Generational distribution of a Candida glabrata population: Resilient old cells prevail, while younger cells dominate in the vulnerable host

Abstract

Similar to other yeasts, the human pathogen Candida glabrata ages when it undergoes asymmetric, finite cell divisions, which determines its replicative lifespan. We sought to investigate if and how aging changes resilience of C. glabrata populations in the host environment. Our data demonstrate that old C. glabrata are more resistant to hydrogen peroxide and neutrophil killing, whereas young cells adhere better to epithelial cell layers. Consequently, virulence of old compared to younger C. glabrata cells is enhanced in the Galleria mellonella infection model. Electron microscopy images of old C. glabrata cells indicate a marked increase in cell wall thickness. Comparison of transcriptomes of old and young C. glabrata cells reveals differential regulation of ergosterol and Hog pathway associated genes as well as adhesion proteins, and suggests that aging is accompanied by remodeling of the fungal cell wall. Biochemical analysis supports this conclusion as older cells exhibit a qualitatively different lipid composition, leading to the observed increased emergence of fluconazole resistance when grown in the presence of fluconazole selection pressure. Older C. glabrata cells accumulate during murine and human infection, which is statistically unlikely without very strong selection. Therefore, we tested the hypothesis that neutrophils constitute the predominant selection pressure in vivo. When we altered experimentally the selection pressure by antibody-mediated removal of neutrophils, we observed a significantly younger pathogen population in mice. Mathematical modeling confirmed that differential selection of older cells is sufficient to cause the observed demographic shift in the fungal population. Hence our data support the concept that pathogenesis is affected by the generational age distribution of the infecting C. glabrata population in a host. We conclude that replicative aging constitutes an emerging trait, which is selected by the host and may even play an unanticipated role in the transition from a commensal to a pathogen state.

Fig 1. Phenotypic characterization of RLS in C. glabrata.

(A) A box-and-whiskers plot demonstrates significant variability in RLSs of C. glabrata strains, shown as box plots with percentiles (n = 20 cells per experiment and 2–6 experiments compared by Log-Rank test). (B) Calcofluor stain reliably identifies generational age of C. glabrata (n = 100 cells). (C) Distinct doubling times during different lifespan phases, and (D) various cell morphologies were captured at time of death. (E) shows thicker cell wall in older BG2 cells by TEM (n = 100 cells compared by Student’s t-test). (F) C. glabrata demonstrated variable virulence in G. mellonella (n = 20 worms compared by Log-Rank test).

Fig 2. Enhanced resilience of older C. glabrata cells.

(A) Older cells contributed to increased virulence in Galleria (n = 20 worms compared by Log-Rank test). (B) CFU counts continuously increased after Galleria mellonella infection with old BG2 cells (10 generations). Young BG2 cells (0–3 generations) were cleared before establishing infection within the first few hours post infection (n = 25 worms per time point). (C) Resistance to neutrophil-mediated killing was observed in older cells (experiments were run in duplicates, with six replicates, and compared by Student’s t-test). (D and E) NET induction was higher by old (14 and 28 generations) compared to young C. glabrata cells, but still significantly lower compared to that seen with hyphal C. albicans. % NETs indicates neutrophil nuclei >1,000 μm² over total neutrophils, and upper panel in D shows sytox staining of nuclei (experiments were run in duplicates, with six replicates, and compared by Student’s t-test). (F) Neutrophil Elastase nuclear localization was higher in old (14 and 28 generations) compared to young C. glabrata cells and unstimulated neutrophils (10 neutrophils were analyzed per condition and compared by one-way ANOVA) (G) H₂O₂ disc diffusion assay shows smaller zone of inhibition for both 14 and 28 generation old cells (experiments were run in triplicates and compared by Student’s t-test). (H) Epithelial cell adhesion assays demonstrate decreased adhesion of older cells (experiments were run in triplicates and compared by Student’s t-test). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig 3. Switch variants have altered lifespan.

(A) Phenotypic switching in BG2 from S to DB colony morphology (inset) increased consistently with aging to 17 fold (experiments were run in duplicates). (B) RLS of DB was shortened over 50% relative to S and reconstituted in the revertant colony (n = 20 cells over two experiments; DB v. S RLS compared by Wilcoxon Rank Sum Test). ***P < 0.001.

Fig 4. Enhanced resilience of older C. glabrata cells stems from cell wall remodeling.

(A) C. glabrata cells aged under sub-therapeutic fluconazole were more resistant to growth inhibition than younger cells when subjected to various concentrations of fluconazole for 4 h (experiments were run in triplicates and compared by Student’s t-test). (B) Older C. glabrata cells contain a slightly higher quantity of β1,3-glucans, but comparable amounts of β1,6-glucans and chitin (experiments were run in triplicates and compared by Student’s t-test). (C) Lipid analysis by GC-MS indicated that younger C. glabrata cells contain a higher amount of ergosterol compared to 14-generation-old cells (experiment was run in triplicates). *P < 0.05, **P < 0.01, ***P < 0.001.

Fig 5. Older C. glabrata cells accumulate in vivo.

(A) Calcofluor stain shows BG2 cells with increased budscars in mouse kidneys (inset). Significantly more BG2 cells with high budscar counts (mean budscars represented by line through plot) were found at day 2 and 4 in kidneys of WT compared to neutropenic mice injected intravenously, and also compared to day 0 in either host. Higher budscar counts were also found in C. glabrata from candiduric patient when compared to the same strain grown in vitro (#42). (B) BG2 cells were larger at days 2 and 4 in kidneys of WT mice compared to neutropenic mice and inoculum (day 0). (C) Model of expansion or contraction of a fungal cell population with five age classes: 0–2, 3–5, 6–8, 9–11 and 12–16 replications, whose populations are given by N₀, N₁, N₂, N₃ and N₄, respectively. r₀…r₃ are replication rates, and m₀…m₄ are mortality rates of each age class. (D) Theoretical model can reproduce experimental results and confirm differential selection hypothesis. Upper panel: Probability distributions of replicative ages found experimentally (shaded areas) overlaid with results from the model parametrized with the most optimal parameters found using optimization (lines). Lower panel: Mortality rate profiles that yielded the best fits of the model to the data for each condition. Precise mortality rates were m = [6.9 3.0 2.8 4.1 4.0] (WT host); m = [2.8 2.3 9.3 2.9 2.8] (neutropenic host); m = [0.0, 10.4, 10.4, 0.0, 0.5] (in vitro).