INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteremia caused by Streptococcus pneumoniae (the pneumococcus) is a major complication of pneumococcal infection in patients with underlying diseases (9, 16). Patients at increased risk of pneumococcal bacteremia include those with impaired production of anticapsular antibody, hypocomplementemia, or defective phagocytosis and killing by neutrophils and macrophages (6, 15). The spleen and liver clear pneumococci from the bloodstream, and abnormal splenic or hepatic function also can predispose to serious and recurrent pneumococcal infection (5, 6).

Patients with alcohol-induced liver cirrhosis are highly susceptible to pneumococcal bacteremia (6, 28, 29). This is thought to be due in part to decreased hepatic production of complement components (10, 24). Patients with alcoholic cirrhosis have lower C3 concentrations in serum and reduced functional complement activity compared to the normal population (10). Our laboratory has developed a rat model of liver cirrhosis in which cirrhotic rats with ascites show a similarly increased susceptibility to pneumococcal infection (18). Cirrhotic rats infected intratracheally with type 3 pneumococci had significantly more organisms in their bloodstreams than did control rats by 4 days postinfection. Cirrhotic rats also had significantly higher mortality and lower hemolytic complement activity in serum (as shown by the 50% hemolytic complement assay) in comparison to controls (18). These results, which mimic those found in humans, suggest that lower activity and reduced levels of complement components during alcoholic cirrhosis may be one reason for the increased susceptibility of cirrhotic hosts to pneumococcal infection.

S. pneumoniae produces a number of virulence factors that contribute to its pathogenesis. One factor that may be of particular importance in the cirrhotic host is pneumolysin (PLY). This 53-kDa protein toxin possesses two properties that contribute to pneumococcal virulence. First, it is cytolytic to several cell types, including endothelial and alveolar epithelial cells, and cells of the immune system, such as neutrophils and monocytes (3, 22, 26). Second, it activates the classical complement pathway (23). As reported previously for an in vitro study, complement activation by PLY led to a decrease in the opsonic activity of serum, thereby reducing the uptake of S. pneumoniae by neutrophils (23). Because cirrhotic hosts are inherently hypocomplementemic, we hypothesized that complement activation by PLY during pneumococcal infection may decrease available complement components to a critically low level, resulting in decreased clearance of pneumococci from the bloodstream.

To study this hypothesis, cirrhotic and control rats were infected with isogenic mutant S. pneumoniae strains expressing no PLY at all or PLY with and without intact complement-activating activity. The clearance of these strains from the bloodstream, the survival rate after infection, and the 50% lethal dose were studied to establish the unique role of PLY's complement-activating activity during pneumococcal bacteremia in the cirrhotic host.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Rat model of cirrhosis. Cirrhosis was induced by a previously published method (18, 25). Briefly, male Sprague-Dawley rats (Charles Rivers, Kingston, N.Y.) were fed rat chow and given water containing 1 mM phenobarbital (Sigma Chemical Co., St. Louis, Mo.) ad libitum. When their weight reached approximately 200 g, they were given an initial 0.04-ml dose of carbon tetrachloride (CCl₄; Sigma) by gastric lavage under light ether anesthesia (Fisher Scientific, Fair Lawn, N.J.). Subsequent doses, determined by calculating weight changes 48 h after the previous dose, were given weekly for a period of 8 to 16 weeks until the rats developed stable ascites for at least 2 weeks. They then were rested for a week without further CCl₄ treatment before being used in experiments. Age-matched control rats also were fed rat chow and phenobarbital water but were subjected to gastric lavage each week with phosphate-buffered saline.

Bacterial strains. To study the role of PLY's complement-activating activity on pneumococcal pathogenesis, three isogenic mutant strains of type 3 S. pneumoniae were used (kindly provided by Mary K. Johnson, Tulane University). All the strains were produced as described previously from the serotype 3 WU2 parent organism in which the PLY gene was excised from the chromosome and then reexpressed from plasmid pVA838 (12, 13, 20). This plasmid also carries a gene for erythromycin resistance. The H+C+ strain expresses wild-type PLY with both hemolytic (cytolytic) and complement-activating activities. The H+C strain expresses PLY with hemolytic but not complement-activating activity due to a point mutation (12). The PLY strain carries the plasmid without the PLY gene and so does not produce PLY.

In vitro growth kinetics of isogenic mutant strains. To ensure that the growth characteristics of the study organisms were equivalent, each mutant strain was grown at 37°C under 5% CO₂ in Todd-Hewitt broth (Difco Laboratories, Detroit, Mich.) supplemented with 5% heat-inactivated normal rabbit serum (GIBCO, Grand Island, N.Y.) and 10 µg of erythromycin (Abbott Labs, North Chicago, Ill.) per ml. Each culture was started at an inoculum of 0.5 × 10⁵ to 2.0 × 10⁵ CFU/ml, and a 1.4-ml sample was collected every 2 h for 24 h. Serial 10-fold dilutions of each sample were plated in duplicate onto blood agar plates (Remel, Lenexa, Kans.), and the mean CFU per milliliter was plotted. A doubling time was calculated for each mutant strain on two separate days to ensure consistency.

Clearance studies. To quantify the clearance of pneumococci from the bloodstream, each mutant strain was grown as described above. The bacteria were collected by centrifugation, washed once in sterile phosphate-buffered saline, and resuspended to 2.0 × 10⁹ CFU/ml as estimated spectrophotometrically at 540 nm and confirmed retrospectively by plate counts. Cirrhotic and control rats were infected with 0.2 ml of inoculum via their tail vein, and blood was drawn immediately from a different site on the tail to confirm that each rat received 1.3 × 10⁷ to 2.5 × 10⁷ CFU/ml of blood. Additionally, blood samples were collected at 2 and 18 h postinfection. All the blood samples were serially diluted, and the dilutions were streaked onto blood agar plates to determine the number of CFU per milliliter of blood.

Retention of plasmid and PLY gene during growth in vivo. To determine if the pVA838 plasmid was retained by the isogenic mutant strains during 18 h of growth in vivo, 50 colonies from each of four to six cirrhotic and control rats were replica plated onto blood agar plates with and without 10 µg of erythromycin per ml. The plates were incubated for 24 h at 37°C under 5% CO₂, and growth on the two types of media was compared.

Survival studies and LD₅₀. The survival of cirrhotic and control rats infected with 2.0 × 10⁷ CFU of the three mutant strains per ml of blood was recorded for 10 days. The 50% lethal dose (LD₅₀) for cirrhotic and control rats was determined by the method of Litchfield and Wilcoxon with rats infected with various doses of each of the mutant strains (17).

Statistical analysis. The results of the growth curve and bloodstream clearance experiments are expressed as the mean ± standard deviation of pneumococcal numbers at each time point. Comparison of clearance data in cirrhotic and control rats was done by Student's t test. Fisher's exact test was performed to determine significant differences in survival after infection, and the 95% confidence intervals for lethal-dose calculations were determined by the method of Litchfield and Wilcoxon (17). A P value of <0.05 was used to determine significance in the mortality and LD₅₀ studies.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

In vitro growth kinetics of isogenic mutant strains. To ensure that the isogenic mutants grew at an equivalent rate, the H+C+, H+C, and PLY strains were grown in vitro and their doubling times were calculated. All three strains demonstrated similar growth curves over a 24-h period, with each reaching maximum stationary phase by 8 h after inoculation (Fig. 1). The calculated doubling times for the H+C+, H+C, and PLY strains were 0.67, 0.74, and 0.54, respectively.

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Growth curves for WU2 pneumococcal mutant strain in vitro. Each strain was grown at 37°C in Todd-Hewitt broth supplemented with 5% rabbit serum and 10 µg of erythromycin per ml. Viable counts over time were determined by a plate count technique.

Retention of plasmid and PLY gene during growth in vivo. To demonstrate that the pVA838 plasmid was retained by the mutant strains during growth in vivo, colonies isolated from the bloodstreams of rats 18 h postinfection were replica plated onto blood agar plates with and without erythromycin. All colonies tested (approximately 50 per rat) from each of the three mutant strains grew on erythromycin-containing medium, demonstrating retention of their erythromycin resistance plasmids.

Bloodstream clearance studies. Clearance of bacteria from the bloodstreams of rats infected with 2.0 × 10⁷ CFU/ml of blood was assessed to demonstrate the importance of PLY's complement-activating activity on resolution of bacteremia. In control rats, the levels of each bacterial strain in the bloodstream decreased within the first 2 h postinfection. By 18 h, the control rats had cleared a mean of 3.5 log units of the H+C+ strain from their bloodstreams (Fig. 2). The levels of the H+C and PLY strains also dropped a mean of 5.0 and 6.2 log units, respectively, by 18 h postinfection. In contrast, cirrhotic rats infected with the H+C+ strain failed to clear the organism during the first 18 h of infection (Fig. 2). The number of CFU per milliliter blood rose by a mean of 0.2 log units in the cirrhotic rats, which was significantly different from the clearance by controls (P = 0.0002). When cirrhotic rats were infected with the H+C or PLY strains, however, the bacteremia fell by a mean of 6.0 and 5.6 log units, respectively, after 18 h, similar to the clearance rats in control rats.

View larger version (44K): [in this window] [in a new window]   FIG. 2.   Clearance of mutant organisms from the bloodstreams of control (top) and cirrhotic (bottom) rats. All rats (n = 5 to 9) were infected intravenously with 2.0 × 107 CFU of the isogenic mutant strains per ml of blood. The number of CFU per milliliter of blood was determined at each time point by a plate count technique. *, significantly higher for cirrhotic than for control animals (P = 0.0002). SD, standard deviation.

Survival studies. All cirrhotic rats died by day 3 after infection with 2.0 × 10⁷ CFU of the H+C+ strain per ml of blood, whereas 45% of the control rats survived the infection (Fig. 3). By contrast, infection of both groups of rats with an equivalent inoculum of the PLY strain resulted in significantly higher survival rates (P < 0.05) than for rats infected with the H+C+ strain. All control rats and 85% of cirrhotic rats infected with the PLY strain survived intravenous challenge. When cirrhotic and control rats were infected with 2.0 × 10⁷ CFU of the H+C strain per ml of blood, 60% of the rats in each group survived.

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Survival of cirrhotic and control rats after infection with the isogenic mutant strains. Rats (n = 5 to 9) were infected with 2.0 × 107 CFU of the isogenic mutant strains per ml of blood as indicated, and survival was determined for 10 days postinfection. Solid symbols represent cirrhotic rats (cir), and open symbols represent controls (con). *, survival was significantly higher for rats infected with the PLY strain than for those infected with the H+C+ strain (P < 0.05).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The complement system is an integral component of the immune response to S. pneumoniae. Deficiencies in functional complement activity and/or levels of complement component C3 in serum are associated with ineffective clearance of the organism from the bloodstream. Injection of guinea pigs with complement-depleting cobra venom factor led to a significant decrease in clearance of type 7 S. pneumoniae from the bloodstream (4, 11). Moreover, compared to controls, complement-depleted guinea pigs were killed by smaller numbers of pneumococci (5). In another study of pneumococcal pneumonia in rats, complement depletion with cobra venom factor led to an increase in the numbers of pneumococci in their lungs at 24 h postinfection (7). Cobra venom factor treatment commonly reduces the hemolytic complement levels in experimental animals to only 0.1 to 1% of normal values (11). Similar investigations have not yet been performed with hosts with more moderately reduced complement levels, such as those with hepatic cirrhosis.

Our laboratory has developed a model of pneumococcal disease in cirrhotic rats, which have complement levels of 37 to 81% of those found in control rats (18). The pneumococcal toxin PLY can activate and consume complement at a distance from the surface of the organisms. Therefore, we hypothesized that production of PLY would contribute significantly to the increased virulence of S. pneumoniae in these hypocomplementemic cirrhotic hosts. To test this hypothesis, we studied the effect of PLY production on the clearance of type 3 S. pneumoniae from the bloodstreams of rats with CCl₄-induced cirrhosis.

Differences in the pathogenicity of the mutant strains were associated with the effects of the PLY they produced rather than with any major differences in their growth kinetics or quantitative expression of the toxin. The strains grew at comparable rates in vitro and retained their plasmids during growth in vivo. The H+C+ and H+C organisms isolated from the rats 18 h postinfection also produced amounts of the toxin comparable to those produced by the infecting strains. This is in contrast to results from a previous study with the same mutant strains in a rabbit corneal-infection model. In that study, 17 of 18 H+C+ organisms recovered from rabbits' eyes 48 h after infection had lost the plasmid insert encoding PLY production. However, those authors suggested that the organisms retained the ability to produce PLY "for a length of time sufficient to produce pathology similar to the wild type strain" (13). This is consistent with the results of the present study, in which the gene was retained for at least 18 h after infection. It is possible that the organisms lost their PLY gene insert during prolonged growth within the rats, since isolates were not tested at a later time point. However, differences in the survival of rats infected with the H+C+ and PLY strains suggest that the effects of PLY were established well before this event, if it did indeed occur.

The reduced clearance of the H+C+ strain from the bloodstreams of cirrhotic rats is associated primarily with the complement-activating activity of the toxin, as shown by both the survival and LD₅₀ studies. Cirrhotic rats were far more likely to die after infection with the H+C+ strain than after infection with the H+C strain. Furthermore, the LD₅₀ of the PLY strain in cirrhotic rats was equal to that of the H+C strain, showing that the detrimental effect of PLY during type 3 pneumococcal infection of cirrhotic animals is predominantly related to its complement-activating activity. We have not performed comparable experiments with less virulent S. pneumoniae serotypes, and so it is unclear whether PLY's contribution is similarly influential during bloodstream infections with all pneumococci. In addition, the fact that the PLY strain is still more virulent for cirrhotic than for control rats suggests that the increased virulence of the type 3 pneumococcus for cirrhotic animals is not due solely to events related to PLY production.

By contrast to what occurred in the cirrhotic animals, the survival of control rats was equivalent after infection with the H+C+ and H+C strains, whereas they were not killed by even very high inocula of the PLY strain. These results indicate that PLY's cytolytic activity plays a more important role in pneumococcal virulence during bloodstream infections of these noncompromised hosts. These results are consistent with those of Berry et al., who found that PLY's cytolytic activity but not its complement-activating activity, contributed to pneumococcal virulence during systemic infection of normal mice (2).

Our results confirm those of previous studies which showed the importance of PLY's complement-activating activity on pneumococcal virulence in situations where concentrations of complement may be limited. In mice, complement activation by PLY contributed to the development of bacteremia during early pneumococcal pneumonia (27). In a rabbit corneal-infection model, PLY contributed to the intense inflammatory response within the rabbits' eyes after pneumococcal challenge (13, 14). In each case, PLY-induced activation of complement was detrimental to the host, resulting in decreased opsonization of the organisms and in induction of the anaphylatoxins C5a and C3a. Unrestricted production of these inflammatory responses can be a major factor in the tissue damage caused by bacterial infections (26).

Although our results emphasize the importance of PLY's complement-activating activity on pneumococcal virulence, they do not diminish the role of the toxin's cytolytic activity. This action of the toxin damages mammalian cell membranes, disrupting the integrity and function of neutrophils and other immune system cells important in defense against pneumococcal disease (8, 19, 21). Survival of both cirrhotic and control rats was significantly higher after infection with the PLY strain than with the H+C+ strain and was intermediate for rats infected with the H+C strain. This shows that the cytolytic activity of the toxin also contributes to the organism's virulence in both groups of animals.

In conclusion, PLY production increases the pathogenicity of S. pneumoniae in both cirrhotic and control rats. It is the complement-activating activity of PLY, however, that appears to be especially important in reducing pneumococcal clearance from the bloodstream and inducing excessive mortality in hypocomplementemic, cirrhotic rats. Studies are under way in our laboratory to determine the specific effects of complement activation by PLY on total complement activity and C3 levels during pneumococcal bacteremia of cirrhotic rats.