Experimental Candida albicans endocarditis: characterization of the ...

5 downloads 349 Views 2MB Size Report
Jan 11, 1977 - Charlottesville, Virginia 22901,* and Microbiology Department, Georgetown University Schools ofMedicine and Dentistry, Washington, D.C. ..... Tested in triplicate by the checkerboard tech- nique, the MIC of amphotericin B ...
Vol. 17, No. 1

INFECTION AND IMMUNITY, July 1977, p. 140-147 Copyright C© 1977 American Society for Microbiology

Printed in U.S.A.

Experimental Candida albicans Endocarditis: Characterization of the Disease and Response to Therapy MERLE A. SANDE,* C. RICHARD BOWMAN,

AND

RICHARD A. CALDERONE

Department of Internal Medicine, Division ofInfectious Diseases, University of Virginia School of Medicine, Charlottesville, Virginia 22901,* and Microbiology Department, Georgetown University Schools of Medicine and Dentistry, Washington, D.C. 20057 Received for publication 11 January 1977

Endocarditis caused by Candida albicans was induced in rabbits after insertion of a catheter across the aortic valve. The mean survival time of 34 rabbits was 26 days. Only 7% of temperature recordings taken were elevated. Candida was recovered from only 9% of blood cultures taken. Precipitating and agglutinating serum antibody was detected after 12 days of infection. Antibody titers rose progressively until death in rabbits with endocarditis, whereas titers peaked early and subsequently decreased in animals that received an intravenous injection of C. albicans without precatheterization. Three groups of rabbits were treated for 6 days with amphotericin B, 5-fluorocytosine, or the two drugs in combination. Amphotericin B alone reduced the mean titer of organisms from logl0 8.79 1.46 to logl0 3.1 + 1.9 colony-forming units/g. 5-Fluorocytosine was less effective (mean titer after 6 days of therapy was logl0 7.4 0.33 colonyforming units/g). The addition of 5-fluorocytosine to amphotericin B did not increase the rate at which Candida cells were eradicated from the vegetations. These in vivo results correlated with the failure to demonstrate an increased rate of fungicidal activity in vitro with the two drugs. ±

±

Organisms of the genus Candida have be- sponse to that of rabbits infected with Candida come a serious cause of endocarditis associated but without endocarditis; and (iii) to compare with prosthetic valves. When infection is the efficacy of amphotericin B, 5-FC, and the caused by Candida species, clinical and labora- combination in treating the experimental distory signs of endocarditis are frequently absent ease. and early diagnosis is difficult (17). Recent MATERIALS AND METHODS studies (3, 5, 10-12, 14, 17, 20-22) have shown that the detection of agglutinating and precipiProduction of Candida endocarditis. Nonbactetating antibody to extracts of Candida species rial thrombotic endocarditis was established in 34 indicates Candida infection. Tests using a cyto- New Zealand white rabbits weighing 2 kg by a modiplasmic antigen preparation seem especially fication of the technique described by Freedman and predictive of systemic candidiasis, including Johnson (4). After intravenous (i.v.) administration of 60 mg of sodium pentobarbital (Abbott LaboratoCandida endocarditis (10, 11, 17, 20). North Chicago, Ill.), the right carotid artery Even with early diagnosis, therapy with am- ries, exposed and ligated, and a sterile polyethylene photericin B alone is rarely sucessful, and valve was (Radio-Opaque, Intramedic, PE-90, Clayreplacement is necessary for cure. Amphoteri- cannula Adams, Parsipanny, N.J.) was inserted across the cin B and 5-fluorocytosine (5-FC) in combina- aortic valve. The catheter was then clamped and tion have been reported to be synergistic in sutured in place, and the skin was closed over the vitro with some strains and have been success- catheter. Candida infection was established by i.v. fully used in several patients without valve irnection of 106 washed C. albicans cells cultured replacement (E. Rubinstein et al., Program from a patient with Candida prosthetic valvular Abstr. Intersci. Conf. Antimicrob. Agents endocarditis. The rabbits were allowed to run the of their disease and either died or were sacriChemother. 13th, Washington, D.C., Abstr. course ficed (by use of intravenous pentobarbital, 100 ml) no. 217, 1973). moribund. The hearts were removed under The purpose of this investigation was three- when aseptic conditions. Cardiac vegetations present on fold: (i) to develop and characterize a rabbit the aortic valve or root were weighed and homogemodel ofCandida endocarditis; (ii) to study the nized in saline. Serial 10-fold dilutions were made in antibody response of rabbits with endocarditis Trypticase soy agar (BBL), and titers were excaused by C. albicans and to compare this re- pressed as the number of colony-forming units 140

VOL. 17, 1977

EXPERIMENTAL C. ALBICANS ENDOCARDITIS

(CFU) per gram of vegetation. A sample was also fixed, sectioned, and stained for microscopy. A small section of the kidney was similarly prepared. Spleens were removed and weights were recorded. Clinical and microbiological procedures. Blood (0.5 to 1 ml) was drawn through either a peripheral ear vein or artery for culture three times per week and was incubated in Trypticase soy agar. Urine was collected in metabolic cages before and after infection and was examined microscopically and also was cultured on Trypticase soy agar. A single rectal temperature was recorded daily from each rabbit. Each animal was closely observed for development of neurological abnormalities. Funduscopic examinations were performed biweekly with a Welch-Allyn ophthalmoscope. Serological studies. (i) Collection of serum. Sera from three i.v.-infected rabbits and nine catheterized and i.v.-infected rabbits were collected prior to infection (zero time) and at time intervals designated in the text. Sera from a noninfected rabbit, collected at the same intervals, served as a negative control. All sera were frozen at -20°C until tested. (ii) Preparation of antigen. A culture filtrate antigen was prepared as follows. Cells of C. albicans were collected in 0.9% saline from a 24-h brain heart infusion (BHI) agar slant grown at 370C. The cell suspension was centrifuged at 5,000 x g for 15 min (400) and was washed twice with saline. Volumes of 10 ml of saline containing 5 x 106 cells/ml were added to 500 ml of BHI broth in 2-liter Erlenmeyer flasks. Cells were grown at 370C in shake culture (150 rpm) for 2 weeks. After centrifugation (7,000 x g for 30 min at 40C), a portion of culture filtrate (500 ml) was concentrated to a volume of 20 ml by use of a Buchi rotary flash evaporator (Brinkmann Instruments Inc., Westbury, N.Y.). The concentrated culture filtrate was sterilized by use of a 0.45-i,m membrane filter (Nalge Corp., Rochester, N.Y.) and was dialyzed against deionized water with frequent changes for 72 h. The nondialyzable material was refiltered and frozen at -20°C until needed. Soluble protein, determined by the procedure of Lowry et al. (9), amounted to 1 mg of protein/ml of culture filtrate. A cytoplasmic antigen was prepared by a modification of the procedures of Reiss et al. (13) and Taschdjian et al. (21). C. albicans was grown at 370C in BHI broth as described above. Cultures were incubated for 48 h, collected by centrifugation at 7,000 x g for 30 min, washed three times with deionized water, and resuspended in phosphate-buffered saline (PBS), 5 x 10-3 M, pH 7.3. The cell suspension was adjusted with PBS so that a 1:1,000 dilution gave an optical density of 0.2 at 550 mm, and 20 ml of this cell suspension containing 48 g of glass beads (0.45 to 0.5 mm; Bronwill Scientific Inc., Rochester, N.Y.) was homogenized with a Braun homogenizer (Bronwill Scientific Inc.) using four 30-s breaking intervals. The cell suspension was cooled during homogenization by a stream of CO2. After cell breakage (90 to 99% breakage as determined by phase microscopy), the cytoplasmic fraction was collected by centrifugation (7,000 x g for 15 min). Soluble protein amounted to 3 g of protein/100 ml of cell-

141

free extract. The extract was frozen at -20°C until needed. (iii) Immunodiffusion procedures. Immunodiffusion was carried out in 90-mm petri dishes containing 0.86% Noble agar (Difco) and 0.02% sodium azide. Antigen preparations were used in a concentrated form or diluted 1:2 or 1:10 with PBS. Reactions were read at 24, 48, and 72 h. Precipitin titers were determined by use of sera diluted in PBS. (iv) Charcoal agglutination. Several concentrations of cytoplasmic antigen were evaluated in charcoal agglutination tests using sera from catheterized rabbits. An antigen concentration of 0.25 mg/ml gave optimal agglutination reactions, and this concentration of antigen was used throughout these studies. The antigen-charcoal conjugate was prepared in the following manner. A charcoal suspension (Hynson, Westcott & Dunning, Inc., Baltimore, Md.) was diluted in sterile, physiological saline (1:3:5, vol/ vol). A portion of this suspension (1.85 ml) was blended in a Vortex mixer with 0.4 ml of antigen for 1 min at room temperature and was mixed overnight at 40C on an Ames Aliquot Mixer (Fisher Scientific Co., Pittsburgh, Pa.). The charcoal-antigen conjugate was then centrifuged (1,800 x g, 40C), washed with cold, sterile saline, and resuspended in a final volume of 1.85 ml of saline. Serial dilutions of sera from infected rabbits were made with heat-inactivated normal rabbit serum in a final volume of 100 ,ul. The antigen-charcoal conjugate was added to each serum dilution on RPR cards (18-mm circles; Hynson, Westcott & Dunning, Inc.), mixed manually, and then rotated (100 rpm) on a New Brunswick shaker (model G-2) for 45 min. Agglutination reactions were read immediately. Evaluation of antifungal agents in vitro. The minimal inhibitory concentration (MIC) and minimal fungicidal concentration (MFC) of amphotericin B (supplied by E. R. Squibb & Sons, Princeton, N.J.) and 5-FC (supplied by Hoffman-La Roche, Inc., Nutley, N.J.) for the test organism were determined by use of serial twofold dilutions. All studies were done in yeast nitrogen base (YNB) (Difco) supplemented with L-asparagine and dextrose by the method of Shadomy (18) with the yeast in the stationary growth phase. Studies of drug synergism were done by a previously reported titration scheme using YNB broth (18) and by a checkerboard microtiter technique (Dynatech Laboratories). The initial number of organisms used was approximately 107/ml. Initial concentrations of amphotericin B and 5-FC were 8 ,ug/ ml and 16 i±g/ml, respectively. Cultures were incubated for 24 and 48 h. The MIC was defined as the lowest concentration of drug preventing visible turbidity after 24 h of incubation. The MFC was defined as the lowest concentration of drug that produced complete sterility on subculture after 48 h of incubation on Sabouraud agar. The rate at which each drug alone and in combination killed the fungus was determined with 106 washed organisms in the YNB broth, rabbit sera, and heparinized rabbit blood. Amphotericin B at a concentration of 2.5 ,ug/ml, 5-FC at 15 ,ug/ml, and a

142

SANDE, BOWMAN, AND CALDERONE

combination of both drugs at these concentrations were used. Samples were removed at 0, 4, 24, 48, and 72 h. Serial 10-fold dilutions were made in sterile saline, and pour plates were made with Trypticase soy media. Colonies were counted and recorded at 48 h. Treatment. One hundred two rabbits were catheterized and 6 days after i.v. injection of 106 C. albicans cells were treated with amphotericin B, 5-FC, or the two drugs in combination. Amphotericin B was administered (2.5 mg/kg) in 20 ml of saline i.v. through an ear vein over a 30-min period every 48 h. This dose assured measurable antifungal activity for the entire 48-h period. Doses of 25 mg of 5-FC/kg were dissolved in 40 ml of sterile water and given intraperitoneally every 8 h. Higher doses of 5-FC produced excessive mortality in this rabbit model. Serum concentrations of both drugs were measured by an agar well diffusion technique with Saccharomyces cerevisiae (ATCC 9763) as the test organism, by methods used by Carrizosa et al. (1). The zones of inhibition were measured with a FisherLilly antibiotic zone reader after 24 h of incubation at 37°C. Levels of antibiotic were then determined by reading the concentration directly from a standard curve derived simultaneously from known concentrations of the antibiotic diluted in rabbit serum. Animals were sacrificed as described above after 2, 4, or 6 days of therapy with 5-FC, amphotericin B, or the combination. All were sacrificed at least 48 h after the last dose of amphotericin B and at least 18 h after the last dose of 5-FC. The vegetations were removed and titrated as described above. The initial titration cultured was a 1:100 dilution, thus ensuring dilution of any residual antifungal activity below the MIC. One rabbit so treated was sacrificed 7 days after therapy ended and another, 21 days after the completion of therapy. RESULTS

INFECT. IMMUN. for

Candida. One rabbit developed positive cul-

tures on day 7 of infection and the other, on day 11; both remained positive until death. Can-

dida was isolated from kidney cultures of all six rabbits postmortem. One animal developed neurological signs of left-sided weakness 2 days prior to death. Biweekly examination with an ophthalmoscope revealed no signs ofretinopathy in any of the 34 animals. Anti-Candida antibody response. A culture filtrate concentrate (25x) and a cytoplasmic extract from C. albicans were used to detect precipitating antibody with sera from infected and catheterized, infected and noncatheterized, and noninfected rabbits. The maximum number of precipitin bands obtained from the serum of each rabbit is given in Table 1. Generally, this serum sample was the last one obtained prior to the death of each animal. The exceptions are the noncatheterized rabbits (no. 1012). With these animals, maximum titers by immunodiffusion were observed early in the infection process. Both the culture filtrate antigen and the cytoplasmic antigen were useful in detecting precipitins to C. albicans except for three of the nine rabbits with endocarditis, which died within 12 days after infection. Although both groups of rabbits produced precipitating antibody, the precipitin response of the TABLE 1. Precipitin reactions of sera from rabbits infected with C. albicans obtained by use of culture filtrate and cytoplasmic antigens No. of precipitin bands

of seRabbit Length of Time Clinical and microbiological characterissurvival rum analy- Cult no.a (day)si (das)b filtrate Cytoa mic antitics of the disease. The disease was characterized in the 34 rabbits who were allowed to run gen antigen the course of their disease. Survival ranged 1 11 6 0 0 from 2 to 58 days, with a mean survival of 26 2 8 7 0 0 days. 3 11 7 0 0 4 14 12 2 3 Fever, defined as a temperature of greater 5 17 15 2 2 than 39.6°C, was a poor clinical sign of active 6 2 18 15 3 infection. Of the 791 temperatures recorded, 7 21 21 2 4 only 53 (7%) were in the febrile range. 8 36 2 29 3 C. albicans was isolated from blood cultures 9 43 2 39 4 from 11 of 34 animals during the course of infec10 43 15 2 2 tion; none of the 11 had persistently positive 11 43 15 2 2 cultures. Only 66 of the total of 731 (9%) blood 12 43 2 15 2 cultures taken grew Candida. There was no 13 43 43 0 0 correlation between duration of survival and a Rabbits 1-9 had an aortic catheter indwelling positive cultures. The mean survival was 27 prior to i.v. infection. Rabbits 10-12 were infected days for rabbits with positive cultures and 24 i.v. without catheterization. Rabbit 13 was nonindays for those with negative cultures. fected. Two of the six rabbits tested with serial urine b Number indicates approximate time in which cultures eventually showed cultures positive greatest number of precipitin bands was observed.

EXPERIMENTAL C. ALBICANS ENDOCARDITIS

VOL. 17, 1977

rabbits with endocarditis was quantitatively and qualitatively greater than that of the rabbits without endoarditis. Precipitin titers tended to increase until the death ofthe rabbits with endocarditis, whereas in the rabbits without endocarditis precipitins peaked early after infection and either remained unchanged or decreased as the infection was eradicated (Table 2). Sera from a normal, noninfected rabbit were negative for precipitins throughout the study. Agglutination titers of sera from infected animals were determined by use ofthe cytoplasmic antigen (Table 3). Rabbits 1-6 had endocarditis, whereas rabbits 7 and 8, although infected, did not have endocarditis. Agglutinins were first TABLE 2. Precipitin titers of sera from infected rabbits as determined with the cytoplasmic antigen Rabbit

Titerb on day 12 21 15 29 39 1 0 (-) (-) 2 0 3 (-) 0 4 (-) 0 1:4 0 1:2 (-) 5 1:2 6 0 1:2 1:4 (-) 7 0 1:2 1:2 (-) 1:4 8 0 1:2 1:4 1:4 1:8 (-) 9 0 1:2 1:2 1:4 1:8 1:16 10 0 1:2 1:4 1:4 1:2 1:2 11 0 1:2 1:4 1:4 1:2 1:2 12 0 1:2 1:4 1:4 1:2 1:2 13 0 0 0 0 0 0 a Rabbits 1-9 had in indwelling catheter prior to i.v. infection. Rabbits 10-12 were infected without catheterization. Rabbit 13 was a noninfected control. b 0, No antibody detected; (-), animal deceased. no."

0

6 0 0 0 0 0 0 0 0 0 0 0 0 0

TABLE 3. Charcoal agglutination titers of catheterized and noncatheterized rabbits Rabbit

no.a

Agglutinin titerb on day

12 15 21 0 35 1 2 4 64 64 256 2 0 32 64 128 512 (-) 3 0 (-) 8 128 4 (-) 0 ND (-) 128 5 0 ND 32 64 128 6 0 ND 32 64 16 7 0 ND 16 128 32 8 0 ND 32 128 16 a Rabbits 1-6 were catheterized; rabbits 7 and 8 were infected i.v. only. b Reciprocal of highest serum dilution which gave visible agglutination. ND, not determined; (-), dead.

143

observed in sera 12 days postinfection. Rabbits 1-6 demonstrated an increasing agglutinin titer over a period of 12 to 35 days postinfection. Rabbits infected without prior catheterization likewise produced detectable antibody at 12 days postinfection. Although an increase in titer was observed between days 15 and 21, levels of agglutinating antibody decreased by day 35 in both of these animals. Pathological findings. All animals sacrified at least 6 days after inoculation had vegetations on or around the aortic valve, and several had extension to the endocardial wall of the left ventricle and aorta. The catheter tip was frequently buried in the vegetations. Outflow obstruction was common with the larger lesions, axd several appeared to have occlusion of the coronary ostia. Vegetations yielded a mean candida titer of log10 8.79 + 1.46 (standard deviation) CFU/g of tissue and weighed from 0.04 to 0.86 g. The size of the vegetation correlated with the duration of infection. Vegetations from rabbits surviving less than 26 days had a mean weight of 0.14 + 0.07 g, and vegetations from those surviving longer than 26 days weighed 0.62 ± 0.18 g (P < 0.001). Microscopically, the vegetations consisted of a tightly packed mass of pseudohyphae, blastospores, and fibrin (Fig. 1). Colonization of the catheter lumen with Candida was also observed. Rabbits that received an i.v. injection of C. albicans without precatheterization showed no signs of aortic valve vegetations. These animals were sacrificed after 43 days, and the fungus was not recovered from either cardiac valves or kidneys. Kidneys from all catheterized animals contained viable C. albicans at postmortem in titers of 104 to 106 CFU/g. Renal cortical infarctions were prominent, and Candida could be demonstrated microscopically throughout the renal parenchyma. There was evidence of infectious glomerulitis and tubular plugging. Mean spleen weight for all animals was 1.51 + 0.54 g (normal, 0.95 + 0.26 g). In vitro activity of antifungal agents. The MIC and MFC of amphotericin B determined in YNB broth were 2 ,ug/ml, whereas 5-FC was only fungistatic at a concentration of 8 ,ug/ml. Tested in triplicate by the checkerboard technique, the MIC of amphotericin B decreased to 0.5 ,ug/ml when combined with 2 ,ug of 5-FC/ml and to 1 ,ug/ml when combined with 0.5 ug of 5FC/ml. However, the MFC of amphotericin B was not altered by the addition of various concentrations of 5-FC. The time kill studies in serum and blood demonstrated a 2-log reduction in colony counts with amphotericin B and with the combination

144

SANDE, BOWMAN, AND CALDERONE

INFECT. IMMUN.

FIG. 1. Section of cardiac vegetation from a rabbit with experimental Candida endocarditis (silver methenamine stain).

of 5-FC and amphotericin B. Killing with 5-FC alone was not observed in serum or blood, but a 4-log reduction in CFU was demonstrated at 4 h with amphotericin B, 5-FC, and the combination when YNB broth was used as the incubating medium. Combination therapy did not increase killing in serum, blood, or YNB broth. The rate of fungicidal activity of amphotericin B alone was equal to or greater than that observed with amphotericin B and 5-FC in combination in all instances. Treatment of Candida endocarditis. One hundred two rabbits were treated with 5-FC, amphotericin B, or the combination. Serum amphotericin B concentrations determined in four rabbits 1 h after i.v. administration of 2.5 mg/kg were 13.0 to 17.0 ug/ml (mean, 15 ,ug/ ml). Serum amphotericin B concentrations at 48 h were 1.9 to 3.2 ,ug/ml (mean, 2.5 ,ug/ml). Serum 5-FC concentrations determined in four rabbits 1 h after 25 mg/kg intraperitoneally were 16.8 to 20 ,ug/ml (mean, 18.0 ,ug/ml), and trough concentrations after 8 h were 1.5 to 2.25 ,ug/ml (mean, 1.9 ,ug/ml). With 4 days of therapy, vegetation colony counts were only slightly lower than in control

animals with all three treatment regimens (range, logl0 6.42 to 7.27 CFU/g) (Fig. 2). However, after 6 days of therapy, 10 of 14 vegetations removed from animals treated with amphotericin B alone were sterile (