Susceptibility to Thrombin-Induced Platelet Microbicidal Protein Is Associated with Increased Fluconazole Efficacy against Experimental Endocarditis Due to Candida albicans

Abstract

Platelet microbicidal proteins (PMPs) are believed to be integral to host defense against endovascular infection. We previously demonstrated that susceptibility to thrombin-induced PMP 1 (tPMP-1) in vitro negatively influences Candida albicans virulence in the rabbit model of infective endocarditis (IE). This study evaluated the relationship between in vitro tPMP-1 susceptibility (tPMP-1^s) or resistance (tPMP-1^r) and efficacy of fluconazole (FLU) therapy of IE due to C. albicans. Candida IE was established in rabbits with either tPMP-1^s or tPMP-1^r strains. Treatment groups received FLU (100 mg/kg/day) intraperitoneally for 7 or 14 days; control animals received no therapy. At these time points, cardiac vegetations, kidneys, and spleens were quantitatively cultured to assess fungal burden. At both 7 and 14 days and in all target tissues, the extent of candidal clearance by FLU was greater in animals infected with the tPMP-1^s strain than in those infected with the tPMP-1^r strain. These differences were statistically significant in the spleen and kidney. Corroborating these in vivo data, FLU (a candidastatic agent), in combination with tPMP-1, exerted an enhanced fungicidal effect in vitro against tPMP-1^s and tPMP-1^r C. albicans, with the extent of this effect greatest against the tPMP-1^s strain. Collectively, these results support the concept that tPMP-1 susceptibility contributes to the net efficacy of FLU against C. albicans IE in vivo, particularly in tissues in which platelets and tPMP-1 likely play significant roles in host defense.

Candida bloodstream infections are increasingly common and often life-threatening, with an estimated frequency of 0.5 to 2.0 per 10,000 hospital admissions (13, 17, 30, 33) and an attributable mortality rate of 38 to 50% (13, 24, 40). Moreover, Candida species now represent up to 10% of all nosocomial bloodstream isolates in the United States (3, 24, 38). Hematogenously disseminated candidiasis is especially prevalent in neutropenic and immunocompromised patients and in postoperative patients receiving long-term intravenous antibiotic therapy or hyperalimentation (17, 24, 30, 40). Specifically, Candida albicans is the most common fungal pathogen associated with endovascular infections, such as infective endocarditis (IE), vascular catheter sepsis, and infections of vascular prostheses (13, 24, 29, 30, 40).

Innate and adaptive immune mechanisms are essential to the success of host defense against C. albicans (3, 4, 4a, 14, 18-21, 27, 28, 35, 44). Platelets are increasingly recognized as integral host defense effector cells against invasive infections (5, 6, 41, 47, 48, 52-55). Platelets exhibit many structural and functional hallmarks of antimicrobial host defense cells that limit the induction and/or progression of endovascular infection. Monocytes and endothelial cells may be stimulated to produce tissue factor, thus catalyzing thrombin generation in the setting of IE (1, 9-11). We have shown that thrombin can prompt the release of low-molecular-weight, cationic antimicrobial peptides from rabbit and human platelets (42, 49, 52, 55). Among these peptides, termed thrombin-induced platelet microbicidal proteins (tPMPs), tPMP-1 exerts potent in vitro activities against Candida species, as well as against Cryptococcus neoformans (47). Platelets have also been shown to protect vascular endothelial cells from damage due to C. albicans corresponding to inhibition of candidal germ tube elongation in vitro (S. G. Filler, M. Joshi, Q. T. Phan, R. D. Diamond, J. E. Edwards, Jr., and M. R. Yeaman, Abstr. 39th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 2163, 1999). The protective effect of platelets was significantly greater against tPMP-1^s C. albicans than against its isogenic tPMP-1^r strain, implicating tPMP-1 as a key effector of the antifungal activity of platelets. In addition, tPMP-1 reduces C. albicans adherence to platelets in vitro (48).

We have previously demonstrated that differences in susceptibility or resistance to tPMP-1 in vitro parallel the severity of IE due to C. albicans (50). These studies demonstrated that tPMP-1^s C. albicans achieves significantly reduced densities within target organs compared to its tPMP-1^r isogenic counterpart. These differences occurred despite equivalence in initial cardiac valve adhesion abilities, fungemia levels, and bloodstream clearance rates of these study strains in this model. These observations further support our hypotheses (7, 8, 50) that (i) the antimicrobial effects of tPMP-1 are likely most important in limiting the proliferation phase of infection, rather than initial adhesion of pathogens to a target tissue, (ii) a tPMP-1^s phenotype in vitro translates to reduced achievable microbial densities in vivo in specific target tissues compared to the tPMP-1^r phenotype, and (iii) the tPMP-1^r phenotype likely confers survival advantages to pathogens in selected target tissues. Similarly, Nail et al. have recently demonstrated that tPMPs likely influence clearance of fungi from the bloodstream in a murine model of infection (31). Together, these data support the idea that platelet release of PMPs in vivo contributes an important host defense role against Candida.

The proven in vivo efficacy of fluconazole (FLU) in treating fungal infections in settings in which fungicidal activity is considered crucial (e.g., neutropenic hosts) is not readily explainable by its in vitro fungistatic activity. This discrepancy has been the basis for an emerging view that antifungal drugs interact favorably with innate and acquired host defense mechanisms (15, 16, 23, 32, 36, 43). Thus, fungistatic agents such as FLU may benefit from synergistic interplay with antifungal host defense components, yielding a net fungicidal effect in vivo. The present study was designed to compare the influence of tPMP-1 susceptibility in vitro on the efficacy of FLU in vivo in IE caused by tPMP-1^s or tPMP-1^r isogenic strains of C. albicans.

(This work was presented in part at the 98th General Meeting of the American Society for Microbiology in Atlanta, Ga., 1999 [abstr. V-83].)

MATERIALS AND METHODS

C. albicans strains.

C. albicans strain ATCC 36082 is a well-characterized clinical isolate that consistently induces experimental IE and is susceptible in vitro to tPMP-1 (36082^s [47, 50]). Previously, we isolated a stable tPMP-1^r counterpart to the 36082^s strain by serial passage in high concentrations of tPMP-1; this tPMP-1^r strain has been termed 36082^r (50). We have demonstrated that, other than differences in their tPMP-1 susceptibility profiles, the 36082^s and 36082^r strains are indistinguishable by genotypic, phenotypic, immunologic, endothelial cell adhesion, metabolic, germination, and growth rate characteristics (50). The tPMP-1^s and tPMP-1^r phenotypes of these strains, respectively, are stable in vitro and on passage through experimental animal models (50). In addition, we have shown that the growth of the 36082^s strain—but not strain 36082^r—is inhibited on Sabouraud dextrose (SAB) agar containing 7 mg of the antimicrobial polypeptide protamine/ml (see below). This difference provided the basis for our use of protamine as a readily available surrogate screen for tPMP-1 susceptibility in the present study (50).

Candida strain 36082^s was stored at 4°C on SAB agar slants, whereas the 36082^r strain was stored at the same temperature on SAB agar containing 7 mg of protamine sulfate/ml. Prior to use in the experimental model, both organisms were routinely cultured to mid-logarithmic phase in yeast-nitrogen broth (Difco Laboratories, Detroit, Mich.) enriched with 0.5% dextrose (wt/vol) at 30°C on a rotating drum as previously described (50). As described above, yeast-nitrogen broth used for culture of the 36082^r strain contained 7 mg of protamine sulfate (see below)/ml. After they were harvested by centrifugation (3,000 × g, 10 min), the yeast cells were resuspended in phosphate-buffered saline (PBS; pH 7.2), sonicated on ice for 2 s to ensure that singlet organisms were obtained (60 Hz, Branson Sonifier model 350; Branson, Danbury, Conn.), washed twice in PBS, and sonicated again. Cells were then counted in a hemocytometer and adjusted to the desired concentration in PBS. Yeast inocula were confirmed by quantitative culture.

tPMP-1.

Native rabbit tPMP-1 was prepared, purified by reversed-phase high-pressure liquid chromatography, and bioassayed as previously described (49, 50). In brief, bioassays were performed with Bacillus subtilis (ATCC 6633), an indicator organism highly sensitive to the bactericidal action of tPMP-1. This organism was added to microtiter wells containing a range of dilutions of tPMP-1 preparations to achieve a final inoculum of 10³ CFU/ml per well and a range of tPMP-1 dilutions from 1:1 to 1:1,024 (final well volume of 200 μl). After 30 min of incubation at 37°C, the specific bioactivity of tPMP-1 (in units/milliliter) was defined as previously described (47, 49).

Screen for in vivo retention of C. albicans susceptibility or resistance to tPMP-1.

The antimicrobial activity of tPMP-1 is reduced in complex solid media, precluding the development of tPMP-1-containing agar. However, we have previously demonstrated that in vitro tPMP-1 susceptibility or resistance phenotypes in strains 36082^s and 36082^r, respectively, are mirrored in response to protamine. Thus, protamine was used as a surrogate marker in a qualitative screen for tPMP-1 susceptibility or resistance (50). Complementary to the tPMP-1 susceptibility assays below, the 36082^s and 36082^r strains were tested to confirm their respective in vitro susceptibility and resistance to protamine pre- and postpassage through the experimental animal model. For these studies, SAB agar was prepared with or without protamine sulfate (Sigma Chemical Co., St. Louis, Mo.) at a concentration of 7 mg/ml. Colonies were obtained from organ cultures of FLU-treated or control animals infected with either the 36082^s or the 36982^r strain. Each C. albicans strain was prepared as described above, and 10⁶ CFU were plated onto SAB with or without protamine. Visible growth after ≥24 h of incubation (35°C) on SAB containing protamine was interpreted as a positive in vitro screening result for putative tPMP-1 resistance. It is important to emphasize that protamine screening was used only as a surrogate for tPMP-1 susceptibility, which was independently confirmed (see below).

In vivo retention of tPMP-1 susceptibility or resistance.

Prior to in vivo infection and after recovery from target tissues in IE, the 36082^s (tPMP-1^s) and 36082^r (tPMP-1^r) Candida strains were tested to confirm retention of respective susceptibility or resistance to tPMP-1. To do so, 10³ CFU of either strain/ml was exposed to a tPMP-1 preparation (5 μg of minimal essential medium [pH 7.2]/ml) for 3 h at 30°C as previously documented (47, 50). The antifungal action of tPMP-1 was stopped at 3 h by the addition of sodium polyanethol sulfonate (0.01% [wt/vol]; Sigma). Control cultures were processed in parallel for each strain. All reaction mixtures were sonicated to ensure that singlet organisms were obtained and quantitatively cultured onto SAB agar. Colonies were enumerated after incubation for ≥24 h at 30°C, and the percent survival of the initial inoculum compared to the controls was determined. The breakpoints for phenotypic susceptibility or resistance to tPMP-1 were considered to be ≤50% or ≥90% survival, respectively (47, 50).

FLU susceptibility testing.

The in vitro susceptibilities of the C. albicans 36082^s and 36082^r strains to FLU were determined by a previously published method (45). In the present study, an inoculum of 10⁶ CFU/ml and an incubation temperature of 30°C were used to simulate readily achievable fungal densities and growth rates relevant to complex biomatrices, such as cardiac vegetations, respectively. FLU powder (a gift from Pfizer Research, Groton, Conn.) was reconstituted according to the manufacturer's instructions. The FLU assay concentrations ranged from 80 to 0.125 μg/ml. After incubation for 24 h, MICs were identified as the lowest FLU concentrations preventing visible growth. Minimum fungicidal concentrations (MFCs) were determined by plating the MIC dilutions onto SAB agar, followed by incubation for another 24 h at 30°C. The MFCs were then identified as the lowest FLU concentrations preventing ensuing growth on SAB agar.

Anticandidal effect of tPMP-1 combined with FLU in vitro.

The potential for enhanced antifungal effects of tPMP-1 in combination with FLU was also assessed. These studies were carried out as described above for the tPMP-1 assays, with the modification that FLU (range, 0 to 2.5 μg/ml [two times the MIC]) was included in exposures containing 5 μg of tPMP-1/ml prior to introduction of the organism. As described above, quantitative cultures were performed after 3 h of exposure at 37°C, and the results were compared to assess candidal survival after exposure to medium alone, tPMP-1, or FLU individually or the combination of tPMP-1 and a range of FLU concentrations. A reduction in surviving C. albicans CFU of ≥50% after exposure to combined tPMP-1 plus FLU, compared to tPMP-1 alone, was considered to represent enhanced candidacidal activity of the combination.

Rabbit model of C. albicans IE.

The rabbit model of experimental aortic valve IE was used in these studies (50, 51). The protocols used in the present study are as previously described (51). The catheter remained in place for the duration of the experiment. IE was produced 48 h after catheterization by intravenous injection of 2 × 10⁷ CFU of either the 36082^s (tPMP-1^s) or the 36082^r (tPMP-1^r) C. albicans strain. This challenge inoculum yields IE in >95% of catheterized animals for both strains, which do not differ in their abilities to adhere to sterile vegetations or in fungemia levels or clearance rates (50).

FLU therapy.

FLU was reconstituted in Cremophor EL vehicle (Sigma) to a concentration of 12.5 mg/ml for intraperitoneal administration. To generate adequate statistical power, four groups containing 20 rabbits each were catheterized and challenged with either the 36082^s or the 36082^r C. albicans strain. These groups received intraperitoneal FLU therapy at 100 mg/kg/day beginning 24 h postinfection and continuing for either 7 or 14 days as previously described (45). This dose regimen has been previously shown to exert good in vivo efficacy in Candida IE caused by strain 36082^s (45). Untreated control groups for each FLU-treated group were comprised of 10 animals with IE.

Fungal burden in tissues.

Animals were sacrificed at 7 or 14 days after induction of IE. In some cases, morbid control animals were sacrificed prior to these time points. Rabbits were euthanized with 200 mg of sodium pentobarbital (Abbott Laboratories, Chicago, Ill.) administered by rapid intravenous injection. All sacrifices were performed at least 24 h after the last FLU administration to minimize antimicrobial carryover effects. At the time of sacrifice, proper catheter position was verified, and the heart, kidneys, and spleen of each animal were removed and quantitatively cultured in parallel onto SAB agar in the presence or absence of 7 mg of protamine sulfate/ml (50). The mean fungal burden, expressed as the log₁₀ CFU/gram of each tissue, was determined. Induction of IE was microbiologically confirmed by finding culture-positive vegetations in untreated controls.

Statistical analyses.

Differences in fungal densities within cardiac vegetations, spleens, and kidney tissues between FLU-treated and control groups (i.e., −Δlog₁₀ CFU/gram of tissue) or between groups infected with 36082^s versus 36082^r were compared by using the Wilcoxon rank-sum test, corrected for multiple comparisons. P values of <0.05 were considered statistically significant.

RESULTS

Candida susceptibility to tPMP-1 and FLU alone and in combination.

The FLU MICs (1.25 μg/ml; considered susceptible) and MFCs (>80 μg/ml) were equivalent for both strains (Table 1) and documented the in vitro fungistatic effect of this agent. As anticipated, strains 36082^s and 36082^r differed significantly in their susceptibility to tPMP-1 and its surrogate peptide, protamine, in vitro (Table 1). For example, strain 36082^s exhibited <50% survival of the initial inoculum after exposure to 5 μg of tPMP-1/ml for 3 h and was unable to grow on SAB containing 7 mg of protamine/ml. In contrast, >90% of the 36082^r C. albicans inoculum survived such tPMP-1 exposure, and this strain was not growth inhibited on SAB-protamine agar. Although FLU alone did not achieve a candidacidal effect at any concentration tested, the combination of tPMP-1 and FLU exerted an enhanced fungicidal effect compared to tPMP-1 alone (Fig. 1). Of note, the combination of tPMP-1 plus FLU at 2.5 μg/ml achieved a candidacidal effect against the tPMP-1^r strain 36082^r that was equivalent to that of tPMP-1 alone against the tPMP-1^s strain 36082^s (see Fig. 1). However, the net extent of this combined candidacidal effect was greater against the tPMP-1^s strain than against the tPMP-1^r strain.

TABLE 1.

Comparison of relevant characteristics of C. albicans 36082^s and 36082^r strains^a

Characteristic	CA 36082s (tPMP-1s)	CA 36082r (tPMP-1r)
FLU MIC (μg/ml)	1.25	1.25
FLU MFC (μg/ml)	>80	>80
Mean % survival in tPMP-1 (5 μg/ml) ± SD	43 ± 4	92 ± 6
Growth on SAB with protamine (7 mg/ml)	−	+
IE induction rate (%)	100	100
Mean IE vegetation wt (g) ± SD	0.1292 ± 0.052	0.1263 ± 0.037

^aSee the text for specific experimental methods for each determination.

FIG. 1.

Effect of FLU in combination with purified tPMP-1 on survival of C. albicans strains 36082^s or 36082^r in vitro. Circles indicate survival of 36082^s (•) or 36082^r (○) after 3 h of exposure to FLU alone at the concentrations indicated. The data confirm that FLU was not candidacidal against either strain at any concentration tested. Bars represent survival of 36082^s (▪) or 36082^r (□) after 3 h of exposure to purified tPMP-1 at a concentration of 5 μg/ml in combination with FLU at the concentrations indicated. Note the progressive dose-response relationship upon combined tPMP-1 and FLU exposure of the tPMP-1^s strain, 36082^s. Error bars correspond to variances in results from three independent experiments performed in triplicate. Asterisks indicate significant differences (P < 0.05) in survival compared to survival in the absence of FLU for each strain.

IE in untreated controls.

Both the 36082^s and 36082^r strains induced IE in all animals challenged. Consistent with our previous findings (50), strains 36082^s and 36082^r exhibited a differential propensity to hematogenously disseminate to and/or proliferate within specific target tissues in untreated controls. For example, strain 36082^r achieved substantially higher fungal densities within cardiac vegetations after 7 days (but not after 14 days) compared to controls challenged with strain 36082^s (Table 2; P = 0.05). Also, at the 7- and 14-day time points, strain 36082^r achieved mean splenic fungal densities that were significantly greater than those achieved by strain 36082^s (P = 0.01 and 0.03, respectively; Table 2). In contrast, there were no significant differences in achievable renal fungal densities of the two strains at either 7 or 14 days (Table 2).

TABLE 2.

Comparison of C. albicans 36082^s and 36082^r tissue densities in distinct target organs after 7 or 14 days with or without FLU therapy^a

Strain and tissueb	Mean CFU/g ± SD at:
	7 days			14 days
	Control	Treated	Difference	Control	Treated	Difference
36082s
Veg	6.0 ± 0.73	4.1 ± 0.32*	−1.9 ± 0.41	6.9 ± 0.21	4.9 ± 0.53*	−2.0 ± 0.32
Spl	3.9 ± 0.83	1.3 ± 0.28*	−2.6 ± 0.55†	4.2 ± 0.74	0.7 ± 0.56*	−3.8 ± 0.18†
Kid	3.1 ± 0.37	2.0 ± 0.35	−1.1 ± 0.02†	5.2 ± 0.20	1.7 ± 0.88*	−3.5 ± 0.68†
36082r
Veg	7.0 ± 0.37	5.2 ± 0.20*	−1.8 ± 0.17	7.3 ± 0.26	5.8 ± 0.74*	−1.5 ± 0.48
Spl	5.0 ± 0.64†	3.9 ± 0.22*	−1.1 ± 0.42†	5.6 ± 0.15	5.2 ± 0.44	−0.42 ± 0.29†
Kid	3.8 ± 0.19	3.9 ± 0.25	+0.1 ± 0.06†	4.3 ± 0.40	4.0 ± 0.17	−0.3 ± 0.23†

^aData are geometric means ± the corresponding standard deviations. −, a reduction in the log₁₀ CFU/g of tissue over the time period indicated; +, an increase in the log₁₀ CFU/g of tissue over the time period indicated; *, a significant difference from the respective control (P < 0.05); †, a significant difference in the log₁₀ CFU/g value in the respective tissue at the corresponding time (P < 0.05).

^bVeg, vegetations; Spl, spleen; Kid, kidney.

FLU therapy for IE. (i) Cardiac vegetations.

Vegetation densities were significantly reduced in animals infected with either Candida strain after 7 and 14 days of FLU therapy versus respective untreated controls (Table 2). The extent of candidal clearance from cardiac vegetations was greater for strain 36082^s than for strain 36082^r, although this difference did not achieve statistical significance (e.g., P = 0.07 at 14 days).

(ii) Splenic infection.

At 7 days of FLU therapy, the extent of candidal clearance from splenic tissue was significantly greater for the 36082^s strain than for the 36082^r strain (P < 0.05; Table 2). Likewise, 14 days of FLU therapy was associated with a significant reduction in the splenic density of strain 36082^s versus that in respective control animals (P < 0.01; Table 2). In contrast, there were no significant reductions in 36082^r strain densities within splenic tissue at 14 days of FLU therapy versus that in respective controls (Table 2).

(iii) Renal infection.

At both 7 and 14 days, FLU therapy substantially reduced the renal densities of strain 36082^s (Table 2). The extent of clearance was significantly greater for this strain than for strain 36082^r. Moreover, FLU did not achieve significant clearance of strain 36082^r from renal infection at either 7 or 14 days versus untreated controls (Table 2).

To assess the reproducibility of the primary results described above, we repeated a key component of these studies, focusing on the impact of the tPMP-1 susceptibility phenotype on the capacity of FLU to reduce fungal burden in cardiac vegetations. We confirmed two major outcomes consistent with our primary results above: (i) in untreated controls (n = 4 for each strain), strain 36082^r proliferated to a greater extent in cardiac vegetations than did its 36082^s counterpart (log CFU/gram = 6.73 ± 0.74 versus 6.25 ± 0.54, respectively); and (ii) FLU therapy exerted a significantly greater impact (n = 9 animals each) in reducing the tissue fungal density of strain 36082^s (mean Δ, −log 1.23; P < 0.05 versus untreated control) than of strain 36082^r (mean Δ, −log 0.56).

Retention of strain susceptibility to tPMP-1 pre- and postpassage in animals.

Isolates of strain 36082^r recovered from animals were culture positive on SAB-protamine agar (7 mg/ml). Therefore, both study strains retained their prepassage protamine susceptibility phenotypes ex vivo. Furthermore, randomly selected 36082^s or 36082^r isolates from target tissues of control or FLU-treated groups retained their tPMP-1 susceptibility or resistance phenotypes ex vivo, respectively (data not shown) (Table 1). Retention of original protamine and tPMP-1 susceptibility phenotypes was observed regardless of time point (7 days versus 14 days) or tissue source (e.g., vegetation, kidney, or spleen [data not shown]).

DISCUSSION

The important contributions of platelets and PMPs to host defense against endovascular infection are evident from numerous studies (7, 41, 50, 52-55). However, in concept, resistance to platelet-mediated antifungal effects may allow C. albicans to exploit the platelet as an adhesive surface, facilitating seeding of the vascular endothelium and ensuing hematogenous dissemination to target organs (4, 4b, 12, 14, 19-21, 27, 28, 37, 39). Thus, antifungal agents alone or in combination with host defense mechanisms may interfere with the establishment or progression of disease. A clearer understanding of the impact of host defense responses, such as tPMPs, on antimicrobial therapy of endovascular infections may yield important new insights and strategies.

We previously established an association between in vitro tPMP-1 resistance and more extensive candidal infection early in the course of experimental IE (72 h) (50). The present study extended these findings and compared the impacts of in vitro tPMP-1 susceptibility phenotypes on FLU efficacy later (7 and 14 days) in the course of infection in this model. The Candida strain pairs used were ideally suited for the present study, since they emanated from an identical genetic background, were indistinguishable by genotypic profiling (e.g., pulsed-field gel electrophoresis), and were identical by detailed phenotypic analyses (47, 50). Moreover, our recent identification of a specific genetic determinant in C. albicans that governs tPMP-1 susceptibility or resistance in vitro (M. R. Yeaman et al., unpublished results) further indicates that these are isogenic strains.

The rabbit model of C. albicans IE is well established and highly relevant, since optimal antimicrobial efficacy of this infection requires a net microbicidal effect. Thus, the use of a fungistatic agent, FLU, in this model enabled the detection of beneficial interactions between FLU and the anticandidal roles of the endogenous host defense effector molecule, tPMP-1. Although the magnitude of FLU efficacy differed in the three target organs studied, the net fungicidal effect of FLU was uniformly greater against the 36082^s strain than against the 36082^r strain at the 7- and 14-day endpoints. Similarly, in untreated animals, the extent to which these strains proliferate parallels their differences in tPMP-1 susceptibility profiles. These data support our hypothesis that tPMP-1 susceptibility is a significant determinant of net FLU efficacy in vivo. The differences in the extent of FLU efficacy in different target organs may indicate different efficacies of tPMP-1 in distinct physiologic settings. For example, many endogenous antimicrobial peptides, including tPMP-1, exhibit an inverse relationship between in vitro microbicidal activity and ionic strength (22); this point is potentially relevant to renal infections (22). Thus, the innate or synergistic efficacy of antimicrobial peptides such as tPMP-1 may be reduced in the context of the kidney compared to spleen or cardiac vegetations. In addition, the vascular endothelium and other endogenous immune mechanisms likely differ significantly in the heart, spleen, and kidney.

As noted above, microbiologic efficacies in endovascular infections such as IE are generally thought to require a microbicidal antibiotic regimen (2). We have previously postulated that local elaboration of tPMPs in vivo at sites of microbial colonization of damaged endothelium synergizes with conventional antimicrobial agents, contributing to the net microbicidal effects against endovascular infections (7, 8). In this respect, recent data from our laboratories strongly suggest that tPMPs and other PMPs amplify the microbicidal activities of conventional antibiotics in vitro and in experimental IE in vivo. Using Staphylococcus aureus as a bacterial prototype associated with endovascular infection, we confirmed the capacity of tPMP-1 to amplify the in vitro and in vivo antibacterial activities of conventional antibiotics, including oxacillin and vancomycin (7, 8, 46), especially against tPMP-1^s strains.

The results of the present investigation extend these paradigms to the in vitro and in vivo efficacy of FLU against C. albicans. We previously demonstrated that native PMP-2 enhances the candidacidal activities of FLU and amphotericin B in vitro against C. albicans strain 36082 used in the present study (A. B. Radner, M. A. Ghannoum, J. E. Edwards, Jr., A. S. Bayer, and M. R. Yeaman, Abstr. 93rd Gen. Meet. Am. Soc. Microbiol. 1994, abstr. F-24, 1994). Our present in vitro findings demonstrate that exposure of tPMP-1 in combination with FLU exerts an enhanced antifungal effect against tPMP-1^s and tPMP-1^r C. albicans. These in vitro observations are concordant with our in vivo findings, demonstrating that FLU therapy is more effective against tPMP-1^s than against tPMP-1^r C. albicans, particularly in target tissues in experimental IE in which tPMP-1 likely accumulates via platelet degranulation. Moreover, FLU efficacy in vivo appears to be uniformly greater against the tPMP-1^s strain than against its tPMP-1^r counterpart and is evident in all target tissues.

Recent studies have identified that human and rabbit platelets contain a family of related antimicrobial peptides. In humans, a predominant member of this family is platelet factor-4 (hPF-4 [42]). Yount et al. recently identified tPMP-1 as a novel rabbit homologue of hPF-4 (N. Y. Yount, K. D. Gank, Y. Q. Xiong, W. H. Welch, A. S. Bayer, and M. R. Yeaman, Abstr. 103rd Gen. Meet. Am. Soc. Microbiol. 2003, abstr. E-013, 2003). Although the mechanism(s) for the favorable interaction between tPMP-1 and FLU has yet to be defined, two reasonable models may be proposed based on previous and current data. Prior studies have shown that tPMP-1 perturbs the membrane structural integrity of Candida species (47). It is possible that in doing so, tPMP-1 potentiates the access or potency of FLU with respect to its mechanism of action. Thus, tPMP-1 and FLU may cooperate in the dysregulation of a single target: sterol synthesis and cytoplasmic membrane function. Alternatively, tPMP-1 may gain access to intracellular targets, such as candidal mitochondria. We have shown that tPMP-1 inhibits intracellular functions in bacteria (46) and may have similar effects in fungi. Thus, tPMP-1 and FLU may exert complementary inhibition of distinct targets, such as membrane structure-function and intracellular activities.

We recognize that FLU or other antifungal agents may act in concert with innate and/or adaptive immune mechanisms in addition to tPMP-1 to yield a net synergistic fungicidal impact. For example, Kuipers et al. (23) recently reported that FLU and the neutrophil antimicrobial peptide, lactoferrin, act synergistically in vitro against clinical isolates of Candida. Garcha et al. found that FLU synergistically increased the candidacidal actions of human monocytes in vitro (15). Interesting studies (16, 26) have suggested that the beneficial effect of antimicrobial or anti-inflammatory agents on cell-mediated immunity correlate with time of macrophage exposure to the agent. Other studies have found that antifungal agents improve the efficacy of phagocyte functions against fungal pathogens in vitro (32, 34, 36, 43). Likewise, Kullberg et al. suggested that FLU and granulocyte colony-stimulating factor achieve an additive effect in modulating the anticandidal functions of neutrophils, possibly accounting for the improved efficacy of this combined regimen in murine candidiasis (25). Collectively, these results provide compelling support for the hypothesis that agents such as FLU and intrinsic antifungal capabilities of the immune system potentiate one another, yielding mutually beneficial effects that amplify overall efficacy in vivo.

In summary, these findings suggest that in vitro tPMP-1 susceptibility or resistance of C. albicans influences the in vivo efficacy of FLU therapy, as well as pathogenesis related to proliferation within specific target organs. These data further suggest that susceptibility to tPMP-1 may significantly potentiate FLU efficacy against C. albicans, particularly in endovascular infections. These findings support our hypothesis that tPMP-1 and related peptides significantly contribute to antimicrobial host defense within and beyond the cardiovascular compartment (52-55) and potentiate the efficacy of several classes of conventional anti-infective agents.