New tissue schizontocidal antimalarial drugs

Bulletin of the World Health Organization, 59 (3): 463-479 (1981) New tissue schizontocidal antimalarial drugs DAVID E. DAVIDSON, JR,1 ARBA L. AGER,2 JOHN L. BROWN,3 FRANK E. CHAPPLE,3 RICHARD E. WHITMIRE,3 & RICHARD N. ROSSAN4 Over 700 causal prophylactic and radical curative antimalarial drugs have been dis- covered during the screening of approximately 4000 chemical compounds in rodent and simian malaria models. Causal prophylactic activity in the Plasmodium berghei-rodent model was demonstrated by 10 distinct groups of chemicals: 1) tetrahydrofolate dehydro- genase inhibitors, 2) naphthoquinones, 3) dihydroacridinediones, 4) tetrahydrofurans, 5) guanylhydrazones, 6) analogues of clopidol, 7) quinoline esters, 8) dibenzyltetrahydro- pyrimidines, 9) 6-aminoquinolines, 10) 8-aminoquinolines. Ofthe causalprophylactic compounds, only the 6- and8-aminoquinolines were capable of curing persistent exoerythrocytic infections of P. cynomolgi in rhesus monkeys. The 6-aminoquinolines were substantially less active than primaquine. This report describes a series of 4-methyl-5-phenoxy-6-methoxy-8-aminoquinolines, which arepotent bloodschizontocides and radicalcurative drugs. The most active memberof this series, 4-methyl-5-(3-trifluoromethylphenoxy)-6-methoxy-8-[(4-amino-J-methylbutyl) aminolquinoline succinate (WR 225448), was 5 times more active thanprimaquine in curing persistent exoerythrocytic infections of P. cynomolgi in rhesus monkeys. Asa bloodschizontocide, WR 225448 was effective in animalmodelsagainst P. berghei, P. cynomolgi, P. vivax, and both drug-sensitive and drug-resistant strains of P. falciparum. WR 225448 was also more toxic than primaquine in rats on subacute (28-day) administration. The number of antimalarial drugs currently avail- able for clinical use, as causal prophylactic or radical curative drugs, is extremely limited. Only primaquine, the 8-aminoquinoline introduced nearly 30 years ago, is clinically effective against the persistent tissue stages of Plasmodium vivax or P. ovale in man. Close analogues, such as pentaquine, isopentaquine, pama- quine, quinocide, etc., are either less effective or more toxic than primaquine in clinical use. The toxicity of primaquine limits its clinical useful- ness in both prophylactic and therapeutic applica- tions. The most serious side-effect is haemolysis, which occurs in individuals who are genetically defi- cient in glucose-6-phosphate dehydrogenase (EC 1.1. 1.49) (1, 2). Methaemoglobinaemia, abdominal cramping, and epigastric distress are also significant side-effects of many 8-aminoquinolines (3). The 8-aminoquinolines as a class are hepatotoxic (4), and, while this is seldom a problem when primaquine is administered in acceptable dosages, it is a potential hazard of overdosing. I Chief, Department of Parasitology, Division of Experi- mental Therapeutics, Walter Reed Army Institute of Research, Washington, DC 20012, USA. 2 Director, Rane Research Laboratory, University of Miami Medical Center, Miami, FL, USA. 3Department of Veterinary Medicine, Armed Forces Research Institute of the Medical Sciences, Bangkok, Thailand. I Gorgas Memorial Laboratory, Panama City, Panama. Drug resistance has developed to almost every blood schizontocidal drug currently in clinical use. Fortu- nately, although there are geographic differences in the susceptibility of persistent tissue stages of vivax malaria to primaquine (3, 5), no true drug resistance has been unequivocally demonstrated. However, the ease with which Arnold et al. (6) were able to induce primaquine resistance experimentally in vivax malaria emphasizes the need for alternative tissue schizonto- cides. Because primaquine is unique as a radical curative drug and because the associated side-effects severely limit its use, it was decided several years ago to initiate a modest screening effort to search for alternative drugs, as part of the US Army Antimalarial Program. Approximately 4000 compounds of diverse structure have been screened, a high proportion of which were analogues of compounds previously reported to have tissue schizontocidal activity. In developing the screening strategy, existing animal models for causal prophylactic and radical curative testing were used. Tissue schizontocidal testing is intrinsically more complex and difficult than blood schizontocidal testing, and none of the existing models were capable of supporting economical, large-scale screening. It was necessary, therefore, to develop models to meet the needs of the programme. Several models were developed and operated by collaborating 4076 - 463 D. E. DAVIDSON, JR ET AL. laboratories, and some of these will be described. The data provided are intended to highlight some of the approaches that have yielded tissue schizontocidal compounds with improved activity. The approaches that were unproductive will also be indicated. METHODS In addition to the causal prophylactic testing against sporozoite-induced P. yoelii malaria in rodents, which was performed by the authors, test data were also provided by Wallace Peters of the Liverpool School of Tropical Medicine, and by Harry Most of New York University. Radical curative testing against sporozoite- induced P. cynomolgi malaria in rhesus monkeys was performed by Leon Schmidt of Southern Research Institute, Birmingham, Alabama, and by the authors. Toxicological studies were performed by C.C. Lee of Midwest Research Institute. The primary mouse prophylactic screening was per- formed using methods described by Rane& Kinnamon (7). In this screen, the test compound is administered subcutaneously to ICR/Ha mice, and followed two hours later by an intraperitoneal inoculation of a lethal dose of P. yoelii sporozoites. If the compound has activity against the pre-erythrocytic stages, or if it is a blood schizontocide and persists long enough to sup- press the blood forms emerging after day 3, survival time of the treated mice is prolonged. In this screening model, there is a 99Vo mortality rate among infected, untreated controls, with deaths occurring between days 6 and 17 (mean survival time, 9.8 days). A test compound is considered active if 2 or more treated mice survive to 30 days at any drug dose. It is to be emphasized that this screening test is not specific, since both causal prophylactic compounds and persistent blood schizontocides give positive results. Nevertheless, it does provide for rapid, inex- pensive screening of large numbers of compounds, and identifies the majority of blood schizontocides that are not persistent. Compounds that were active in the primary prophy- lactic mouse screen were evaluated in one of the more definitive rodent tests, which are capable of distin- guishing causal prophylactic compounds from those possessing only residual blood schizontocidal activity. The mouse causal prophylactic test described by Gregory & Peters (8) uses mice inoculated with sporo- zoites of P. yoelii, with parasitized blood, and with both. By a mathematical analysis of the effects of the test compound on the subsequent parasitaemias, it is possible to identify compounds that are truly causal prophylactics, and to estimate the level of activity. A second causal prophylactic model developed by Most& Montouri (9) also distinguishes between causal prophylactic and suppressive prophylactic activity. In this model, rats are given the test drug on two consecu- tive days, and, on the second day, are also inoculated intravenously with either 10 000 or 250 000 P. berghei sporozoites. Those receiving the higher sporozoite inoculum are sacrificed 43-45 h after inoculation and their livers are removed, sectioned, stained, and ex- amined microscopically for exoerythrocytic forms. Significant reduction or absence of exoerythrocytic forms is indicative of causal prophylaxis. The course of the parasitaemia following drug administration is monitored in the rats that received the lower sporozoite inoculum. Absence of parasitaemia on examination of stained blood films, confirmed by subinoculation of blood into mice, indicates a parasitological cure, which, in the absence of a cure of exoerythrocytic forms in the liver at 43-45 h, is indicative of suppres- sive prophylaxis. Radical curative testing against persistent tissue stages of P. cynomolgi bastianelli malaria in rhesus monkeys was performed using the methods described by Schmidt et al. (10). In this model, the test com- pound is administered orally for 7 consecutive days, beginning 10-12 days after intravenous inoculation of 0.5-1.5 x 106 sporozoites (2-4 days after the appear- ance of parasitaemia). Chloroquine phosphate (5 mg/kg of body weight per day) is administered con- comitantly with the test compound to eliminate blood forms and to permit assessment of the drug's activity against tissue stages, which are unaffected by chloro- quine. Relapse of parasitaemia after completion of the drug regimen is indicative of the failure of the test compound to eliminate all exoerythrocytic parasites. If parasitaemia does not reappear within 100 days (or within 30 days in the experiments in which splenectomy was carried out), the test compound is considered to have cured the exoerythrocytic infection. In the prophylactic rodent models, blood schizonto- cidal activity has an important impact on the interpre- tation of results because the tissue stages of rodent malaria are present for such a short time. For this reason, data on the blood schizontocidal activity of compounds were also obtained using the Rane model (11), which is a primary blood schizontocidal screen in mice. In this test, infected mice surviving for 60 days after administration of a single subcutaneous dose of the test compound are considered cured. Deaths occurring before day 6 are considered to be a result of drug toxicity. The blood schizontocidal activities of 7-day regimens of WR 225448 and primaquine were also assessed against trophozoite-induced infections of P. cynomolgi in rhesus monkeys (12) and against P. vivax inAotus trivirgatus monkeys of Panamanian origin, using methods described by Schmidt for Colombian owl monkeys (13). 464 NEW TISSUE SCHIZONTOCIDAL DRUGS 465 RESULTS The results of tissue schizontocidal tests for non- aminoquinolines are presented in Table 1. Only com- pounds that have been sufficiently well tested to provide assurance of true tissue schizontocidal activity are included. Furthermore, for each class of com- pound, only the most active representatives are mentioned. The compounds in Table 1 are grouped according to chemical class. Each compound is identified by its Walter Reed (WR) accession number. The corres- ponding structural formulae are given in Annex 1. Primary prophylactic mouse screening results are ex- pressed as a 50%o effective dose (ED50), which is defined as the lowest test dose (administered orally or subcutaneously) that permitted at least 5007o of the ani- mals to survive the otherwise lethal sporozoite inocu- lum. In each case in which true causal prophylactic activity was confirmed in one or both of the secondary rodent models (8, 9), this is indicated. A primaquine index, indicating the activity of the test compound relative to primaquine, is provided for the cases where the data obtained in the Peters model were adequate. Results of radical curative testing in the rhesus monkey-P. cynomolgi model are also provided, and again expressed as a primaquine index. The test com- pound is considered inactive ifno monkey was cured at the maximum dose tested (generally 10 mg/kg of body weight per day for 7 days). Results of blood schizontocidal testing in the Rane mouse model are presented as an approximate 50%o curative dose (CD50), which is defined as the lowest dose at which 50%0 or more of the mice were cured. Compounds producing no increase in survival time at the highest dose tested (640 mg/kg ofbody weight) are scored as inactive. For compounds that significantly increased survival time, but that were not sufficiently active to effect cures, the highest dose tested is indicated. Tetrahydrofolate dehydrogenase inhibitors Because of the many reports in the scientific litera- ture indicating causal prophylactic activity of this class ofcompound both in animal models and in man, many diaminopyrimidines, triazines, quinazolines, pteri- dines, and related inhibitors of tetrahydrofolate dehydrogenase (dihydrofolate reductase) (EC 1.5.1.3) have been screened for tissue schizontocidal activity. A few examples are presented in Table 1. In general, these compounds, which are blood schizontocides, are also potent causal prophylactic agents in the rodent models. None of the compounds, however, have ever exhibited activity against persistent tissue forms in the rhesus monkey model. Naphthoquinones A number of analogues of the well-known causal prophylactic compound menoctone (14) have been tested. Many, including WR 6012 and WR 25175, have exhibited causal prophylactic activity in the rodent models, but none has been more active than menoctone. Furthermore, neither menoctone nor its analogues have exhibited radical curative activity in the rhesus model. Because naphthoquinones are not readily absorbed in rhesus monkeyswhen administered orally, the intramuscular route was used for the radical curative tests. Dihydroacridinediones A number of analogues of the Hoechst compound Floxacrine (WR 233602)a have been screened for tissue schizontocidal activity. Like Floxacrine (15), the analogues listed in Table 1 exhibited potent causal prophylactic activity in the rodent models, but none had radical curative activity against P. cynomolgi in rhesus monkeys. As yet, only Floxacrine has been tested for causal prophylactic activity in the rhesus model (15). Tetrahydrofurans The causal prophylactic activity of this class ofcom- pounds has been described by Peters (16). On this basis, a number of analogues have been screened, and have exhibited causal prophylactic activity. These compounds reversibly inhibit the growth of folate- dependent bacteria (C.C. Smith, personal communi- cation, 1971), and may be tetrahydrofolate dehydro- genase inhibitors. Like other compounds with this mechanism of action, they have causal prophylactic activity in the rodent models, but have no radical curative activity against P. cynomolgi. Curiously, the compounds with a complete furan ring system are more potent blood schizontocides, while those that may be viewed as an opened ring (e.g., WR 179305 and WR 199334), appear to be more potent causal prophylactics. Guanylhydrazones A number of guanylhydrazones have been found to have modest blood schizontocidal and causal prophy- lactic activity in the rodent models. One of these, WR 9792, has been tested against P. cynomolgi, but had no radical curative activity. Clopidol analogues Clopidol (WR 61112) and one of two analogues tested had weak causal prophylactic activity in the a 7-chloro-3,4-dihydro-10-hydroxy-3-[4-(trifluoromethyl)phenyl] -1,9 (2H, lOH)-acridinedione. 466 D. E. DAVIDSON, JR ET AL. Table 1. Antimalarial activity of non-aminoquinolines Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED50 (mg/kg of radical curative (mg/kg of body weightWR body weight) test subcutaneous) comnpound no. __________________ ___________ Minimum Subcutaneous Oral Primaquine index CD50 toxic dose Tetrahydrofolate dehydrogenase inhibitors 2978 1.25sb 2.5 inactive 80 160 CP activeC 5473 5s,b 2.5 inactive > 640 > 640d CP activec 38839 0.63k 0.63 inactive° 80 > 640 159412 1.258 1.25 inactive" 10 > 640 206891 2.5a,b 2.5 inactive 10 > 640 Naphthoquinones 49808 101 108 inactive 320 > 640 (intramuscular) [inactive orally] 6012 40, b 408 inactive 320 > 640 (intramuscular) 25175 40ab 408 inactive > 640 > 640 (intramuscular) Dihydroacridinediones 233602 2.5 2.5 inactive 20 640 (intramuscular) CP activec 226626 1 0a,b 108 inactive 160 > 640 (P = 20.6)f (P = 0.6)f 226970 2.5 10 inactive 160 > 640 234062 160 40 inactive > 640 > 640 Tetrahydrofurans 93133 408 408 inactive° 80 640 190729 160b 160 inactive 80 640 179305 1 a,b 10 inactive9 > 640 320 199334 40 10 > 640 640 Guanylhydrazones 9792 40b 10 inactive 640 160 99682 160 40 80 > 640 91808 408 40 80 > 640 Clopidol analogues 61112 160b 160 inactive > 640 > 640 156949 160 160 inactive > 640 167655 inactive inactive inactive° > 640 > 640 Quinoline esters 7295 160 160 > 640 > 640 194905 0.638 0.63w inactive > 640 > 640 Dibenzyl pyrimidines 158124 40 40 inactive inactive > 640 214235 160 160 214705 160 160 Causal prophylactic activity confirmed in secondary rodent model by Peters et al. (8). b Causal prophylactic activity confirmed in secondary rodent model by Most & Montouri (9). c CP = Causal prophylactic test: P. cynomolgi in rhesus monkey. Data provided by Leon H. Schmidt. d No toxic deaths at highest dose tested, i.e., 640mg/kg of body weight. ° Data provided by Leon H. Schmidt, Southern Research Institute, Birmingham, Alabama. f P = Primaquine index in secondary rodent model of Peters et al. (8). NEW TISSUE SCHIZONTOCIDAL DRUGS rodent models. Radical curative activity in the rhesus model was not detected. Quinolones Two analogues of the quinolones and acetoxy deri- vatives described by Ryley & Peters (17) were tested. One of these, WR 194905, was among the most active causal prophylactics in the rodent model, although its blood schizontocidal activity was weak. In the rhesus model, there was no evidence of radical curative activity. Dibenzylpyrimidines Three compounds of this class have exhibited acti- vity in the causal prophylactic mouse screen. While this has not as yet been confirmed in secondary rodent testing, results in mice suggest that the activity is not due to residual blood schizontocidal activity. Only one of these compounds, WR 158124, has been tested for radical curative activity and it was ineffective. 6-Aminoquinolines The results of testing of a group of 5, 8-dimethoxy- 6-aminoquinolines are presented in Table 2. All of the compounds listed were tested in the rhesus monkey radical curative test, but rodent causal prophylactic and blood schizontocidal data are not available for several compounds. Two compounds in this group, WR 188438 and Ni-147/36, exhibited modest radical curative activity in the rhesus model, although both were considerably less active than primaquine. WR 188438, WR 203766, and Ni-147/36 also exhibi- ted causal prophylactic activity in the rodent systems. 7-Aminoquinolines The results of testing a group of 5, 8-dimethoxy-7- aminoquinolines are presented in Table 3. None of these compounds had any activity in either the prophy- lactic or blood schizontocidal rodent models. One compound, WR 213640, had weak radical curative activity in the rhesus model. Table 2. Antimalarial activity of 6-aminoquinolines5 Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED,0 (mg/kg of radical curative (mg/kg of body weight body weight) test subcutaneous) Compound R2 R4 R no. Sub- Minimum cutaneous Oral Primaquine index CD,0 toxic dose CH, 181614 -CH3 H -CH(CH2)3NH - Inactiveb CH3 182144 -CH3 -CH3 -CH(CH2)3NH - Inactiveb CH3 182146 -CH3 -CH3 -CH(CH2)3NH2 Inactiveb CH3 188438 -CH3 -CH3 -CH(CH2)3NH-iJ> 160 44 Inactive 640 C2H5 199065 -CH3 -CH3 -[(CH2)2N(C2H,)12h Inactive Inactiveb Inactive > 640 CH3 203766 -H -CH3 -CH-(CH2 ) 3NH2 40 40 Inactiveb CH3 Ni-147/36 -CH3 -CH3 -CH-(CH2)3N(C2H.)2 < 0.5b a Basic chemical structure given in Annex 1. b Data provided by L. H. Schmidt, Southern Research Institute, Birmingham, Alabama. 467 D. E. DAVIDSON, JR ET AL. Table 3. Antimalarial activity of 7-aminoquinolinesa Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED,0 (mg/kg of radical curative (mg/kg of body weight body weight) test subcutaneous) Compound R2 R4 R no. Sub- Minimum cutaneous Oral Primaquine index CD,0 toxic dose CH3 207766 -CH3 -CH3 -CH-(CH2)3NIC2H1-2 Inactive Inactiveb Inactive > 640 CH3 213640 -H -H -CH-(CH2)3N(C2H-)2 Inactive Inactive < 0.1b CH3 217270 -CH3 -CH3 -CH-(CH2)3NH2 Inactiveb CH3 218336 -H -H -CH-ICH2)3NH2 Inactive Inactiveb CH3 218677 -H -CH3 -CH-(CH2)3N(C2H,)2 Inactiveb CH3 218948 -CH3 -CH3 -CH-(CH2 )3NH 7) Inactive Inactiveb -N C2H- 219008 -H -H -ICH2)4-N N-CH2CH2OH Inactive Inactiveb 8 Basic chemical structure given in Annex 1. b Data provided by L. H. Schmidt, Southern Research Institute, Birmingham, Alabama. 1-A minonaphthalenes The results of testing a group of l-aminonaphtha- lenes analogous to the 6-methoxy-8-aminoquinolines are presented in Table 4. None of these compounds exhibited blood or tissue schizontocidal activity. 8-Aminoquinolines The results of testing of an interesting class of 5-aryloxy-8-aminoquinolines are presented in Tables 5-10. Although side-chain variants of this class have been synthesized and tested, only compounds with the primaquine side-chain are included in the tables. The results for 5-phenoxy compounds without methyl substitution on the quinoline ring are listed in Table 5. It is particularly notable that none of the com- pounds in this group had causal prophylactic activity in the rodent models, yet all of them had radical cura- tive activity in the rhesus model. With the exception of WR 215295 (primaquine index 3.2), the radical cura- tive activity was modest, with primaquine indices ranging from 0.2 to 1.4. Blood schizontocidal activity, like that of primaquine, was low in the mouse model. In general, these compounds extended survival time, but did not cure even at the highest dose tested. WR 216100, WR 235720, and WR 215295 were slightly more active than primaquine as blood schizon- tocides, curing at 320-640 mg/kg of body weight. It is also notable that all of these 5-phenoxy compounds were substantially less toxic than primaquine, and only WR 235724 produced toxic deaths at 640 mg/kg of body weight, whereas primaquine is toxic at 160 mg/kg of body weight. The results for 5-phenoxycompounds with 2-methyl substitution on the quinoline ring are listed in Table 6. These were also inactive in the causal prophylactic mouse screen, but had radical curative activity in the rhesus model. Again, all the 2-methyl compounds were less toxic than primaquine in mice, and had only weak blood schizontocidal activity. The 2-methyl-4- chlorophenoxy and 4-fluorophenoxy compounds had substantially stronger radical curative activity than their non-methyl substituted analogues (3.1 and 3.0 versus 1.0 and 1.4, respectively). The 3-trifluoro- 468 469NEW TISSUE SCHIZONTOCIDAL DRUGS Table 4. Antimalarial activity of aminonaphthalenesa Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED50 (mg/kg of radical curative (mg/kg of body weight body weight) test subcutaneous) Compound Rl R R3 no. Sub- Minimum cutaneous Oral Primaquine index CD50 toxic dose CH3 180128 -OCH3 -H -CH(CH2)3N(C2H5)2 Inactive CH3 218575 -OCH3 -OCH3 -CH(CH2)3N(C2H5)2 Inactiveb CH3 232036 -OCH3 -OCH3 -CH(CH2)3NH2 Inactive Inactive Inactive 640 CH3 232143 -OCH3 -OCH3 -(CH2)3CH-NH2 Inactive Inactive Inactive 160 232439 -OCH3 -OCH3 -CH-(CH2)3-NH2 Inactive Inactive Inactive 320 C2H5 a Basic chemical structure given in Annex 1. b Data provided by L. H. Schmidt, Southern Research Institute, Birmingham, Alabama. Table 5. Antimalarial activity of 5-phenoxy-8-aminoquinolinesa Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED50 (mg/kg of radical curative (mg/kg of body weight Compound no. Radical Salt body weight) test subcutaneous) Minimum Subcutaneous Oral Primaquine index CD50 toxic dose 225374 -H H3PO4 inactive inactive 1.3 > 640 > 640 182232 4-Cl H20 inactive inactive 1.0 > 640 > 640 216100 4-F Succinate inactive inactive 1.4 320 > 640 235720 4-OCH3 Citrate inactive 0.2 640 > 640 235724 4-OCF3 Citrate inactive 0.2 > 640 640 215295 3-CF3 Succinate inactive inactive 3.2b 640 > 640 233878 2,4-Cl H3P04 inactive inactive 0.4 > 640 > 640 233881 3,4-Cl Fumarate inactive 0.4 > 640 > 640 234738 3,5-CF3 2H3P04 inactive 0.6 > 640 > 640 Primaquine 2H3PO4 50c,d 50c 1.0 > 640 160 a Basic chemical structure given in Annex 1. b Data provided by L. H. Schmidt, Southern Research Institute, Birmingham, Alabama. c Causal prophylactic activity confirmed in secondary rodent model by Peters et al. (8). d Causal prophylactic activity confirmed in secondary rodent model by Most E Montouri (9). D. E. DAVIDSON, JR ET AL. Table 6. Antimalarial activity of 2-methyl-5-phenoxy-8-aminoquinolines8 Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED,0 (mg/kg of radical curative (mg/kg of body weight Compound no. Radical Salt body weight) test subcutaneous) Minimum Subcutaneous Oral Primaquine index CD,0 toxic dose 211532 4-Cl Fumarate inactive inactive 3.1 b 320 > 320 224097 4-F Fumarate inactive 3.0 320 > 640 224486 3-CF3 Fumarate inactive inactive 1.7 > 640 320 2-Methyl- 2HCI 40C,d 40 1.0 > 640 320 primaquine (P = 3.8)e 8 Basic chemical structure given in Annex 1. b Data provided by L. H. Schmidt, Southern Research Institute, Birmingham, Alabama. c Causal prophylactic activity confirmed in secondary rodent model by Peters et al. (8). d Causal prophylactic activity confirmed in secondary rodent model by Most & Montouri (9). * Pnmaquine index in secondary rodent model of Peters et al. (8). methylphenoxy analogue wasless active as the2-methyl quinolines appeared to enhance substantially both substituted compound. blood and tissue schizontocidal activity, without Only one 3-methyl compound has been tested so far increasing toxicity (Table 8). All the 4-methyl (Table 7). In the radical curative test, its primaquine compounds listed had potent radical curative activity, index was only 1.5, but it had a high level of blood with primaquine indices ranging from 4.2 to 4.8, and schizontocidal activity and was less toxic than all were potent blood schizontocides, curing mice at primaquine in the mouse. It was also effective in the doses of 5-40 mg/kg of body weight. All the mouse prophylactic screen, although residual blood compounds were highly active in the primary prophyl- schizontocidal activity has not yet been ruled out. It is actic mouse screen, but secondary test in the Peters notable that 3-methylprimaquine itself is not a potent model showed that this was attributable to residual blood schizontocide, although its radical curative blood schizontocidal activity. However, the Most potency is comparable with that ofWR 235485, and it technique showed that WR 225448 was a true causal is a powerful causal prophylactic in the mouse model. prophylactic in the rat. By comparison, 4-methyl- 4-Methyl substitution on 5-phenoxy-8-amino- primaquine itself was a true causal prophylactic in the Table 7. Antimalarial activity of 3-methyl-5-phenoxy-8-aminoquinolines8 Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED50 (mg/kg of radical curative (mg/kg of body weight Compound no. Radical Salt body weight) test subcutaneous) Minimum Subcutaneous Oral Primaquine index CD50 toxic dose 235485 3-CF3 Succinate 20 10 1.5 20 > 640 3-Methyl- 2HCI 40b 40b 1.3d > 160 80 primaquine (P = 37.0)c (P = 3.8)c 9 Basic chemical structure given in Annex 1. b Causal prophylactic activity confirmed in secondary rodent model by Peters et al. (8). c Primaquine index in secondary rodent model. d By both oral and subcutaneous administration. 470 NEW TISSUE SCHIZONTOCIDAL DRUGS Table 8. Antimalarial activity of 4-methyl-5-phenoxy-8-aminoquinolinesa Mouse causal Rhesus/ Mouse blood prophylactic screen P. cynomolgi schizontocidal test ED,0 (mg/kg of radical curative (mg/kg of body weight Compound no. Radical Salt body weight) test subcutaneous) Minimum Subcutaneous Oral Primaquine index CD,0 toxic dose 232956 4-F H3PO4 40b 40b 4.2 20 320 232584 4-OCH3 H3PO4 40b 40b 4.3 40 640 225448 3-CF3 Succinate 20b job 4.8 20 640 233195 2,4-Cl H3PO4 40b 40b 4.6 20 > 640 233078 3,4-Cl H3PO4 40b 40b 4.6 5 > 640 4-Methyl- 2H3P04 50c,d 25c, d 2.1 640 640 primaquine (P= 0.7)e (P = 2.3)e a Basic chemical structure given in Annex 1. b Activity in Peters model (8) attributed to residual blood schizontocidal properties. c Causal prophylactic activity confirmed in secondary rodent model by Most & Montouri (9). d Causal prophylactic activity confirmed in secondary rodent model by Peters et al. (8). e Primaquine index in secondary rodent model of Peters et al. mouse, a weak blood schizontocide, and a radical curative agent in the rhesus monkey with a primaquine index of 2.1. The results of studies on the blood schizontocidal activity ofWR 225448 are presented in Tables 9 & 10. Trophozoite-induced P. cynomolgi parasitaemias were consistently cured by a 7-day oral regimen of 1.0 mg/kg of body weight per day or more of the com- pound, and it was clearly superior to primaquine in this blood schizontocidal model. While transient Table 9. Blood schizontocidal activity of a 7-day course of WR 225448 against trophozoite-induced P. cynomolgi malaria in rhesus monkeys WR 225448 Primaquine Daily oral dose Total dose No N. N. N.(mg/kg of (mg/kg of cleared cured cleared cured body weight) body weight) 31.6 220 2/2 0/2 10.0 70 2/2 2/2 2/2 0/2 3.16 22 2/2 2/2 2/2 0/2 1.00 7 2/2 2/2 2/2 0/2 0.316 2.2 2/2 0/2 2/2 0/2 0.100 0.7 2/2 0/2 0/2 0/2 0.0316 0.22 2/2 0/2 clearing of parasitaemia was regularly obtained after administration of a total of 2.2 mg of primaquine/kg ofbody weight or more, blood schizontocidal cure was never attained even at a total dose of 220 mg/kg of body weight. WR 225448 was also effective as a blood schizonto- cide against trophozoite-induced Chesson strain P. vivax malaria in Panamanian Aotus trivirgatus monkeys. It cleared parasitaemia in all monkeys at the Table 10. Blood schizontocidal activity of a 3-day course of WR 225448 against trophozoite-induced P. vivax malaria (Chesson strain) in Aotus trivirgatus Daily oral dose Total dose (mg/kg of (mg/kg of No. cleared No. cured body weight) body weight) WR 225448 16 48 3/3 3/3 4 12 4/4 4/4 2 6 5/5 0/5 1 3 3/3 0/3 Primaquine 53 160 3/3 0/3 27 80 3/3 0/3 13 40 3/3 0/3 3.3 10 0/4 0/4 471 D. E. DAVIDSON, JR ET AL. lowest dose tested (1 mg/kg ofbody weight per day for 3 days, orally), and was fully curative at a total dose of 12 mg/kg of body weight (4 mg/kg per day for 3 days). Primaquine, tested simultaneously, cleared parasitaemia only in a total 3-day regimen dosage of 40 mg/kg of body weight or more, and was not curative, even at a total dose of 160 mg/kg of body weight. DISCUSSION Over a four-year period, approximately 4000 com- pounds have been screened for causal prophylactic and radical curative activity in one or more animal models. Because compounds have been selected for test on the basis of structural or functional analogy with known causal prophylactic compounds, a large number of them have exhibited such activity. To date, over 700 active compounds, the majority of which are 8-aminoquinolines, have been identified. There were few active non-aminoquinolines, except for tetrahydrofolate dehydrogenase inhibitors, of which there are many active examples. Compounds identified as having causal prophylactic activity in the rodent models may be grouped into eight general classes: 1. Tetrahydrofolate dehydrogenase inhibitors 2. Naphthoquinones 3. Dihydroacridinediones 4. Tetrahydrofurans 5. Guanylhydrazones 6. Clopidol analogues 7. Quinoline esters 8. Dibenzyltetrahydropyrimidines For each of these classes, reports of causal prophyl- actic activity were available prior to testing, and thus no new chemical classes have been identified in this screen. However, specific compounds within these classes have been demonstrated to have activity which has not previously been reported. It is of interest to note that of the listed non-amino- quinolines exhibiting causal prophylactic activity in rodents, none were active against persistent tissue stages of P. cynomolgi in the radical curative test. There are, however, published accounts for a number of the compounds listed in this report indicating activity against pre-erythrocytic tissue forms of P. cynomolgi in rhesus monkeys, or pre-erythrocytic forms of P.falciparum and P. vivax in man. This work has been reviewed by Peters (18). In man, tetrahydro- folate dehydrogenase inhibitors, combinations of these with sulfones or sulfonamides, and the 8-amino- quinolines are effective in causal prophylaxis of falciparum or vivax malaria. Limited clinical trials of several non-aminoquinolines have been reported by Canfield et al. (19), but none were successful. Menoc- tone produced no causal prophylactic effect in volun- teers receiving 500 mg daily for 3 days after challenge with P.falciparum. Clopidol produced unacceptable neurological side-effects, did not control P.falci- parum parasitaemia, and was excluded from further consideration before causal prophylactic studies could be performed. The pyrocatechol RC-12 (WR 27653), a compound reported to have causal prophylactic activity against P. cynomolgi in rhesus monkeys, but ineffective in rodent malaria models (20), also failed to exhibit causal prophylactic activity against P. vivax challenge. The newer dihydrotriazine, WR 38839 (clociguanil), was effective in causal prophylaxis against P.falciparum only when administered in combination with sulfadiazine. To our knowledge, no representatives of the dihydroacridinedione class, the tetrahydrofurans, the guanylhydrazones, the quinoline esters, or the dibenzylpyrimidines have been evaluated clinically. A quinoline ester (ICI 56780) has been reported to have causal prophylactic activity against P. cynomolgi in rhesus monkeys (18), and Floxacrine, the dihydro- acridinedione, was also active in this simian model (15). Fink (21) has suggested that the rodent models may exaggerate the activity of compounds that interfere with nucleic acid synthesis because of the high rate of synthesis in P. berghei exoerythrocytic forms. This might also account, in part, for the moderate sensiti- vity of pre-erythrocytic stages of simian and human malarias to compounds such as the tetrahydrofolate dehydrogenase inhibitors, and the total lack of activity of these compounds against persistent exoerythrocytic forms. Clearly, there is an important need for research into the biochemistry and physiology of exoerythro- cytic forms. Efforts tofindaminoquinolines, otherthan 8-amino- quinolines, with tissue schizontocidal activity have not been successful. A few 6-aminoquinolines have exhi- bited weak causal prophylactic activity in the rodent models, and two analogues (WR 188438 and Ni-147/ 36) had weak radical curative activity. The 7-amino- quinolines tested were virtually devoid of tissue schizontocidal activity. Several 3-, 4-, and 5-amino- quinolines have also been synthesized and tested, and these exhibited no tissue schizontocidal activity. Aminonaphthalenes were also devoid of antimalarial activity. A variety of isoquinolines and azoquinolines, analogous to the 8-aminoquinolines, have also been tested and were inactive. The activity of the 4, 5-disubstituted 6-methoxy-8- aminoquinolines is, we believe, an important observa- tion. The marked enhancement of radical curative activity of these disubstituted compounds would not have been expected from the activity of either the 472 NEW TISSUE SCHIZONTOCIDAL DRUGS 4-methyl or the 5-phenoxy-8-aminoquinolines. The compounds listed in Table 8 have the most potent radical curative activity of all the compounds ex- amined. The inactivity of the 5-phenoxy-8-aminoquinolines in the rodent causal prophylactic test is notable. We are unaware of any other compound, except the pyro- catechol RC-12 (20), with tissue schizontocidal acti- vity against P. cynomolgi in rhesus monkeys that is not a causal prophylactic in the rodent models. The reason for this species difference is unknown, although it may be a result of differences in metabolism between mice and monkeys, but this has not yet been tested experi- mentally. Greenberg (22) demonstrated that prim- aquine and pentaquine were inactive against P. gallinaceum in vitro while metabolites of the drugs were active. It has been suggested by Greenberg that 6-quinone and 5, 6-quinone may be the active metabo- lites, and Smith (23) has identified a 5, 6-quinoline- quinone metabolite of pentaquine in the rhesus monkey. The rationale for developing these 5-aryloxy analogues was to block metabolism. It will be interest- ing to determine whether quinone is formed during the metabolism of these new compounds in mice or monkeys. It should also be noted that WR 225448 was a causal prophylactic in the Most rat model (Table 8). If differences in host metabolism are responsible for the variation in activity, then the rat and monkey are different from the mouse. The potent blood schizontocidal activity of 3- and 4-methyl-8-aminoquinolines was also unexpected, since 3- and 4-methylprimaquine are weak blood schizontocides, as are the 2-methyl and non-methyl substituted 5-phenoxy-8-aminoquinolines. This acti- vity was observed in mice as well as in rhesus and owl monkeys, and was observed against P. berghei, P. cynomolgi, and P. vivax. WR 225448 was also effective as a blood schizontocide against drug-sensi- tive and drug-resistant strains of P.falciparum in owl monkeys (R.N. Rossan, personal communication, 1980), but was inactive against P.falciparum in vitro (R. Desjardins, personal communication, 1978). Some preliminary toxicological studies have been performed withWR 225448 (C.C. Lee, personal com- munication, 1980). On single-dose subcutaneous administration in mice, it was less toxic than prima- quine (Table 8). WR 225448 induced methaemoglo- binaemia in the dog, but its potency in this respect relative to primaquine is not yet known. In 28-day oral toxicity studies in rats, WR 225448 was found to be more toxic than primaquine. Hepato- toxicity was more severe, and in addition, WR 225448 produced renal tubular degeneration and lymphoid depletion. BothWR 225448 and primaquine produced degenerative changes in the heart and diaphragmatic muscle. While WR 225448 administered subacutely is qualitatively and quantitatively more toxic than primaquine, it is still not known whether the thera- peutic index is better or worse. Further toxicity studies with WR 225448 and other analogues are planned. Annex 1 CHEMICAL STRUCTURES OF COMPOUNDS TESTED NON-AMINOQUINOLINES (TABLE 1) CH 3CH2 C1 ( /N\ NH 2 -N NH2 (Pyri methami ne) CH3 Ci 0 N \ NH2 -N NH2 (Cycloguanil pamoate) 5473 473 2978 D. E. DAVIDSON, JR ET AL. 0 OH (CH2)33 25175 0 0 C1 (Fl oxacrine) 233602 ci ~ N CH2 LH (CH3 )2 N 4 206891 0 OH (CH2)8-Q (Menoctone) 49808 0 OH ,r (CH 2 0K 'NH2 o 0 Cl 226626 N OH CF3 226970 0 0 N1 OCHCH N(C2H5 )2 OH 234062 NH2 (Cl oci guani 1 ) 38839 NH2 159412 474 6012 NEW TISSUE SCHIZONTOCIDAL DRUGS NH N-NH-C-NH2 CF3 m C - C1 99682 NH N-NH-C-NH2 C C-CH=C- C CH3 91808 OH CH3 CH3 (Clopidol ) 61112 - Cl C1 CH3 156949 NH N-NH-C-NH2 CF3 Oe C F C1 I CH3 167655 - Cl 93133 C1 190729 CH ' 3 CH3CH20 179305 CH3CH20- 199334 'C1 CH3 475 9792 D. E. DAVIDSON, JR ET AL. OH 3110 O (CH2)6CH3 CH30 N CH3 7295 OCOCH3 CH3(CH2)90 COOCH3 (CH3)2CHO N Ci NH N Cl "CH2 2 194905 214235 214705 AMINOQUINOLINE DERIVATIVES CH30 R4 R-NH CH30 CH30 R4 R-HN N2 CH30 6-Aminoquinoline (Table 2) ci Cl 158124 476 7-Aminoquinoline (Table 3) NEW TISSUE SCHIZONTOCIDAL DRUGS R -0 CH33 N NHCH(CH3) (CH2)3NH2 Aminonaphthalene (Table 4) R NJ 0 CH30 N CH3 NHCH(CH3) (CH2)3NH2 2-Methyl-5-phenoxy-8-aminoquinoline (Table 6) R 0 5-Phenoxy-8-aminoquinoline (Table 5) R- 0 CH3 0 CH 3~~~~ NHCH(CH3) (CH2)3NH2 3-Methyl-5-phenoxy-8-aminoquinoline (Table 7) 4-Methyl-5-phenoxy-8-aminoquinoline (Table 8) R2 R1N NH-R3 477 478 D. E. DAVIDSON, JR ET AL. RLSUMP NOUVEAUX SCHIZONTOCIDES TISSULAIRES CONTRE LE PALUDISME Afin de rechercher des schizontocides tissulaires plus surs et plus efficaces, on a procede au criblage d'environ 4000 composes contre Plasmodium berghei yoelii dans des modeles murins adaptes aux etioprophylactiques, ou contre Plasmodium cynomolgi bastianelli dans un modele de traitement radical chez le singe rhesus. Les composes selec- tionnes pour le criblage etaient en majeure partie des analogues structuraux ou fonctionnels d'etioprophylac- tiques connus; environ 700 d'entre eux, pour la plupart des amino-8 quinoleines, se sont reveles actifs. Des composes de 10 familles chimiques differentes etaient doues d'activite etioprophylactique dans les modeles murins: 1) inhibiteurs de la tetrahydrofolate deshydroge- nase, 2) naphtoquinones, 3) dihydroacridinediones, 4) tetra- hydrofurannes, 5) guanylhydrazones, 6) analogues du clopidol, 7) esters de la quinoleine, 8) dibenzyltetrahydro- pyrimidines, 9) amino-6 quinoleines, 10) amino-8 quino- 1mines. L'activite pour la guerison radicale des infections a P. cynomolgi chez le singe rhesus n'a et observee qu'avec les composes appartenant aux familles des amino-6 et amino-8 quinoleines. Toutes les amino-6 quinoleines testees etaient sensiblement moins actives que la primaquine dans l'epreuve de guerison radicale. Une famille de methyl-4 phenoxy-5 methoxy-6 amino-8 quinoleines etait douee d'une activite exceptionnelle pour la guerison radicale des infections a P. cynomolgi chez le singe rhesus. De plus, ces composes etaient fortement actifs comme schizontocides sanguins. L'activite optimale comme schizontocides tissulaires a et observ&e avec le substituant methyl-4; les analogues non substitues, ou substitues en methyl-2 et methyl-3, etaient moins efficaces. L'activite optimale comme schizontocides sanguins s'observait egalement avec les analogues substitues en methyl-3 et methyl-4. Le representant le plus efficace de cette nouvelle famille de schizontocides tissulaires etait leWR 225448 (succinate de methyl-4 (trifluoromethyl-3 phenoxy)-5 methoxy-6 (amino-4 methyl-1 butylamino)-8 quinoleine). Dans le modele a singe rhesus, ce compose etait cinq fois plus efficace que la primaquine dans la guerison des infections exoerythrocytaires persistantes a P. cynomolgi. Dans ce cas, on ne connait aucun compose plus actif. Comme schizonto- cide sanguin, le WR 225448 est tres actif contre P. berghei, P. cynomolgi et P. vivax, et contre les souches tant resistantes que sensibles de P.falciparum dans les modeles animaux. Des etudes preliminaires de toxicite sur 28 jours chez le rat montrent que le WR 225448 est plus toxique que la prima- quine. Comme cette derniere, il est hepatotoxique, induit une methemoglobinemie, et provoque une degenerescence dans les muscles cardiaque et diaphragmatique. En outre, il induit des modifications d6g6n6ratives du rein et une hypo- plasie lymphocytaire. Des etudes complementaires destinees A evaluer et A chiffrer cette toxicite sont en cours. Des etudes toxicologiques sur les autres analogues sont prevues. REFERENCES 1. CARSON, P. E. ET AL. Enzymatic deficiency in prima- quine-sensitive erythrocytes. Science (New York), 124: 484-485 (1956). 2. ALVING, A. S. ET AL. Malaria, 8-aminoquinolines and haemolysis. In: Goodwin, L. G. & Nimmo-Smith, R. H. ed., Drugs, parasites and hosts, London, J. A. Churchill Ltd, 1962, pp. 83-97. 3. ALVING, A. S. ET AL. Mitigation of the haemolytic effect of primaquine and enhancement of its action against exoerythrocytic forms of the Chesson strain of Plasmodium vivax by intermittent regimens of drug administration. A preliminary report. Bulletin of the World Health Organization, 22: 621-631 (1960). 4. SCHMIDT, L. H. ET AL. Comparison of the curative anti- malarial activities and toxicities of primaquine and its d and I isomers. Antimicrobial agents and chemotherapy, 12: 51-60 (1977). 5. BLACK R. H. Results of the clinical use of primaquine for the eradication of relapsing malaria of South-West Pacific origin. Australasian annals of medicine, 1: 259 (1958). 6. ARNOLD, J. ET AL. Induced primaquine resistance in vivax malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 55: 345-350 (1961). 7. RANE, D. S. & KINNAMON, K.E. The development of a high volume tissue schizontocidal drug screen based upon mortality of mice inoculated with sporozoites of Plasmodium berghei. American journal of tropical medicine and hygiene, 28: 937-947 (1979). 8. GREGORY, K. G. & PETERS, W. The chemotherapy of rodent malaria, IX. Causal prophylaxis, Part I: A method for demonstrating drug action on exo-erythro- cytic stages. Annals of tropical medicine and parasit- ology, 65: 15-24 (1970). 9. MOST, H. & MONTOURI, W. A. Rodent systems (Plas- modium berghei-Anopheles stephensi) for screening compounds for potential causal prophylaxis. American journal of tropical medicine and hygiene, 24: 179-182 (1975). 10. SCHMIDT, L. H. ET AL. The activity of a repository form of 4, 6-diamino-l-(p-chlorophenyl)-1, 2-dihydro-2, 2- dimethyl-s-triazine against infections with Plasmodium cynomolgi. American journal of tropical medicine and hygiene, 12: 494-503 (1963). 11. OSDENE, T. S. ET AL. 2,4,7-triamino-6-ortho-substi- tuted aryl-pteridines. A new series of potent antimalarial agents. Journal of medicinal chemistry, 10: 431-434 (1967). NEW TISSUE SCHIZONTOCIDAL DRUGS 479 12. DAVIDSON, D. E. ET AL. Evaluating new antimalarial drugs against trophozoite-induced Plasmodium cynomolgi malaria in rhesus monkeys. American journal of tropical medicine and hygiene, 25: 26-33 (1976). 13. SCHMIDT, L. H. Infections with Plasmodium falci- parum and Plasmodium vivax in the owl monkey- Model systems for basic biological and chemothera- peutic studies. Transactions of the Royal Society of Tropical Medicine and Hygiene, 67: 446-474 (1973). 14. BERBERIAN, D. A. ET AL. Causal prophylactic effect of menoctone (a new hydroxynaphthoquinone) against sporozoite-induced Plasmodium berghei infection in mice. Journal ofparasitology, 54: 1181-1189 (1968). 15. SCHMIDT, L. H. Antimalarial properties of floxacrine, a dihydroacridinedione derivative. Antimicrobial agents and chemotherapy, 16: 475-485 (1979). 16. PETERS, W. Substituted tetrahydrofurans, a new chemi- cal family of antimalarials. The action of 2-(p-chloro- phenyl)-2-(4-piperidyl) tetrahydrofuran against Plas- modium berghei and Plasmodium chabaudi. Annals of tropical medicine and parasitology, 64: 189-202 (1970). 17. RYLEY, J. F. & PETERS, W. The antimalarial activity of some quinolone esters. Annals of tropical medicine and parasitology, 64: 209-222 (1970). 18. PETERS, W. Chemotherapy and drug resistance in malaria. London, Academic Press, 1970. 19. CANFIELD, C. J. & ROZMAN, R. S. Clinical testing of new antimalarial compounds. Bulletin of the World Health Organization, 50: 203-212 (1974). 20. SCHMIDT, L. H. ET AL. Studies on the antimalarial activity of 1,2-dimethoxy-4-(bis-diethylaminoethyl)- amino-5-bromobenzene. Bulletin of the World Health Organization, 34: 783-788 (1966). 21. FINK, E. Assessment of causal prophylactic activity in Plasmodium bergheiyoelii and its value for the develop- ment of new antimalarial drugs. Bulletin of the World Health Organization, 50: 213-222 (1974). 22. GREENBERG, J. ET AL. Studies on Plasmodium gallinaceum in vitro. II. The effects of some 8-amino- quinolines against the erythrocytic parasites. Journal of infectious diseases, 88: 163-167 (1951). 23. SMITH, Metabolism of pentaquine in the rhesus monkey. Journal of pharmacology and experimental thera- peutics, 116: 67-76 (1956).