Enhancing Effect on Alkalinization of the Medium

Appum MICROBIOLOGY, Sept. 968, p. 88-9 Copyright @ 968 American Society for Microbiology Vol. 6, No. 9 Printed in U.S.A. Enhancing Effect on Alkalinization of the Medium on the Activity of Erythromycin Against Gram-negative Bacteria LEON D. SABATH,' VICTOR LORIAN, DEBORAH GERSTEIN, P. BRONWEN LODER, AND MAXWELL FINLAND Thorndike Memorial Laboratory, Second and Fourth (Harvard) Medical Services, Boston City Hospital; Department ofmedicine, Harvard Medical School; and Mattapan Chronic Diseases Hospital, Boston, Massachusetts 8 Received for publication 5 June 968 The antibacterial activity of erythromycin was markedly enhanced by alkalinization of the culture medium or urine within the clinical range (ph 6. to 8.). This effect was demonstrated against recent isolates of Escherichia coli, Klebsiella pneumoniae, Enterobacter sp., and Pseudomonas aeruginosa, as well as against Staphylococcus aureus and Streptococcus faecalis. The urine of normal volunteers was made alkaline by ingestion of sodium bicarbonate or acetazolamide (Diamox) during administration of. g of erythromycin every 8 hr; such urine was capable of inhibiting E. coli and K. pneumoniae even when diluted up to (in one instance) 8 times with broth of the same ph as the urine. Undiluted urine of the same subjects, without alkalinization, was seldom capable of inhibiting these organisms. The range of ph (6.6 to. 8.6) over which the antibacterial effect was enhanced coincided with that over which there was decreasing ionization of a basic group. Studies on erythromycin in 95 (6) and later (, ) indicated that its activity is much greater in an alkaline than in an acid medium. The test organisms upon which this observation was originally based were Streptococcus hemolyticus and Sarcina lutea (and later also Staphylococcus aureus). Zagar (5) and later J. C. Sylvester (personal communication) suggested that the susceptibility of gram-negative bacilli to erythromycin might be similarly affected by changes in ph. Subsequently, Sylvester (unpublished data) reported strains of three different gram-negative bacilli, Escherichia coli, Proteus vulgaris, and Pseudomonas aeruginosa, for which the minimal inhibiting concentration (MIC) of erythromycin was.5,.5, and.5,ug/ml, respectively, at ph 8.; at ph 6., the MIC for each was,ug/ml. The present study was designed to explore the clinical potential of this phenomenon. Urine from seven normal volunteers receiving erythromycin with and without alkalinizing medication was studied for activity against some pathogenic gramnegative bacilli in vitro. The pharmacological effects of sodium bicarbonate and acetazolamide ' Recipient of a Career Development Award, National Institute of Allergy and Infectious Diseases. Present address: Bronx-Lebanon Hospital, Fulton Avenue, Bronx, N.Y. 56. on the concentration and antibacterial activity of erythromycin in serum and urine will be reported separately (L. D. Sabath et al., submitted for publication). MATERIALS AND METHODS Organisms. Eighteen clinical isolates of gram-negative bacilli from patients with urinary-tract infections, three clinical isolates of gram-positive cocci, and a laboratory strain of Staphylococcus aureus were studied. The clinical strains were isolated and identified at the Mattapan Chronic Diseases Hospital. Sensitivity testing. Twofold dilutions of erythromycin base (supplied by Abbott Laboratories) were made in -ml amounts of Difco Nutrient Broth (NB) that had been adjusted to various levels of ph (6. to 8.) with dibasic sodium phosphate, sodium hydroxide, and (or) citric acid. One drop of a : dilution of a -hr culture (about, viable units) of the organism to be tested was then added to each tube and incubated at C. The lowest concentration of erythromycin that prevented visible growth after hr was taken as the MIC for that ph. Subcultures were made by withdrawing. ml of culture from each of the tubes showing no growth at hr and flooding it over the surface of antibioticfree nutrient agar, which was then incubated for 8 hr. The surviving cell count was calculated on the basis of the number of colonies appearing; the lowest concentration of erythromycin that produced a 88 Downloaded from http://aem.asm.org/ on November, 8 by guest

VOL. 6, 968 ENHANCING ERYTHROMYCIN ACTIVITY BY ALKALINIZATION 89 99.9% or greater reduction in bacterial cell count was considered the minimum bactericidal concentration (MBC). Activity oferythromycin in urine. Seven adult volunteers took erythromycin estolate (kindly supplied by R. S. Griffith of Lilly Laboratories for Clinical Research),. g orally every 8 hr for four doses on two separate occasions, week apart, once with and once without alkalinizing treatment (NaHCO or acetazolamide). Each specimen of urine was tested for activity against E. coli 5 and Klebsiella pneumoniae. The ph of the urine was measured (Radiometer ph meter) at the end of each collection period. In the tests of antibacterial activity, the urine was diluted in NB that had been adjusted to the same ph as that of the voided urine. One drop of a : dilution of a -hr culture was inoculated into each tube of diluted urine, and the maximum dilution of urine that prevented growth after hr of incubation at C was noted. Ionization of erythromycin. The titration curve of erythromycin base (Upjohn) was determined by using a Radiometer Titrator and Titrigraph (type SBRc; Copenhagen), correcting for a water blank. RESULTS All strains tested except those of Proteus showed moderate to marked increases in susceptibility to erythromycin at ph 8. when compared with their susceptibility at more acidic ph levels (Table ). Erythromycin also decreased the number of viable cells, the MBC generally being to 6 times the corresponding MIC. In some instances, there was a -fold or greater increase in activity over the range of ph tested. Urine from subjects receiving erythromycin without alkalinizing medication failed to inhibit the test strain of K. pneumoniae in every instance and rarely inhibited E. coli 5. Identical doses of erythromycin given with NaHCO or acetazolamide produced urine that inhibited the two test organisms in all samples (Table ). The increase in activity produced with alkalinization was highly significant (P <.). The extent to which the urines could be diluted and still inhibit the test organisms was clearly related to the degree of alkalinization achieved. Urine from subject, at ph 8., inhibited the test E. coli even at a dilution of :8, whereas the second specimens from subjects 6 and, with a ph of, inhibited the strain only at dilutions of : and :, respectively. The titration curve of the erythromycin base (Fig. ) shows one basic group titrating between ph 6 and with apparentpk of 8.6, as previously reported (). DIscussIoN The fact that some antibiotics show changes in antibacterial activity at different hydrogen ion Downloaded from http://aem.asm.org/ TABLE. Changes in minimal inhibiting concentration (MIC) and minimal bactericidal concentration (MBC) of erythromycin with increasing ph Organism MIC (MBC)a ph 6. ph. ph. ph.8 ph 8. Escherichia coli.. > 5 ().5 (>) 6. (5). (.5) E. coli 5.5 5 (5).5 (.5). (.).6 (.6) E. coli 8.> 5 (> ). (6.) E. coli 9.> 5 ().6 (.) E. coli. > >.5 (5) E. coli. > 5 ().5 (5). (.). (.) E. colil9... > 5 (>). (> ) E. coli 5... > 5 (5).5 (5). (.5).8 (6.) Klebsiella pneumoniae > 5 ().5 (6.) K. pneumoniae. > (>) 5 (>) 6. (.5) 6. (.5) Enterobacter.> >.5 (5) Enterobacter 5.> 5 (>) 6. (5) Enterobacter. 5 (>).5 (5) 6. (.5). (.5) Proteus sp.. > 5 (>) P. mirabilis 9. > > 5 (>) P. vulgaris. > > (>) > Pseudomonas aeruginosa. > > 5 (5) P. aeruginosa. > (>) 5 (5) Staphylococcus aureus A.... (6.). (.8). (.) S. aureus9... (6.). (.6). (.) Streptococcus faecalis... 5 (5). (6.).6 (.). (.8). (.) S. faecalis....5 (.5).6 (.6). (.6) In nutrient broth; expressed as micrograms per milliliter. on November, 8 by guest

9 SABATH ET AL. APPiL. MICROBIOL. TABLE. Effect ofalkalinizing medication on antibacterial activity ofurine from normal subjects receiving erythromycin estolate Erythromycin alone Erythromycin and alkalinization Period of urine collection (hr after Subject MID (MBD)- MID (MBD) first dose) ph of urine ph of urine E. coli 5 K. pneumoniae E. coli 5 K. pneumoniae -6 6- - All periods 56 56 5 6 J9b 66 9 5.. 5. 6. 5. 5. 5.5 5. 6. 6. 6. 6. 5.5 5. 6. 5. 5.-6. 6.-. < (<) () < (). () 8. 8. 8. 8. 8..8. 8..8.6.9.9..5 8. 8.8 8. 8....8.-..6-.9 8.-8.8.5 (.9) (6) 6 (8) 6 (6) 6 (6) () 8. (6.) 8 (6) (6) (6) 6 () 9. (.9).5 (.). (.). (8.) () () (). (.6) 8 () () 8 () (). (.6) (8) 8 () () (.8).8 (.). (.8).6 (.9) a Reciprocals of maximal inhibiting dilution (maximal bactericidal dilution) : = growth in undiluted serum; = no growth in undiluted sera only; etc. b Number of urine collections at all periods; the ph of these specimens and the geometric means of the MID (MBD) are shown in the other columns. concentrations has been known for over years (). The general principle appears to be that some acidic antibiotics are more active at acidic ph levels, whereas the reverse is true with some basic antibiotics. Although erythromycin, a weekly basic substance, showed profound changes in antibacterial activity with both gram-positive and gram-negative bacteria, over the range from ph 6 to 8, not all acidic and basic antibiotics show such effects. Benzylpenicillin, an acidic substance, shows only minimal changes in antibacterial activity in this range (). A currently popular explanation for the effect of changes in ph on antibacterial activity is that change in ionization of the antibiotic is the crucial factor, and the supposition is made that, as more of the antibiotic becomes ionized (as with erythromycin at progressively more acidic ph levels), it becomes increasingly difficult for it to permeate the bacterial cytoplasmic membrane or cell wall; this results in an apparent decrease in antibacterial activity. This could explain the striking ph effect shown by streptomycin (, ), since the marked (> -fold) increase in activity betweenph 5 and 8 coincides with the decrease in ionization of one of its guanido groups (8). However, the concurrence of such changes in ionization and in biological effects has not been demonstrated in all instances. Furthermore, Tallgren and von Bonsdorf () reported strains of urinary- Downloaded from http://aem.asm.org/ on November, 8 by guest

VOL. 6,968 ENHANCING ERYTHROMYCIN ACTIVITY BY ALKALINIZATION 9 z w ceptibility testing was in medium that was much more alkaline (ph.) than that of normal urine (ph 5-6), in which the erythromycin activity would have been considerably lower. In the present study, the subjects receiving erythromycin alone had acid urines (ph <6.) which failed to inhibit the test organisms (Table ); with alkalinization, all samples contained from to 8 times the MIC of erythromycin when tested at the ph of the urine. Although, with usual dosage regimens, concentrations of erythromycin achieved in serum and tissues will rarely inhibit the common urinary pathogens, the urine, if alkalinized, certainly will. Other studies (, 9) have shown that favorable results were usually obtained in treating bac-.- o.5- >.5- a.s - pk 8.6 TABLE. Comparative effects ofph on activity of total (T) and un-ionized (U) erythromycin against three organisms MIC (Ag/mI) ph % Ua E. coli 5 Enterobacter S. faccalis T Ub T U T U 6#. 5.5,. 5.5. 5.8 5.6.. 5. 5.65 5.6..6. 8.5..5..6..6 6..5 6..6.8..8 6..5 6.... 8..6...68.. 5 6 8 9 a Estimated from Fig.. p H b MIC of un-ionized erythromycin is product of FIG.. Ionization oferythromycin. %U X MIC of T. tract pathogens that failed to show an effect of ph on activity over ranges in which other strains of the same species had shown profound effects with the same antibiotics. Some strongly basic acridines are fully ionized throughout a ph range over which they show marked changes in antibacterial effect, indicating that a change in ionization of the acridine cannot account for the change in antibacterial effect (). Other suggested explanations are that a receptor on the bacterial cell may be changing its charge and, therefore, its affinity for the antibacterial substance, over the critical ph range (), or that charged acidic antibiotics compete with hydroxyl ions for a receptor and have fewer hydroxyl ions for competition at the more acidic ph. Similarly, ionized basic antibacterial substances may compete with hydrogen ions (). However, convincing experimental documentation of a single mechanism for the ph effect is lacking. Two pieces of evidence from the present study are consistent with the hypothesis and it is the neutral or un-ionized molecules that exert the antibacterial effect: (i) a decrease in ionization (Fig. ) occurred over the same ph range in which there was a striking increase in antibacterial activity; and (ii) the amount of un-ionized erythromycin required to inhibit three of the organisms (Table ) varied only three- to five-fold between ph 6 and 8, whereas the amount of total (ionized plus un-ionized) erythromycin required to inhibit the same organisms at the same range of ph varied 5- to -fold, suggesting that the increasing amount of ionized erythromycin at the more acid ph is without antibacterial effect. However, these observations do not exclude the other two possible mechanisms cited. The in vitro effect of changes in ph activity of erythromycin indicated that antibiotic concentrations effective against many gram-negative bacilli should be achievable in human urine providing that the usually reported (5) fraction ( to 5%) of ingested erythromycin appears in the urine of subjects who are also taking alkalinizing medication. The present study shows that up to 8 times the concentration of erythromycin required to inhibit the test E. coli appeared in the urine of the subjects receiving both erythromycin and NaHCO or acetazolamide. In fact, more than % of ingested erythromycin may appear in the urine after repeated doses; this was confirmed in concurrent studies in the present subjects (Sabath et al., submittedfor publication). Although a significant portion of urinary pathogens are reported as being susceptible to erythromycin (, ), it should be noted that the sus- Downloaded from http://aem.asm.org/ on November, 8 by guest

9 SABATH ET AL. APPL. MICROBIOL. teriuria when adequate antibacterial activity has been achieved in the urine, though not in the serum. It therefore seems desirable to investigate the possibility of using erythromycin along with alkalinization for the treatment of bacteriuria, particularly when alternative therapy requires the use of more toxic antibiotics. ACKNOWLEDGMENTS This investigation was supported by Public Health Service research grants AI- and TOI-AI-86 from the National Institute of Allergy and Infectious Diseases. LITERATURE CITED. Abraham, E. P., and E. S. Duthie. 96. Effect of ph of the medium on activity of streptomycin and penicillin and other chemotherapeutic substances. Lancet :55-59.. Albert, A. 965. Selective toxicity. Methuen and Co., Ltd., London, p. 8-.. Flynn, E. H., M. V. Segal, P. F. Wiley, and K. Gerzon. 95. Erythromycin. I. Properties and degradation studies. J. Am. Chem. Soc. 6:-.. Garrod, L. P., and P. M. Waterworth. 956. Behavior in vitro of some new antistaphylococcal antibiotics. Brit. Med. J. :6-65. 5. Griffith, R. S. 959. Laboratory and clinical studies with erythromycin propionate. Antibiotics Ann. 958-59, p. 6-. 6. Haight, T. H. and M. Finland. 95. Observations on mode of action of erythromycin. Proc. Soc. Exptl. Biol. Med. 8:88-9.. McCabe, W. R., and G. G. Jackson. 965. Treatment of pyelonephritis: bacterial, drug, and host factors in success and failure among 5 patients. New Engl. J. Med. :-. 8. Sabath, L. D. 968. Synergy of antibacterial substances by apparently known mechanisms. Antimicrobial Agents and Chemotherapy- 96, p. -. 9. Sabath, L. D., H. A. Elder, C. E. McCall, and M. Finland. 96. Synergistic combinations of penicillins in the treatment of bacteriuria. New Engl. J. Med. :-8.. Tallgren, L. G., and C. H. von Bonsdorf. 965. The effect of varying the ph level upon the sensitivity of urinary bacteria to antibiotics. Acta Med. Scand. 8:5-55.. Tunevall, G., and P. Hedenius. 95. Laboratory and clinical studies with erythromycin. Antibiot. Chemotherapy :68-68.. Waterworth, P. M. 96. The antibacterial properties of leucomycin. Antibiot. Chemotherapy :-8.. Williamson, G. M., and K. Zinnemann. 96. The susceptibility of coliform bacilli to erythromycin. Antibiot. Chemotherapy :69-.. Wolinsky, E., and W. Steenken, Jr. 96. Streptomycin and penicillin resistant staphylococci. I. Influence of ph body fluids on streptomycin action. Proc. Soc. Exptl. Biol. Med. 6:6-65. 5. Zagar, Z. 96. Sensitivity of E. coli, Ps. aeruginosa, and B. proteus to erythromycin in various ph culture media. Chemotherapia 6:8-89. Downloaded from http://aem.asm.org/ on November, 8 by guest