Phenotypic Variability Correlates with Clinical Outcome in Cryptococcus Isolates Obtained from Botswanan HIV/AIDS Patients

Abstract

Pathogenic species of Cryptococcus cause hundreds of thousands of deaths annually. Considerable phenotypic variation is exhibited during infection, including increased capsule size, capsule shedding, giant cells (≥15 μm), and micro cells (≤1 μm). We examined 70 clinical isolates of Cryptococcus neoformans and Cryptococcus tetragattii from HIV/AIDS patients in Botswana to determine whether the capacity to produce morphological variants was associated with clinical parameters. Isolates were cultured under conditions designed to simulate in vivo stresses. Substantial variation was seen across morphological and clinical data. Giant cells were more common in C. tetragattii, while micro cells and shed capsule occurred in C. neoformans only. Phenotypic variables fell into two groups associated with differing symptoms. The production of "large" phenotypes (greater cell and capsule size and giant cells) was associated with higher CD4 count and was negatively correlated with intracranial pressure indicators, suggesting that these are induced in early stage infection. "Small" phenotypes (micro cells and shed capsule) were associated with lower CD4 counts, negatively correlated with meningeal inflammation indicators, and positively correlated with intracranial pressure indicators, suggesting that they are produced later during infection and may contribute to immune suppression and promote proliferation and dissemination. These trends persisted at the species level, indicating that they were not driven by association with particular Cryptococcus species. Isolates possessing giant cells, micro cells, and shed capsule were rare, but strikingly, they were associated with patient death (P = 0.0165). Our data indicate that pleomorphism is an important driver in Cryptococcus infection.IMPORTANCE Cryptococcosis results in hundreds of thousands of deaths annually, predominantly in sub-Saharan Africa. Cryptococcus is an encapsulated yeast, and during infection, cells have the capacity for substantial morphological changes, including capsule enlargement and shedding and variations in cell shape and size. In this study, we examined 70 Cryptococcus isolates causing meningitis in HIV/AIDS patients in Botswana in order to look for associations between phenotypic variation and clinical symptoms. Four variant phenotypes were seen across strains: giant cells of ≥15 µm, micro cells of ≤1 µm, shed extracellular capsule, and irregularly shaped cells. We found that "large" and "small" phenotypes were associated with differing disease symptoms, indicating that their production may be important during the disease process. Overall, our study indicates that Cryptococcus strains that can switch on cell types under different situations may be more able to sustain infection and resist the host response.

Keywords: Cryptococcus; capsule; cell size; meningitis; pleomorphism; yeasts.

FIG 1

Genetic diversity and relatedness among clinical isolates belonging to each genotype. Minimum spanning network analysis of the clinical isolates based on the concatenated sequences of seven MLST loci. Each unique sequence type is represented by a circle, with the size of the circle proportional to the number of isolates belonging to that sequence type. Isolate names in black type correspond to patients who were alive at the time of the analysis, while isolate names in white type on darker backgrounds correspond to patients who died. C. neoformans VNI (n = 17), VNII (n = 2), VNBI (n = 25), VNBII (n = 9), C. gattii VGI (n = 1), and C. tetragatiii VGIV (n = 16) are shown. MLST alleles for each strain can be found in Table S1 in the supplemental material.

FIG 2

Capsule thickness and yeast cell diameter vary among the clinical isolates following growth under capsule-inducing conditions. (A and B) Capsule thickness (A) and yeast cell diameter (B) of clinical isolates from each genotype grown on DMEM with 5% CO₂ at 37°C for 5 days. Each data point represents the average of 100 cells measured for a single isolate. Error bars show the means ± 95% confidence interval. C. neoformans (Cn) VNI (n = 17), VNII (n = 2), VNBI (n = 25), VNBII (n = 9), C. gattii VGI (n = 1) (dark purple), and C. tetragattii (Ct) VGIV (n = 16) (light purple) are shown. (C) Indian ink preparations of representative strains from each genotype showing variation in capsule and cell size. Scale bar = 30 µm.

FIG 3

Morphological variants produced by clinical isolates following growth under capsule-inducing conditions. (A to C) Indian ink preparations, capsule and cell size associations, and species associations of morphologically variant phenotypes including giant cells > 15 µm (A), micro cells < 1 µm (B), and shed capsule (C). Scale bars = 20 µm. ns, not significantly different. Error bars show the mean ± 95% confidence interval. (D and E) Indian ink preparations were further stained with DAPI (D) or calcofluor white (E) to examine nuclei and cell walls, respectively. In isolates that were scored as having shed capsule, this was occasionally present as an irregular cluster around cells (Clus) as well as released into the media in small clumps (Rel). Scale bars = 15 µm.

FIG 4

Correlation analysis of clinical and phenotypic variables. (A to C) Correlation plots showing the strength and direction of associations between clinical variables (A), phenotypic variables (B), and between clinical and phenotypic variables (C) for all isolates in the data set. The size of the circle corresponds to the strength of the association, and statistically significant associations (P < 0.05) are highlighted in green. The P values for each correlation including species breakdowns can be found in Table S2 in the supplemental material. (D and E) PCA biplots of the first two significant dimensions obtained using principal-component analysis of clinical data accounting for 25.4% of variation in the data set (D) and phenotypic data accounting for 61.9% of variation in the data set (E). Large diamonds represent genotype averages, the length and opacity of arrows represent the degree of contribution of that variable to the model, and genotype ellipses represent the 80% confidence interval.

FIG 5

A summary model of the phenotypic variants seen in this study and the direction of their associations with the C. tetragattii or C. neoformans species and clinical variables. “Large” phenotypes include larger yeast cell size, larger capsule size, and giant cells; these are prevalent in C. tetragattii and are generally correlated with symptoms indicating a less suppressed immune system and low intracranial pressure. “Small” phenotypes include micro cells and shed capsule; these are prevalent in C. neoformans and are generally correlated with symptoms indicating a more suppressed immune system, high intracranial pressure, and low inflammation. Overall, a higher capacity for variation may play a role in increased virulence in Cryptococcus.