Journal of Jlcdicinal Cheinislry, 1970, Vol. 1.3, Ll'o. 3 511
LIPOPHILIC CHARACTER
Influence of Lipophilic Character on the Antibacterial Activity of Cephalosporins and Penicillins GIANLurar BIAGI,XARIA CLELIAGUER~IA. AXNAMAMABAIIBIIW, AND J I A R I A FRA4XCESCAGA431B-4
Islituto di Farmucologia e Farniacognosiu tlell'C'nrversdd
tli
Uolognu, I l d i l
lieceizetl .Yotember 3, 1969
A relationship between lipophilic character, as expressed by the chromatographic IL, value, a i d biological activity of cephalosporins and penicillins was shown in Escherichia c o l i , Staphylococcus aureus, and Treponeniu palliduna. The Compounds most act,ive agaiust E . coli are more hydrophilic than those most active again*t S. aureus. This was interpreted on the basis of the different lipid composition of the cell walls of Gram-positive and Gram-negative microorganisms. T h e high lipid content of the cell wall of a Gram-negative microorganism such as E. coli could retain the most lipophilic molecules, which would not reach t,heir site of action. Therefore only the most hydrophilic compounds could be able to cross the cell wall and would be the most active ones. 111 the case of a Gram-positive microorganism such as 8. a w e u s the low lipid content of its cell wall would permit the high activity of more lipophilic compounds. T . palliduin, xhich is lacking a cell wall, behaves like S . aweus. I n fact there is an overlapping of the compounds most active against both microorganisms. The compounds most active against both E . coli and S . aureus have X, values which are between those of the compounds most active against Gram-negative and Gram-positive microorganisms, respectively. It is therefore suggested that compounds x i t h intermediate R, values could be characterized by a broad-spectrum antibacterial activity. The regression analysis of the relatiouvhip between R, valrie; and biological activity of penicillins in 8. aureus arid .'2 pallidurn showed evident deviations from the parabolic curve for methicillin, cloxacillin, aud dicloxacillin. It is suggested that, this could be due t o the presence of ortho substituellts on the aromatic ring of the penicilliii side chain.
Hansch, et al.,l and Lien, et ~ l . by , ~ means of two substituent constants T and u found very good correlations betn een the chemical structure and the biological activity of several sets of antibacterial agents. The substituent constant n was defined as log (PxIPH) u here PH arid P x are the octanol-water partition coefficients for a parent compound H and its derivative x. The lipophilic character, as expressed by the partition coefficient log P or by Bn, both calculated from T , ~ - ~ n as shown to be the most important factor in determining the biological activity. I n particular, the compounds effective against Gram-negative microorganismi I\ ere more hydrophilic than those effective against Gram-positive ones. The electron density on aromatic ring. as measured by u or 2~ !vas shown to play a -igriific:int role in several series of antibacterial agents,2 electron vithdrawal promoting activity.2 I n the field of aritibiotic- the activity against Staphylococcus aureus i n derivative< of phenoxymethj lpenicillin and methicillin incre lied with the hydrophilic character of the hubstituent-.G While in the in vivo and in vitro experiments with derivatives of phenoxymethylpenicillin the electronic effects of substituents were of very little or no importance,Gin the case of methicillin derivatives, they seemed to play a role.6 I n both cases a positive value of u seemed to promote activity. While Hansch, et al.,' pointed out that the calculated log P Z T cannot completely replace the experimental determination of the partition coefficient because of (1) (a) C . iIanscli, R . A I . M u i r . T. Fujita, P. I?. hlaloney, F. Geiger, and A I . Streich, J . Amer. Chem. Soc., 85, 2817 (1963); (b) C. Hansch a n d T. Fujita. i b i d . , 86, 1616 (1964). (2) E. J. Lien, C. Hansch, and S. RI. Anderson, J. M e d . Chem., 11, 430 (1968). (3) T. Fujita, J. Iwasa, and C. Hansch, J . Amer. Chem. Soc., 86, 5175 (1964). (4) J. Iwasa. T . Fujita, and C . Hansch, J . M e d . Chem. 8, 150 (1965). ( 5 ) C. Hansch, .I R. .Stea-ard. and J. Iwasa, i b i d . , 8, 868 (1965). (6) (a) C. H a n d ) and A. R . Steward, ibid., 7, 6'31 (1964); (b) C. Hansch a n d E. \\:. Deutsch, ibid,. 8, 705 (1965). fi) C. Flanscli, A . R. Sten.ard, .I. In-aaa, and E. W.Deutsch, M o l . Piurnuicul.. 1, 2 0 3 (1!165).
possible group interactions, Bird and Marshall8 found some anomalies in the calculated Za values of penicillins. On the other hand, in order to avoid the practical difficulties of the direct determination of a partition COefficient, Boyce and 3Iilborrow-9 suggested the use of the chromatographic R,, value, shown to be related to the partition coefficient lo and calculated from the formula :
Rm = log
(i;
- 1)
I n previous papers it was shown that a reversed-phase tlc method was a suitable technique for the determixition of the R,, value of penicillins and cephalosporins." The purpose of the present norli was to show that there is a relationship betn-een partition data and antibacterial activity of cephalosporins and penicillins.
Materials and Methods The antibiotics used in the present study are indicated in Tables I and 11. Glaxo Laboratories Ltd. and Eli Lilly and Co. are gratefully acknowledged for the supply of the noncommercial cephalosporins. Some cephalosporins and the penicillins were available in the form of water-soluble compounds. The remaining cephalosporins, in acidic form, were dissolved in 0.1 N XaHC03. The compounds \\-ere assayed againt S. aureus Oxford strain arid Escherichia cola 0-25 by means of the cylinder-plate method. l 2 The microorganisms ere inoculated into a culture medium consisting of brain(8) E. A . Bird and A. C. A'Iarshall, Biochem. Pharmncol., 16, 2275 (1967). (9) C. B. C . Bosce a n d B. V. Milborrow, Nature, 208, 537 (1965). (10) E. C. Bate-Smith and R. G . Westall. Biochim. B i o p h y s . Acta. 4, 427 (1950). (11) (a) G. L. Hiagi and RI. F. Gamba. Boll. SOC.I t d . B i d . S p e r . , 44, 189 (1967); (b) 0 . L. Biagi, .1.A I . Barbaro. M. C . Guerra, and 11.F. Garnlla, J . Chromatoyr., 41, 371 (1969); (c) G . L. Biagi, .\, M . Barharo. M. C. Guerra, and hl. F. Gamba, ibid.. 44, 195 (1969). (12) I). C. Grove and W..1. Randall. ".\soay Rletliodv of .\ntil>iotics." Rledical I'hckcluiiedia Inc., Ne\\. Y o r k , S . I-,, 1955.
p
,
c
8
f 7.
7-s. tiLiSd
log l / c obsd
1.63
0.66
0.4"
4.23
1.39
0 72
0.74
3.72
1 .34
0 . 86
0.80
4.10
1 ,03
0.Y3
1.18
3.98
4.48
4.84
5 . 13
1.03
1.06
1.21
4.80
4.52
5.10
5.18
0.8Y
1.72
1.39
5 14
4.70
5 78
3.37
0.33
2.0Y
1,83
4,S.j
4.8Y
5.76
5.64
0.47
0 83
1.94
3 . 13
0.08
3.10
2.43
4.3.5
4..57
3.43
5.69
-0.29
2.94
2.93
3.87
3.85
5.14
5.26
-0.46
3.12
3.1.5
3.4.5
3.37
.j.16
.i 0U .
Coinpd
Uiclvsacilliii
Xaicilliii
Osacilliii
chr
N,O,CCH:,
Pheiiethicilliii
Beiixylpeiiicilliii
log 1 /c
lluI'CuI'----.
log 1,'C calcd
li,,,
value
CsHjCHs
llethicilliii
1% l / c calcd
lug 1 c obsd
lop 1 L calcd
4.89
3.77
4.34
4.51
4.83
4.30
OCH, O C H -
- I
Met hyleiiampicilliii
KH,
QCH-
- I 9=CH2
Cnrbeiiecilliii COOH
heart tgar from Difco. The antibiotic solutions (0.1ml) I\ ere allowed to diffuse from holes cut in the agar layer. At the end of a 16-hr incubation period a t 37" the diameter of the zone of inhibition was measured. Each antibiotic \vas tested at various concentrations. There \vas a range of linear relationship between inhibition diameters arid the log of drug concentrations. From the equations of the straight lines the log of the concentration giving an inhibition diameter of 20 mm was calculated. The significance of b for each equation n as shown by means of a t test. The biological activity \!as finally expressed ac log l / c nhere c i h the molar (m.11 X IO-*) concentration of each antibiotic. which gives ail inhibition diameter of 20 mm. Penicillins arid cephalosporins were a5sayed against Treporzema pallidunz Reiter strain by means of the turbidimetric method. Reiter strain was grown in Fluid Thioglycollate 1Iedium (Difco) with addition of 1% human serum. 1Iinimal inhibitory concentrations (niic) 11 ere deterniiried after u 3-day iricubatiori :tnd rcported :IS log 1 / e . where c i i the molar concentration (m.11 X lo+) of each antibiotic, which prevents the growth of ipirocheteh.
The lipophilic character of the molecules \ \ a i expressed as R , value. This was measured by means of a reversed-phase tlc. The polar mobile phase was represented by acetone arid water in various proportions. The nonpolar stationary phase consisted of silicone oil, which impregnated a silica gel G layer. The R, values were linearlj related to the concentration of acetone in the mobile phase. I n this n a y it was possible to obtain, by interpolation or extrapolation, an R, value for each compound in a standard system, i e . , silicone oil-H,O. Higher and/or positive R , values indicate compounds more lipophilic than those represented by a lon-er and/or negative R, value. The detail< of the method have been described previously. l 1 The structure-activity relationships I\ ere analyzed by means of multiple regression analysis. The multiple correlation coefficient 1' was obtained by computing the correlation between observed and calculated log l / c values. The fraction of the variance of log l / c attributable to the multiple regre4on i i indic:ited by
,
2 Id (13) G.
ions-State
\V Snedecor and W. G Cocliran ' Statiztical Metlitids ' The Universit\ Press, . h i e s , I o n a , l Y 6 7 ,
402
log 1 ti7 f n, tiiiinber of the tested compounds; r, multiple correlatiou coeiticieiit ; ,xt st:iiiditrd dcviatioii
+
Results Antibacterial Activity of Cephalosporins against 6. coli and S. aweus.-The R,n values arid the ob-ervetl log 1, c values of cephalosporins against E . coli aiid S. aweus are indicated in Table I. The equation-. \\ Iiich correlate the itructure activity relationship i i i E. coli and S. a u t w p , were calculated from the data of Table I by means of multiple regrebsion aria1 reported in Table 111. I n the Gram-negative E. coli : i d in the Gram-positive S.aurcus the best rationalizat ion of the relationship bet\\ eeri structure and activity ih ptaovided by eq 2 arid 1,respectivel>-. These hhon a 1):irabolic relatioridlip betn eeii lipopliilic charitcter and 1)iologic:il activity id esplairi 7 3 arid S.15; , re;.pec*i ivcdy, of tlie variance iii the bio1ogic:tl of the 23 :md (32% of eq 1 :iiid :3. An 1' t the introductiori of the K,,?term into ccl 2 aiitl 1. [rotti nhicli tlie calculated log 1 vdues of Table I I\ ere obtairietl, significantly iniprovcd the correlationprovided by cq 1 arid 3. T h e negative sign associatctl uith the IInI2term in both equations meanb that tlic. activity of cephalosporins against E. coli :tiid S. aweus increases and decreases :xh t l i t > R,,, values change :mtl pass through an optimum.
Antispirochetal Activity of Cephalosporins against l'repoiienza palliduwz Reiter Strain.-The observed Discussion log 1,'c value- of cephalosporin.: agairi.:t T . paZZir1ui)i :ire reported in Table I. .I parabolic relationship The data of t lie preseiit i\ork hlio\v the iilfluciict: of h t v eeri R,, values :tnd biological activity, is shown by lipophilic cliarncter o i i tlic &)logical activity o f eq 13,11hich explains of the variance in the biologicephalosporini :uid peiiicillins. There is :L parabolic cal activity. The calculated log 1 'c values of Table I dependence of log 1 r O I L tlie E,,, values of cephalo\\ere obtained from eq 0. hporiiis for E. coli. A'. m r e u s , arid T . palliduin. This is Antibacterial Activity of Penicillins against E . coli and iri agreement \\ itli tlic postulate of the parabolic S . aureus.-The R,,, values and the observed log 1 c relatioii3hip bet\\ (wi t l i c petletration rate of conivdues ( i f penicillinr against E. coZi and S. aweus are pouiids througli hiological membranes sild Idieir repcirted in Table 11. The structure-activity relationlipophilic c 1 i a r : ~ t e r l. 4~ The penetration rnte iricrww\ ship in E . roll is beht rationalized by eq 7 , from which the ca1oul:itetl Lig 1 c value- i ) f Table I1 \$ere obtaiiitd. I'art (Sl[';) of the wriatice i t t thth biological activitj i\ c,\pl;iiiictl. 'l'lie iicgativc higii abhocinted \\ ith tlic /I,,,
LIPOPHILIC CHAILACTKR
and decreases as the lipophilic character increases progressively and passes through an optimum. I n the case of chemotherapeutic agents the penetration rate is related to their antibacterial activity, as the penetration through the cell wall or the cell membrane is a prerequisite for their activity. The log l / c values of Table I show that the cephalosporins most active against E. coli are more hydrophilic (lon.er R, values) than those most active against S. aureus. This confirms the results of Hansch, et d . , I and Lien, et U Z . , ~ and supports the view that also in the field of antibiotics the compounds active against Gramnegative microorganisms are more hydrophilic than those active against Gram-positive ones. As suggested by Lien, et aL2 the reason of the different lipophilic character of the molecules active against Gram-positive or Gram-negative organisms may be in the diff ererit lipid composition of the cell wall. It is known’j that the cell wall of the Gram-negative microorganisms is richer in lipid than that of the Gram-positive ones. The lipophilic molecules could be retained by the cell wall of Gram-negative microorganisms more strongly than the hydrophilic ones. Therefore only the latter could be able to reach the cell and exert their toxic effects. The calculated log l / c values of Table I show that cephaloridine and compounds 111 and I V are the most active compounds against S. aureus. This can be considered to be in agreement with the results of several investigators, as reported by Van Heyningen. l6 I n particular it was pointed out that cephaloridine shows a greater activity than cephalothin against staphylococci. l7 The same data show that cephaloglycin and compound X with lower R, values are the most active cephalosporins against E . coli. The products of the calculated log l / c values for each compound on S . aureus and E . coli indicate that cephaloram and cephalothin are the most active cephalosporins against both microorganisms. This agrees M ith the data of other investigators pointing out the broadIt spectrum antibacterial activity of cephalothin. is interesting to note that the R, value of cephalothin is between those of cephaloridine and cephaloglycin. This could suggest that above a certain degree of lipophilicity or hydrophilicity the compounds are mainly active against Gram-positive microorganisms or, respectively, Gram-negative ones. A compound M ith an intermediate R, value could be the most active against both type of microorganisms. However, if the present data seem to support such a suggestion, it must be pointed out that they were obtained only from E . coli and S. aureus. Moreover a survey of the literature concerning the assay of the above cephalosporins against several species of Gram-positive and Gram-neg,‘L t’we organisms seems to indicate a broad-spectrum antibacterial activity also for cephaloridine and cephaloglycin. I 6 , l 8
(15) $1. R. J. Salton. “ T h e Bacterial Cell Wall,” Elsevier Publishing Co., Amsterdam, 1964. (16) E. Van Heyningen, Aduan. Drug Res., 4. 1 (1967). (17) M. Barber a n d P. &I. U’atermorth, Brit. .Wed. J . . 1, 344 (1964). (18) 1.I. White, in “Textbook of Organic hledioinal a n d Pharmaceutical Cherniatry,” C . 0. Wilson, 0. Gisrold, a n d R . F. Doerge, Ed.. J. U. Lippincutt Co.. Philadelphia a n d Toronto, 1966. p 333.
Journal of J l ~ d i c m u lChwtistry, 19Y0, Vol. 13, Yo. 3 515
The calculated log l / c valueb regarding the activity of the tested cephalosporins ttgainst 7’. pallidurn show that the most active compounds arc ccphaloram and cephaloridine. This result, R hicli practically confirms the data of Ferrari, el U Z . , ’ ~ means that there is an overlapping of the R,,, values of tlie compounds most active against S. aureus and T . pallzclurn. This supports the hypotlieiib that tlic different lipophilic character of tlie compounds active against Gram-positive or Gram-negative microoi-ganihms may depend on the different lipid composition of the cell wall. In fact, T . pallidurn, which lacks a cell n-all, seemq to be most sensitive to those lipophilic compounds which are not trapped by the low lipid content of the cell n-all of a Gram-positive microorganism such a i X. aureus. Finally 7-aminocephalosporanic acid is consistently lass active than expected by itb R,,, value. This could indicate that, as regards the antibiotic activitl of cephalosporins, the side chain exerts a steric effect. The data regarding the activity of penicillins against E. coli show a significant linear relationship between R,,, values and antibacterial activity. In particular the most active compounds against E. colt are the most hydrophilic ones. The absence of a parabolic de~ R,,,values for E . coli is likely to pendence of log l / on be due to the lack of more hydrophilic compounds, which should show a progressive decrease in activity. I n the case of S. aweus arid T . pallidum the interpretation of the data is much more difficult. Equations 9 and 12 qeem to indicate a parabolic relationship between R, values and biological activit) . However they provide very low correlation coefficients ( r = 0.468 in eq 9 and 0.458 in eq 12). An explanation could be in the fact that all the data of the present work were obtained from experiments carried out in the absence of serum in the culture medium ( T . palliduin was only a partial exception because it n-as groir 11 in the presence , ~examining the of 1YG bovine serum). Hansch, et ~ l . by relationship between T values and antibacterial activity of penicillins on S. aur eus, found very poor correlations when human serum had riot been added to the culture medium. A striking improvement in tlie correlations TV:~S obtained when the serum binding was taken into consideration.6 I n any case the present data n ould suggest the influence of serum only in the case of penicillins on S. aureus arid T . pallicluin. Equations 10 and 13, calculated without methicillin, cloxacillin, and dicloxacillin, show a better correlation. The fact that methicillin, cloxacillin, and dicloxacillin have substituent groups on the ortho positions of the aromatic ring might suggest a role of these groups. They could exert an electronic effect and this nould be in agreement with the findings of Lien, et al.,* that electron nithdrawal promoted activity in their series of antibacterial agentq. I n the field of penicillins Hanscli, et ~ l . also , ~ noted some evidence indicating the role of positive cr values. However in the present case, if it \\as povible to express the electronic effects of the ortho substituents by means of the Hammett constant ( u values), the lack of available data in the literature did not permit evaluations of the electronic effects of substituents in other sites of the side chain, as expressed by the Taft constant (*. values).
'l'he oi tho substituerith could nlso exei t :i -teric effect. the evaluation of u hich is verj difficult a t the pre-ent time. l h a l l y both electronic :ind steric cff e c t i ma\ play a role. I t ii3 ire11 knoitii that nietliicilliri, dthougli not xuvxptible to penicillinase, i i much I benz) lperiicilliii and ampicillin :tg arid Gram-riegati\ c inicroorgitni.m comparative studv of Barber and W'it c ~ n o r t l i on ? ~ tlit. activity of b penicillin\ againat .i G I ani-pohitir e inicroorg:iniini- :tiid 1.; Gram-iiegativc oiie. cle:irli h o \ \ ec-l, :is reported ba C;:irrod,L2 t1i:it metliirlllin I - 1)) far t h e leit5t actiLc compound. Thr dcviatiori- of inetliicillin. cloxacilliri, arid dicloxacillin tram t tie p i diolic rurvc could nieaIi that the enzj niic h~ -tern 01 nieinbrarie em with uhich t h e penicillini interact iri tlie S . au,c u s arid 7'. pallzduiu :we different OI tliiit niet:tboli~m i i irivolvrd in home J + ~ Jnhicli c:tu-es tlic ciifieiwice. The calculated log 1 e vcilu(l- foi i$ alii'( U Y m d 7'. palldurn h o \ + that a1.0 i i i t lie ca'e of periicillm- tlie E,,, VBII~C. of tlir :tcti\ e compound. O I I t lie-(' org:tnl\n~. (9Oj G . .I7, Stewart.. "l'rie Penicillin Griiuii of D r u g s . " I < l w v i r r I'iil,lishinr ('ompany, Amsterdam, London a n d New York. ll(65. 1,p 13-30, (21) hl. Barber and P. 51. Waternortti. Brit. .llrd. J . . 1, 1 t5!4 !1!16'2j, (22) I,. P. Garrod in "Experimental C'liemotherapv." R . .I. SchnitLer a n d F. Hawking, FX.. Vol. 111. .irademir Press, S e i v Y o r k and London. 1964, p p 28. 29.
Agents Acting on the Central Nervous SysLem. X111.
2,3,4,4a,5,6-Hexahydro-l(H)-pyrazino [1,2-a]quiiiolines. A New Class of Hypotensive Agents'
The synthesis and pharmacological evaluation of 2t iiimiber of 3-substit iited %,3,j,4~,.j,ti-hesah?-dl.o-l ( H )pyrazino [ 1,2-a]qainoliries are reported. These compounds in general show hypot eiisive and adi,ciiei,gi(.-iecepl.tiI. blocking activity. The hypotensive activity is particularly marked in :3-&phenetIiyl- :tiid :',-~-ip-flti~,t.ot,ellzc,!.i 1piopy1-2,3,4,4a,5,6-hexahydro-liH)pgrazino [ 1,2-a]quiriolinch.
N-Phenylpiperaziries possess C'SS arid cardiovascular activities, and substit,utiori of the second imino group greatly modulates and modifies these act)ivities.2 N-Phenylpiperazines have also served as a side chain in a number of pharmacologically import'ant molecules. I n general 0-alkyl substitution in the phenyl residue of these N-phenylpiperazines greatly enhances the effect 011 the cardiovascular a c t i v i t i e ~and, , ~ in fact, ii number of 1-substituted 4-o-tolylpiperaxincs are known to be strong :idrenolytics.j It therefore ,seemed of interest ( 1 ) ('omrnunication N o . 14;3!1 from t h e Central Drug Hesearcli Institute. L u c k n o w India. ( 2 ) IX. G. Aiorren. V. Bieneft. and .I.11. Reyntjens. "I'sicho P h a r i n a rological Agents." Vol. I , & Gordon, I. Ed.. .\cademic Press, I n c . , S e w P o r k and London, 1Y64, p 251. (3) (a) R. P. hfuli. C . Tannenbaum. M . K. Dapero, AI. Bernier. W .Yost. a n d C;. d e Stevens. J. .hied. Chcrn.. 8 , 332 (1965): ( b ) P.C, Jain. 1.. K a p o o r , N . Anand. .\. Ahmad, and G . K. Patnaik, ibid.. 10,812 (1967). ( 4 1 ( a ) V. P.Arya. G . S. Grewal, J . David, and c'. L. Kaul. I : r p r r i i 93, 514 (1967); ( h ) G. d e Stevens a n d R. P. 311111,I.'rpni,ti P a t e n t . 31 [Nor-. 15. 1965); Cttmn. A h t r . . 64, 9746 (196tj).
