity of in vitro tests can explain the discrepancies observed in the diagnosis procedures used for North-American and European specialists [32]. While the former do not consider in vitro tests, European allergist propose their use as a preliminary step to skin tests, mainly in high risk patients [33, 34]. We strongly believe that the problem is related to the lack of knowledge about the immunological mechanisms involved in the -lactam allergy. What is the inducer of the hypersensitivity reaction besides the penicilloyl determinant, and how it has been formed are questions that nowadays are still without convincing answers. Within this work we have tried to clarify a little bit more the complicated world of the penicillin allergy. 1. Materials and methods 1.1. Chemistry. The Instrumentation and reagents section is described in the supporting information (SI). The synthesis and characterization of the different peptide probes is also described in the SI. 1.1.1. Synthesis of the adducts. General procedure. The peptide probes were dissolved in 10 mM buffer solution at three different pHs (4.6, 7 and 11). Previously, all buffers were degassed under reduced pressure to remove the oxygen dissolved. Then, 10 eq. of amoxicillin were added into the solution. The optimization of the reaction between the amoxicillin and the peptide probes is described in the SI. Simultaneously, an amoxicillin and a peptide solution were separately prepared (blanks) and treated at the same way. The 6
solutions were stirred at 37ºC under argon atmosphere during 4 days. All the reactions were followed by UPLC-MS every day, monitoring the wavelength at 470 nm. and the MS at [MWpeptide + MWamoxicillin + H]+. The crudes were purified by preparative HPLC, collecting those peaks that did not appear in the blank samples (see Figure 2). The purified compounds were characterized by MS, and those ones that have the required molecular weight were lyophilized and redissolved with a certain amount of milliQ water. The final concentration of the adducts was calculated by fluorescent spectroscopy. The preparation and characterization of all the adducts is described in the SI. 1.2. Immunochemistry. Evaluation of sera from allergic patients by magnetoimmunoassay (see Figure 4). The Reagents and Instrumentation section is described in the SI. Into vials of 2 ml capacity were added 25 μg of the antibody conjugated magnetic particles (anti-IgE-MP) and washed three times with 400 μL PBST. Next, human sera from penicillin allergic patients and blank serum (100 μL per vial) were incubated for 1h. at 37ºC under agitation. The particles were washed again three times with 400 μL PBST. Then, the adducts and the peptide probe PP1 NBD-N(Me)-DKK(Biotin)βA-OH (200 nM in PBST, 200 μL per vial) were added and incubated for 1h. at r.t. under agitation. After another washing step (400 μL PBST x 3 times), the particles were incubated with a solution of the STV-HRP (0.05 μg mL-1 in PBST, 200 μL per vial) for 30 min. at r.t. under agitation. Particles were washed following the standard procedure and changed to different containers to avoid signal noise due to the STV-HRP adsorbed on the vial surface. Finally, the particles were treated with 200 μL of the substrate solution and incubated for 30 min. at r.t. and light protected. Color development was stopped with 4 N H2SO4 (100 μL per vial), and the absorbances were read at 450 nm. 2. Results and discussion The main objective of this work was the identification of new antigenic determinants of β-lactams. Within this work, amoxicillin was the antibiotic used as a model for the βlactams reactivity. And we chose the approach based on the use of peptide probes because of the possibility to isolate and characterize the amoxicillin adducts. One of the first challenges was the design of the peptide probe, which had to have three differentiate components. The probe should include a lysine as a nucleophilic residue able to react with 7
the drug, a chromophore group for allowing the monitoring and purification of the adducts, and a biotin derivative in order to use the antigenic determinants in an immunoassay format (see Figure S5). Peptide probes were synthesized by solid phase using the Fmoc/tBu strategy. Basically, peptides were built through the N-terminal chain. The polymeric support selected was the NovaPEG Wang resin, which allows the preparation of peptides with a carboxylic acid group at the C-terminal side. The cleavage of the lateral chain protecting groups and the peptide from the resin was done by acidolysis with TFA. And after purification by preparative HPLC, the peptides were lyophilized. The general procedure for the peptide probes synthesis, and all the details of the preparation and characterization of each peptide can be found in the SI. Four different peptides probes were synthetized (Figure S5), one of them (PP1) for use in the immunochemical assays, and the others for the characterization of the adducts. In all cases, the first amino acid (AA1) was β-alanine in order to avoid the formation of a diketopiperazine (DKP) derivative during the peptide synthesis [35]. Peptide PP2 differs from PP1 in the biotin building block, while PP3 only contains the fluorophore building block. The fluorescent dye chosen for PP1, PP2 and PP3 was nitrobenzoxadiazole (NBD) [36] because of its absorbance wavelength (470 nm), far from the wavelengths of the amoxicillin byproducts, the nonreactivity of its heteroatoms, the simplicity of the NMR signals and the low price. Due to the fluorescent building block was not commercially available, the derivatized amino acid NBD-N(Me)-Asp(OtBu)-OH (AA4) was previously synthesized from NBD-Cl and protected aspartic acid (see Section S1.2 at SI), which was chosen for conferring an increase of solubility of the final peptide probe. Aspartic acid had to be N-methylated in order to avoid the peptide probe degradation in alkaline aqueous media [37]. Peptide probe PP4 is very similar to PP2 and only differs in the fluorophore. In that case a dinitrophenol derivative (DNP) was used, and the fluorescent building block was made through a glycine instead from the aspartic acid. Previously to the formation of the adducts, the reaction between the peptide probe and amoxicillin was optimized (Table S5 and Graph S1, in SI). The working concentrations selected were 0.5 mM of the peptide probes and 5 mM of the amoxicillin (10 eq.). All buffers were degassed, the reaction times were 4 days, at 37ºC and under argon atmosphere. Finally, reactions with the peptide probe (PP1) and the β-lactam were carried out at three different pHs (4.6, 7 and 11). Although “physiological conditions” are approximately pH 7 at 37ºC, we decided to work at several pHs because endogenous 8
proteins (HSA) contain polar amino acids (arginine, lysine or aspartic and glutamic acids) which are able to create localized acidic/alkaline environments. The evolution of the reactions was monitored by UPLC-MS, looking for chromatographic peaks with an absorption wavelength of 470 nm, and a MW of 1228 g mol-1 (MWPP1 + MWamoxicillin). Preparative HPLC was used for purification of the adducts. The results are shown in Figure 2, where it can be observed the HPLC chromatograms of the peptide probe PP1 alone (graphs A, C and E) and in combination with the amoxicillin (graphs B, D and F), at different pHs. At pH 4.6 two different adducts (NKAx1 and NKAx2) were found and isolated, while at pH 7 and 11 another one (NKAx3) was obtained. Curiously, the third adduct was obtained in a very low concentration, even less at pH 7. Two other peaks were able to be identified, the degradation product of the chromophore NBD-OH (4-hydroxy7-nitrobenzofurazan) and an oxidation byproduct of the peptide probe (MWPP1 + 16). Characterization of all the adducts is found in the SI. Using this strategy we were able to obtain three different amoxicillin adducts. It is well known and repeatedly reported the synthesis of the amoxicilloyl derivative [38], which is prepared in alkaline conditions with amoxicillin and an amine group, obtaining the formation of an amide bond and the cleavage of the β-lactam ring (Figure 3). We have assumed that NKAx3 is the adduct containing the amoxicilloyl group as it has been made in the same pH conditions reported in the literature. However, any reference has been found related to the adducts obtained at slightly acidic pH conditions. In order to figure out the configuration of the other two adducts, three new peptide probes with different modifications, mentioned above, were synthesized. All the reactions between these probes and amoxicillin were carried out at pH 4.6 and followed by UPLC-MS. Two different adducts (A4 and A5) were isolated using PP2, two other adducts were identified using PP3 (A6 and A7) and only one adduct was obtained with PP4 (A8). The first remark is the reaction between the β-lactam and PP3, where two adducts were obtained without the participation of any nucleophilic group, concluding that the amino group of the lysine is not the responsible of the adduct formation in acidic media. A second remark is the chromophore, which is changed in PP4 and an adduct is still obtained. Summarizing, only one group is common in all the peptides probes, the carboxylic system. And the number of adducts correlates exactly with the number of carboxylic groups of the peptide probes since PP1, PP2 and PP3 have two carboxylic groups and two adducts were obtained, while PP4 has one carboxylic group, and only one adduct was observed. In order to elucidate the mechanisms involved in the formation of the adducts, two of them (A4 and A5) were scaled up and purified by 9
preparative HPLC. The proposed structure for both adducts and their NMR assignations are described in the SI. The resonance confirmed the hypothesis since its
13
C-NMR
chemical shift changed significantly (Adduct A4, C1 from 177.0 ppm to 175.38 ppm. Adduct A5, C2’’ from 175.2 ppm to 173.75 ppm). Moreover, the shifts of the carbons involved in the β-lactam ring were also displaced form the original amoxicillin structure, while the carbons of the lateral chain did not changed. A possible explanation could be a direct transamidation between the nitrogen of the tiazolidinic ring and the acid group of the peptide probe in acidic media, but we could not find any example in the literature. Probably, a more accurate hypothesis was described by Sharma et al. [39, 40] in 1987. The authors postulated the mechanism in two steps (Figure 3). First, a mixed anhydride is likely to be formed by lysis of the β-lactam linkage with an acid group. Second, an intramolecular O→N acyl transfer between the secondary amino group and the anhydride. This mechanism was similarly postulated earlier and described in the book “The Chemistry of Penicillin” [41]. In conclusion, this strategy allowed the preparation of three β-lactam adducts (NKAx1, NKAx2 and NKAx3), two of them presenting a new antigenic determinant never described before in the literature. Casually, the structure of the “penamidyl” group (Figure 1) is very similar to allergenic determinants currently used in skin tests (penicilloyl and penicilloate), since the backbone of the penicillin is preserved and most of its functional groups have not changed. Adducts were synthetized including a biotin group for being applied in the development of immunoassays, since it allows the validation of the adducts and the evaluation of the samples from the allergic patients. We must take into account that the final objective is the detection of specific IgEs using the antigenic determinants obtained with the peptide probes. Initially, the three adducts were validated using both the ELISA and the magnetoELISA formats (Section S2.2 and Figure S9 in SI). Working with magnetic particles can help to minimise the background signal and allow the preconcentration of the serum samples. In both assays, the surfaces were coated with anti-AX-IgG, a polyclonal antisera (As218) produced against the open-ringed form of the amoxicillin [42]. In the magnetoELISA format, the magnetic particles were previously covalently linked to a commercial anti-IgG. Different concentrations of the adducts and the peptide probe were added. The signal was obtained enzymatically using streptavidin linked to HRP (STV-HRP). The results of these validations are shown in Figure S9. In conclusion, in both assays only the adducts were recognized by the specific antibody As218, but not the peptide probe. These 10
results demonstrate that the three adducts contain an active form of the amoxicillin which is able to interact with an antibody raised against this drug. Finally, both assays were applied for the detection of specific IgEs from sera of penicillin allergic patients. In that case, a commercial anti-IgE was used to capture the IgEs, which was absorbed onto the plate wells in the ELISA format or covalently linked to the nanoparticles in the magnetoELISA format (Figure 4). In early experiments, two different types of samples were used, a pool of specific sera and a pool of non-specific sera (blank samples). Samples were tested against the three adducts and the peptide probe PP1. The ELISA format was quickly discarded because of the background signal obtained (data not shown). Basically, the nonspecific absorption onto the wells of the microtiter plates was observed when highly concentrated serum samples were used. Thus, the assays were finally carried out using the magneto-ELISA format, as the particles could be transferred to different containers in order to avoid background noise. The three adducts and PP1 were tested against 15 different serum samples (100 μL per assay) from allergic patients and a blank serum. The clinical history of all the patients is shown in the Table S7. All patients were admitted to hospital due to a clinical picture of hypersensibility to penicillins but, although all patients were positively diagnosed, only 4 of them were positive to a commercial specific IgE test (25% of effectiveness). The results obtained in the evaluation of the samples using the magneto-ELISA are represented in the Figure 4. The normalized signal is calculated as the difference between the signal obtained from each adduct minus the signal obtained from the peptide probe. The continuous horizontal line marks the non-specific signal, which is the average of 18 measurements using the three adducts with the blank serum, while the horizontal dotted lines represent the standard deviation of the mean. Therefore, the samples that show signals over the dotted top line should contain specific IgEs against amoxicillin. Looking at the results, all samples show a specific answer to at least one of the three adducts. More concretely, adduct NKAx1 is recognized by the 87% of the samples, adduct NKAx2 by the 80% and adduct NKAx3 by the 40%. Curiously, the antigenic determinant of the third adduct is the amoxicilloyl (see Figure 1), which has been considered as the major one in the haptenization process and the real cause of βlactam hypersensibility. In the supporting information, the evaluation of sera of allergic patients against the three adducts and PP1 separately (Figure S10) is represented. 3. Conclusions
11
The main contribution of this work has been the isolation and identification of a new antigenic determinant of penicillins, which has been named the “penamidyl” epitope. The strategy has been based on the use of peptide probes, small labeled and chemically active peptides, structurally modified for reacting with the β-lactam molecule. The probes have been synthesized with a biotin group as a linker to be applied in the development of immunoassays. Through the reaction of these peptide probes with amoxicillin, we have been able to isolate three different derivatives or “adducts”. One of them has been obtained at basic media, most probably the “penicilloyl” determinant, and the other two in slight acidic media. It has been proposed that the formation of these two new derivatives occurs through the carboxylic group of the peptide probe and the nitrogen of the β-lactam ring. A magneto-ELISA has been developed and validated, and the immunoassay has been applied as a diagnostic technique to the detection of specific IgEs from sera of penicillin allergic patients. We believe that the strategy of using peptide probes can improve understanding of the mechanisms involved in adverse reactions to drugs. And, although we still have to work more along in this way, we are convinced that the discovery of this new antigenic determinant will improve significantly the diagnosis of hypersensitivity reactions to β-lactams. Acknowledgement This work has been supported by the Ministry of Science and Innovation (MICINN) with the β-Array project (SAF2008-03082). The Nb4D group is a consolidated research group (Grup de Recerca) of the Generalitat de Catalunya and has support from the Departament d’Universitats, Recerca i Societat de la Informació de la Generalitat de Catalunya (expedient: 2014 SGR 1484). CIBER-BBN is an initiative funded by the Spanish National Plan for Scientific and Technical Research and Innovation 2013-2016, Iniciativa Ingenio 2010, Consolider Program, CIBER Actions are financed by the Instituto de Salud Carlos III with assistance from the European Regional Development Fund. The Custom Antibody Service (CAbS) is acknowledged for their assistance and support during antibody production. Marta Broto wishes to thank the FPI-fellowship from the Spanish Ministry of Economy and Competitiveness. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version.
12
References [1] V. Economou, P. Gousia, Agriculture and food animals as a source of antimicrobial-resistant bacteria, Infect. Drug Resist. 8 (2015) 49-61. [2] J. Uetrecht, D.J. Naisbitt, Idiosyncratic adverse drug reactions: Current concepts, Pharmacol. Rev. 65(2) (2013) 779-808. [3] T.V. Rajan, The Gell-Coombs classification of hypersensitivity reactions: a re-interpretation, Trends Immunol. 24(7) (2003) 376-379. [4] http://www.eaaci.org/organisation/about-us/introduction.html. [5] R.P. Elander, Industrial production of b-lactam antibiotics, Appl. Microbiol. Biot. 61(5-6) (2003) 385-392. [6] A.L. Demain, S. Sanchez, Microbial drug discovery: 80 years of progress, J. Antibiot. 62 (2009) 5-16. [7] E. Macy, Penicillin and beta-lactam allergy: Epidemiology and diagnosis, Curr. Allergy Asthm. R. 14(476) (2014) 1-7. [8] P. Demoly, D. Hillaire-Buys, Classification and epidemiology of hypersensitivity drug reactions., Immunol. Allergy Clin. N. Am. 24(3) (2004) 345-356. [9] A.I. Neugut, A.T. Ghatak, R.L. Miller, Anaphylaxis in the United States: an investigation into its epidemiology, Arch. Intern. Med. 161 (2001) 15-21. [10] P. Demoly, A. Romano, Update on beta-lactam allergy diagnosis, Curr. Allergy Asthm. R. 5 (2005) 9-14. [11] M.J. Torres, M. Blanca, J. Fernandez, A. Romano, A. de Weck, W. Aberer, K. Brockow, W.J. Pichler, P. Demoly, Diagnosis of immediate allergic reactions to beta-lactam antibiotics, Allergy 58 (2003) 961-972. [12] A. Ariza, D. Garzon, D.R. Abánades, V. de los Ríos, G. Vistoli, M.J. Torres, M. Carini, G. Aldini, D. Pérez-Sala, Protein haptenation by amoxicillin: High resolution mass spectrometry analysis and identification of target proteins in serum, J. Proteomics 77 (2012) 504-520. [13] E. Meschkat, M.D. Barratt, J.-P. Lepoittevin, Studies of the Chemical Selectivity of Hapten, Reactivity, and Skin Sensitization Potency. 2. NMR Studies of the Covalent Binding of the 13CLabeled Skin Sensitizers 2-[13C]- and 3-[13C]Hex-1-ene- and 3-[13C]Hexane-1,3-sultones to Human Serum Albumin, Chem. Res. Toxicol. 14 (2001) 118-126. [14] M. Aleksic, C. Pease, D. Basketter, M. Panico, H. Morris, A. Dell, Mass spectrometric identification of covalent adducts of the skin allergen 2,4-dinitro-1-chlorobenzene and model skin proteins, Toxicol. In Vitro 22 (2008) 1169-1176. [15] C. Jenkinson, R.E. Jenkins, J.L. Maggs, N.R. Kitteringham, M. Aleksic, B.K. Park, D.J. Naisbitt, A Mechanistic Investigation into the Irreversible Protein Binding and Antigenicity of pPhenylenediamine, Chem. Res. Toxicol. 2009(22) (2009) 6. [16] J. Mutschler, E. Giménez-Arnau, L. Foertsch, G.F. Gerberick, J.-P. Lepoittevin, Mechanistic assessment of peptide reactivity assay to predict skin allergens with Kathon® CG isothiazolinones, Toxicol. In Vitro 23 (2009) 439-446. 13
[17] J. Eilstein, E. Giménez-Arnau, D. Duché, N. Cavusoglu, G. Hussler, F. Rousset, J.-P. Lepoittevin, Sensitization to p-amino aromatic compounds: Study of the covalent binding of 2,5dimethyl-p-benzoquinonediimine to a model peptide by electrospray ionization tandem mass spectrometry, Bioorg. Med. Chem. 16 (2008) 5482-5489. [18] J.A. Troutman, L.M. Foertsch, P.S. Kern, H.J. Dai, M. Quijano, R.L.M. Dobson, J.F. Lalko, J.-P. Lepoittevin, G.F. Gerberick, The incorporation of lysine into the peroxidase peptide reactivity assay for skin sensitization assessments, Toxicol. Sci. 122(2) (2011) 422-436. [19] P. Vemuri, K.E. Harris, L.A. Suh, L.C. Grammer, Preparation of benzylpenicilloyl-polylysine: A preliminary study, Allergy Asthma Proc. 25 (2004) 165-168. [20] H.-P. Ho, R.-J. Lee, C.-Y. Chen, S.-R. Wang, Z.-G. Li, M.-R. Lee, Identification of new minor metabolites of penicillin G in human serum by multiple-stage tandem mass spectrometry, Rapid Commun. Mass Sp. 25 (2011) 25-32. [21] F. Sánchez-Sancho, E. Pérez-Inestrosa, R. Suau, C. Mayorga, M.J. Torres, M. Blanca, Dendrimers as carrier protein mimetics for IgE antibody recognition. Synthesis and characterization of densely penicilloylated dendrimers, Bioconjugate Chem. 13 (2002) 647-653. [22] D. Garzon, A. Ariza, L. Regazzoni, R. Clerici, A. Altomare, F.R. Sirtori, M. Carini, M.J. Torres, D. Pérez-Sala, G. Aldini, Mass spectrometric strategies for the identification and characterization of human albumin covalently adducted by amoxicillin: Ex Vivo studies, Chem. Res. Toxicol. 27 (2014) 1566-1574. [23] R. Mirakian, S.C. Leech, M.T. Krishna, A.G. Richter, P.A. Huber, S. Farooque, N. Khan, M. Pirmohamed, A.T. Clark, S.M. Nasser, Management of allergy to penicillins and other betalactams, Clin. Exp. Allergy 45 (2015) 300-327. [24] N.C.G.C. (UK), Drug Allergy: Diagnosis and Management of Drug Allergy in Adults, Children and Young People, London: National Institute for Health and Care Excellence (UK) (2014). [25] E. Macy, Penicillin allergy: optimizing diagnostic protocols, public health implications, and future research needs, Curr. Opin. Allergy Cl. 15 (2015) 308-313. [26] R. Solensky, D.A. Khan, Evaluation of antibiotic allergy: the role of skin tests and drug challenges, Curr. Allergy Asthma Rep. 14(459) (2014) 1-9. [27] R.I. Siles, F.H. Hsieh, Allergy blood testing: A practical guide for clinicians, Clev. Clin. J. Med. 78 (2011) 585-592. [28] D.G. Ebo, J. Leysen, C. Mayorga, A. Rozieres, E.F. Knol, I. Terreehorst, The in vitro diagnosis of drug allergy: status and perspectives, Allergy 66 (2011) 1275-1286. [29] A.P. Uyttebroek, V. Sabato, M.A. Faber, N. Cop, C.H. Bridts, H. Lapeere, L.S. De Clerck, D.G. Ebo, Basophil activation tests: time for a reconsideration, Expert Rev. Clin. Immunol. 10 (2014) 1325-1335. [30] M. Zidarn, M. Silar, M. Vegnuti, P. Korosec, M. Kosnik, The specificity of tests for anti-blactam IgE antibodies declines progressively with increase of total serum IgE, Wien. Klin. Wochenschr. 121 (2009) 353-356. [31] A. Vultaggio, A. Matucci, G. Virgiliw, O. Rossi, L. Filı´z, P. Parronchi, S. Romagnani, E. Maggi, Influence of total serum IgE levels on the in vitro detection of b-lactams-specific IgE antibodies, Clin. Exp. Allergy 39 (2009) 838-44. 14
[32] M.J. Torres, A. Romano, G. Celik, P. Demoly, D.A. Khan, E. Macy, M. Park, K. Blumenthal, W. Aberer, M. Castells, A. Barbaud, C. Mayorga, P. Bonadonna, Approach to the diagnosis of drug hypersensitivity reactions: similarities and differences between Europe and North America, Clin. Transl. Allergy 7 (2017) 1-13. [33] K. Brockow, H. Garvey, W. Aberer, M. Atanaskovic-Marckovic, A. Barbaud, M.B.e.a. Bilo, E.E.D.A.I. Group, Skin test concentrations for systemically administered drugs - an ENDA/EAACI Drug Allergy Interest Group position paper, Allergy 68 (2013) 702-712. [34] C. Fontaine, C. Mayorga, P.J. Bousquet, B. Arnoux, M.J. Torres, M. Blanca, P. Demoly, Relevance of the determination of serum-specific IgE antibodies in the diagnosis of immediate beta-lactam allergy, Allergy 62 (2007) 47-52. [35] E. Pedroso, A. Grandas, X. de las Heras, R. Eritja, E. Giralt, Diketopiperazine formation in solid phase peptide synthesis using p-alkoxybenzyl ester resins and Fmoc-amino acids, Tetrahedron Lett. 27 (1986) 743-746. [36] Z.-D. Shi, R.G. Karki, S. Oishi, K.M. Worthy, L.K. Bindu, P.G. Dharmawardana, M.C. Nicklaus, D.P. Bottaro, R.J. Fisher, T.R.J. Burke, Utilization of a nitrobenzoxadiazole (NBD) fluorophore in the design of a Grb2 SH2 domain-binding peptide mimetic, Bioorg. Med. Chem. Lett. 15 (2005) 1385-1388. [37] A.A. Aboderin, R.E.K. Semakula, Base-promoted cleavage of a-N-NBD-peptides, FEBS Lett. 34 (1973) 90-94. [38] www.iedb.org/epId/116053 (Immune epitope database and analysis resource). [39] R. Sharma, R.J. Stoodley, A. Whiting, Total synthesis of analogues of the b-lactam antibiotics. Part 3. 2-Ethoxycarbonyl derivaties of carbapen-1-em-exo-carboxylates and carbapenam-3-exocarboxylates, J. Chem. Soc. Perkin Trans. 111 (1987) 2361-2369. [40] J.R. Everett, J.S. Davies, J.W. Tyler, A 1H, 13C, and 19F nuclear magnetic resonance study of rotational isomerism and long-range coupling in methyl (1S,5R,7R)-1-ethyl-3-oxo-6trifluoroacetyl-2,8-dioxa-6-azabicyclo[3.2.1]octane-7-carboxylate, J. Chem. Soc. Perkin Trans. 2 2 (1986) 349-353. [41] H.T. Clarke, J.R. Johnson, R. Robison, The chemistry of penicillin, Princeton University Press1949. [42] M. Broto, S. Matas, R. Babington, M.-P. Marco, R. Galve, Immunochemical detection of penicillins by using biohybrid magnetic particles, Food Control 51 (2015) 381-389.
15
Figures Figure 1. A) Antigenic determinants of penicillin G. Three of them (penicilloyl, penicanyl and penicoyl) are covalently bound through the amino group of the lysine of the protein, while the other three (penicillenyl, penicillamine and penamaldate) through a thiol group of the cysteine. The framed structure is the proposed antigenic determinant, named the “penamidyl” epitope, obtained through the carboxylic group of the aspartic or glutamic acid of the protein. B) Structures of the penicillin derivatives used as minor allergenic determinants in the skin tests. Figure 2. HPLC chromatograms at three different pHs (4.6, 7 and 11) of the peptide probe PP1 with (B, D and F) and without (A, C and E) amoxicillin. In blue is represented the PP1, while the by-products of the peptide probe are coloured in grey (Ox.PP1 has MSPP1+16 corresponding to an oxidation by-product. NBD-OH is 4-hydroxy-7nitrobenzofurazan). The pics of the adducts obtained from the reaction between the PP1 and amoxicillin are painted in red. Figure 3. Proposed structures of the three adducts obtained at different pHs from the reaction between the peptide probe PP1 and amoxicillin. Within the square the possible mechanisms of reaction are represented (R1 is the variable side chain of the β-lactam, R2 is the rest of the peptide probe). Figure 4. Graphical representation of the evaluation of the serum samples using the magneto-ELISA. All samples were tested against the three adducts and the peptide probe PP1. The continuous line shows the non-specific signal (n=18), while the dotted lines represent the standard deviation
16
Figr-1
17
Figr-2
18
Figr-3
19
Figr-4
20
S0731-7085(17)31594-7 http://dx.doi.org/10.1016/j.jpba.2017.08.024 PBA 11470
To appear in:
Journal of Pharmaceutical and Biomedical Analysis
Received date: Revised date: Accepted date:
20-6-2017 10-8-2017 17-8-2017
Please cite this article as: Sonia Matas, Marta Broto, Merc`e Corominas, Ramon Lleonart, Ruth Babington, M.-Pilar Marco, Roger Galve, Immediate hypersensitivity to penicillins.Identification of a new antigenic determinant, Journal of Pharmaceutical and Biomedical Analysishttp://dx.doi.org/10.1016/j.jpba.2017.08.024 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Immediate hypersensitivity to penicillins. Identification of a new antigenic determinant. Sonia Matas1,2, Marta Broto1,2, Mercè Corominas3, Ramon Lleonart3, Ruth Babington2,1, M.-Pilar Marco1,2, Roger Galve1,2,# 1
Nanobiotechnology for Diagnostics (Nb4D), Department of Chemical and Biomolecular
Nanotechnology, Institute for Advanced Chemistry of Catalonia (IQAC) of the Spanish Council for Scientific Research (CSIC). 2CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Jordi Girona 18-26, 08034 Barcelona, Spain. 3
Allergology-Internal Medicine, Bellvitge University Hospital, IDIBELL, Feixa Llarga
sn, 08907 L’Hospitalet de Llobregat, Barcelona. CORRESPONDING AUTHOR #
To whom correspondence should be addressed. Phone: 934006100 (Ext. 5117). Fax:
932045904. E-mail: [email protected]
1
Graphical Abstract
Highlights:
A strategy based on the use of peptide probes, small labeled and chemically active peptides, has been used for identification of new allergenic epitopes. A new antigenic determinant of penicillins has been isolated and identified using these peptides probes. A magneto-ELISA has been developed for the detection of specific IgEs from sera of penicillin allergic patients.
Abstract The study of adverse drug reactions (ADRs) constitutes a challenge in the area of Medicine. Drugs generate a large number of the total registered hypersensitivity reactions, where penicillins are responsible for more than half of them. In vitro tests in the market are not efficient enough since they lack in sensitivity and specificity. This is the reason why in vivo tests are carried out, with the subsequent danger to the patient’s life. It is essential to discover new β-lactam antigenic determinants to develop more 2
effective detection systems and thus, obtain better explanations of the allergic mechanisms related to these drugs. We propose a strategy based on the use of “peptide probes”, small labeled and chemical active peptides which have been structurally modified for reacting with the β-lactam moiety at different conditions. The probes also contain a biotin group for application in an immunoassay format. Three different amoxicillin adducts have been obtained, purified and characterized by HPLC-MS and NMR techniques. These results have helped us to elucidate and propose a new antigenic determinant for β-lactams, named the “penamidyl” epitope. All the adducts have been validated and evaluated with sera from different penicillin allergic patients by means of a Magneto-ELISA, immunochemical technique that has allowed us to detect specific IgEs in a very high percentage of the serum samples. An immunoassay has been developed, validated and applied as a diagnostic tool for the detection of specific IgEs in the sera of penicillin allergic patients using a new antigenic determinant.
Abbreviations SI (supporting information), NBD-Cl (4-chloro-7-nitrobenzofurazan), NBD-OH (4hydroxy-7-nitrobenzofurazan), dimethylaminopyridine),
DIC
DIEA
(N,N’-diisopropylcarbodiimide), (N,N-diisopropylethylamine),
DMAP HATU
[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium
(4(1-
3-oxid
hexafluorophosphate), DKP (diketopiperazine), Fmoc (fluorenylmethoxycarbonyl group), tBu (tert-butyl group), DNP (2,4-dinitrophenyl group), TIS (triisopropylsilane), TFA (trifluoroacetic acid), r.t. (room temperature), STV-HRP (streptavidin linked to horseradish peroxidase), MP (magnetic particles), TMB (3,3’,5,5’-tetramethylbenzidine), PVP (polyvinylpyrrolidone), PP (peptide probe), MW (molecular weight), AX (amoxicillin), AP (ampicillin), PG (penicillin G), PV (penicillin V), MDM (minor determinant mixture), PPL (penicilloyl polylysine), HSA (human serum albumin), BSA (bovine serum albumin), PLys (PLL or PL, polylysine), IgE (immunoglobulin E), IgG (immunoglobulin
G),
RP-HPLC
(reverse-phase
high-performance
liquid
chromatography), UPLC-MS (ultra-performance liquid chromatography tandem massspectrometry), NMR (nuclear magnetic resonance).
3
Keywords: β-Lactam Hypersensitivity. Amoxicillin. Antigenic Determinant. Adverse Drug Reactions (ADRs). PenamidylIntroduction Since Alexander Fleming discovered penicillin in 1928, antibiotics have been widely used to treat humans against bacterial infections. The presence of these drugs has been linked to many undesirable consequences such as bacterial resistance [1] and adverse drug reactions (ADRs) [2]. Many of ADRs are predictive reactions, but in some cases drugs can also induce, through immunological mechanisms, hypersensitivity reactions with symptoms that can lead to severe reactions and anaphylaxis. According to the modified classification of Gell-Coombs [3], there are mainly 7 types of hypersensitivity reactions based on different immunological mechanisms and involving the participation of immunoglobulins, complement activation and T cells. Allergy is often associated with a type I hypersensitivity reaction, which is mediated by IgE antibodies produced against an external agent such as pollen, mites, food or drugs that acts as antigen [4]. The most widely used antibiotics are β-lactams due to their broad antibacterial spectrum and low toxicity. These include penicillins, cephalosporins, carbapenems and monobactams, and represent 65% of the world antibiotic market [5, 6]. Allergy to lactams represents about 30% of the total hypersensitivity drug reactions [7, 8], and penicillin and its derivatives are responsible for 75% of death from anaphylaxis [9]. Allergic reactions to -lactam drugs can be classified according to their temporal profile in immediate, accelerated and delayed, which is related with the clinical manifestations and the pathogenic mechanisms [10]. Most of the immediate and accelerated reactions are mediated by IgEs, with clinical manifestations of immediate hypersensitivity [11]. Unlike other allergens, most drugs are low molecular weight compounds incapable of eliciting an immune response. Although the mechanisms involved are complex, for several drugs the hypersensitivity phenomenon has been based mainly on the "hypothesis of the hapten", in which the drug itself or some metabolites covalently binds with an endogenous molecule, usually a blood protein (i.e. HSA, human serum albumin), forming an antigen capable of initiating an immunological reaction [12]. The sites for the binding are not well known and much research has been done to determine the chemical reactions of the haptens. The nature of the drug and the reaction mechanisms influence the outcome of the haptenized protein [13-15]. We can find several examples in the literature of the 4
behavior of this hapten reactivity. For instance, the use of peptide probes has been frequently applied in the study of allergic contact dermatitis (ACD) [16, 17]. In this context, Troutman et al. [18] described the utility of lysine as a nucleophile for screening around 70 contact allergens. The studies done for the identification of antibiotic allergens, and particularly for β-lactams, have been based on similar strategies, either working with whole proteins (HSA, BSA or PLys [19]) or with selected amino acids [20] or dendrimers [21]. Few antigenic determinants have been described until now (Figure 1A), but only the major one (penicilloyl) has been postulated to be the real precursor for this type of allergies. Concerning the specific example of the amoxicillin, one of the most worldwide used β-lactams, only the amoxicilloyl derivative has been isolated. An article recently published by Garzón et al. [22] describes the identification of a possible adduct between amoxicillin and the lysine 190 of the HSA. The diagnostic of penicillin allergy represents in many cases a serious challenge. This is because the potential risk of in vivo methods and the lack or low performance of in vitro tests. Initially, when a patient reports a clinical picture of hypersensitivity, a comprehensive study of the clinical history is essential to judge if the patient has had a reaction related to the drug. The study aims to determine whether the reaction produced is due to a selective allergic reaction against this drug or due to another idiosyncratic adverse reaction [23, 24]. Basically, skin tests and exposure controlled tests using increasing amounts of drug doses are performed [25, 26]. Penicillin skin testing is the most reliable way to evaluate patients for IgE-mediated penicillin allergy. Skin testing is conducted using benzylpenicilloyl polylysine (PPL, major determinant), penicillin G and some penicillin analogues (minor determinants, Figure 1B). Testing with both the major and minor determinants can be identified up to 97% of patients with an immediate hypersensitivity to these drugs. However, in vivo tests are invasive techniques with a certain risk for the patient. Thus, several in vitro diagnostic tests have been developed for the diagnosis of drug allergies, being the immunochemical techniques the most used ones. These methods are based on the detection of the specific IgEs using coated antigenic determinants. Examples of immunoassays are the RAST (radioallergosorbent assay), ELISA (enzyme-linked immunosorbent assay) or FIA (fluorescent immunoassay) [27, 28]. Other in vitro techniques based on cellular studies are the LTT (lymphocyte transformation test), the CAST (cellular allergen stimulation test) and the BAT (basophil activation test) [29]. However, nowadays the use of in vitro diagnostic tests is far from 5
representing an effective and reliable alternative to the in vivo assays. The low sensitivity and the lack of specificity of in vitro tests are due to several causes. Likely, the low concentration of specific IgEs in the serum could be one of the reasons. However, the main cause is probably related to the antigenic determinants used in these techniques. All of the developed immunoassays, including the commercial ones (i.e. 3gAllergy test from Siemens, Immuno CAP test from Thermo-Fisher), are based exclusively on the detection of specific IgEs by means of the penicilloyl determinants. These companies provide "specific tests" for penicillin G, penicillin V, ampicillin and amoxicillin, showing poor sensitivities and low true positives, so currently their values are only complementary to the in vivo tests [30, 31]. Concerning cell tests, they are not well standardized, and their use is exclusively reduced to individual laboratories that apply their own methodologies. The low sensitivity of in vitro tests can explain the discrepancies observed in the diagnosis procedures used for North-American and European specialists [32]. While the former do not consider in vitro tests, European allergist propose their use as a preliminary step to skin tests, mainly in high risk patients [33, 34]. We strongly believe that the problem is related to the lack of knowledge about the immunological mechanisms involved in the -lactam allergy. What is the inducer of the hypersensitivity reaction besides the penicilloyl determinant, and how it has been formed are questions that nowadays are still without convincing answers. Within this work we have tried to clarify a little bit more the complicated world of the penicillin allergy. 1. Materials and methods 1.1. Chemistry. The Instrumentation and reagents section is described in the supporting information (SI). The synthesis and characterization of the different peptide probes is also described in the SI. 1.1.1. Synthesis of the adducts. General procedure. The peptide probes were dissolved in 10 mM buffer solution at three different pHs (4.6, 7 and 11). Previously, all buffers were degassed under reduced pressure to remove the oxygen dissolved. Then, 10 eq. of amoxicillin were added into the solution. The optimization of the reaction between the amoxicillin and the peptide probes is described in the SI. Simultaneously, an amoxicillin and a peptide solution were separately prepared (blanks) and treated at the same way. The 6
solutions were stirred at 37ºC under argon atmosphere during 4 days. All the reactions were followed by UPLC-MS every day, monitoring the wavelength at 470 nm. and the MS at [MWpeptide + MWamoxicillin + H]+. The crudes were purified by preparative HPLC, collecting those peaks that did not appear in the blank samples (see Figure 2). The purified compounds were characterized by MS, and those ones that have the required molecular weight were lyophilized and redissolved with a certain amount of milliQ water. The final concentration of the adducts was calculated by fluorescent spectroscopy. The preparation and characterization of all the adducts is described in the SI. 1.2. Immunochemistry. Evaluation of sera from allergic patients by magnetoimmunoassay (see Figure 4). The Reagents and Instrumentation section is described in the SI. Into vials of 2 ml capacity were added 25 μg of the antibody conjugated magnetic particles (anti-IgE-MP) and washed three times with 400 μL PBST. Next, human sera from penicillin allergic patients and blank serum (100 μL per vial) were incubated for 1h. at 37ºC under agitation. The particles were washed again three times with 400 μL PBST. Then, the adducts and the peptide probe PP1 NBD-N(Me)-DKK(Biotin)βA-OH (200 nM in PBST, 200 μL per vial) were added and incubated for 1h. at r.t. under agitation. After another washing step (400 μL PBST x 3 times), the particles were incubated with a solution of the STV-HRP (0.05 μg mL-1 in PBST, 200 μL per vial) for 30 min. at r.t. under agitation. Particles were washed following the standard procedure and changed to different containers to avoid signal noise due to the STV-HRP adsorbed on the vial surface. Finally, the particles were treated with 200 μL of the substrate solution and incubated for 30 min. at r.t. and light protected. Color development was stopped with 4 N H2SO4 (100 μL per vial), and the absorbances were read at 450 nm. 2. Results and discussion The main objective of this work was the identification of new antigenic determinants of β-lactams. Within this work, amoxicillin was the antibiotic used as a model for the βlactams reactivity. And we chose the approach based on the use of peptide probes because of the possibility to isolate and characterize the amoxicillin adducts. One of the first challenges was the design of the peptide probe, which had to have three differentiate components. The probe should include a lysine as a nucleophilic residue able to react with 7
the drug, a chromophore group for allowing the monitoring and purification of the adducts, and a biotin derivative in order to use the antigenic determinants in an immunoassay format (see Figure S5). Peptide probes were synthesized by solid phase using the Fmoc/tBu strategy. Basically, peptides were built through the N-terminal chain. The polymeric support selected was the NovaPEG Wang resin, which allows the preparation of peptides with a carboxylic acid group at the C-terminal side. The cleavage of the lateral chain protecting groups and the peptide from the resin was done by acidolysis with TFA. And after purification by preparative HPLC, the peptides were lyophilized. The general procedure for the peptide probes synthesis, and all the details of the preparation and characterization of each peptide can be found in the SI. Four different peptides probes were synthetized (Figure S5), one of them (PP1) for use in the immunochemical assays, and the others for the characterization of the adducts. In all cases, the first amino acid (AA1) was β-alanine in order to avoid the formation of a diketopiperazine (DKP) derivative during the peptide synthesis [35]. Peptide PP2 differs from PP1 in the biotin building block, while PP3 only contains the fluorophore building block. The fluorescent dye chosen for PP1, PP2 and PP3 was nitrobenzoxadiazole (NBD) [36] because of its absorbance wavelength (470 nm), far from the wavelengths of the amoxicillin byproducts, the nonreactivity of its heteroatoms, the simplicity of the NMR signals and the low price. Due to the fluorescent building block was not commercially available, the derivatized amino acid NBD-N(Me)-Asp(OtBu)-OH (AA4) was previously synthesized from NBD-Cl and protected aspartic acid (see Section S1.2 at SI), which was chosen for conferring an increase of solubility of the final peptide probe. Aspartic acid had to be N-methylated in order to avoid the peptide probe degradation in alkaline aqueous media [37]. Peptide probe PP4 is very similar to PP2 and only differs in the fluorophore. In that case a dinitrophenol derivative (DNP) was used, and the fluorescent building block was made through a glycine instead from the aspartic acid. Previously to the formation of the adducts, the reaction between the peptide probe and amoxicillin was optimized (Table S5 and Graph S1, in SI). The working concentrations selected were 0.5 mM of the peptide probes and 5 mM of the amoxicillin (10 eq.). All buffers were degassed, the reaction times were 4 days, at 37ºC and under argon atmosphere. Finally, reactions with the peptide probe (PP1) and the β-lactam were carried out at three different pHs (4.6, 7 and 11). Although “physiological conditions” are approximately pH 7 at 37ºC, we decided to work at several pHs because endogenous 8
proteins (HSA) contain polar amino acids (arginine, lysine or aspartic and glutamic acids) which are able to create localized acidic/alkaline environments. The evolution of the reactions was monitored by UPLC-MS, looking for chromatographic peaks with an absorption wavelength of 470 nm, and a MW of 1228 g mol-1 (MWPP1 + MWamoxicillin). Preparative HPLC was used for purification of the adducts. The results are shown in Figure 2, where it can be observed the HPLC chromatograms of the peptide probe PP1 alone (graphs A, C and E) and in combination with the amoxicillin (graphs B, D and F), at different pHs. At pH 4.6 two different adducts (NKAx1 and NKAx2) were found and isolated, while at pH 7 and 11 another one (NKAx3) was obtained. Curiously, the third adduct was obtained in a very low concentration, even less at pH 7. Two other peaks were able to be identified, the degradation product of the chromophore NBD-OH (4-hydroxy7-nitrobenzofurazan) and an oxidation byproduct of the peptide probe (MWPP1 + 16). Characterization of all the adducts is found in the SI. Using this strategy we were able to obtain three different amoxicillin adducts. It is well known and repeatedly reported the synthesis of the amoxicilloyl derivative [38], which is prepared in alkaline conditions with amoxicillin and an amine group, obtaining the formation of an amide bond and the cleavage of the β-lactam ring (Figure 3). We have assumed that NKAx3 is the adduct containing the amoxicilloyl group as it has been made in the same pH conditions reported in the literature. However, any reference has been found related to the adducts obtained at slightly acidic pH conditions. In order to figure out the configuration of the other two adducts, three new peptide probes with different modifications, mentioned above, were synthesized. All the reactions between these probes and amoxicillin were carried out at pH 4.6 and followed by UPLC-MS. Two different adducts (A4 and A5) were isolated using PP2, two other adducts were identified using PP3 (A6 and A7) and only one adduct was obtained with PP4 (A8). The first remark is the reaction between the β-lactam and PP3, where two adducts were obtained without the participation of any nucleophilic group, concluding that the amino group of the lysine is not the responsible of the adduct formation in acidic media. A second remark is the chromophore, which is changed in PP4 and an adduct is still obtained. Summarizing, only one group is common in all the peptides probes, the carboxylic system. And the number of adducts correlates exactly with the number of carboxylic groups of the peptide probes since PP1, PP2 and PP3 have two carboxylic groups and two adducts were obtained, while PP4 has one carboxylic group, and only one adduct was observed. In order to elucidate the mechanisms involved in the formation of the adducts, two of them (A4 and A5) were scaled up and purified by 9
preparative HPLC. The proposed structure for both adducts and their NMR assignations are described in the SI. The resonance confirmed the hypothesis since its
13
C-NMR
chemical shift changed significantly (Adduct A4, C1 from 177.0 ppm to 175.38 ppm. Adduct A5, C2’’ from 175.2 ppm to 173.75 ppm). Moreover, the shifts of the carbons involved in the β-lactam ring were also displaced form the original amoxicillin structure, while the carbons of the lateral chain did not changed. A possible explanation could be a direct transamidation between the nitrogen of the tiazolidinic ring and the acid group of the peptide probe in acidic media, but we could not find any example in the literature. Probably, a more accurate hypothesis was described by Sharma et al. [39, 40] in 1987. The authors postulated the mechanism in two steps (Figure 3). First, a mixed anhydride is likely to be formed by lysis of the β-lactam linkage with an acid group. Second, an intramolecular O→N acyl transfer between the secondary amino group and the anhydride. This mechanism was similarly postulated earlier and described in the book “The Chemistry of Penicillin” [41]. In conclusion, this strategy allowed the preparation of three β-lactam adducts (NKAx1, NKAx2 and NKAx3), two of them presenting a new antigenic determinant never described before in the literature. Casually, the structure of the “penamidyl” group (Figure 1) is very similar to allergenic determinants currently used in skin tests (penicilloyl and penicilloate), since the backbone of the penicillin is preserved and most of its functional groups have not changed. Adducts were synthetized including a biotin group for being applied in the development of immunoassays, since it allows the validation of the adducts and the evaluation of the samples from the allergic patients. We must take into account that the final objective is the detection of specific IgEs using the antigenic determinants obtained with the peptide probes. Initially, the three adducts were validated using both the ELISA and the magnetoELISA formats (Section S2.2 and Figure S9 in SI). Working with magnetic particles can help to minimise the background signal and allow the preconcentration of the serum samples. In both assays, the surfaces were coated with anti-AX-IgG, a polyclonal antisera (As218) produced against the open-ringed form of the amoxicillin [42]. In the magnetoELISA format, the magnetic particles were previously covalently linked to a commercial anti-IgG. Different concentrations of the adducts and the peptide probe were added. The signal was obtained enzymatically using streptavidin linked to HRP (STV-HRP). The results of these validations are shown in Figure S9. In conclusion, in both assays only the adducts were recognized by the specific antibody As218, but not the peptide probe. These 10
results demonstrate that the three adducts contain an active form of the amoxicillin which is able to interact with an antibody raised against this drug. Finally, both assays were applied for the detection of specific IgEs from sera of penicillin allergic patients. In that case, a commercial anti-IgE was used to capture the IgEs, which was absorbed onto the plate wells in the ELISA format or covalently linked to the nanoparticles in the magnetoELISA format (Figure 4). In early experiments, two different types of samples were used, a pool of specific sera and a pool of non-specific sera (blank samples). Samples were tested against the three adducts and the peptide probe PP1. The ELISA format was quickly discarded because of the background signal obtained (data not shown). Basically, the nonspecific absorption onto the wells of the microtiter plates was observed when highly concentrated serum samples were used. Thus, the assays were finally carried out using the magneto-ELISA format, as the particles could be transferred to different containers in order to avoid background noise. The three adducts and PP1 were tested against 15 different serum samples (100 μL per assay) from allergic patients and a blank serum. The clinical history of all the patients is shown in the Table S7. All patients were admitted to hospital due to a clinical picture of hypersensibility to penicillins but, although all patients were positively diagnosed, only 4 of them were positive to a commercial specific IgE test (25% of effectiveness). The results obtained in the evaluation of the samples using the magneto-ELISA are represented in the Figure 4. The normalized signal is calculated as the difference between the signal obtained from each adduct minus the signal obtained from the peptide probe. The continuous horizontal line marks the non-specific signal, which is the average of 18 measurements using the three adducts with the blank serum, while the horizontal dotted lines represent the standard deviation of the mean. Therefore, the samples that show signals over the dotted top line should contain specific IgEs against amoxicillin. Looking at the results, all samples show a specific answer to at least one of the three adducts. More concretely, adduct NKAx1 is recognized by the 87% of the samples, adduct NKAx2 by the 80% and adduct NKAx3 by the 40%. Curiously, the antigenic determinant of the third adduct is the amoxicilloyl (see Figure 1), which has been considered as the major one in the haptenization process and the real cause of βlactam hypersensibility. In the supporting information, the evaluation of sera of allergic patients against the three adducts and PP1 separately (Figure S10) is represented. 3. Conclusions
11
The main contribution of this work has been the isolation and identification of a new antigenic determinant of penicillins, which has been named the “penamidyl” epitope. The strategy has been based on the use of peptide probes, small labeled and chemically active peptides, structurally modified for reacting with the β-lactam molecule. The probes have been synthesized with a biotin group as a linker to be applied in the development of immunoassays. Through the reaction of these peptide probes with amoxicillin, we have been able to isolate three different derivatives or “adducts”. One of them has been obtained at basic media, most probably the “penicilloyl” determinant, and the other two in slight acidic media. It has been proposed that the formation of these two new derivatives occurs through the carboxylic group of the peptide probe and the nitrogen of the β-lactam ring. A magneto-ELISA has been developed and validated, and the immunoassay has been applied as a diagnostic technique to the detection of specific IgEs from sera of penicillin allergic patients. We believe that the strategy of using peptide probes can improve understanding of the mechanisms involved in adverse reactions to drugs. And, although we still have to work more along in this way, we are convinced that the discovery of this new antigenic determinant will improve significantly the diagnosis of hypersensitivity reactions to β-lactams. Acknowledgement This work has been supported by the Ministry of Science and Innovation (MICINN) with the β-Array project (SAF2008-03082). The Nb4D group is a consolidated research group (Grup de Recerca) of the Generalitat de Catalunya and has support from the Departament d’Universitats, Recerca i Societat de la Informació de la Generalitat de Catalunya (expedient: 2014 SGR 1484). CIBER-BBN is an initiative funded by the Spanish National Plan for Scientific and Technical Research and Innovation 2013-2016, Iniciativa Ingenio 2010, Consolider Program, CIBER Actions are financed by the Instituto de Salud Carlos III with assistance from the European Regional Development Fund. The Custom Antibody Service (CAbS) is acknowledged for their assistance and support during antibody production. Marta Broto wishes to thank the FPI-fellowship from the Spanish Ministry of Economy and Competitiveness. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version.
12
References [1] V. Economou, P. Gousia, Agriculture and food animals as a source of antimicrobial-resistant bacteria, Infect. Drug Resist. 8 (2015) 49-61. [2] J. Uetrecht, D.J. Naisbitt, Idiosyncratic adverse drug reactions: Current concepts, Pharmacol. Rev. 65(2) (2013) 779-808. [3] T.V. Rajan, The Gell-Coombs classification of hypersensitivity reactions: a re-interpretation, Trends Immunol. 24(7) (2003) 376-379. [4] http://www.eaaci.org/organisation/about-us/introduction.html. [5] R.P. Elander, Industrial production of b-lactam antibiotics, Appl. Microbiol. Biot. 61(5-6) (2003) 385-392. [6] A.L. Demain, S. Sanchez, Microbial drug discovery: 80 years of progress, J. Antibiot. 62 (2009) 5-16. [7] E. Macy, Penicillin and beta-lactam allergy: Epidemiology and diagnosis, Curr. Allergy Asthm. R. 14(476) (2014) 1-7. [8] P. Demoly, D. Hillaire-Buys, Classification and epidemiology of hypersensitivity drug reactions., Immunol. Allergy Clin. N. Am. 24(3) (2004) 345-356. [9] A.I. Neugut, A.T. Ghatak, R.L. Miller, Anaphylaxis in the United States: an investigation into its epidemiology, Arch. Intern. Med. 161 (2001) 15-21. [10] P. Demoly, A. Romano, Update on beta-lactam allergy diagnosis, Curr. Allergy Asthm. R. 5 (2005) 9-14. [11] M.J. Torres, M. Blanca, J. Fernandez, A. Romano, A. de Weck, W. Aberer, K. Brockow, W.J. Pichler, P. Demoly, Diagnosis of immediate allergic reactions to beta-lactam antibiotics, Allergy 58 (2003) 961-972. [12] A. Ariza, D. Garzon, D.R. Abánades, V. de los Ríos, G. Vistoli, M.J. Torres, M. Carini, G. Aldini, D. Pérez-Sala, Protein haptenation by amoxicillin: High resolution mass spectrometry analysis and identification of target proteins in serum, J. Proteomics 77 (2012) 504-520. [13] E. Meschkat, M.D. Barratt, J.-P. Lepoittevin, Studies of the Chemical Selectivity of Hapten, Reactivity, and Skin Sensitization Potency. 2. NMR Studies of the Covalent Binding of the 13CLabeled Skin Sensitizers 2-[13C]- and 3-[13C]Hex-1-ene- and 3-[13C]Hexane-1,3-sultones to Human Serum Albumin, Chem. Res. Toxicol. 14 (2001) 118-126. [14] M. Aleksic, C. Pease, D. Basketter, M. Panico, H. Morris, A. Dell, Mass spectrometric identification of covalent adducts of the skin allergen 2,4-dinitro-1-chlorobenzene and model skin proteins, Toxicol. In Vitro 22 (2008) 1169-1176. [15] C. Jenkinson, R.E. Jenkins, J.L. Maggs, N.R. Kitteringham, M. Aleksic, B.K. Park, D.J. Naisbitt, A Mechanistic Investigation into the Irreversible Protein Binding and Antigenicity of pPhenylenediamine, Chem. Res. Toxicol. 2009(22) (2009) 6. [16] J. Mutschler, E. Giménez-Arnau, L. Foertsch, G.F. Gerberick, J.-P. Lepoittevin, Mechanistic assessment of peptide reactivity assay to predict skin allergens with Kathon® CG isothiazolinones, Toxicol. In Vitro 23 (2009) 439-446. 13
[17] J. Eilstein, E. Giménez-Arnau, D. Duché, N. Cavusoglu, G. Hussler, F. Rousset, J.-P. Lepoittevin, Sensitization to p-amino aromatic compounds: Study of the covalent binding of 2,5dimethyl-p-benzoquinonediimine to a model peptide by electrospray ionization tandem mass spectrometry, Bioorg. Med. Chem. 16 (2008) 5482-5489. [18] J.A. Troutman, L.M. Foertsch, P.S. Kern, H.J. Dai, M. Quijano, R.L.M. Dobson, J.F. Lalko, J.-P. Lepoittevin, G.F. Gerberick, The incorporation of lysine into the peroxidase peptide reactivity assay for skin sensitization assessments, Toxicol. Sci. 122(2) (2011) 422-436. [19] P. Vemuri, K.E. Harris, L.A. Suh, L.C. Grammer, Preparation of benzylpenicilloyl-polylysine: A preliminary study, Allergy Asthma Proc. 25 (2004) 165-168. [20] H.-P. Ho, R.-J. Lee, C.-Y. Chen, S.-R. Wang, Z.-G. Li, M.-R. Lee, Identification of new minor metabolites of penicillin G in human serum by multiple-stage tandem mass spectrometry, Rapid Commun. Mass Sp. 25 (2011) 25-32. [21] F. Sánchez-Sancho, E. Pérez-Inestrosa, R. Suau, C. Mayorga, M.J. Torres, M. Blanca, Dendrimers as carrier protein mimetics for IgE antibody recognition. Synthesis and characterization of densely penicilloylated dendrimers, Bioconjugate Chem. 13 (2002) 647-653. [22] D. Garzon, A. Ariza, L. Regazzoni, R. Clerici, A. Altomare, F.R. Sirtori, M. Carini, M.J. Torres, D. Pérez-Sala, G. Aldini, Mass spectrometric strategies for the identification and characterization of human albumin covalently adducted by amoxicillin: Ex Vivo studies, Chem. Res. Toxicol. 27 (2014) 1566-1574. [23] R. Mirakian, S.C. Leech, M.T. Krishna, A.G. Richter, P.A. Huber, S. Farooque, N. Khan, M. Pirmohamed, A.T. Clark, S.M. Nasser, Management of allergy to penicillins and other betalactams, Clin. Exp. Allergy 45 (2015) 300-327. [24] N.C.G.C. (UK), Drug Allergy: Diagnosis and Management of Drug Allergy in Adults, Children and Young People, London: National Institute for Health and Care Excellence (UK) (2014). [25] E. Macy, Penicillin allergy: optimizing diagnostic protocols, public health implications, and future research needs, Curr. Opin. Allergy Cl. 15 (2015) 308-313. [26] R. Solensky, D.A. Khan, Evaluation of antibiotic allergy: the role of skin tests and drug challenges, Curr. Allergy Asthma Rep. 14(459) (2014) 1-9. [27] R.I. Siles, F.H. Hsieh, Allergy blood testing: A practical guide for clinicians, Clev. Clin. J. Med. 78 (2011) 585-592. [28] D.G. Ebo, J. Leysen, C. Mayorga, A. Rozieres, E.F. Knol, I. Terreehorst, The in vitro diagnosis of drug allergy: status and perspectives, Allergy 66 (2011) 1275-1286. [29] A.P. Uyttebroek, V. Sabato, M.A. Faber, N. Cop, C.H. Bridts, H. Lapeere, L.S. De Clerck, D.G. Ebo, Basophil activation tests: time for a reconsideration, Expert Rev. Clin. Immunol. 10 (2014) 1325-1335. [30] M. Zidarn, M. Silar, M. Vegnuti, P. Korosec, M. Kosnik, The specificity of tests for anti-blactam IgE antibodies declines progressively with increase of total serum IgE, Wien. Klin. Wochenschr. 121 (2009) 353-356. [31] A. Vultaggio, A. Matucci, G. Virgiliw, O. Rossi, L. Filı´z, P. Parronchi, S. Romagnani, E. Maggi, Influence of total serum IgE levels on the in vitro detection of b-lactams-specific IgE antibodies, Clin. Exp. Allergy 39 (2009) 838-44. 14
[32] M.J. Torres, A. Romano, G. Celik, P. Demoly, D.A. Khan, E. Macy, M. Park, K. Blumenthal, W. Aberer, M. Castells, A. Barbaud, C. Mayorga, P. Bonadonna, Approach to the diagnosis of drug hypersensitivity reactions: similarities and differences between Europe and North America, Clin. Transl. Allergy 7 (2017) 1-13. [33] K. Brockow, H. Garvey, W. Aberer, M. Atanaskovic-Marckovic, A. Barbaud, M.B.e.a. Bilo, E.E.D.A.I. Group, Skin test concentrations for systemically administered drugs - an ENDA/EAACI Drug Allergy Interest Group position paper, Allergy 68 (2013) 702-712. [34] C. Fontaine, C. Mayorga, P.J. Bousquet, B. Arnoux, M.J. Torres, M. Blanca, P. Demoly, Relevance of the determination of serum-specific IgE antibodies in the diagnosis of immediate beta-lactam allergy, Allergy 62 (2007) 47-52. [35] E. Pedroso, A. Grandas, X. de las Heras, R. Eritja, E. Giralt, Diketopiperazine formation in solid phase peptide synthesis using p-alkoxybenzyl ester resins and Fmoc-amino acids, Tetrahedron Lett. 27 (1986) 743-746. [36] Z.-D. Shi, R.G. Karki, S. Oishi, K.M. Worthy, L.K. Bindu, P.G. Dharmawardana, M.C. Nicklaus, D.P. Bottaro, R.J. Fisher, T.R.J. Burke, Utilization of a nitrobenzoxadiazole (NBD) fluorophore in the design of a Grb2 SH2 domain-binding peptide mimetic, Bioorg. Med. Chem. Lett. 15 (2005) 1385-1388. [37] A.A. Aboderin, R.E.K. Semakula, Base-promoted cleavage of a-N-NBD-peptides, FEBS Lett. 34 (1973) 90-94. [38] www.iedb.org/epId/116053 (Immune epitope database and analysis resource). [39] R. Sharma, R.J. Stoodley, A. Whiting, Total synthesis of analogues of the b-lactam antibiotics. Part 3. 2-Ethoxycarbonyl derivaties of carbapen-1-em-exo-carboxylates and carbapenam-3-exocarboxylates, J. Chem. Soc. Perkin Trans. 111 (1987) 2361-2369. [40] J.R. Everett, J.S. Davies, J.W. Tyler, A 1H, 13C, and 19F nuclear magnetic resonance study of rotational isomerism and long-range coupling in methyl (1S,5R,7R)-1-ethyl-3-oxo-6trifluoroacetyl-2,8-dioxa-6-azabicyclo[3.2.1]octane-7-carboxylate, J. Chem. Soc. Perkin Trans. 2 2 (1986) 349-353. [41] H.T. Clarke, J.R. Johnson, R. Robison, The chemistry of penicillin, Princeton University Press1949. [42] M. Broto, S. Matas, R. Babington, M.-P. Marco, R. Galve, Immunochemical detection of penicillins by using biohybrid magnetic particles, Food Control 51 (2015) 381-389.
15
Figures Figure 1. A) Antigenic determinants of penicillin G. Three of them (penicilloyl, penicanyl and penicoyl) are covalently bound through the amino group of the lysine of the protein, while the other three (penicillenyl, penicillamine and penamaldate) through a thiol group of the cysteine. The framed structure is the proposed antigenic determinant, named the “penamidyl” epitope, obtained through the carboxylic group of the aspartic or glutamic acid of the protein. B) Structures of the penicillin derivatives used as minor allergenic determinants in the skin tests. Figure 2. HPLC chromatograms at three different pHs (4.6, 7 and 11) of the peptide probe PP1 with (B, D and F) and without (A, C and E) amoxicillin. In blue is represented the PP1, while the by-products of the peptide probe are coloured in grey (Ox.PP1 has MSPP1+16 corresponding to an oxidation by-product. NBD-OH is 4-hydroxy-7nitrobenzofurazan). The pics of the adducts obtained from the reaction between the PP1 and amoxicillin are painted in red. Figure 3. Proposed structures of the three adducts obtained at different pHs from the reaction between the peptide probe PP1 and amoxicillin. Within the square the possible mechanisms of reaction are represented (R1 is the variable side chain of the β-lactam, R2 is the rest of the peptide probe). Figure 4. Graphical representation of the evaluation of the serum samples using the magneto-ELISA. All samples were tested against the three adducts and the peptide probe PP1. The continuous line shows the non-specific signal (n=18), while the dotted lines represent the standard deviation
16
Figr-1
17
Figr-2
18
Figr-3
19
Figr-4
20