Werner J. Pichler, M.D., Adverse Drug Reactions – Analysis and Consulting (ADR-AC), 91 Holligenstrasse, CH-3008 Bern, Switzerland. Email: firstname.lastname@example.org
In drug hypersensitivity, change of drug treatment and continuation with a new drug may result in reappearance of drug hypersensitivity symptoms. This is not uncommon in patients with chronic infections requiring continued and long-lasting antibiotic treatments. For the clinician, the question arises whether these symptoms are due to cross-reactivity, are due to a new sensitization or are a reflection of a multiple drug hypersensitivity syndrome. Based on the p-i concept (pharmacological interaction with immune receptors), we propose that the efficient stimulation of T cells by a drug is the sum of drug–T-cell receptor affinity and readiness of the T cell to react, and therefore not constant. It heavily depends on the state of underlying immune activation. Consequently, drug hypersensitivity diseases, which go along with massive immune stimulations and often high serum cytokine values, are themselves risk factors for further drug hypersensitivity. The immune stimulation during drug hypersensitivity may, similar to generalized virus infections, lower the threshold of T-cell reactivity to drugs and cause rapid appearance of drug hypersensitivity symptoms to the second drug. We call the second hypersensitivity reaction a “flare-up” reaction; this is clinically important, as in most cases the second drug may be tolerated again, if the cofactors are missing. Moreover, the second treatment is often too short to cause a relevant sensitization.
Drug hypersensitivity (DH) reactions are unusual and complicated diseases, both clinically and mechanistically: they are not elicited by an infectious agent, but by immune reactions to small molecules. However, small molecules like drugs, which often have a size below 1000 D, are actually too small to be immunogenic per se and are assumed to be unable to elicit an immune reaction. How is it then possible that immune reactions to drugs develop? One hypothesis postulates that drugs causing allergies have the ability to bind covalently to larger proteins.1–3 This makes them immunogenic and antigenic. The resulting immune mechanisms are very heterogeneous and result in a wide variety of diseases.4,5
A further level of complexity of DH is based on the pharmacological activity of the drug itself: it was found that even drugs unable to bind covalently to proteins may bind to T-cell receptors (TCR) and/or major histocompatibility complex (MHC) molecules, as they bind to other receptors.6,7 This pharmacological interaction model does not require that the drug binds first covalently to proteins/peptides. Instead, this drug binds directly to immune receptors by van der Waals forces, electrostatic and hydrogen bonds among others. This binding is actually similar to TCR interactions with peptides presented by MHC molecules. Under certain circumstances this rather labile interaction of a small drug with a certain TCR or MHC molecule stimulates T cells. Stimulation may occur only if: (i) the drug has a certain affinity for the particular TCR; (ii) a supplementing interaction of the TCR with the MHC molecule takes place; and (iii) if the T cell is ready to react to such a minor signal given by the drug–TCR interaction. Details of this so-called p-i concept (pharmacological interaction with immune receptors) has been described and reviewed previously.6–8
For the following discussion, one has to keep in mind that:
1 Only pre-activated T cells (memory and effector T cells) react to drugs directly, as these memory T cells have a lower threshold for reactivity.9 Factors increasing the T-cell reactivity may also increase reactivity to drugs.
2 Such p-i concept-based T-cell stimulations result in an isolated T-cell response, as the drug selectively interacts with T cells only, but may not stimulate dendritic cells or form antigenic determinants for B cells.8 This is in contrast to classical hapten responses, which often elicit simultaneously a T- and B-cell immune response.
3 It is actually not an own immune response to the drug, but some kind of pharmacological stimulation of T cells, although the clinical picture of such reactions is similar to classic drug allergies.
Pursuing the idea of the p-i concept consequently leads to additional considerations:
4 The direct drug–immune receptor interaction may not always be stimulatory; actually, the majority of drug–TCR interactions may be functionally irrelevant, and some may even act more through blocking the “normal” interaction of specific TCR with peptide–MHC complexes.
5 Some drug–TCR interactions may actually be incomplete, and also the arising symptoms may be incomplete (or minor/transient).
6 Other functionally relevant surface structures may also be targeted by some drugs, although their heterogeneity is low in comparison to immune receptors. The consequences of interactions of drugs with these surface structures is unclear.
Are these considerations important for the clinician? They may well be so. Actually, many symptoms and complaints of patients following drug therapy like a rash, pruritus or a flash reaction are often rather elusive, transient and not always reproducible. Skin or in vitro tests may be negative in spite of a very suggestive history and symptoms. The reproducibility is low, possibly because the underlying immune reactivity alters. If immune-stimulatory cofactors (e.g. a virus infection) are missing, the T-cell reactivity may alter as well: a drug, previously causing some exanthema, may be tolerated again. Such reactions may contribute to confusion and misinterpretation.
Here, we will briefly present the concept of “flare-up” reactions, which is related to the immune stimulation by drugs and differentiate these reactions from cross-reactivity and multiple drug hypersensitivity.
Drug hypersensitivity reactions are facilitated by generalized immune stimulations: typical examples are herpes virus infections like Epstein–Barr virus (EBV), cytomegalovirus (CMV) or other herpes virus infections (e.g. HHV-6) as well as HIV infections.4,10 These infections appear to increase the incidence of DH, and may also increase the severity of reactions, as severe cutaneous DH like Stevens-Johnson syndrome (SJS)/toxic epidermal necrolysis (TEN) appear substantially more in HIV-infected than non-infected persons.10
The reason for this enhanced incidence of drug hypersensitivity is not clear. One hypothesis postulates that the generalized immune activation facilitates the immune response to drugs.4 It is well known that the virus infections mentioned stimulate the immune system massively; they go along with elevated cytokine levels in the serum and tissues;11 in addition, an expansion of activated T cells is observed in the circulating blood. Interestingly, systemic drug hypersensitivity reactions like severe maculopapular drug rashes as well as drug (induced) hypersensitivity syndrome (D[i]HS) (drug rash with eosinophilia and systemic symptoms [DRESS]) show a similar immune activation and high cytokine levels.12–14 Are severe drug hypersensitivity reactions also a risk factor for DH? Also, do DH patients have a lower threshold to react to a new drug?
Some reports14,15 and our own clinical experience support this assumption. The early switch of an antibiotic therapy in a patient with DH symptoms to another antibiotic can be associated with a flare-up of the DH. The second drug may exacerbate the existing drug allergy (more skin symptoms in exanthema, or liver enzyme elevations, flare-up of body temperature, often already after 1 or 2 days of treatment with the new drug). Often the new or – more frequently – enhanced “rash” is interpreted as a new allergy to the novel antibiotic. The features of such a DH flare-up reaction are: (i) treatment with a potentially allergenic substance in a patient who already shows signs of an ongoing DH (mainly exanthemas); and (ii) rapid appearance of symptoms (1–2 days) or actual aggravation of pre-existing symptoms (Fig. 1).
The new symptoms often prompt stopping the second drug as well, and the patient is assumed to be allergic to two drugs. However, two possibilities may result. The patient may be sensitized to the first drug, to which he/she was exposed longer, but not the flare-up-causing second drug. Testing after remission might reveal only a sensitization to the first antibiotic, while the second is well tolerated in provocation test or later re-exposure (own observation). Alternatively, and more rarely, the allergic reactivity may have developed to both drugs. Which factors determine the difference between double and single sensitization in flare-up reactions after DH is at present unclear. Basically, one could assume that a longer exposure to the second antibiotic would favor a second sensitization.
The situation is similar to the maculopapular exanthems which can be observed in patients with acute EBV infection and receive aminopenicillins. Most of the time no sensitization can be found to the aminopenicillins and these drugs are later tolerated again; only occasionally the EBV and amoxicillin exposure suffices to cause a persistent sensitization.
The detailed mechanism of this flare-up reaction is not clear. The clinical observation of a flare-up reaction and the previously mentioned cofactors for DH by viral infections and so forth suggest that the efficient stimulation of T cells by a drug is the sum of drug–TCR affinity and readiness of the T cell to react (Fig. 2). The p-i concept, namely that drugs directly stimulate T cells, may help to understand this flare-up reaction, which often appears quite rapidly and probably too fast for allowing an own immune response to develop.
Applying the p-i concept to flare-up reaction would mean that the new drug would interact with some of the pre-activated T cells with an affinity sufficient to cause a T-cell activation, expansion and exanthema. Many TCR have the principal ability to interact with drugs, probably with different affinities. If the immune system is resting, only drug–TCR interactions of high affinity may be able to stimulate T cells sufficiently to cause a T-cell expansion. If sufficient T cells react, symptoms might arise. But even such a high affinity interaction may remain unnoticed if too few cells are stimulated. If the immune system is activated and the readiness to react increased (lower threshold), even relatively low affinity drug–TCR interactions might suffice to activate many T cells and symptoms arise. After the cofactor (i.e. stimulation of immune system by a virus or prior DH) is eliminated, the low affinity binding drug remains negative in skin tests and is later well tolerated, when the co-stimulatory conditions, which reduced the threshold level to react, are not present any more. This hypothesis might also explain the transient nature of many exanthems to antibiotics in childhood, as the presumably viral cofactors are missed during testing, provocations or re-exposure.
Can one prevent this flare-up reaction after DH? One possibility is to delay the implementation of the second drug for a few days to allow the immune system to calm down. Second, one may more aggressively treat the first DH with, for example, systemic steroids (0.5 mg/kg prednisolone for ∼3 days) and thereby suppress the immune reactivity. Third, one may introduce the second drug together with moderate doses of steroids (0.1–0.2 mg/kg) for a few days, to prevent an additional reaction.
Many drugs belong to a certain drug class, which have structural similarities and differ in side chains and so forth. The successful use of a drug like ciprofloxacin may promote the launch of similar drugs like moxifloxacin or gemifloxacin, some of which may have advantages in terms of efficacy, side-effects, dosage or kinetic. A reactivity of the immune system to one compound of the drug class may result in a certain cross-reactivity to related compounds. This cross-reactivity of the immune system with structurally similar drugs depends on: (i) structural similarity; and (ii) precursor frequency of drug-specific T (or B) cells. In patients with rather severe, delayed appearing DH, the T-cell precursor frequency is approximately 1:2000–1:10 000 T cells, mostly detectable for years after the acute event.16 Drugs which might differ in only an OH group are more cross-reactive than drugs which are different in a longer side-chain. In older studies, we observed already that all T-cell clones (TCC) reactive with amoxicillin do also react with ampicillin, which differs in only an OH group.17 But only a fraction cross-reacted with penicillin G and even less with piperacillin. No cross-reactivity of these amoxicillin-specific TCC was seen with cephalosporins.
Clinical symptoms arise when drug-specific T cells expand sufficiently to reach a certain number at the site of inflammation and which is sufficient to orchestrate an inflammation. This relationship between reactivity and number of reactive T cells is also detectable in skin tests where, for example, in an lidocaine-sensitized patient, mepivacaine (highly cross-reactive) induced a positive patch test reaction simultaneously with lidocaine (48 h), while bupivacaine, which is less cross-reactive with lidocaine, stimulated a skin-reaction after 72–96 h.18,19 Expansion of the few bupivacaine cross-reactive T cells needed more time than the many lidocaine- or mepivacaine-specific T cells to cause skin test reactivity.
If only a small fraction of drug-specific T cells do cross-react, then these cells will be stimulated and expanded during the drug treatment. Depending on the initial amount of drug-specific T cells and the speed of cell expansion, hypersensitivity symptoms will arise: the lower the starting precursor frequency is, the later the symptoms may arise. If in patients with low precursor frequency the treatment with a cross-reactive drug is stopped early enough before a certain amount of cells is reached, no symptoms may become apparent (but the patients may be broader sensitized than before).
However, one has to be aware that these considerations for T-cell reactions do not apply to immunoglobulin (Ig)E-mediated reactions. Cross-reactivity in IgE-mediated reactions is more problematic, as in sensitized individuals already small amounts of drugs (which are normally haptens) may suffice to cross-link drug-specific IgE on mast cells by hapten–carrier complexes. Therefore, acute symptoms, reflecting cross-reactivity may arise even if the amount of drug is very low and the level of cross-reactivity is limited and restricted to a small part of drug-specific IgE.
Multiple drug Hypersensitivity
The term multiple drug hypersensitivity is used for different forms of side-effects to various drugs.20 Some use it to characterize patients with multiple drug intolerance (e.g. “pseudoallergy” to various non-steroidal anti-inflammatory drugs), others reserve this term for well-documented, repeated and clearly immune-mediated reactions to structurally unrelated drugs.21,22 Cross-reactivity due to structural similarity is not included.
According to our experience, approximately 10% of patients with well-documented drug hypersensitivity (skin and/or lymphocyte transformation test positive) have multiple drug allergies;23 for example, a patient reacts to injected lidocaine with a massive angioedema and, years later, the same patient develops a contact allergy to corticosteroids. Alternatively, a patient reacts to amoxicillin, phenytoin and sulfamethoxazole within a few months, but with different symptoms (exanthema, DiHS/DRESS, erythrodermia). Most patients have had rather severe reactions to at least one drug. An IgE-mediated reaction might be followed by a T-cell-mediated reaction. However, in our experience, multiple T-cell reactions with distinct clinical pictures developed, but not IgE and T-cell reactions together.
The reason for this accumulation of DH in one individual is unknown. A possible explanation might be a deficient tolerance mechanism against small chemical compounds.24 An immune reaction to a drug – be it initiated through hapten or p-i mechanism – can be seen as a failure of tolerance, and the same patient might not only be prone to develop other drug allergies as well but also autoimmunity. Indeed, a previous drug allergy might be a risk factor for the development of a delayed hypersensitivity reaction to contrast media.25
Preliminary data suggest, however, that no defect of regulatory T cells (Foxp3+) can be detected in patients with multiple drug hypersensitivity, The drug reactive CD4+ T cells show, however, signs of a pre-activated state (Daubner et al., unpubl. data, 2010). This pre-activation seems to persist for years in spite of drug avoidance and may explain the enhanced readiness to react to new drugs, which are not cross-reactive.
In conclusion, the efficient stimulation of T cells by a drug can be seen as the sum of drug–TCR affinity and readiness of the T cell to react. This readiness of T cells to react to drugs is influenced by the general level of immune activation, and thus highly variable. The variability of these two features is well documented in the “flare-up” reactions after a prior DH (level of T-cell activation), as well as in cross-reactivity (level of affinity and number of precursor cells) and to a certain extent multiple drug hypersensitivity (level of cell activation).
The drug hypersensitivity group at the University of Bern is supported by SCAHT and Pfizer.