REVEWS OF NFECTOUS DSEASES VOL. 5, SUPPLEMENT 3 JULY-AUGUST 1983 1983 by The University f Chicag. All rights reserved. 0162-0886/83/0504-0027$02.00 Penetratin f Antibitics int Human Leukcytes and Dermal Suctin Blisters c. O. Slberg, A. Halstensen, A. Digranes, and K. B. Helium Frm Medical Department B and the Department f Micrbilgy and mmunlgy, Schl j Medicine, University j Bergen, Bergen, Nrway Staphylcccus aureus phagcytized by leukcytes frm healthy dnrs and frm patients with chrnic granulmatus disease were prtected frm the antibacterial effect f gentamicin. Cnsiderable numbers f phagcytized bacteria remained viable after expsure fr 20 hr t antibitic cncentratins that killed >990/0 f extracellular bacteria in <4 hr. A higher prprtin f intracellular bacteria were killed by rifampin; this finding indicated that rifampin penetrates better int the phagcytic vacule than des gentamicin and/r is mre active against phagcytized bacteria than is gentamicin.after ral adminstratin f 450 mg f rifampin t three healthy vlunteers, cncentratins f the antibitic in serum and skin blister fluid were measured. Cncentratins in serum peaked within 3 hr f ral administratin (mean peak level, 13.2 J,lg/ml). Cncentratins in blister fluid peaked between 6 hr and 9 hr (mean peak cncentratin, 2.7 J,lg/ml). Between 9 hr and 12 hr, the cncentratins f rifampin in serum and blister fluid were similar; later, levels in blister fluid were higher than thse in serum. The mean eliminatin half-life f rifampin was 2.5 hr in serum and 6.0 hr in blister fluid. Plymrphnuclear leukcytes prvide early, nnspecific defense against infectin. Mst micrrganisms pathgenic t' humans are rapidly ingested by the cells, smetimes within a few minutes [1]. Therefre, unless antibitics can enter the phagcytes and penetrate int the phagcytic vacule surrunding the micrrganism, the time available fr actin n the micrbe may be very shrt. Penetratin f antibitics int the phagcytic vacule seems imprtant since several pathgens can survive within the phagcytes fr a lng time and can even destry the cell and prpagate infectin. A primary cncern f the clinician is the antibitic cncentratin at the site f infectin. Samples f serum are easily btained, but antibitic levels in serum d nt necessarily reflect the activity f the drug in extravascular fluids and fci f inflammatin. Hwever, several mdels have been develped fr study f the penetratin f nfrmed cnsent was btained frm the patients and vlunteers, and the guidelines fr human experimentatin f the Schl f Medicine, University f Bergen, were fllwed in the cnduct f the clinical research. This study was supprted by grants frm the Nrwegian Research Cuncil fr Science and Humanities. Please address requests fr reprints t Dr. C. O. Slberg, Medical Department B, Schl f Medicine, University f Bergen, 5016 Haukeland Sykehus, Bergen, Nrway. antibitics int tissues. One f these mdels, the dermal suctin blister, seems t prvide valuable infrmatin n the kinetics f antibitics in small abscesses and the edema f sft-tissue infectin. n the present study, the activity f rifampin against phagcytized bacteria was cmpared with that f gentamicin in an in vitr test system with leukcytes and phagcytized Staphylcccus aureus, which can survive within the phagcytes. Furthermre, the penetratin f rifampin int dermal suctin blisters was studied. Materials and Methds Leukcytes. Heparinized bld (lo units f heparin/ml) frm healthy human dnrs and frm patients with chrnic granulmatus disease was allwed t settle fr 30-60 min at 37 C. After the leukcyte-rich plasma had been remved, the leukcytes were sedimented, washed twice in sterile phsphate-buffered saline (PBS) by centrifugatin at 100 g fr 10 min, and resuspended in PBS t a cncentratin f 1.0 x 10 7 plymrphnuclear leukcytes/ml. Bacteria. S. aureus "Oxfrd" (Heatley strain, btained frm the Natinal Cllectin f Type Cultures, Clindale, Lndn, England, 1958) was cultured vernight in Penassay brth (Difc, Detrit, Mich.), washed twice in 0.45070 NaC, and S468
Rifampin in Leukcytes and Skin Blisters S469 suspended in PBS t a cncentratin f 35-44 x 10 7 cfu/ml. The MC f gentamicin fr this rganism was 0.04 ug/ml; that f rifampin was 0.002 ug/rnl, Serum. Fresh, pled human serum frm six adults was stred at -70 C in l-ml aliquts. mmediately prir t each experiment, 1 ml f serum was thawed and added t 3 ml f PBS. Leukcyte-bacterial suspensins. Fr preparatin f leukcyte-bacterial suspensins, 0.5 ml f the leukcyte suspensin, 0.1 ml f the bacterial suspensin, and 0.4 ml f diluted serum were placed in dispsable plastic tubes and incubated at 37 C. This mixture cntained abut eight bacteria (i.e., cfu) per leukcyte and a final cncentratin f serum f 10070. After incubatin f the tubes fr an interval adequate fr phagcytsis (see Results), cell-assciated bacteria (i.e, bacteria either lcated intracellularly r attached t the external cell wall) were separated frm extracellular bacteria by centrifugatin at 500 g fr 5 min and were washed twice in PBS. The cellular pellet was resuspended in 1 ml f PBS cntaining 10070 serum. This suspensin was referred t as the leukcyte-bacterial suspensin. Skin blister fluid. Three healthy medical students (ne wman and tw men) weighing 62, 68, and 70 kg, respectively, participated in the study. nfrmed cnsent was btained frm each student. N student tk ther drugs during the study. The students fasted fr 8 hr befre and 1 hr after antibitic administratin. Skin blisters were prduced accrding t the methd f Helum et al. [2]. n brief, a perspex ( C, England) blck with 10 bres (diameter, 8 mm) was tightly strapped t the midvlar area f the frearm, and cntrlled suctin (- 0.3 kg/em") was applied t each bre. After 2 hr a half-spherical dermepidermal blister cntaining 0.10-0.15 ml f serus fluid had been created at each bre (figure 1). After the prductin f blisters, each individual received 450 mg f rifampin (Rimactan", Ciba Geigy, Basel, Switzerland) as a single ral dse with 100 ml f water. Samples f venus bld were btained befre and at serial intervals after antibitic administratin. At the time when each pstadministratin sample was btained, fluid frm ne skin blister was als btained by puncture with a cannula and was cllected in micrcapillary tubes. The serum and blister fluid were Figure 1. Dermal suctin blister n the midvlar area f the frearm. immediately frzen at - 20 C and were assayed fr rifampin within tw weeks. A standard dilutin f antibitic in serum was frzen simultaneusly fr tests f drug inactivatin during strage. Assay fr Rifampin. The cncentratins f rifampin in serum and blister fluid were determined by an agar well diffusin methd [3]. Fr determinatin f serum cncentratins, standard curves were prepared n the basis f undiluted, pled human sera; fr determinatin f blister fluid cncentratins, curves were prepared n the basis f pled human -sera diluted 1:1 in PBS. S. aureus strain 209 was used as the indicatr strain and PDM-Antibitic Sensitivity Medium (AB Bidisk, Slna, Sweden) as the test medium. The plates were incubated at 37 C fr 18 hr, and the znes f inhibitin were read. Each determinatin was made in duplicate, and the means were calculated. The lwer limit f sensitivity was 0.05 ug/ml. Assayfr antibitic killing fphagcytized bacteria. Gentamicin (5 ug/ml) and rifampin (0.5 ug/rnl) were added t tubes cntaining 1 ml f leukcyte-bacterial suspensin that had been preincubated fr 20 min and washed (as described previusly) and t tubes cntaining 1 ml f bacterial suspensin. The tubes were incubated at 37 C. Samples (0.01 ml) were remved peridically, washed twice in PBS, and resuspended in 1 mlf distilled water (fr facilitatin f the smtic disruptin f the leukcytes). Viable bacteria were
S470 Slberg et al. quantitated by a standard pur-plate technique with use f Penassay agar (Difc). Results n the bacterial suspensin, bacteria were rapidly killed. Very few micrbes remained viable after incubatin fr 4 hr, particularly in the suspensin with gentamicin (figure 2). n the leukcyte-bacterial suspensin cntaining gentamicin, the reductin in the number f viable bacteria was slw, and a cnsiderable number remained viable after incubatin fr 20 hr. The mre prnunced reductin during the early phase f incubatin may have been due partly t killing f cntaminating extracellular bacteria by the antibitics [4]. Only minr differences in numbers f viable bacteria were bserved when antibitic cncentratins three times higher r lwer were added t the leukcyte-bacterial suspensin; this finding indicated that the decrease in the number f viable bacteria was due t killing by bactericidal mechanisms f leukcytes rather than t inactivatin by gentamicin. The bactericidal effect was mre prnunced, hwever, in the leukcyte-bacterial suspensin --- Leuccyte-bacteria suspensin Bacteria suspensin cntaining rifampin. This result suggested that rifampin penetrates better int the phagcytic vacule then des gentamicin and/r is mre active against phagcytized bacteria than is gentamicin. n an experiment designed t test this hypthesis, leukcytes with a defective bactericidal capacity were btained frm patients with chrnic granulmatus disease, expsed t bacteria fr 10 min (an interval adequate fr phagcytsis), and incubated with 5 JJg f gentamicin/ml r 0.5 JJg f rifampin/ml. Only a minr reductin in the number f viable bacteria was nted in the test with gentamicin (figure 3). This result was prbably due in part t a lack f killing fintracellular bacteria by leukcyte enzymes but als t slw penetratin f gentamicin int the phagcytes and/r pr activity f the drug against phagcytized bacteria. n the test with rifampin, a marked reductin in the number f viable bacteria again indicated that this drug penetrates better int phagcytes than des gentamicin and/r is mre active against phagcytized bacteria than is gentamicin. Cncentratins f rifampin in serum peaked within 3 hr f ral administratin; the mean peak eṉ 'c :J 0).~ 10 6 ' ->. c -0 Gentamicin 5~g/ml Rifampin 0.5 jjg/ml ncubatin time (hurs) Rifampin 0.5 jjg/ml 20 Figure 2. Effect f gentamicin and rifampin n phagcytized and nnphagcytized Staphylcccus aureus (mean f three experiments). 2 4 ncubatin time (hurs) 12 Figure 3. Survival f phagcytized Staphylcccus aureus expsed t gentamicin and rifampin. The bacteria were phagcytized by plymrphnuclear leukcytes frm patients with chrnic granulmatus disease.
Rifampin in Leukcytes and Skin Blisters S471 level was 13.2 JLg/ml (range, 10.0-16.0 ug/ml) (figure 4). The mean antibitic eliminatin halflife in serum was 2.5 hr (range, 1.8-3.4 hr), as calculated by linear regressin analysis f the data btained 6-21 hr after drug intake. Cmpared with levels in serum, peak cncentratins f rifampin in blister fluid were delayed, and ccurred between 6 hr and 9 hr. The mean peak cncentratin was 2.7 ug/ml (range, 2.7-2.8 ug/rnl), r 20010 f the mean peak level in serum (range, 17% 28070). Between 9 hr and 12 hr, cncentratins in serum and blister fluid were similar; later, levels in blister fluid were higher than thse in serum. The mean eliminatin half-life fr rifampin in blister fluid was 6.0 hr, as calculated by regressin analysis f data btained 9 hr and 30 hr after drug intake. Discussin Several types f bacteria, including Mycbacterium species, Brucella species, Shigella species Salmnella species, and even S. aureus, may survive ingestin by leukcytes, mncytes, r macrphages despite the bactericidal activity f these cells. Whether antibitics can penetrate int phagcytes and kill intracellular bacteria has been the subject f several studies. Shaffer et al. [5] shwed that cells f Brucella inside rat peritneal phagcytes were nt killed by high cncentratins f streptmycin but were killed by chlrtetracycline, xytetracycline, and chlramphenicl at cncentratins f 50-100 JLg/ml. Mackaness [6] and Suter [7] fund that tubercle bacilli phagcytized by macrphages were sensitive t high cncentratins f streptmycin, and Jenkin and Benacerraf [8] believed that streptmycin cntributed t the killing f Salmnella enteritidis ingested by muse peritneal macrphages. Mre recently, Alexander and Gd [9], Hlmes et al. [10], Lb and Mandell [11], Mandell and Vest [12], and Slberg and Helum [13] shwed that S. aureus [9, 10, 12, 13], Pseudmnasaeruginsa [9], and Escherichia cli [11] surviving intracellularly within leukcytes [9, 10, 12, 13] r macrphages [11] were nt killed when incubated with high cncentratins f varius antibitics, including penicillins and aminglycsides. Hwever, rifampin had a significant effect n E. cli [11] and S. aureus [12, 13] phagcytized by macrphages and leukcytes, respectively. Als, in the 10.0 Ẹ... 01 =1.0 c ~ C euuc u 0.1 c 00.. E ~ ~ 0, 0,, ';&.., ",,~,!,,0, ', fj 0...',' "-, "'" '', ' " ',,,1, 01,, '~, 1 0, 0 ' " -, " ' -, " -,, 0, b Serum 1 3 6 9 12 15 18 21 24 27 30 hurs Figure 4. Cncentratins f rifampin in serum and in fluid frm dermal suctin blisters after ral administratin f a single dse f 450 mg t three vlunteers. The results fr each vlunteer are shwn separately. present study, rifampin was mre effective against intraleukcytic bacteria than was gentamicin. Little is knwn abut the reasn fr the pr activity f gentamicin against intracellular bacteria. Hwever, studies with cultured fibrblasts have shwn that the intracellular accumulatin f aminglycsides is a slw prcess and that drug cncentratins stabilize nly after fur days f incubatin [14]. Furthermre, aminglycsides are lcalized predminantly in the lyssmes, where the prevailing acid ph may inhibit the antibacterial activity f the drugs (the ptimal ph fr gentamicin being 'V8). The reasn fr the gd activity f rifampin against phagcytized bacteria in the present investigatin and in previus studies [11-13] is nt bvius. Hwever, agents that are highly sluble in lipids usually penetrate cell membranes easily. Since rifampin is a highly lipid-sluble substance, this prperty may explain its ability t penetrate bth the phagcyte and the phagsme membrane and t kill bacteria that have survived phagcytsis. t is well knwn that bacteria may survive in tissue abscesses and ther purulent lesins despite lng-term treatment with large dses f antibi-
S472 Slberg et al. tics. This phenmenn may be due partly t the prtectin f phagcytized bacteria against antibitics. The in vitr finding that rifampin can penetrate int phagcytic cells and kill intracellular bacteria seems t be f clinical significance in the treatment nt nly f tuberculsis but als f ther infectins (i.e., cases f staphylcccal arthritis [15] in which seemingly adequate therapy fails). n ur experience, rifampin in cmbinatin with ther antibitics seems t be f particular value fr the treatment f infectins caused by rifampin-sensitive rganisms in patients with defective leukcyte bactericidal activity (e.g., thse with chrnic granulmatus disease r leukemia). Levels f antibitics in serum may serve as a gd guide t the cncentratins t be expected in extravascular fluids and fci f inflammatin [16]. Hwever, the multiplicity f pharmaclgic factrs that perate in varius clinical situatins have led t an array f mdels in humans and animals fr the study f the penetratin f tissues by antimicrbial agents with different physicchemical prperties [16-18]. N single mdel is preferable t the thers [16]. Cmparisn f the mdels in humans indicates that the dermal suctin blister methd represents a "small-reservir" mdel f tissue penetratin and apprximates the kinetics in small abscesses and the edema f sfttissue infectin [16-19]. n pharmackinetic terms, skin blister fluid is a deeper cmpartment than tissue fluid btained frm sc-implanted threads; the latter mre clsely resembles interstitial fluid [18,,20]. The prtein cntent and the diffusinal barrier are the majr determinants f antibitic penetratin int blister fluid. Accrding t ther studies [18-20], the prtein cntent in suctin blisters is 25-45 Ag/ml, with albumin predminating. The binding f serum prteins has a restraining effect n the passage f drug int the tissues; recently, an inverse relatin was demnstrated between the degree f prtein binding and the cncentratin f ttal and nnprtein-bund antibitics in tissue fluids [16-18]. Accrdingly, an 80% rate f prtein binding t rifampin [21] culd explain, at least in part, the lw level f ttal drug in skin blister fluid as well as the mdest area under the cncentratin-vs-time curve. The timing and height f the serum peaks and the serum elimina- tin half-life are als in accrd with previus results btained after single ral dses fthis drug [21, 22]. Finally, the cncentratin ratis fr serum and blister fluid cmpare favrably with thse fr serum and skin, muscle, spleen, cavernae and pleural fluid [21, 22]. The assessment f antibitic levels in tissues is imprtant fr ptimal dsing schedules. n the present study, the eliminatin f rifampin frm blister fluid was slw, and, even 24 hr after drug administratin, the antibitic cncentratin in blister fluid was as high as 0.4 Ag/m!. Accrdingly, tissue-fluid levels f rifampin abve the MC [22] fr mst relevant pathgens were maintained fr 24 hr after a single ral dse f 450 mg. This result supprts the cmmn practice f administering nce- r twice-daily dsage regimens. References 1. Slberg CO. Enhanced susceptibility t infectin. A new methd fr the evaluatin f neutrphil granulcyte functins. Acta Pathl Micrbil Scand [B] 1972;80: 10-8 2. Hellum KB, Schreiner A, Digranes A, Bergmann 1. Skin blisters prduced by suctin: a new mdel fr studies f penetratin f antibitics in humans. n: Siegenthaler W, Luthy R, eds. Current Chemtherapy. Vl. Washingtn, DC: American Sciety fr Micrbilgy, 1978: 620-2 3. Bennett JV, Brdie LJ, Benner EJ, Kirby WMM. Simplified, accurate methd fr antibitic assay f clinical specimens. Appl Micrbil 1966;14:170-7 4. Slberg CO. Prtectin f phagcytized bacteria against antibitics. A new methd fr the evaluatin f neutrphil granulcyte functins. Acta Med Scand 1972; 191:383-7 5. Shaffer JM, Kucera CJ, Spink 'WW. The prtectin f intracellular Brucella against therapeutic agents and the bactericidal actin f serum. J Exp Med 1953;97:77-90 6. Mackaness GB. The actin f drugs n intracellular tubercle bacilli. J Pathl Bacteril 1952;64:429-46 7. Suter E. Multiplicatin f tubercle bacilli within phagcytes cultivated in vitr, and effect f streptmycin and isnictinic acid hydrazide. American Review f Tuberculsis 1952;65:775-6 8. Jenkin C, Benacerraf B. n vitr studies n the interactin between muse peritneal macrphages and strains f Salmnella and Escherichia cli. J Exp Med 1960;112: 403-17 9. Alexander JW, Gd RA. Effect f antibitics n the bactericidal activity f human leukcytes. J Lab Clin Med 1968;71:971-83 10. Hlmes B, Quie PG, Windhrst DB, Pllara B, Gd RA. Prtectin f phagcytized bacteria frm the killing actin f antibitics. Nature 1966;210:1131-2
Rifampin in Leukcytes and Skin Blisters S473 11. Lb MC, Mandell GL. The effect f antibitics n Escherichia cli ingested by macrphages. Prc Sc Exp Bi Med 1973;142:1048-52 12. Mandell GL, Vest TK. Killing f intraleukcytic Staphylcccus aureus by rifampin: in-vitr and in-viv studies. J nfect Dis 1972;125:486-90 13. Slberg CO, Hellum KB. Prtectin f phagcytsed bacteria against antimicrbial agents. Scand J nfect Dis [Suppl] 1978;14:246-50 14. Tulkens P, Truet A. The uptake and intracellular accumulatin f aminglycside antibitics in lyssmes f cultured rat fibrblasts. BichemPharmacl 1978; 27:415-24 15. Beam TR Jr. Sequestratin f staphylccci at an inaccessible fcus. Lancet 1979;2:227-8 16. Barza M. Principles f tissue penetratin f antibitics. J Antimicrb Chemther 1981;8(Suppl C):7-28 17. Bergan T. Pharmackinetics f tissue penetratin f antibitics. Rev nfect Dis 1981;3:45-66 18. Hffstedt B. Penetratin f antibitics int subcutaneus tissue fluid with special reference t a new mdified thread sampling technique. Scand J nfect Dis [Suppl]: 1981;30:1-33 19. Schreiner A, Bergan T, Hellum KB, Digranes A. Pharmackinetics f ampicillin in serum and in dermal suctin blisters after ral administratin f bacampicillin. Rev nfect Dis 1981;3:125-31 20. Helum KB, Lehmann V. Penetratin f benzyl penicillin t human tissue fluid in subcutaneus wicks and skin blisters [abstract]. Acta Med Scand [Suppl]: 1978;621:35 21. Accella G. Clinical pharmackinetics f rifampicin. Clin Pharmackinet 1978;3:108-27 22. Walter AM, Heilmeyer L. Antibitika-Fibel. Stuttgart: Gerg Thieme Verlag, 1975:551-74