The Discovery of Ivermectin and Other Avermectins

The Discovery of Ivermectin and Other Avermectins 1 W. C. CAMPBELL, R. W. BURG, M. H. FISHER, and R. A. DYBAS Merck Institute for Therapeutic Research, Rahway, NJ 07065 The avermectins were initially detected in a program in which thousands of microbial fermentation products were tested in mice for activity against the nematode, Nematospiroides dubius. Among the few preparations showing activity in this assay, was the product of a microorganism isolated from soil by workers at The Kitasato Institute. The microorganism was classified as a new species of actinomycete, Streptomyces avermitilis. Its anthelmintic activity was shown to reside in 8 closely related macrocyclic lactones, named avermectins, which were also found to possess activity against free-living and parasitic arthropods. One of the natural components, avermectin B 1, is now being evaluated as a pesticide for the control of mites of citrus and cotton crops and control of the Red Imported Fire Ant. A chemical derivative, 22,23-dihydroavermectin B 1, or ivermectin, has been developed as an antiparasitic agent. It is being marketed for use in cattle, horses and sheep and is expected to become available for swine and dogs. The title of this session, Guided and Serendipitous Discovery (within the symposium, Approaches to Rational Synthesis of Pesticides) focuses attention on the nature of the discovery, rather than the outcome. The discovery of the avermectin family of compounds was by no means serendipitous; those who were seeking found what they sought. It is the purpose of this paper to record the manner of the seeking and the manner 0097-6156/ 84/ 0255-0005S06.00/ 0 1984 American Chemical Society

6 PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES of assessing and enhancing what was found. The discovery of ivermectin is part of the larger story of the discovery of the avermectin family of compounds: the avermectins are produced by a microorganism, and ivermectin is a chemical modification of one of those substances. The initial objective of the search was an anthelmintic with properties radically different from those of known anthelmintics. What was found was an anthelmintic whose properties were indeed markedly different not only in terms of chemical structure and efficacy against helminths, but also in the extension of the potential utility of the class to the control of arthropod parasites of animals. Further, the compounds turned out to have striking activity against a variety of free-living and plant-parasitic nematodes and arthropods and so it has come about that a new livestock anthelmintic has become an agenda item in this symposium on agricultural pesticides. Primary Screening and Microbiology Several in vitro assays for detecting fermentation products with anthelmintic activity had been run without success, primarily because of the large number of toxic compounds which had to be eliminated. Finally, the decision was made to use an in vivo assay in mice with the hope that the mice would screen out the non-specific toxic compounds. The nematode, Nematospiroides dubius, was selected for the assay. Infected mice were fed for six days with milled Purina Lab Chow which had been mixed with the fermentation product to be tested. The mice were then fed a normal diet and, at 14 days postinfection, fecal pellets were examined for the presence of eggs. If eggs were absent on three successive days, the mice were sacrificed and their small intestines examined for the presence of worms. This assay was relatively successful in achieving the goal of eliminating nonspecific toxic activities. Among the many thousands of cultures tested in this assay, about 1% were active in the first test. All active cultures were regrown, and about 20% of the cultures initially selected were active. Among the cultures confirmed active, most were still extremely toxic and were not pursued further. However, during the testing of a large collection of cultures which had been received from The Kitasato Institute of Tokyo, Japan, the choice of the in vivo assay was finally justified. Among one

1. CAMPBELL ET AL. Discovery of Ivermectin 1 group of 50 Kitasato cultures submitted for assay was a culture bearing the number 0S-3153. The results of the early assays of this culture are summarized in Table I. It was scored as fully active (no eggs, no worms) in the first test. Marked reduction in mouse body weight indicated that the fermentation product was either unpalatable or toxic. Culture 0S-3153 was regrown on the original medium (designated KH) and on a second, unrelated medium (JH) (Table I, Experiment No. 2). The culture grown on the second medium was very toxic, resulting in host death on the day that the first fecal pellet was examined for eggs. That grown on the original medium was also toxic, causing severe suppression of weight gains; but, again, there were no eggs or worms. In the third test, the activity was titrated using serial 2-fold dilutions. All four levels were fully active, and this time there was little sign of toxicity. A few cultures had been confirmed active previously, but none had exhibited a separation of toxicity from activity, let alone activity over at least an 8-fold range. After an inauspicious start, the activity of culture OS-3153 had been firmly established. Success had been achieved in the quest for a fermentation product with anthelmintic activity. In fact, as events would soon prove, the newly found product had an even broader activity than had been anticipated. The culture, now bearing the product number C-076 and the Merck culture collection number MA-4680, was submitted for taxonomic studies. Its characteristics, including a brownishgray spore mass color, smooth spore surface, spiral sporophores born as side branches on the aerial mycelia and the production of melanoid pigments, were unlike those of any previously described species of Streptomyces. The culture was named Streptomyces avermitilis, the Streptomyces "capable of separating from worms". Based on a correlation of anthelmintic activity and HPLC analysis of the total avermectin complex, it was estimated that the third fermentation contained a minimum of 9 yg/ml. Improvement of the medium increased the yield by the original culture (MA-4680) to 120 yg/ml. A high-producing isolate (MA-4848) obtained from this culture produced nearly 500 yg/ml of total avermectins. Thus, this culture yielded readily to medium improvement and isolate selection to produce relatively large amounts of the avermectins. Accounts of the early fermentation studies and taxonomy have been published (1-2).

8 PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES Table I. Summary of the First Assays of Culture OS-3153. Feed Mouse Expt. Eaten Weight No. Medium Dose ' (g) (g) Activity 1 KH 50 13 14 A 2 KH 50 25 15 A JH 50 13 - Dead 3 KH 50 25 22 A 25 25 25 A 12 25 29 A 6 25 28 A JH 50 25 22 SA 25 25 29 SA 12 25 ΝΑ» Ml of broth per 25 g of feed. (2) A, active - no eggs or worms; SA, slightly active - no eggs, worms present; NA, not active.

1. CAMPBELL ET AL. Discovery of Ivermectin 9 Chemistry Isolation. The avermectin complex, consisting of four major components designated A^a, A 2 a, ^la* ^2a anc * ^ o u r l w e r homologs designated A^, A 2^, ^ib* ^2b* w a s (^ e x t r a c t e with acetone from the mycelia of Streptomyces avermitilis (3). Solvent partition and adsorption on granular carbon produced an oily residue containing 5% 16% A 2, 20% Β χ and 15% Β 2 Separation of the A components from the Β components was achieved by partition chromatography with hexane-methylene chloride-methanol (10:10:1) over Sephadex LH-20. A^ was separated from A 2 using Sephadex LH-20 and a solvent system containing hexane-toluene-methanol (6:1:1). B^ was separated from Β2 either by crystallization from ethylene glycol or by Sephadex LH-20 chromatography using hexane-toluenemethanol (3:1:1). Similar chromatography was also used to separate the lower homologs and purity was established by reverse-phase HPLC analysis. Structure Determination. The structures of a new group of pesticidal sixteen-membered lactones named milbemycin B^, B 2, and Ββ were described in 1975 (4) on the basis of X-ray analysis. A later publication (5) gave details of isolation and structures of thirteen milbemycins, with spectral data. The avermectins were discovered in 1975 and part-structures, 1 ο deduced from proton and C NMR spectra and their mass J spectral fragmentation patterns, suggested a close relationship with the milbemycins (6). Methanolysis of avermectin A 2 gave an aglycone and a 6:1 mixture of -and -methyl-loleandroside. The recovery of this glycoside in more than 100 mol. % yield demonstrated the presence of two identical sugars in the molecule. Further spectral examination indicated the attachment of an a-l-oleandrosyl-ot-loleandrosyloxy disaccharide to the 13-ct-position of the macrolide ring. A chemical proof for the point of attachment and identity of the disaccharide was provided by ozonolysis and isolation of the disaccharide attached to the fragment C l l through C 14. The structures of the eight components are shown in Fig. 1 and all contain the same disaccharide substituent at the 13-α-position. They vary at C-5 with hydroxy or methoxy groups, at C-23 with an axialmiydroxy group on a 22,23-olefin and at C-25 with isopropyl or sec-butyl groups in contrast to the methyl and ethyl substituents at the 25-position of the milbemycins. X-ray analysis (7) of B 2 a aglycone and B^a both

v m H a m Η Χ m H a: Ο C ο X 25 δ > r > no η χ m in QCH, Δ, R 5 = CH 3 Β, R 5 = Η α series R 25 = CH(CH 3)C 2H 5 b series R 25 = CH(CH 3) 2 H 2 (1 atm) (0 3P) 3 RhCI 0CH 3, 25, 18 h 85% 4' v C H 3 A 2 R 5 = CH 3 B 2 R 5 = H 22, 23-Dihydroavermectin B } ivermectin Figure 1. Structures of the Avermectins.

1. CAMPBELL ETAL. Discovery of Ivermectin 11 confirmed the structure and, through the L-oleandrose, established the absolute stereochemistry. Parasitological Evaluation Following the demonstration of efficacy in the Nematospiroides-mouse assay, and the associated microbiological and chemical research described above, much work was needed to determine the relative anthelmintic efficacy of both the natural avermectin components and the derivatives. For this purpose two types of bioassay were employed. In one, compounds were tested in small laboratory animals infected with nematodes other than N. dubius. For example, test materials were given to jirds (Meriones unguiculatus) infected with Trichostrongylus colubriformis, and the animals were subsequently killed for determination of worm burden (8). The use of that host-parasite combination for anthelmintic testing had been reported by Panitz and Shum (9) and has proved useful in the evaluation of a variety of anthelmintics, including the avermectins. In the other, compounds were tested against a variety of nematodes in sheep. These were small scale tests, done in conjunction with the small-animal testing and providing important information on the efficacy of the test substances in a ruminant host. Compounds of special interest were similarly tested against helminths in sheep, and occasionally in other hosts, using larger numbers of test animals (10-12). As the remarkable potency and unique structure of the avermectins became apparent, testing was extended to organisms other than helminths. The first test against an insect was done using the Confused Flour Beetle, Tribolium confusum, and the incorporation of the test substance into the flour in which the beetles lived. In this fashion the insecticidal activity of the avermectins was demonstrated (13) and was followed independently and almost immediately by the demonstration of efficacy against a parasitic insect (14). Efficacy against parasitic insects was further established by tests using the rodent bot, Cuterebra sp., in mice (D. A. Ostlind, unpublished) and was shown to extend to some parasitic acarines (15-16). Tests against the trematodes Schistosoma mansoni and Fasciola hepatica and the cestode Hymenolepis diminuta in laboratory animals failed to show efficacy (D. A. Ostlind, unpublished data). This is in accord with reports that the avermectins disrupt GABA-mediated nerve

PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES transmission in nematodes and arthropods and that flukes and tapeworms do not employ GABA as a neurotransmitter (17). Agricultural Chemical Evaluation The results (above) against the Confused Flour Beetle (Tribolium confusum), rodent bot (Cuterebra spp.), and the ectoparasitic larva of the sheep blowfly (Lucilia cuprina) were sufficiently encouraging to suggest that the avermectins may possess general biological activity against arthropod pests and, in particular, those of importance in crop protection (18). To investigate this potential and to expand our in-house capabilities, a research program was established with the Boyce Thompson Institute for Plant Research to test the avermectin derivatives in their miticide and insecticide screens. In all, nearly seventy-five related structures, natural products and semisynthetic derivatives, were evaluated in the greenhouse for toxicity to a spectrum of arthropod pests. Of these, avermectin B^, the major component of the fermentation process, was determined to be the most promising candidate as an agricultural pesticide. Results from these laboratory studies demonstrated that avermectin B^ had high toxicity for the twospotted spider mite (Tetranychus urticae) on bean plants. When applied in solution directly onto adult and nymphal spider mite populations on foliage, avermectin B^ was shown to be 50-200 times as potent as commercially available acaricides, with an LC 9 0 of 0.02-0.03 ppm. Additional tests on foliage with insects in the order Lepidoptera, Coleoptera, Homoptera, Orthoptera, Diptera, Isoptera and Hymenoptera confirmed the broad spectrum activity and potency of the avermectin family of compounds and avermectin B^ in particular. Table II provides LC(JQ values for avermectin B^ for the control of larval forms of several of these insects in foliar residue assays (18).

1. CAMPBELL ET AL. Discovery of Ivermectin 13 Table II. Efficacy of Foliar Residues of Avermectin B^a Against Adult mites and Larval Insects. Insect L C 90 (PP m ) Twospotted spider mite 0.02-0.03 Tomato hornworm 0.02 Colorado potato beetle 0.03 Mexican bean beetle 0.2 Cabbage looper 0.75-1.2 Southern armyworm 6.0 On the basis of the efficacy demonstrated in the greenhouse and laboratory studies avermectin was selected for development and assigned the Merck development code number MK-936. Avermectin B^ has been evaluated worldwide for efficacy against mites and insects affecting a number of agricultural crops including citrus, cotton, apples, pears, vegetables, potatoes, tree nuts, and grapes. Under field use conditions it has been observed that excellent control of a number of economically important pests including the citrus rust and red mite, twospotted spider mite, broad mite, Colorado potato beetle, diamond back moth, pear psylla, and Liriomyza leafminers can be achieved at extremely low application rates of MK-936 in the range of 0.005-0.03 lb per acre (5.5-33 g per hectare). For foliage applications a 0.15 EC (1.8% w/v) emulsifiable concentrate formulation has been developed. Field studies have shown that the formulation is non-phytotoxic to all target crops on which it has been evaluated including many varieties of sensitive ornamental plants. During the course of the development program, samples of avermectin B^ were provided to a number of outside agencies for evaluation in specialized assays. As a consequence, it was discovered in testing conducted by the USDA laboratory for Insects Affecting Man and Animals, Gainesville, Florida, that the red imported fire ant (Solenopsis invicta) is among the most susceptible species of insects to the toxic action of avermectin B^. When applied in a corn grit bait, avermectin

14 PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES B^ at rates as low as in 25 to 50 mg per acre has been effective in controlling fire ant infestations in large scale trials in the southern United States. A submission for registration of MK-936 for this application has been made. The avermectin natural products are pesticides possessing novel chemistry and mode of action. Cross-resistance has not been observed in laboratory or field studies with mites andinsects tolerant to commercially available organophosphate, carbamate, chlorinated hydrocarbon and pyrethroid pesticides. Synthetic Program Structure-Activity Relationships. Compounds of the Β series were generally more potent than those of the A series. Thus an unsubstituted hydroxy group at the 5-position is activity enhancing (19). Differences in potency between the 1- and 2- series varied among parasites, but in most instances the 1- series was more potent. Reduction of the 22,23-olefin had little effect on activity but further reduction caused a substantial decrease in activity. The monosaccharides were two- to fourfold less active than the parent compounds while the aglycones were more than thirtyfold less active, Table III. Acetylation at the 4"-position caused no change in activity whereas acetylation at the 5- or 23- position caused a considerable decrease in activity. Diacetates and triacetates showed similarly reduced activity, Table IV (20). Ivermectin. Early biological studies demonstrated that while avermectin B^ was more active than avermectin B 2 by oral administration, the converse was true when the compounds were given parenterally. Furthermore, avermectin B^ was much less effective against Cooperia species when given parenterally than by oral treatment. Avermectin B 2 had generally lower activity against Haemonchus species. Examination of the B^ and B 2 structures revealed that the differences centered on the 22,23-position. Avermectin B^ is a 22,23-olefin whereas in avermectin B 2 this bond is hydrated with the hydroxyl group at the 23-position. The conformation of the ring bearing these functionalities is different and it was reasoned that bioactivity might be linked to conformation. It therefore became an important objective to synthesize 22,23- dihydroavermectin B^ which required for its synthesis the selective reduction of one of five olefins. However, only the

1. CAMPBELL ET AL. Discovery of Ivermectin 15 Table III. Activity of Avermectin Derivatives against Adult Gastrointestinal Helminths of Experimentally Infected Sheep on Oral Administration. Efficacy a Structure Dose, mg/kg H.c. O.c. T.a. T.c. C. spp. O.c. Al 0.1 2 2 0 0 2 0 A 2 0.1 3 3 3 3 0 3 Bl 0.05 3 3 3 3 3 3 B 2 0.1 0 3 3 3 3 3 H2A1 0.3 3 2 0 1 0 3 H 2 Bi 0.1 3 3 3 3 3 3 BiMS 0.15 2 2 3 3 3 0 B 2 MS 0.2 1 1 3 3 3 3 H2B 1 MS 0.3 3 3 3 3 2 3 H 2 BiAG 3.0 1 2 3 3 1 3 H4B1 0.2 0 0 1 0 0 3 a 0 = < 50%, 1 = 50-74%, 2 = 75-90%, 3 = > 90% efficacy. Abbreviations used: H.c. Haemonchus con tortus; O.c, Ostertagia circumcincta; T.a., Trichostrongylus axei; T.c, Trichostrongylus colubriformis; C. spp., Cooperia spp.; O.c, Oesophagostomum columbianum. *MS = monosaccharide, AG = aglycon, H 2 = 22,23-dihydro derivative, H4 = 3,4,22,23-tetrahydro derivative. Benzimidazole resistant.

16 PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES Table IV. Derivatives of Avermectin A~ and B^a and Anthelmintic Activity against Trichostrongylus colubriformis in Gerbils H 0R 5 anthelmintic R 5 R23 acta H CH 3 H 0.05 H H H 0.0125 CH3CO CH 3 H 0.0625 H CH 3 CH 3CO 0.25 CH3CO CH 3 CH 3CO 0.5 CH3CO CH 3CO CH 3CO 0.5 aminimal dose(mg/kg) needed to remove > 83% of the worm burden.

1. CAMPBELL ET AL. Discovery of Ivermectin 17 22, 23-olefin is cis-substituted, suggesting the use of Wilkinson's homogenous catalyst (Ph^P^RhCl known to be very sensitive to the steric environment of an olefin. Hydrogénation of avermectin B^ for 20 hours using Wilkinson's catalyst in benzene or toluene at 25 C under one atmosphere of hydrogen gave 22,23-dihydroavermectin B^ in 85% yield (Figure 1). This compound was selected for development as a broadspectrum antiparasitic agent for animals on the basis of its overall efficacy by oral and parenteral routes and for its improved safety profile (19-21). 22,23-Dihydroavermectin B^, containing at least 80% of 22,23-dihydroavermectin B l a and not more than 20% of 22,23-dihydroavermectin B^ has been assigned the non-proprietary name ivermectin. The compound was subjected to a large international program of development, which lies beyond the scope of this paper, and which included efficacy trials and safety assessment in sheep, cattle, horses, swine and dogs. This development program resulted in the introduction of ivermectin as a commercial antiparasitic agent in 1981. For cattle, sheep and horses, the dosage recommended for general antiparasitic use is 0.2 mg/kg; for swine the dosage is 0.3 mg/kg. The compound is used both orally and parenterally the formulation and route of administration depending on the host species being treated. Discussion The discovery of the avermectins, by virtue of the wide spectrum of the compounds, and their extreme potency and novel mode of action, met the initial objective of finding an anthelmintic with radically different characteristics. The avermectins are not active against all groups of helminths they have not been reported active against flukes or tapeworms but they are active against all nematode groups that have been tested, and indeed there is no clear evidence that any species of any genus of nematode is refractory to the action of ivermectin. In at least one instance (adult Dirofilaria immitis) a particular life cycle stage is refractory while other stages of the same species are susceptible. The occurrance of antinematodal and antiarthropod activity in a single chemical class, is not entirely unprecedented. The organophosphates are active against parasites of both groups, but their spectrum of activity against nematodes is relatively narrow. The salicylanilide compounds are active against certain nematodes and arthropods but are used primarily against flukes.

PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES The discovery of avermectins resulted from the deliberate choice of fermentation products as the prime source of substances to be submitted for anthelmintic screening. Many factors were critical to the success of the venture, including the submission of novel actinomycte isolates by workers at The Kitasato Institute in Japan, the selection of the in vivo screen, the rapid isolation and identification of the active principle, the assessment of antiparasitic properties, and the enhancement of biological properties by synthetic chemical modification of the structure. The names of those responsible for these contributions may be found in the early papers published on the subject, and are listed elsewhere according to scientific discipline (2). The discovery of the avermectins thus rested on empirical testing as did the discovery of all other successful anthelmintics and ectoparasiticides. Such discoveries are nevertheless attributable to deliberate and far-from-arbitrary choices made during the initial conception and subsequent operation of the screening program. The identification of a biochemical mode of action that appears to differ profoundly from that of previous antiparasitic agents has provided a tool for new approaches to understanding and exploiting the basic biochemical pathways of animal and plant parasites. The biological properties of the avermectins have opened new possibilities for the study of low-dose drug delivery systems, and many aspects of nematodology, entomology and acarology, as well as contributing directly to the control of many livestock parasites and agricultural pests. Literature Cited 1. Burg, R.W., Miller, B.M., Baker, E.E., Birnbaum, J., Currie, S.A., Hartman, R., Kong, Y-L., Monaghan, R.L., Olson, G., Putter, I., Tunac, J.B., Wallick, Η., Stapley, E.O., Oiwa, R., and Omura, S. Antimicrob. Agents Chemother. 1979, 15, 361-7. 2. Stapley, E.O. and Woodruff, H.B., in "Proceedings, An International Conference on Trends in Antibiotic Research"; Umezawa, H., Demain, A.L., Hata, T., and Hutchinson, C.R., Eds.; Japan Antibiotics Research Association, Tokyo, 1982; pp. 154-170.

CAMPBELL ET AL. Discovery of Ivermectin 19 3. Miller, T.W., Chaiet, L., Cole, D.J., Cole, L.J., Flor, J.E., Goegelman, R.T., Gullo, V.P., Kempf, A.J., Krellwitz, W.R., Monaghan, R.L., Ormond, R.E., Wilson, Κ.Ε., Albers-Schonberg, G., and Putter, I. Antimicrob. Agents Chemother. 1979, 15, 368. 5. Takiguchi, Y., Mishima, H., Okuda, M., Terao, M., Aoki, Α., and Fukuda, R. J. Antibiot. 1980, 33, 1120. 6. Albers-Schonberg, G., Arison, B.H., Chabala, J.C., Douglas, A.W., Eskola, P., Fisher, M.H., Lusi, Α., Mrozik, H., Smith, J.L., and Tolman, R.L. J. Am. Chem. Soc.1981, 103 4216. 7. Springer, J.P., Arison, B.H., Hirshfield, J.M., and Hoogsteen, K. J. Am. Chem. Soc. 1981, 103, 4221. 8. Ostlind, D.A., and Cifelli, S. Research in Veterinary Science, 1981, 31, 255-6. 9. Panitz, E., and Shum, K.L. J. Parasit. 1981, 67, 135-6. 10. Egerton, J.R., Ostlind, D.A., Blair, L.S., Eary, C.H., Suhayda, D., Cifelli, S., Riek, R.F., and Campbell, W.C. Antimicrob. Agents Chemother. 1979, 15, 372-8. 11. Blair, L.S., and Campbell, W.C. J. Helm. 1978, 52, 305-307. 12. Blair, L.S., and Campbell, W.C. Journal of Parasitology, 1978, 64(6), 1032-4. 13. Ostlind, D.A., Cifelli, S., and Lang, R. Vet. Rec., 1979, 105, 168. 14. James, P.S., Picton, J., and Riek, R.F. Vet. Rec. 1980, 106, 59. 15. Wilkins, C.A., Conroy, J.A., Ho, P., O'Shanny, W.J., Malatesta, P.F., and Egerton, J.R. Am. J. Vet. Res. 1980, 41, 2112-13. 16. Wilkins, C.A., Conroy, J., Ho. P., and O'Shanny, W.J. Proc. 25th Annual Mtg. Am. Assoc. Vet. Parasitol., Washington, 1980, p. 18. 17. Wang, C.C., and Pong, S.C. Progress in Clinical and Biological Research 1981, 97, 373-95 18. Putter, I., MacConnell, J.G., Preiser, F.Α., Haidri, A.A., Ristich, S.S. and Dybas, R.A. Experientia 1981, 37, 963-964. 19. Chabala, J.C., Mrozik, H., Tolman, R.L., Eskola, P., Lusi, Α., Peterson, L.H., Woods, M.F., Fisher, M.H., Campbell, W.C., Egerton, J.R., and Ostlind, D.A. J. Med. Chem. 1980, 23, 1134.

PESTICIDE SYNTHESIS THROUGH RATIONAL APPROACHES 20. Mrozik, Eskola, P., Fisher, M.H., Egerton, J.R., Cifelli, S., and Ostlind, D.A. J. Med. Chem. 1982, 25, 658. 21. Egerton, J.R., Birnbaum, J., Blair, L.S., Chabala, J.C., Conroy, J., Fisher, M.H., Mrozik, Η., Ostlind, D.A., Wilkins, C.A., and Campbell, W.C. Br. Vet. J. 1980, 136, 88-97. RECEIVED April 10, 1984