Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

3.1 Introduction

Adverse drug reactions are a major health problem worldwide, but they are difficult to predict. The propensity for an individual to develop a reaction depends on the chemistry of the drugs or chemicals, on environmental factors, and on the biology of the patient. A good proportion of these reactions are immune mediated, which are also called allergic drug reactions or drug hypersensitivity reactions. These drug hypersensitivity reactions are due to several distinct immune mechanisms, but all types eventually involve stimulation of T cells by the drug.

For a T-cell-dependent immune response to occur, the drug must stimulate the T-cell antigen receptor (TCR). Drugs, which are usually small chemical compounds of low molecular weight (less than 1,000 Da) and metal ions, are thought to be too small to be antigenic per se. How they are able to stimulate an immune response often is an important open question left to answer.

The TCR is composed of the TCRαβ (or TCRγδ), CD3εδ, CD3εγ, and CD3ζζ dimers (Alarcon et al. 2003) (Fig. 3.1a). TCRαβ possesses variable immunoglobulin domains that bind the ligand, an antigenic peptide bound to major histocompatibility complex (pMHC) molecules (Garboczi et al. 1996; Garcia et al. 1996) (Fig. 3.1b). Foreign peptides, such as derived from viruses or bacteria, presented by MHC have a high affinity for their appropriate TCR and thus elicit an immune response. Peptides derived from endogenous proteins (self-peptides) are also presented. However, due to negative selection processes in the thymus, mature T cells only show weak affinity to self-peptide-MHC, so that T cells are not stimulated and autoimmune diseases are prevented. In addition, superantigens can bridge MHC and TCR peptide independently and stimulate T cells (Fig. 3.1c).

Fig. 3.1
figure 1

The T-cell antigen receptor (TCR). (a) The TCR comprises the pMHC-binding TCRαβ and the signal-transducing CD3εδ, CD3εγ, and CD3ζζ dimers. (b) Foreign peptide-MHC has a high affinity for the responding TCR. The peptide and the MHC molecule have contacts with TCRαβ, triggering intracellular signaling events, such as the activation of PLCγ and other signaling proteins, leading to T-cell activation. (c) Superantigens simultaneously bind to MHC in a peptide-independent manner and to the constant regions of TCRαβ. Thus, pMHC is bridged to the TCR largely independent of pMHC-TCRαβ contacts. Superantigen stimulation leads to the activation of PLCβ and other signaling proteins, resulting in T-cell activation

Full size image

The CD3 chains contain tyrosine residues in their cytoplasmic tails, that are phosphorylated upon successful ligand binding to TCRαβ and that transmit the signal inside the cell. Phosphorylation of the tyrosines in these cytoplasmic tails is the critical event initiating signaling cascades. Phosphotyrosines serve as binding sites for signaling proteins with Src-homology 2 (SH2) domains.

The molecular mechanism as to how ligand binding induces CD3 phosphorylation is still a matter of debate. In the conformational change models (Aivazian and Stern 2000; DeFord-Watts et al. 2011; Gil et al. 2002; Minguet et al. 2007; Schamel et al. 2006; Xu et al. 2008), pMHC binding induces a structural change in CD3 that enables phosphorylation of the cytoplasmic domains. In other models the kinase-phosphatase equilibrium is disturbed in the vicinity of the TCR by excluding phosphatases (Choudhuri and van der Merwe 2007; Choudhuri et al. 2005; Davis and van der Merwe 2006; James and Vale 2012) or by enhancing the concentration of kinases (Boniface et al. 1998; Cochran et al. 2000; Irvine et al. 2002). These models are discussed in detail in the preceding review “Activation of the TCR complex by peptide-MHC and superantigens.”

Here, we want to discuss models, which have emerged in the recent years to explain how small chemicals or metals can stimulate the TCR.

3.1.1 Chemicals Acting as Haptens

The term “hapten” was introduced by Landsteiner and Jacobs in 1935. Haptens are chemically reactive compounds that form a covalent bond with endogenous proteins. In 1992 it was demonstrated that hapten recognition by T cells required covalent hapten attachment to MHC-associated peptides (Ortmann et al. 1992). In general hapten-modified proteins are processed by the antigen-presenting machineries and haptenated peptides displayed on MHC class I or II to T cells. In contrast to self-pMHC, which only weakly binds to the TCR and does not stimulate T cells (Fig. 3.2a), haptenated self-pMHC can possess strong binding to an appropriate TCR, and the T cell is stimulated (Fig. 3.2b). These haptenated self-pMHCs were absent during thymocyte development and negative selection, so that the specific TCRs reacting to these haptenated self-pMHCs were not removed from the T-cell population. One example for haptens are β-lactam antibiotics, such as penicillin, which covalently bind to lysine residues of many proteins, such as serum albumin (Batchelor et al. 1965; Jenkins et al. 2009; Levine and Ovary 1961; Schneider and De Weck 1965). One study, using the synthetic penicillin Flucloxacillin, showed that the modification of the lysine residues in human serum albumin occurs in a dose-, time-, and site-dependent manner (Jenkins et al. 2009). However, the exact mechanisms for the immune responses to penicillin are still not clear and create difficulties in the development of effective diagnostics methods (Blanca et al. 2009).

Fig. 3.2
figure 2

pMHC-TCRαβ interactions mediated by small chemical compounds I. (a) Negative selection in the thymus ensures that self-peptide-MHC only has a weak affinity for TCRs in peripheral T cells. Self-pMHC does not perfectly fit to TCRαβ, thus not triggering their TCR. (b) Haptens bind covalently to endogenous proteins, thus generating haptenated self-peptides that together with MHC can form high affinity ligands for the TCR. (c) Transitional metal ions or drugs can non-covalently bind to the common self-peptide-MHC surface and thus potentially generate a high affinity ligand for the TCR. (d) Metal ions or drugs might also non-covalently bind to the MHC only. This could also form a complementary structure to some TCRαβ, forming a high affinity ligand. (e) Similar to superantigens, Fig. 3.1c, metal ions or drugs might bridge MHC with TCRαβ with high affinity without the involvement of the peptide. Thus, a large fraction of MHC might become competent in stimulation of TCRs. The stimulation of PLCγ (or PLCβ in case of superantigen-like ions/drugs) and other intracellular signaling molecules induced by high affinity TCR binding is indicated. This leads to activation of the T cell

Full size image

3.1.2 Chemicals Acting as Prohaptens

Prohaptens are chemicals that only become active after a metabolic reaction. Often bioactivation of prohaptens involves oxidative processes, with the cytochrome P450-dependent metabolism playing a major role. In the liver, for example, sulfamethoxazole is converted by P450 to sulfamethoxazole hydroxylamine (Cribb and Spielberg 1992), and autooxidation of the latter compound generates the metabolite nitroso sulfamethoxazole (SMX-NO), which reacts with cysteine residues of proteins (Callan et al. 2009). Indeed, the peptides derived from the SMX-NO-modified proteins after antigen processing can possess high affinity for the appropriate TCR and thus are potent antigens to the specific T cells (Castrejon et al. 2010). Prohaptens are a major problem for drug development as the different metabolic systems in a whole organism have to be investigated to make sure that an initially nonreactive drug does not become reactive upon metabolism.

3.1.3 Haptens as Tools in Basic Research

Traditionally haptens have been an important tool in studying the immune response. For example, trinitrophenol and dinitrophenol have been used to demonstrate the exquisite specificity of the immune system (Weltzien et al. 1996). More recently, haptenated peptides were synthesized to generate TCR antigens with defined properties. For example, photoreactive 4-azidobenzoic acid on the lysine of the peptide SYIPSAEKI was used and the haptenated peptide bound to MHC class I. Specific TCRs could bind, and using a short pulse of UV light, a covalent link between the peptide-MHC molecule and the bound TCR was made, thus “freezing” the otherwise transient peptide-MHC-TCR interaction (Gregoire et al. 1996; Luescher et al. 1995). Using this system we could show that bivalent peptide-MHC binding was required to induce an activatory conformational change in the CD3 subunits of the TCR (Minguet et al. 2007).

3.1.4 Transitional Metal Ions

Nickel (Ni) allergy is the most common and best studied of all metal hypersensitivities, again stimulating T cells. Reactions to other metals such as cobalt (Co), chromium (Cr), platinum (Pt), beryllium (Be), mercury (Hg), and gold (Au) are also reported. These transitional metals are only active as ions (e.g., Ni2+) after their oxidation that can take place on the skin. Metal ions cannot form covalent bonds to peptides and thus are not classical haptens. Instead, metal ions form geometrically highly defined non-covalent coordination bonds with four or six electron donors. The electron donors are mainly nitrogen or oxygen in the amino acid side chains of proteins (Zhang and Wilcox 2002). Thus, metal ions can form very specific complexes with proteins. Ni is studied best (Thierse et al. 2005) and thus serves as an example here.

T-cell activation by Ni-treated APCs did not require antigen processing in some cases (Moulon et al. 1995), being in contrast to the classical haptens. In some cases, Ni bound to the MHC molecule and the peptide, thus forming a new surface to be recognized by the TCR (Fig. 3.2c) (Lu et al. 2003). This might require a certain MHC haplotype, a limited set of peptides and some specific TCR. However, if the hypervariable regions of the TCR are not required to make the contact, then a large number of different TCRs (e.g., those containing a certain Vα and/or Vβ segment) might be stimulated (Vollmer et al. 1997). Unfortunately, a crystal structure of a pMHC-Ni-TCRαβ complex does not exist. However, mimotopic peptides have been used to substitute for the Ni and the self-peptide. The structure shows that the same diagonal orientation of the pMHC-TCRαβ interaction as for the classical pMHC-TCRαβ exists (Yin et al. 2012). This might suggest that the canonical PLCγ pathway is used (Fig. 3.2c). Unfortunately, it is unknown whether metal ions also allow other orientations than the diagonal one. In this case the interaction might resemble more the one of superantigen-mediated TCR activation and thus also activate PLCβ (Bueno et al. 2006, 2007).

It could also be possible that the metal ion only contacts MHC and TCR (Fig. 3.2d).

In another case, Ni activated the TCR with the correct MHC haplotype, but independent of any peptide (Gamerdinger et al. 2003). Thus, it was proposed that Ni acted in a similar manner as superantigens (Fig. 3.2e). However, in this case the hypervariable regions of the TCR were used as a contact site; thus, only very few TCRs might be activated. This is in contrast to superantigens that can activate all TCRs of a given Vβ region (Fraser and Proft 2008; Petersson et al. 2004). It was suggested that self-pMHC first binds to the TCR and that Ni then stabilizes the complex, in order to generate a high affinity interaction (Thierse et al. 2005). Whether in these cases PLCβ is simulated is unknown.

3.1.5 The “Pharmacological Interaction of Drugs with Immune Receptors (p-i) Concept”

The mechanisms of generating high affinity pMHC ligands for the TCR that we have just discussed for metal ions (Fig. 3.2) might also hold true for small organic compounds that do not covalently bind to peptides but still activate T-cell responses. These mechanisms were proposed in the “pharmacological interaction of drugs with immune receptors (p-i) concept” (Adam et al. 2011; Pichler 2005) and have gained much experimental support in the last years.

This model accounts for the observation that T cells could be activated, in an MHC-dependent mechanism, even in the presence of glutaraldehyde-fixed APCs (unable to process antigens) and the nonreactive drug (Schnyder et al. 1997; Zanni et al. 1998). Furthermore, the speed in which T cells could be activated (visualized by calcium influx) after introduction of the drug was too fast for antigen processing to have occurred (Zanni et al. 1998). Thus, these chemicals form reversible, non-covalent bonds (electric, van der Waals, hydrophobic, and hydrogen bonding forces) with pMHC and TCRαβ.

3.1.6 Drugs Binding Non-covalently to pMHC

Strong associations between drug hypersensitivity reactions and specific HLA alleles (human leukocyte antigen, HLA, is the name for human MHC) have been described, although the exact mechanisms for the TCR-stimulating activity of the drugs are often unclear. For example, strong associations between HLA-B*5801 and allopurinol-induced or HLA-B*1502 and carbamazepine-induced Stevens-Johnson syndrome or between HLA-B*5701 and flucloxacillin-induced reactions have been reported (Chung et al. 2004; Daly et al. 2009; Tassaneeyakul et al. 2009). In fact, carbamazepine seems to bind directly but non-covalently to endogenous peptide-loaded HLA-B*1502 (Wei et al. 2012; Yang et al. 2007). Thus, the pMHC-binding mechanisms shown in Fig. 3.2c, d might hold true. The 5-carboxamide chemical moiety of carbamazepine was seen to be required for presentation with HLA-B*1502, and three key residues in the peptide-binding groove of HLA-B*1502 were identified (Wei et al. 2012).

3.1.7 Drugs Altering the Self-Peptide Repertoire Bound to MHC

The antiviral drug abacavir causes severe adverse reactions exclusively in HIV-infected individuals expressing HLA-B*5701. As one example for personalized medicine, it is now common practice to genotype a patient for HLA-B*5701 before prescribing abacavir. The adverse reactions are mediated by the activation of cytokine-producing cytotoxic CD8+T cells, and the specificity of the abacavir-HLA interaction was mapped to the F pocket of the peptide-binding cleft of the MHC molecule (Chessman et al. 2008). Several groups showed recently that abacavir binds non-covalently to amino acid residues within the F pocket and thus changes the shape and chemistry of the cleft (Illing et al. 2012; Ostrov et al. 2012), changing the repertoire of self-peptides bound to and presented by HLA-B*5701 (Fig. 3.3a). The X-ray crystal structure of HLA-B*5701 bound to pep-V in the presence of abacavir suggests that the peptide is bound in the MHC molecule in a normal antigen conformation allowing for conventional pMHC-TCR interaction (Ostrov et al. 2012). So, self-peptides, that can bind to HLA-B*5701 only in the presence of abacavir, will then form new pMHC complexes not present in the thymus during negative selection of T cells. Thus, the abacavir hypersensitivity syndrome is a model of drug-induced autoimmunity in which the drug alters the self-peptide repertoire presented by MHC and so drives responses of T cells recognizing these neo-self epitopes (Illing et al. 2012; Norcross et al. 2012; Ostrov et al. 2012).

Fig. 3.3
figure 3

pMHC-TCRαβ interactions mediated by small chemical compounds II. (a) In at least one documented case, the drug binds to the peptide-binding cleft of the MHC molecule. This allows peptides that otherwise cannot bind to the MHC, to bind and be presented. Since these new self-peptides are not presented in the thymus, reactive T cells are present in the periphery. (b) A drug or metal ion might influence proteins of the antigen processing pathways, such that new self-peptides are made and presented on MHC. (c) As seen in Fig. 3.2, metal ions or drugs can bind to pMHC and thus create a high affinity docking site for specific TCRs. (d) The same pMHC-drug-TCRαβ complex as in (c) can be accomplished, if the drug (or metal ion) binds to TCRαβ. TCR-induced downstream signaling via PLCγ and other pathways is indicated

Full size image

A different mechanism for the presentation of new self-peptides was suggested for metal ions, in cases where antigen processing was required to stimulate Ni-specific T cells by APCs. It was suggested that the presence of Ni altered antigen processing, so that new self-peptides are presented (Fig. 3.3b). Thus, the T cells might recognize these Ni-induced cryptic self-peptides, but not Ni itself (Griem et al. 1998).

3.1.8 Drugs Binding to TCRαβ

In principle drugs that bridge pMHC and TCRαβ (Fig. 3.2c, d, e) could have a higher affinity for pMHC or a higher affinity for TCRαβ, thus either first binding to pMHC or to TCRαβ before the pMHC-drug-TCRαβ complex forms. At first sight, it might seem irrelevant for the complex, and thus for T-cell activation, whether pMHC or TCRαβ binds first (Fig. 3.3c, d). However, the flexibility of the large TCRαβ hypervariable loops reduces the on-rate of the pMHC-TCR interaction and thus might act negatively on T-cell activation (Aleksic et al. 2010). If the drug binds first to TCRαβ and thereby fixes the conformation of the hypervariable loops in the pMHC-permissive binding state, then the on-rate would be enhanced and T-cell activation would be favored. Thus, although hypothetical, drug binding to TCRαβ first might have different effects than binding to pMHC first.

The antibacterial sulfonamide sulfamethoxazole (SMX) is one of the drugs suggested to bind mainly or firstly to the TCR (in addition, to be a prohapten). TCRs from several SMX-specific T-cell clones isolated from patients with hypersensitivity to SMX reacted to SMX only in the presence of APCs indicating that the TCR stimulation by the drug was MHC dependent but antigen processing independent (Depta et al. 2004). Recent modeling of the drug-TCRαβ interaction suggested several candidate SMX binding sites on the CDR2 and CDR3 hypervariable loops of the TCRα and TCRβ chains (Pichler et al. 2011). Thus, SMX might be one example in which the drug binds stably to TCRαβ (Fig. 3.3d). Since the MHC-bound peptide did not affect the reactivity of SMX-specific T cells (Burkhart et al. 2002), the drug-TCRαβ complex might bind to the MHC molecule without the involvement of the peptide, thus resembling the way that superantigens use to bride TCRs to MHCs (Figs. 3.1c and 3.2e) with the difference that superantigens do not bind to the CDR loops of TCRαβ.

3.1.9 Drug-Induced MHC-Independent TCR Triggering

TCRs can be activated independently of pMHC, such as multivalent anti-TCRαβ and anti-CD3 antibodies (Chang et al. 1981; Kaye and Janeway 1984) or activation of a chimeric TCR by artificial ligands (Minguet et al. 2007). Even in vivo MHC-independent TCR activation was shown. In mice deficient for MHC class I, MHC class II, the coreceptors CD4 and CD8 thymocyte selection produced mature αβ T cells recognizing ligands independently of MHC (Van Laethem et al. 2007). The TCRs of two T-cell clones derived from these mice recognized glycosylation-dependent epitopes of the self-protein CD155 (Tikhonova et al. 2012). Thus, it is possible that drugs could bind to some endogenous proteins that are not pMHC, thereby creating new ligands for the TCRαβ or CD3 that could stimulate T cells in an MHC-independent manner.

3.1.10 The Second Signal of T-Cell Activation

The theory stipulating that an immune cell needs two signals to be activated was first developed in 1970 (Bretscher and Cohn 1970). TCR triggering is an important step in naïve T-cell activation, but it is usually not sufficient to release the T cells (αβ T cells) from their quiescent state. If TCR triggering (signal 1) is the only signal, the naïve cells usually become anergic and cannot be stimulated further (Jenkins et al. 1990; Schwartz 2003). To avoid the anergic, nonresponsive state, naïve T cells require signal 2, provided by a costimulatory receptor such as CD28 (Rudd et al. 2009). In addition, other members of this family (CD2, ICOS) or of the tumor necrosis factor receptor family (including OX40 and 4-1BB) have been shown to be critical as stimulatory co-signals for T-cell activation (Sharpe 2009). So, a small chemical should require the presence of a costimulatory signal, in order to stimulate naïve T cells inducing proliferation and effector functions. Very often, in drug/metal hypersensitivities, the drug or metal also activates the innate immune system leading to the expression of costimulatory ligands by the APCs. Alternatively, the second signal could come from an ongoing infection or autoimmune reaction that takes place at the same time.

However, it is quite possible to imagine that some chemicals would directly stimulate effector or memory T cells where the requirement for a costimulatory signal is smaller (Boesteanu and Katsikis 2009; Kannan et al. 2012), thus overcoming the requirement for simultaneous stimulation of the innate immune system. In addition, effector or memory T cells have a lower affinity threshold for activation than naïve T cells (Bachmann et al. 1999; Cho et al. 1999; Iezzi et al. 1998; Kedl and Mescher 1998; Zimmermann et al. 1999) and respond to lower amounts of antigen than naïve T cells (Kimachi et al. 1997; London et al. 2000; Pihlgren et al. 1996; Rogers et al. 2000). The increased sensitivity to antigenic stimulation by effector and memory cells is partly caused by enhanced pre-clustering of the TCR (Kumar et al. 2011). This pre-clustering increases the avidity towards pMHC (Molnar et al. 2010, 2012). So an effector or memory T cell could be activated by a chemical without the need for a high affinity pMHC interaction with the TCR or even the need for costimulation.

3.2 Conclusion

The molecular mechanisms that small chemical compounds or metal ions use, in order to generate novel peptide-MHC surfaces that can bind with high affinity to TCRs, is an important topic in pharmacologic research. Beginning with the discovery that haptens can covalently bind to proteins whose peptides are presented by MHC in the last century, up to the recent demonstration that a drug altered the kind of self-peptides that are presented by a certain MHC haplotype, a number of different mechanisms have emerged. We believe that novel mechanisms will be discovered in the next years and decades. Detailed insight into these mechanisms might help in treating drug hypersensitivity reactions.