Cholinesterase Inhibitor Therapy in Alzheimer’s Disease: The Limits and Tolerability of Irreversible CNS-Selective Acetylcholinesterase Inhibition in Primates

INTRODUCTION

Inhibition of cholinesterase, primarily acetylcholinesterase (AChE; EC 3.1.1.7), is a mainstay strategy for treating dementing disorders that involve a critical loss of central nervous system (CNS) acetylcholine [1]. This approach is well grounded in the findings that dementia in Alzheimer’s disease (AD) is due, at least in part, to a severe loss of acetylcholine in the nucleus basalis of Meynert, as well as in other midbrain nuclei, cortex, and hippocampus [2–4]. Recent evidence also suggests that cholinesterase inhibitors may be useful in treating vascular cognitive impairment [5–7], Parkinson’s disease dementia and Lewy body dementia [8, 9], and for age-related memory impairment as suggested by animal experiments [10]. In addition, a wide variety of preclinical and clinical studies show that AChE inhibitors may have neuroprotective effects that can delay or modify the course of the disease [11–17]. In view of the key role that cholinergic dysfunction plays in AD and related diseases, it is disappointing that the development of AChE inhibitors has focused on only those with competitive or pseudo-irreversible mechanisms of action as, for example, with donepezil and rivastigmine. These have produced only marginal clinical improvement, far below what might be expected from treating such a well-validated therapeutic target.

For the purposes of this paper, a “competitive” AChE inhibitor occupies the catalytic site temporarily (without a covalent bond) and is readily established and reversed according to concentration-driven kinetics. “Pseudo-irreversible” inhibitors such as the carbamates (e.g., rivastigmine) form a covalent enzyme-inhibitor bond at the catalytic site, which then undergoes spontaneous hydrolysis according to pseudo-first-order kinetics [18], eventually allowing the recovery of enzyme activity. Neither competitive nor pseudo-irreversible inhibitors require the new synthesis of enzyme for the recovery of activity. An “irreversible” AChE inhibitor, on the other hand, forms a permanent covalent enzyme-inhibitor bond at the catalytic site that does not undergo spontaneous hydrolysis to restore enzyme activity. Thus, recovery from irreversible inhibition requires the synthesis of new, uninhibited enzyme.

The main problem with AChE inhibitors that rely on competitive or pseudo-irreversible mechanisms of action is that they do not deliver the high level of CNS selectivity required for successful therapy. Besides being necessary in the CNS, acetylcholine is also the major neurotransmitter in all autonomic ganglia, in the nerve/muscle junction on somatic striated muscle, and in the parasympathetic control of smooth muscle, cardiac muscle, and glands. When pharmacological levels of AChE inhibition begin to be achieved in the CNS, drugs without sufficient CNS selectivity wreak havoc in these other critical peripheral tissues, especially in the gastrointestinal system. This is not a trivial problem. Gastrointestinal toxicity (nausea, vomiting, and diarrhea) has been an impenetrable barrier to using the high doses of conventional AChE inhibitors that are required to correct severe acetylcholine deficits in the brain [19]. To realize the full, as yet untested therapeutic benefit of AChE inhibition for CNS disorders, exceptionally high CNS selectivity must be obtained.

One strategy for establishing and maintaining the necessary high levels of AChE inhibition in the brain, but without dose-limiting nausea, vomiting, and diarrhea, is to use an AChE inhibitor with an irreversible mechanism of action. Importantly, irreversible AChE inhibitors are inherently CNS-selective because the new synthesis of AChE in the brain occurs very slowly, with a half-time (t_½) of ∼12 days [20, 21]. In contrast to the CNS, new synthesis of AChE in peripheral tissues is, by comparison, very rapid, with a t_½ of only 1 day in intestines [20]. Slow new synthesis of AChE in the brain thus allows irreversible inhibition to carry over, dose after dose, stacking inhibition on top of accumulating inhibition, until high levels of enzyme blockade are produced and maintained. In peripheral tissues, on the other hand, AChE is replaced by 10+ times more rapid new synthesis. When AChE in peripheral tissues is blocked by an irreversible inhibitor, most of the inhibited enzyme is immediately replaced with newly synthesized active enzyme during the time between doses. The rapid replacement of AChE in peripheral tissues prevents the accumulation of inhibition such that these tissues remain unaffected.

The idea of using an irreversible AChE inhibitor for AD is not new. Metrifonate, an organophosphate tested for the treatment of AD, is often described as being “irreversible” [19, 22]. However, metrifonate is a pro-drug, a slow release formulator, that is converted non-enzymatically to DDVP (O,O-dimethyl O-(2,2,-dichlorovinyl) phosphate), its active AChE inhibitor intermediate [23, 24]. However, in vivo administration of metrifonate, or DDVP itself, results in mainly pseudo-irreversible AChE inhibition which then undergoes rapid spontaneous reactivation within a few hours, like that seen with rivastigmine, a classic pseudo-irreversible inhibitor [18]. Only the lesser residual part of metrifonate and DDVP-induced AChE inhibition is thus truly irreversible [21, 24]. For example, after a single injection of 10 mg/kg DDVP in mice, a dose which markedly suppressed locomotor activity, rotarod performance, and rectal temperature, produced a peak of 88% CNS AChE inhibition at 2 h which then underwent spontaneous reversal to only 20% inhibition within 24 h, showing a t_1/2 of ∼12 h like that found with a pseudo-irreversible inhibitor [21]. However, in spite of the rapid spontaneous reversal of most of the daily DDVP-induced AChE inhibition, ten daily injections in that study accumulated ∼65% inhibition of cortical AChE which did not undergo spontaneous reactivation. In the absence of further DDVP treatment, cortical enzyme activity recovered with a t_1/2 of several days as is characteristic of irreversible inhibition [21]. This outcome would occur if as little as 15% of the DDVP-induced inhibition produced by each daily dose “ages” into the more stable, irreversible organophosphate bond that resists reactivation [25]. Thus, because of its predominantly pseudo-irreversible mechanism of action, metrifonate was not an adequate test of the real therapeutic potential of an irreversible AChE inhibitor for treating AD.

Methanesulfonyl fluoride (MSF), in contrast to metrifonate, forms an extremely stable covalent bond at the catalytic site that is totally refractory to reactivation [20, 26]. MSF-induced AChE inhibition can only be overcome by the new synthesis of uninhibited enzyme. Moreover, it is important to note that MSF is a sulfonyl fluoride and does not interact with the neuropathy target enzyme associated with organophosphate-induced delayed neuropathy [27]. MSF is, therefore, free of the risk of life-threatening respiratory paralysis and muscle weakness that contributed to the termination of metrifonate development [28]. Moreover, the organophosphates have the disadvantage of producing a wide range of serious non-cholinergic side effects [29]. Because of its truly irreversible mechanism of action, and its freedom from organophosphate-induced toxic effects, MSF was selected to determine the uppermost practical and theoretical limits of AChE inhibition that can be tolerated in primates, the purpose of this study.

The critical importance of an irreversible mechanism of AChE inhibition to the eventual success in treating AD cannot be overemphasized. Figure 1 presents a side-by-side comparison of the results that can be expected with an AChE inhibitor with either a competitive or pseudo-irreversible mechanism of action as compared to an AChE inhibitor with an irreversible mechanism of action. The left side of Fig. 1 (A) is modeled after a competitive inhibitor with a clearance t_1/2 of 70 h and strong CNS selectivity, producing 1.25 times more inhibition in the brain than in intestines (e.g., Aricept/donepezil). As shown in Fig. 1A, there is considerable overlap between AChE inhibition in the CNS and intestines. When brain inhibition is maintained above the 50% level required for a therapeutic effect, inhibition in the intestines is also at or near 50%, a level that causes gastrointestinal toxicity [19]. The unavoidable overlap between brain and intestinal AChE inhibition by inhibitors without sufficient CNS selectivity (e.g., competitive or pseudo-irreversible inhibitors) is the cause of the unbearable nausea, vomiting, and diarrhea that has limited such AChE inhibitors to ineffective doses. In sharp contrast to Fig. 1A, Fig. 1B shows the separation of brain versus intestinal AChE inhibition that can be obtained with an irreversible inhibitor. MSF-induced AChE inhibition is expected to accumulate to a high level in the CNS (>60%, estimated t_1/2 for AChE replacement = 12 days) without producing clinically significant inhibition of AChE in the periphery (<25%, estimated t_1/2 for AChE replacement = 1 day). The clear separation between brain and peripheral AChE inhibition in response to MSF is based solely on the powerful effect of the 10x slower replacement of AChE that occurs in the brain.

A prior study has empirically validated the theoretical advantage of an irreversible inhibitor, the clear separation of AChE inhibition in the brain and peripheral tissues, in an experiment in which young rats were treated with MSF over a period of three weeks [30]. According to a priori computations like those shown in Fig. 1, the dose of MSF and the dosing schedule used during the three weeks were expected to produce an average of 70% AChE inhibition in the brain (using t_1/2 = 12 days) and 25% inhibition in the intestines (using t_1/2 = 1 day). The ex vivo results shown in Fig. 2 show excellent agreement with the a priori estimates. The 75% CNS AChE inhibition (Fig. 2) was achieved without toxicity and is more than the 50% minimum required for a therapeutic effect [19]. Less than 30% AChE inhibition is sufficient to avoid gastrointestinal as well as cardiac and skeletal muscle toxicity [19]. As is expected from such powerful AChE inhibition, MSF also produces larger and longer lasting increases in brain acetylcholine than donepezil [31].

The absence of drug-induced nausea, vomiting, and diarrhea presents a unique challenge to estimating the dose of MSF that should be tested in AD patients. The dose cannot be established by simple dose escalation, going up until higher doses cannot be tolerated, as with competitive and pseudo-irreversible inhibitors (e.g., donepezil, rivastigmine, etc.).

The CNS selectivity shown in Fig. 2 has real-world significance. In a recent study of MSF in normal humans, reports of nausea were, as expected, uncommon with only 5 reports of mild, transient nausea out of 56 administrations of the highest dose. Diarrhea and vomiting were even more infrequent and inconsequential [30]. Without dose-limiting gastrointestinal toxicity, however, the starting point for what might be an effective clinical dose of MSF in humans had to be estimated a priori based on animal experiments [32]. The absence of dose-limiting nausea, vomiting, or diarrhea in the humans suggests that the doses of MSF tested in the clinical trials conducted thus far may not have, in fact, been anywhere close to the limits of tolerability or clinical effectiveness that might be possible [32].

In view of the untested bounds of MSF tolerability or potential clinical efficacy, the purpose of the present study was, therefore, to determine the upper practical and theoretical limits of CNS AChE inhibition and tolerability in primates.

MATERIALS AND METHODS

Four three-year-old male Macaca fascicularis (Crab Eating Cynomolgus) monkeys (Charles River Primate Breeding Facility, Port Washington, NY) served as subjects in all of the following protocols which were approved by the U.T. El Paso Institutional Animal Care and Use Committee and following guidelines recommended by NIH. Methanesulfonyl fluoride was purchased from Aldrich Chemical Company (Milwaukee, WI).

RESULTS

DISCUSSION

To provide a human clinical reference for the escalating MSF doses shown in Table 1, the clinically effective dose of MSF in AD patients was 180 μg/kg given three times per week [30, 32]. Using these human data as a reference point, the MSF-treated monkeys in Experiment One were receiving more than 5 times, and then more than 8 times, the human clinical dose of MSF when they received 1.0 mg/kg and 1.5 mg/kg, respectively. The ability of the monkeys to tolerate these high doses may be due, at least in part, to the prior escalation from lower doses [19].

As shown in Table 1, the repeated escalating doses produced significant accumulated inhibition of RBC AChE with an asymptote at ∼90% inhibition during the final weeks of repeated 1.5 mg/kg doses of MSF. The dose/response equation for the effect of single doses of MSF on human RBC AChE derived from Moss et al. [30] is: % INH = ( 11 . 17 ) ( ln [ mg / kg MSF ] ) + 38.715 ( r = + 0.9964 )

Using the above equation to estimate the effects of MSF on monkey RBC AChE, it is expected that 1.5 mg/kg MSF would produce 43% inhibition of the remaining active enzyme at the time of each new dose. Given that M. fascicularis RBCs, like those in humans, have a 70–90 day lifespan [36], it is expected that the dosing schedule used in this experiment would produce a mean of ∼90% RBC AChE inhibition, in good agreement with the results obtained (Table 1). The accumulated monkey RBC AChE inhibition in this experiment was limited by a ceiling effect and it is not expected to be sensitive to dose-dependent differences like those that are found at more clinically relevant doses [30, 32].

Besides showing that ∼80% cortical AChE inhibition can be produced and maintained by repeated doses of 1.5 mg/kg MSF, the data in Table 2 can also be used to estimate the rate at which monkey brain AChE is replaced by new synthesis, a factor that is critical for predicting the cumulative effect of MSF in the brain when it is administered repeatedly during clinical use (e.g., computations as shown in Fig. 1).

Even though the level of brain AChE inhibition was determined by biopsies at only one time point (2.5 days after the last MSF injection), the lower limit of the t_1/2 for de novo AChE synthesis in monkey cortex can be estimated by proposing, for purposes of calculations only, that the last injection of 1.5 mg/kg MSF produced a peak of 100% inhibition. If the brain showed 100% AChE inhibition at the time of the final injection (time 0 for calculations of enzyme recovery), and it still retained 80% inhibition 2.5 days later, the t_1/2 for new synthesis of cortical AChE could not be less than 8 days. It is, of course, highly unlikely that the brain had 100% inhibition at the end of MSF treatment. With anything less than 100% inhibition, the t_1/2 will be longer, very likely similar to the 12 days that is observed in mice and rats [20, 21]. Using a proposed t_1/2 of 12 days for AChE recovery in monkey cortex and estimating the effect of the repeated doses of 1.5 mg/kg at 43% inhibition with each dose (from human RBC AChE dose-response function shown above [30]), retrospective calculations indicate that a range of 84% (peak) declining to 73% AChE (trough) inhibition between doses would be expected from the series of injections shown in Table 1. The actual finding of ∼80% cortical inhibition 2.5 days after the last injection is within the expected range, confirming that a t_1/2 of 12 days is a useful estimate for primate brain AChE turnover.

Whether or not the estimated t_1/2 of 12 days or the calculated maximum AChE inhibition of 92% are accurate is of only theoretical interest. The empirically established level of 80% cortical AChE inhibition actually achieved and confirmed by enzyme assays far exceeds the minimum of 50% inhibition estimated to be required for clinically significant therapeutic effects [19].

Experiment Two showed that monkey CSF AChE recovered with a t_1/2 of 2.4 days, which is faster than the minimum possible t_1/2 of 8 days for cortical AChE recovery, and especially the more likely t_1/2 of 12 days. Monkey CSF AChE recovery with a t_1/2 of 2.4 days (Fig. 3) compares favorably to the 2.2 day t_1/2 of CSF AChE reported in humans [37]. The high turnover rate of CSF AChE, as compared to brain tissue, suggests that monitoring CSF AChE is not a useful indicator of the accumulated effects of repeated doses of an irreversible AChE inhibitor. It is important to note that two monkeys in this experiment had never received MSF, the controls in the previous experiment, and the other two had not received MSF for seven months. In spite of this, the abrupt administration of 1.5 mg/kg, without prior escalating doses, did not produce any discernable side effects. The above monkey experiments using MSF demonstrate that dose-limiting side-effects need not be major barriers to obtaining the clinical doses required to effectively treat AD or other CNS disorders. Indeed, an irreversible AChE inhibitor can produce and maintain substantially more inhibition than is probably needed for effective treatment of CNS disorders. An irreversible inhibitor circumvents the major barrier of unbearable dose-limiting nausea, vomiting, and diarrhea that is experienced by patients given AChE inhibitors that have competitive or pseudo-irreversible mechanisms of action. While the use of an irreversible inhibitor appears to effectively reduce the risk of peripheral toxicity, there is an as yet inadequately tested possibility of limiting centrally-mediated hypercholinergic side effects.

MSF was used in these experiments because it is the only irreversible inhibitor for which a comprehensive series of experiments, from mice to monkeys and humans, is available. Similar results would be expected from any truly irreversible AChE inhibitor that crosses the blood-brain barrier and has a suitable partitioning coefficient into mainly lipid tissues.

If toxic side-effects do not determine the upper limit of doses of an irreversible AChE inhibitor that might be used to treat dementia, what then can be the basis for a guideline for optimal dosing?

One of the problematic pharmacological realities of irreversible AChE inhibitors is that they produce decreasing returns with increasing doses. This is especially evident at the higher end of the dosing scale. Accumulated AChE inhibition produced by an irreversible inhibitor eventually approaches an asymptotic level that is a function of the percent inhibition produced by each dose, the frequency with which the doses are administered, and the t_1/2 for the synthesis of new enzyme. An example of this is shown in Fig. 1B wherein the asymptotic accumulated AChE inhibition in the intestines was achieved in 4 days, yet even after 21 days of dosing the brain was still approaching, but had not yet achieved, the final asymptotic accumulated AChE inhibition even at 65%.

Figure 4 shows the estimated asymptotic level of accumulated AChE inhibition that can be expected from increasing doses of an irreversible inhibitor in brain and intestine. These computations assume a t_1/2 of 12 days for brain and 24 h dose-to-dose intervals. As shown in Fig. 4, doses of an irreversible inhibitor that produce as little as 10% and 20% inhibition of the remaining active AChE with each dose, will eventually produce an expected asymptotic inhibition of 65% or 80%, respectively. Figure 4 also shows that higher doses produce the sharply decreasing returns as discussed above. Even though the monkey experiments show that very high doses are well tolerated, it seems unlikely that the clinical use of a daily dose that produces more than 20% inhibition with each dose would be useful.

As a basis for estimating the effect of an irreversible inhibitor in a clinical trial, the asymptotic inhibition of brain AChE that can be expected with extended daily administration of any particular dose of an irreversible inhibitor (at equilibrium between drug-induced inhibition and new enzyme synthesis) can be estimated from a double reciprocal plot which gives a straight line with the equation: 1 / Asym % = ( 0 . 0576 / % INH ) + 0.01

In this equation, Asym% is the expected asymptotic accumulated AChE inhibition in the brain expected after extended daily administrations of an irreversible inhibitor; and, % INH is the percent inhibition of the remaining active enzyme at the time of each dose. The slope of this equation is unique to a t_1/2 of 12 days (brain) and will not predict asymptotic accumulated AChE inhibition in other tissues. It is not surprising that with a t_1/2 of 1 day (intestine) and a schedule of 1 day between doses, that there is an almost linear increase in asymptotic AChE inhibition in the intestines with increasing dose size (Fig. 4).

As mentioned above, a cursory examination of Fig. 4 shows that there is little benefit to be gained by daily doses of an irreversible inhibitor that produce more than 20% inhibition. Extended use of a 20% dose will eventually produce and maintain ∼80% and 25% AChE inhibition in the brain and intestines, respectively. The maximum practical dose of an irreversible inhibitor is not only limited by decreasing returns in brain AChE inhibition. It is also subject to limitations imposed by the accumulated asymptotic inhibition in the intestines. For example, Fig. 4 shows that a daily dose which produces 30% inhibition with each administration will approach ∼40% accumulated AChE inhibition in the intestines, a level that will possibly allow the reappearance of gastrointestinal toxicity [19]. Satisfactory results may even be obtained by a daily dose of drug that produces as little as 10% inhibition. This dosing would result in an estimated 65% inhibition in the brain, within the therapeutic range, with as little as 15% inhibition in the intestines (Fig. 4), well below levels expected to be toxic [19]. In the specific case of MSF, the percent inhibition produced by single doses can be determined from the human RBC AChE dose-response equation from Moss et al. [30] shownabove.

The above computations are designed to estimate a practical schedule of daily dosing that will produce an adequate clinical effect in the brain with minimum peak levels of inhibition and risk of gastrointestinal side-effects. Other dosing schedules such as twice- or three-times per week, as used in the present monkey experiments, can be satisfactory, but will produce a wider range of peaks and valleys of AChE inhibition. Higher peak levels of inhibition may increase the risk of gastrointestinal toxicity. New estimated effects would need to be computed for different dosing schedules.

In summary, a myriad of in vitro and in vivo experiments in mice, rats and monkeys clearly demonstrate, as a proof-of-concept, that irreversible AChE inhibition, like that produced by MSF, can produce and maintain high levels of CNS inhibition without interference associated with AChE inhibition in peripheral tissues. The high CNS selectivity of an irreversible inhibitor can be understood within a simple biochemical model based on pseudo first-order kinetics. Irreversible inhibitors are unique in that they alone exploit the slow de novo synthesis of AChE in the brain to increase therapeutic benefit and decrease the risk of peripheral toxicity. Indeed, the promise of strong efficacy and excellent tolerability of an irreversible AChE inhibitor has already been realized in human clinical trials [30, 32].

Irreversible AChE inhibitors thus open the door to the full potential of CNS cholinesterase inhibitor therapy for the treatment of memory loss. Because of the important clinical implications for effective CNS cholinesterase inhibitor therapy, irreversible AChE inhibitors, especially MSF, deserve serious evaluation in humans experiencing cognitive loss, especially when the aging population around the world is growing exponentially.