Factors involved in the absence of sedative effects by thesecond‐generation antihistamines

Histamine is a monoamine with a wide cellular distribution and diverse functions in man. In the central nervous system (CNS), histamine is synthesized by a restricted population of neurons originating in the tubermammillary nucleus of the posterior hypothalamus, where it acts as a neurotransmitter via G-protein-coupled receptors. These neurons project diffusely to most cerebral areas and are thought to be involved in brain functions such as sleep/wakefulness, hormonal secretion, and cardiovascular control (1). Histamine is the predominant constituent of the mediators stored in the granules of mast cells and basophils and is released after appropriate allergen stimulation of specific membrane-bound IgE. The released histamine then binds to the H1 receptors. Histamine is also stored in the enterochromaffin-like cells and mast cells of the gastric mucosae and is released from these cells upon stimulation with gastrin and acetylcholine. It then stimulates gastric acid secretion from parietal cells by interaction with the specific H2 receptors of the gastric mucosae. A third H3 histamine receptor has been described which is presynaptically located and functions as a feedback system that is important in histamine synthesis and release, acting as a general regulatory system in both the peripheral and central nervous systems. There is some evidence for a fourth cellular receptor for histamine, the HIC receptor, although this concept has not yet been widely accepted. However, certain antihistamines, particularly the H1 class, have been shown to react with the HIC system, and it is thought that this receptor might be important in tissue growth. Histamine exerts its functions postsynaptically by interaction with H1 and H2 receptors, both of which belong to the family of G-protein coupled receptors. In addition to the H3 autoreceptors, H3 receptors also occur as heteroreceptors, being present presynaptically on several other neurons (e.g., dopaminergic and serotoninergic). The prevalence of asthma, rhinitis, and dermatitis has increased remarkably in many countries over the past 30 years (2), and this has led to a greater awareness of the social and economic costs of allergic diseases. Although histamine is not the only mediator involved in allergic reactions, its release from mast cells and basophils makes a major contribution to the symptoms of allergic disease. Therefore, therapeutic intervention has often focused on antagonizing the effects of histamine. Accordingly, early research led to the development of several potent antihistamines for the treatment of certain symptoms of allergic diseases. However, histamine also functions as a neurotransmitter, being particularly important in maintaining a state of arousal or awareness within the CNS (3). The classical H1 receptor antagonists are lipophilic drugs with a marked tendency to cross the blood–brain barrier, resulting in the well-documented sedative effects of the early antihistamines. Moreover, the old antihistamines have rather potent muscarinic receptor-blocking properties, which are responsible for the often observed anticholinergic side-effects of these compounds. These properties have greatly limited their use in the treatment of allergic symptoms. The second-generation H1 antagonists have been developed over the past 15 years and have major advantages over the earlier drugs, most notably their lack of significant CNS and anticholinergic side-effects. They are therefore considered to be a major pharmaceutic advance over their predecessors and have proved to be important therapeutic tools in the treatment of atopic disease (4, 5). To date, the classification of antihistamines into older or first-generation drugs and newer or second-generation drugs has been made on the basis of a much reduced tendency to cause sedation combined with an absence of antimuscarinic activity. Since the early observations that first-generation antihistamines induce sedation, debate has continued on the mechanisms involved. These discussions intensified with the introduction of the nonsedating H1 blockers – the second-generation antihistamines. This paper will discuss the equivalence of CNS H1 receptors and those occurring in peripheral systems, and will also attempt to explain the poor CNS penetration of the second-generation derivatives, which is responsible for the absence of or weak CNS effects of these newer compounds. In the 1980s, it was established that the sedation induced by antihistamines was a consequence of their blockade of central H1 receptors. Evidence for this was provided by a mouse model in which the sedative properties of antagonists significantly correlated with occupancy of the H1 receptor (6). In man, administration of both enantiomers of chlorpheniramine and dimethidene resulted in sedation limited to the enantiomer with higher H1 affinity (7). The reduced ability of the second-generation H1 antagonists to cause sedation was thought to be related to the previously noted “existence of differences” between CNS and peripheral H1 receptors (8). Evidence for this notion was provided by studies which demonstrated that mequitazine (9) and loratadine (10) have a somewhat lower affinity for CNS H1 receptors than peripheral H1 receptors. In fact, there is evidence that there are no differences between CNS and peripheral organ H1 receptors. For example, in a study which examined a large selection of both first- and second-generation antihistamines, Ter Laak et al. (11) showed that there is a highly significant correlation between the affinities of these compounds for H1 receptors obtained from either lung or cerebellum tissue. Furthermore, Northern blot analysis with mRNA derived from a H1 receptor cloned from bovine adrenal medulla did not yield any evidence of a difference between CNS and lung receptors (12). It is therefore unlikely that the reduced ability of the second-generation antihistamines to cause sedation is a consequence of differences between CNS and peripheral H1 receptors. The reason that some authors have described small differences in affinities between central and peripheral sites is probably the difference in the methods used (e.g., influence of buffers) or the use of nonequilibrium conditions. It is now well established that certain compounds do not reach the CNS. In general terms, compounds that are highly hydrophilic or that have a high molecular weight do not readily pass into the brain. For these compounds (e.g., the hydrophilic L-dopa or the high-molecular-weight analgesic peptides, respectively), an active transport system is required for their passage into and from the CNS (13). The reason for this important point is found in the morphologic and anatomical differences between capillaries in peripheral organs and those in the brain. The differences between peripheral and brain capillaries are known as the blood–brain barrier. In the peripheral tissues, compounds not exceeding a certain molecular weight (e.g., proteins) easily pass into the surrounding tissue. In the absence of active transport, the presence of tight junctions and the lack of fenestrae in the capillaries of the CNS mean that only those compounds with certain chemicophysical properties can penetrate the brain. In order to pass the blood–brain barrier, a compound must enter the capillary membrane from the aqueous phase (blood) and leave the fatty environment before passing into an aqueous medium (the extracellular space in the CNS). Thus, polar compounds will not enter the membrane whereas highly lipophilic compounds have a high affinity for the fatty environment of the lipid bilayer and will therefore be retained by the membrane. It has been observed that in the absence of active transport mechanisms, compounds with an appropriate degree of lipophilicity enter the CNS while highly hydrophilic compounds do not (14). Lipophilicity can be expressed as log P, which is defined as the negative logarithm of the distribution coefficient of a compound between an organic solvent – for example, n-octane – and water; thus, a high log P denotes a high level of lipophilicity. There is no linear relationship between the central effect of a compound and log P; instead, this relationship is parabolic. In 1987, Hansch et al. defined the principle of “optimal hydrophobicity”, finding that the log P value associated with optimal penetration of the blood–brain barrier has a value of around 2 (15). For several compounds, such an optimal value of log P for CNS activity has been found; for example, for a series of hypnotic amides, anesthetic ethers, or barbiturates (16–18). The issue is further complicated when, in addition to its lipophilicity, the influence of the ionization of a compound is taken into account. In this situation, the ratio of a neutral to a protonated charged compound depends on the pKa of the molecule and is defined as log D. For basic compounds, a similar parabolic relationship for log D as seen for log P in cases of neutral com-pounds should be expected; for example, Timmerman et al. (19) observed a log D value for optimal CNS activity of 2.16 for a series of basic imidazolines. When the log D of several antihistamines are examined at physiologic pH (log D7.4=log P+log (1+10(pKa–7.4)), notable differences can be observed (Table 1). For example, the sedating antihistamine azatadine has a log D7.4 of 1.88, while its analog loratadine has a log D7.4 of 4.4. In contrast, cetirizine has a log D7.4 of 1.04, while the parent sedating compound hydroxyzine has a higher log D7.4 of 2.87 (20). Another useful measurement parameter uses the difference in solubility of a given compound between two solvents which are soluble in each other; e.g., n-octanol and cyclohexane. This is termed Δlog P, and it is measured as the difference between log Poct−H2O and log Pcycl.hex−H2O. Thus, a high Δlog P identifies a compound which dissolves more readily in octanol than cyclohexane, a feature which is dependent on the hydrogen-bonding capacity of octanol as opposed to the neutral behavior of cyclohexane. Therefore, a high Δlog P is associated with the high hydrogen-bonding capacity of a compound, a fact which in turn may result in enhanced binding to serum proteins and thus restrict penetration of the blood–brain barrier (21). The data presented in Table 1 show that Δlog P alone is not a reliable predictor of the ability of the first- and second-generation antihistamines to penetrate the CNS. However, when the data in Table 1 are used in conjunction with the hypothetic model of the brain-penetration capacity of antihistamines shown in Fig. 1 and applied to those compounds with an intermediate log Doct, the Δlog P concept becomes more attractive. While a Δlog P value of less than 2 denotes good brain penetration, values in the region of 4 indicate poor capacity to reach the CNS. It is clear that lipophilicity is an important factor in brain penetration, but that the rule of thumb that lipophilicity correlates positively with penetration into the brain is too simple to be generally applicable. Although the precise significance of Δlog P remains to be elucidated, use of Table 1 in combination with Fig. 1 may be helpful in predicting the ability of a given antihistamine to penetrate the CNS. Hypothetic model for brain-penetration capacity of antihistamines. See text for explanation. Numbers refer to compounds listed in Table 1. It has generally been assumed that a high level of hydrogen bonding may be correlated with poor blood–brain barrier passage (21). Consequently, hydrogen bonding is associated with serum protein binding: a high hydrogen-bonding capacity leads to high protein binding, which in turn results in poor penetration. The most important protein fraction in the blood is albumin, which could be considered an extra lipophilic compartment. Any fraction of a compound bound to albumin would not be available for CNS penetration; therefore, high protein binding may explain the low incidence of CNS effects of some antihistamines. However, in these circumstances, we may ask how antihistamines retain their peripheral effects. The most likely explanation is that penetration into the CNS is diffusion velocity-controlled, whereas perfusion is the major determinant factor for peripheral organs. Thus, the free fractions of the H1 antagonist readily leave the circulation in peripheral tissue; competition then follows between albumin and receptor for the antihistamine, in which the receptor prevails due to its much greater affinity. Moreover, peripheral capillaries show an increased permeability by large molecules in the presence of inflammation. Recent developments have shown that the CNS concen-tration is also under the control of an exporter system, the P-glycoprotein transporter. First observed in cancer cells with a reduced sensitivity to certain chemotherapeutics, the P-glycoprotein has the ability to pump substrates actively against a higher concentration gradient. For instance, the system is present in mem-branes of cancer cells, the membranes of the capillaries of the gut epithelium, and the membranes of the capillaries of the blood–brain barrier. In P-glycoprotein knockout mice, certain compounds reach high CNS concentrations and/or high serum levels; by contrast, in wild-type mouse controls, these compounds are exported from the CNS or the blood into the gut by the exporter system (22). What is the evidence that the P-glycoprotein influences CNS concentrations of antihistamines? Before this question can be answered, it is important to distinguish between substrates and blockers of the exporter. A substrate would not reach an appreciable CNS concentration. Examples of substrates from other classes of drugs include loperamide and ivermectine; to date, there is no evidence of any H1-antagonist being such a substrate. In the CNS, blockers of P-glycoprotein cause enhanced concentration of the compound itself when CNS penetration occurs, as well as of other compounds. A large number of compounds have been found to be blockers; however, most of them are poor substrates rather than pure blockers. Many Ca2+-channel blockers have also been found to block the exporter system. Since there are structural similarities between the Ca2+-entry blockers and the H1 antagonists, one might expect certain antihistamines to be blockers of the P-glycoprotein transporter. Indeed, terfenadine, which was originally described as a calcium-entry blocker (calcium antagonist), has been found to restore the sensitivity of multidrug-resistance cancer cells to chemotherapeutics (23). Moreover, a P-glycoprotein-blocking activity has also been reported for acrivastine in cancer cells (24). It must be emphasized that an effect on the CNS concentration of the compound itself could be expected only in cases where that compound crossed the blood–brain barrier, an event which is unlikely in the case of acrivastine. It would, however, be very beneficial to obtain more information on the interaction of antihistamines with the P-glycoprotein system. The observation in a 1996 position paper (8) that the second-generation compounds “are relatively large, have long side-chains, and are poorly soluble in lipid” is a misunderstanding and contains a contradiction. Nor is this paper correct in stating that these compounds “have higher affinity than the older ones”. On the contrary, some of the second-generation medicines are relatively poor antihistamines. Another overlooked issue is whether the function of the blood–brain barrier might be influenced by certain diseases or pathologic processes. Indeed, inflammatory events may change the properties of the blood–brain barrier, and histamine and several other mediators associated with the allergic response may have profound effects on the permeability of the blood–brain barrier (25, 26). Thus, histamine opens the blood–brain barrier through a H2-receptor-mediated effect which can be blocked by H2 antagonists. Therefore, it is possible that antihistamine concentrations in the CNS are higher when given to patients during an allergic episode than those seen in healthy volunteers. This is a cause for concern because many studies of the sedative effects of antihistamines use normal volunteers. The second-generation drugs have very different chemical structures (27) and consequently show important variations in therapeutic profiles, side-effects, half-life, tissue distribution, metabolism, and degree of antiallergic effects. Thus, with the passage of time, it has become apparent that the newer antihistamines in fact represent a heterogeneous group of compounds. In this review, only the CNS effects of H1 blockers have been discussed; on this basis, it is possible to classify two generations of antihistamines. There has also been a great deal of concern regarding the possible cardiotoxic effects associated with some of the newer medications; these, however, are in no way related to the antihistaminic properties of the individual compounds (28). However, for some compounds, there is an overlap between the structural requirements for H1-blocking properties and those involved in cardiac side-effects. It is therefore possible to obtain antihistamines which do not induce cardiac side-effects. Can all the currently available antihistamines be considered to be the same as far as their prime H1 effects are concerned? This does appear to be the case for their ultimate therapeutic effect when used at recommended doses. However, this does not mean that all H1 blockers interact with the H1 receptor in the same way. Some years ago, we showed that selected agonists of the H1 receptor may react with different amino-acid moieties of the receptor protein while causing exactly the same pharmacologic effects (29), and we have more recently demonstrated that antihistamines also differ in the way they react with the receptor. From modeling studies, we concluded that a lysine in transmembrane domain 5 of the receptor could play a role in binding the H1 ligands. These unpublished studies demonstrate that only the zwitterionic H1 antagonists, which do not show important CNS effects (30), are capable of interacting with the lysine moiety via an effect mediated by their carboxyl group. We have further demonstrated that zwitterionic antagonists lose much of their affinity for H1 receptor mutants in which the lysine moiety has been exchanged for an alanine, whereas other H1 blockers still exhibit full affinity under the same conditions. This discriminative behavior may permit a further classification of the H1 antagonists to be made. Antihistamines may be classified in several ways. The nonsedating or second-generation H1 blockers do not reach the CNS due to their physicochemical characteristics. The possible role of the interaction of the H1 blockers with the P-glycoprotein exporter system deserves further investigation. Cardiac side-effects are not related to the antihistamine properties of the second-generation drugs. A future reclassification of third-generation drugs may be possible if it were based on the way(s) in which certain compounds react at the molecular level with the H1 receptor.