Variants of a Cryptococcus neoformans Strain Elicit Different Inflammatory Responses in Mice

Abstract

The virulence of Cryptococcus neoformans isolates with high and low extracellular proteolytic activity was investigated in mice. No consistent relationship between proteolytic activity and virulence was observed, but isolates derived from one strain were shown to elicit different inflammatory responses.

Cryptococcus neoformans infections are notorious for eliciting very different types of inflammatory responses in both immunocompromised and immunocompetent hosts (5, 11). The host and fungus variables responsible for the protean inflammatory responses to C. neoformans infection are not understood. C. neoformans produces a variety of extracellular enzymes and proteinases (1, 3, 4) which may be associated with virulence and could conceivably influence the inflammatory response. Histological studies have shown degeneration of collagen fibrils in the tissue of mice infected with C. neoformans, suggesting that proteolytic activity could promote tissue destruction (10). Proteases have been suggested as possible virulence factors for C. neoformans (2), but their role in virulence has not been experimentally investigated. C. neoformans has extracellular proteolytic activity associated with proteins of ∼200, 100, and 50 kDa and can grow in media containing protein as the sole source of nitrogen and carbon (3). Here, we investigate the relationship between inflammatory response, virulence, and proteolytic activity for seven isolates of one strain which arose through spontaneous microevolution in different laboratories (7).

Strain 24067 originated from the American Type Culture Collection. Isolates 24067-A, 24067-B, 24067-E, and 24067-G were obtained from four different laboratories and differ in extracellular proteolytic activity as well as other characteristics (Table 1). Isolates 5B, 29A, and 29B are mutants of strain 24067-A generated by irradiation with 65,000 μJ of 254 nm-light (UV Stratalinker 1800; Stratagene, La Jolla, Calif.) and selected after replica-plating onto protein agar plates (3). Capsule size and proteolytic activity of the strain 24067 isolates were measured and normalized for colony size, and melanization was determined as previously described (7). Doubling times were calculated by plotting CFU (1 colony = 1 CFU) data as a function of time and curve fitting the data using SigmaPlot version 2.03 (Jandel Scientific, San Rafael, Calif.). Adult female A/J Cr mice were purchased from the National Cancer Institute (Frederick, Md.) and infected by either intraperitoneal (i.p.) or intratracheal (i.t.) inoculation with 0.5 × 10⁸ to 1 × 10⁸ or 4 × 10⁵ cells, respectively, as described previously (6, 8). At various times after infection, mice were killed and organs were removed for CFU counts and histopathology as described (7).

TABLE 1.

Characteristics of isolates of strain 24067 used in this study

Isolate	Laboratorya	Clearance halob at 9 days	Clearance halob at 14 days	Capsule thickness/cell diameterc	Pigment formation on l-dopa plates	Doubling timed at 37°C (h)	Log10 of CFUs/organe
							Lung	Liver
24067-A	A. Casadevall	0.037 ± 0.040	0.631 ± 0.093	0.347 ± 0.140	Yes	3.94	4.466 ± 0.363	5.843 ± 0.317
5B	A. Casadevall	Not detected	0.070 ± 0.076	0.229 ± 0.099	Yes	3.61	3.768 ± 0.367	5.174 ± 0.565
29A	A. Casadevall	0.142 ± 0.095	1.061 ± 0.109	0.267 ± 0.130	Yes	3.84	3.199 ± 0.267	4.901 ± 0.372
29B	A. Casadevall	0.195 ± 0.184	1.084 ± 0.115	0.399 ± 0.190	Yes	3.43	4.093 ± 0.494	5.117 ± 0.387
24067-B	M. Scharff	0.342 ± 0.042	0.780 ± 0.056	0.236 ± 0.104	Yes	3.67	7.180 ± 0.197	3.563 ± 0.854
24067-E	G. Huffnagle	0.258 ± 0.064	0.671 ± 0.051	0.195 ± 0.074	Yes	3.02	6.997 ± 0.621	3.539 ± 0.729
24067-G	J. Murphy	0.015 ± 0.020	0.445 ± 0.042	0.122 ± 0.050	Yes	3.09	3.836 ± 1.534	3.732 ± 1.563

^aThe laboratory from which the strain was obtained for this study. For additional details of the origin and handling of these strains, see Franzot et al. (7). 

^bClearance halo is measured as the thickness of the halo divided by the colony diameter. The numbers are averages ± standard deviations (n = 30). At day 9 clearance was significant at P < 0.05 for all pairwise comparisons except 24067-A versus 5B or 24067-G and 5B versus 24067-G (Bonferroni t test). At day 14, clearance was significant at P < 0.05 for all pairwise comparisons except 29A versus 29B and 24067-A versus 24067-E (Bonferroni t test). 

^cCapsule and cell measurements were done when cultures had a density of ∼10⁸ cells/ml. The numbers are averages ± standard deviations (n = 20). Statistically significant differences were seen for the following pairwise comparisons: 29B versus 5B, 29A, 24067-B, -E, and -G; 24067-A versus 24067-G; and 24067-A versus 5B (P < 0.05; Bonferroni t test). 

^dDoubling time was calculated from the logarithmic phase growth rate by curve fitting the CFU data and applying the formula N = N₀ × e^{(a × t)} (SigmaPlot version 2.03; Jandel Scientific, San Rafael, Calif.) with t defined as time, N as the CFU at t, N₀ as the initial CFU, and a as the growth factor of the isolate. Errors are the standard errors of the means of the calculated a values. 

^eInfection with isolates 24067-A, 5B, 29A, and 29B was done via i.p. injection of C. neoformans. Isolates 24067-B, -E, and -G were administered via i.t. infection. Statistical significance within the groups of infections was seen only for pairwise comparison between lungs infected with 24067-B or -E versus -G (P < 0.05; Bonferroni t test). 

Extracellular proteolytic activity was measured at days 7 and 14 of growth in agar plates (Table 1). At day 14, the relative activity was 29B ≈ 29A > 24067-B > 24067-A ≈ 24067-E > 24067-G > 5B, levels which were statistically significant (P < 0.05; Bonferroni t test). Mice infected i.t. with isolates 24067-B and 24067-E had higher lung fungal burdens than those infected with isolate 24067-G, but there was no significant difference in liver CFUs (Table 1). Isolates 24067-E and -G had comparable mean survival times and extrapulmonary dissemination despite significant differences in extracellular proteolytic activity (Fig. 1; Table 1). Mice infected with isolates 24067-B or -E had comparable mean survival times following i.p. infection (7) but i.t. infection revealed significant mean survival differences (Fig. 1). The UV-induced mutants were generated before we learned that isolate 24067-A had reduced virulence (7), and we limited our virulence studies to measuring liver and lung CFUs after i.p. infection. The relative order of organ fungal burdens at day 4 of i.p. infection was 24067-A > 29B > 5B > 29A for both lung and liver tissue CFUs (Table 1). For isolate 24067-A and its variants derived by UV mutagenesis, differences in capsule thickness, doubling time, and extracellular proteolytic activity did not correlate with organ CFU burden.

FIG. 1.

Survival of A/J Cr mice after i.t. infection with 4 × 10⁵ cells of isolate 24067-B, -E, or -G. Mean survival times of mice infected with 24067-B, -E, and -G were 39.5, 94, and 94 days, respectively. The mean survival of mice infected with 24067-B was significantly different from the others (P = 0.0033 versus 24067-E; P = 0.0257 versus 24067-G [log rank test]). Seven and five mice remained alive from groups E and G, respectively, at the termination of the experiment and were censored. There was no difference between strains 24067-E and -G (P = 0.36; log rank test).

Histological examination of lung tissue from mice infected with isolates 24067-B, -E, and -G revealed surprising differences in the inflammatory response. Infection with 24067-G elicited significantly stronger granulomatous response with better containment of infection in lung tissues than isolates 24067-B and -E (Fig. 2) and was associated with lower lung CFU counts (Table 1). For 24067-B there was widespread dissemination of cryptococci throughout the alveoli with little inflammatory response (Fig. 2; Table 2). The extensive involvement of lung tissue after infection with 24067-B is not likely to be due to faster growth since this isolate has the longest doubling time of the three (Table 1). Overall, the intensity of the granulomatous inflammation in response to infection at day 7 with the three isolates was 24067-G > 24067-E > 24067-B. Later the inflammatory response to 24067-G decreased as the infection was cleared in the lung (Table 2). Since granuloma formation is a hallmark of cell-mediated immunity, this result strongly suggests that strain microevolution can result in variants that trigger different immune responses.

FIG. 2.

Histology of lung infection in A/J Cr mice at day 28 of i.t. infection. Panels A, B, and C are photographs of the inflammatory response to isolates 24067-B, -E, and -G, respectively. Tissue sections were stained with mucin and counterstained with methyl yellow and iron hematoxylin. For all panels, the tissues were photographed at a magnification of ×50. Similar results were seen at day 7 of infection.

TABLE 2.

Summary of histopathological characteristics of infected lung tissues

Day of infection	Strain	No. of mice analyzed	Inflammatory Responsea	Patternb
7	B	5	2	Diffuse
	E	5	2.6	Focal
	G	5	2.6	Focal and localized
28	B	8	1.69	Diffuse
	E	9	4.33	Focal
	G	8	2.88	Focal and localized

^aInflammatory response was graded for each mouse on a scale of 1 to 5. A rating of 1 corresponded to no inflammation. A rating of 5 corresponded to intense inflammation. The inflammatory response score is the average for the mice analyzed. 

^bPattern refers to a qualitative assessment of the extent and type of involvement in lung section. For representational pictures, see Fig. 2. 

We found no major differences in virulence among two independent sets of isolates that had large differences in extracellular proteolytic activity. For example, isolates 24067-E and -G had comparable virulence despite major differences in proteolytic activity (Table 1). Consistent with this finding, extracellular proteolytic activity is not a constant characteristic of virulent clinical strains. For example, strain 371 (serotype A) has almost no extracellular proteolytic activity (3) yet is highly virulent in mice (9). Another recent clinical isolate, J32, has no detectable extracellular proteolytic activity yet is virulent in humans and mice (unpublished observations). The apparent lack of association between extracellular proteolytic activity and virulence must be considered in the context of the limitations of this study which include (i) the analysis of only a single strain; (ii) the possibility that the animal model used may not have been sufficiently sensitive to demonstrate differences in survival for mice infected with isolates having high and low extracellular proteolytic activity; and (iii) the possibility that the isolates may not be isogenic because of microevolution or mutagenesis. Nevertheless, we note that isolate 24067-G had the lowest extracellular proteolytic activity and elicited the most intense inflammatory response. Conversely, isolate 24067-B had the highest extracellular proteolytic activity and elicited the least inflammation. Proteolytic activity could conceivably interfere with inflammatory responses by destroying molecules necessary for cellular recruitment, such as cytokines, chemokines, and adhesion molecules. However, any inferences regarding the relationship between isolate characteristics and virulence properties must be made cautiously because these isolates manifested several differences (Table 1). Our data show that isolates from a single strain can elicit different inflammatory responses and suggest the need for additional studies to explore the role of proteases in cryptococcal pathogenesis.