Systems biology of Host-Candida interactions: understanding how we shape each other.

Abstract

Candida albicans is both a member of the human mucosal microbiota and a common agent of invasive fungal disease. Systems biology approaches allow for analysis of the interactions between this fungus and its mammalian host. Framing these studies by considering how C. albicans and its host construct the niche the other occupies provides insight into how these interactions shape the ecosystems, behavior, and evolution of each organism. Here, we discuss recent work on multiscale systems biology approaches for examining C. albicans in relation to the host ecosystem to identify the emergent properties of the interactions and new variables that can be targeted for development of therapeutic strategies.

Introduction

Systems biology is an integrative approach for studying biology that combines information from multiple levels to allow for the identification of emergent properties of the system [1,2]. A challenge of applying systems analysis to understanding biology is that organisms are not closed systems. Iterated interactions between organisms and their environments shape both the traits that organisms express in the short term and their evolutionary trajectories. When the environments that organisms respond to include other individuals or species, the potential for complex ecological and evolutionary feedback loops arises [3].

Microbe-host interactions provide a rich opportunity to study dynamics of eco-evolutionary feedback loops and their consequences for human health and disease. Candida albicans is an example of a eukaryotic microbe found in association with mammalian hosts at multiple body sites, and it is capable of both commensal colonization and invasive disease [4–7]. Fully understanding the complexity of C. albicans biology requires studying the interactions between this microbe and its complex ecosystem.

Niche construction is one framework for analyzing how the interactions among C. albicans, host, and microbiota co-determine the trajectory of the systems they all share. Niche construction is the idea that organisms both inhabit and modify their environments with consequences at ecological and evolutionary timescales [8] (Figure 1 and Table 1). Examples of niche construction include engineering of the abiotic environment, as in Candida alkalinization of extracellular pH during carbon limitation [9] or extrusion of extracellular matrix during biofilm formation [10,11], and biotic interactions, for example, the effects of C. albicans intestinal colonization on mediating immune regulation of the microbiota [4,12–14]. A key feature of niche construction is that changes produced in the environment are non-random and generate systematic selection pressures that lead to adaptation [8]. Systems biology approaches can uncover the diverse relationships between C. albicans and their hosts and provide a framework for analyzing these changing relationships across niches and time (Figure 1 and Table 1). Here we review recent literature elucidating Candida albicans interactions with host ecosystems at multiple scales with special attention to novel methods that will contribute to an understanding of the dynamic interactions, eco-evolutionary feedback loops, and emergent properties at work in these systems.

Figure 1: Schematic of Candida albicans and the host ecosystem.

Niche construction is the idea that organisms systemically remodel their environments by living in them, which can change the selection pressures operating in those environments. C. albicans (represented here by the purple yeast) is a eukaryotic microbe associated with mammalian hosts (large blue circle) that can inhabit many different body sites aka ecological niches in the host. Each of these niches has a characteristic set of abiotic conditions (e.g. pH, humidity, nutrients) and biotic conditions (e.g. microbial flora, immune cells) that can influence C. albicans growth and behavior. In response, C. albicans can shape these niches by altering local chemistry, competing with microbes, driving immune responses, and influencing immune surveillance. The effects of these niche modifications, whether that is the establishment of a persistent multispecies biofilm or training of host immunity, have the potential to outlast the life cycle of any single cell by many generations and generate systematic selection pressures on all the players in the niche. Thus, the interconnected feedback loops between C. albicans and the host demonstrate the importance of studying the interactions of the organism to understand their behavior and evolution.

Table 1: Multi-scale systems biology.

Systems biology approaches can capture a snapshot of the interactions between organisms. By using multi-scale approaches, it is possible to build a fuller picture of the relationships. Each set of tools can uncover different aspects of the interactions.

Timescale	Evolutionary Time	Generation Time	Organismal Time
Unit of measure	Genomes	Communities	Molecules/Cells
Experimental Approaches	Experimental Evolution, Competition Assays, Natural variation	Microbial community composition, Competition assays, Functional Genomic Screens	Transcriptomics, Proteomics, Microscopy

Microbial community dynamics

An integrative understanding of how C. albicans fits into its ecological context takes into account both the host ecosystem and the varied impacts of diverse microbes that C. albicans interacts with in situ [15]. These interactions between microbes can be antagonistic, mutualistic, or neutral, with varying impacts on host health. Better understanding of these interactions, and how these microbial players influence each other and the host, will be helpful in identifying variables that can be selectively perturbed to improve clinical outcomes.

One outcome of microbial interactions is synergism or mutualism between microbes. For example, C. albicans produces biofilms [16] that can offer surfaces for bacterial growth [17] and shelter from antibiotics [10]. Biofilms also form surfaces where cross-feeding interactions take place between Candida and bacteria [11]. In these examples, the host is damaged by microbial synergism. However, synergistic interactions may also offer new targets for clinical intervention when microbial communities depend on one another: targeting one member--the “weakest link”--could interfere with the entire relationship or community [18]. This strategy highlights the importance of understanding interdependent relationships in microbial communities.

Interactions among microbes can also antagonize pathogenicity. Bacterially-produced metabolites and peptides can interfere with C. albicans virulence [19–23]. For example, Enterococcus faecalis EntV peptide inhibits filamentation and biofilm development [19,24]. As E. faecalis and C. albicans co-occur in the GI, this interaction may promote C. albicans commensal growth and limit hyphal invasion [25]. Examining the interaction between C. albicans and additional gut commensal microbes identified Roseburia spp. and Bacteroides ovatus as organisms that can limit C. albicans growth [20]. Other microbial interactions can directly kill fungal cells, such as when Serratia marcescens deploys the Type VI secretion system [23], or Pseudomonas aeruginosa secretes phenazines [26,27]. Competition for resources or host niches may drive this antagonism, though bacterial use of fungal cell wall components as a nutrient [28–31] suggests the potential for parasitic or predatory interactions.

In vivo mouse studies provide an opportunity to understand microbial interaction in the context of a host, however, murine models do not always fully recapitulate human microbial communities. Lab mice do not natively possess C. albicans in their GI tracts and generally require antibiotic depletion or dietary changes to allow for stable fungal colonization [6,32], highlighting the role of lab-adapted microbiota in limiting C. albicans colonization. This colonization resistance has been attributed to Bacteroides regulation of HIF-1 [25,33]. A recent exciting development is that lab mice introduced to naturalistic contexts (“re-wilded”) acquire microbiomes that both include more fungi and induce immune states more closely resembling those of wild mice and humans [13]. Moving towards models like re-wilded mice [13] or more complex models of human tissues in culture with microbiota [22] that more faithfully reproduce the relevant systems will be an important source of new insight.

Transcriptomics

Profiling Candida transcriptomes in different environmental contexts and time points in colonization and infection has offered useful information on the key mechanisms this microbe uses to persist and thrive in diverse niches. Transcriptional analysis of fungi during infection can identify proteins that are specifically used to respond to and manipulate the host, for example the putative effector protein, Sel1, which is sufficient to induce an inflammatory response via signaling through TLR2 and TLR4 [34]. Dual RNA-seq, in which transcripts from the microbe and host are captured simultaneously, permits analysis of the interaction during infection. Dual RNA-seq during a time course of infection showed an upregulation of genes involved in glycolysis in both species with the ability of fungal cells to outcompete macrophages for glucose resulting in macrophage starvation and death [35]. Previous work on C. albicans growth in macrophages demonstrated the importance of the glyoxylate cycle and metabolic plasticity for virulence [36,37]; this work shows that fungal growth requires this pathway and that the competition impairs host cell survival. To further dissect the interplay between macrophage and C. albicans gene expression during infection, Munoz and colleagues [38] performed dual RNA-seq on host and C. albicans cells separated into subpopulations based on infection stage. Stratifying transcriptional signatures over the course of phagocytosis and infection allowed the authors to describe the trajectory of macrophage and microbe transcriptomes. Although the bulk of Candida-host experiments have been performed in macrophages, neutrophils also play a major role in anti-Candida responses. In neutrophils, the signature of infection is changes in cell signaling pathways [39], with enrichment in expression of genes related to the NLRP3 inflammasome. Current single-cell sequencing approaches to fungal infection have focused on host responses during gastrointestinal colonization [40] or at lymph nodes [41]. Future work using single-cell transcriptional analysis of both the fungus and the host during more realistic models will be vital for understanding the dynamic interactions that drive infection.

Proteomics

Proteomic studies complement transcriptomic studies by providing direct measures of the proteins present in a cell and can be used to measure protein abundance changes upon infection. In concordance with transcriptional data, the proteomic signature upon infection includes an increase in fatty acid beta oxidation and glyoxylate cycle proteins [42,43]. Proteomic analysis also revealed an increase in C. albicans phosphatidylethanolamine synthesis proteins [42] that are required for systemic infections [44]. Potentially, additional processes regulated upon infection can be targeted for novel antifungal development. A limitation of proteomic approaches is that lower abundance proteins are difficult to identify and quantify, although these proteins might be crucial players in mediating the interaction.

In addition to quantifying the proteins present under a particular condition, proteomic studies can be used to identify protein-protein interactions (PPIs), thus giving insight into functional relationships between proteins. One approach is to use affinity-purification mass spectrometry (AP-MS) to identify changes in abundance of an interaction upon an environmental stress, such as treatment with an antifungal drug [45], or during an infection. Another recently-developed approach for identifying protein complexes is to use non-denaturing co-fractionation followed by mass spectrometry to identify proteins that co-purify with each other [45,46]; when this is combined with stable isotope labeling, it is possible to obtain quantitative changes in interactions. Protein-protein interactions can also occur between microbial and host proteins (as reviewed in [47] for bacterial pathogens), and the development of C. albicans proteomic resources will soon allow for analysis of these interactions.

Functional genomics approaches

Functional genomics couples forward and reverse genetics by screening libraries of defined mutants for phenotypes of interest. In vivo functional genomics screens can reveal mutants that only display a phenotype during infection. The pioneering in vivo screen focused on C. albicans mutants with defects in systemic infection and identified that fungal sphingolipids are required during systemic disease but dispensable for in vitro growth, highlighting the utility of this approach [48]. However, hosts contain many microenvironments, and pinpointing which processes are required during a particular interaction can be challenging. This is particularly true for C. albicans, which can occupy many niches within a host and have a variety of relationships with that host.

To answer this challenge, researchers have designed screens to isolate particular components of in vivo fitness. For example, during invasive disease C. albicans must translocate across epithelial barriers. Using epithelial cells as a model of infection, Allert and colleagues screened fungal mutants for defects in translocation; by examining both host and fungus in parallel, they were able to show that translocation is a fungus-driven process [49]. The unbiased screen implicated multiple genes with unknown function in regulating this in vivo phenotype, suggesting new mechanisms of action. Additional models of in vivo infection will allow for a deeper understanding of how specific C. albicans genes contribute to fitness during systemic infection.

Candida albicans is also a colonizer of the gut, and the processes required for colonization are not necessarily the same processes required for causing invasive disease. Therefore, multiple studies have used mutant screens to identify genes required for colonization in the murine gut [50,51]. One surprising result was that ablating expression of several C. albicans genes conferred increased fitness over wild type during gut colonization, suggesting that C. albicans maintains genes whose fitness benefit in another environment trades off with fitness in the murine gut, or, possibly, that these genes function differently in the animal model than in the human GI tract. In deletion mutant screens, multiple yeast-locked mutants were hypercompetitive in the GI tract [32,50]; one hypothesis is that hyphal cells express proteases and other virulence factors that are a liability during colonization [50]. Alternately, bacteria could drive filamentation in the gut: C. albicans adopts a yeast form when colonizing gnotobiotic mice [32] but exists as a mixture of yeast and hyphae in antibiotic-treated mice [50,52]. Notably, ume6 and tec1 deletion mutants show morphogenesis defects during in vitro but not in vivo gut colonization, demonstrating that the signals that drive filamentation in the host are not fully recapitulated by in vitro conditions. Additionally, the GI tract comprises multiple environments potentially requiring differential fungal responses: overexpression of CRZ2 permits higher colonization of the stomach and upper GI presumably due to improved adaptation to acidic pH and bile salts [51]. These studies highlight the importance of multi-scale systems biology approaches that identify host environments and microbiota contributing to changes in fitness among mutant strains.

Experimental evolution

Experimental evolution approaches complement functional genomic experiments by identifying the genetic and phenotypic changes resulting from applying a selection pressure of interest over multiple generations. Selecting on naturally occurring variation in a population or among defined sets of genotypes permits unbiased identification of variants that increase fitness in a particular context. To examine adaptation to host environments, three recent studies described the genetic and phenotypic changes resulting from serial passaging of C. albicans through mice [7,53]. Consistent with the functional screen reported above, Tso and colleagues showed that evolving strains through antibiotic-treated mouse GI tracts for 10 weeks resulted in strains with defects in filamentation on agar and reduced virulence in systemic infection [7]. The authors propose that filamentation is costly but required for competition with gut-resident bacteria; the mechanism by which bacteria select for hyphae is currently unknown. In addition to mutations in transcription factors known to regulate filamentation, they observed multiple instances of loss of heterozygosity and aneuploidy, especially on Chromosome 7. While strains evolved in mice show up to 11x higher mutation rates than those evolved in vitro, a second study also recovered Chromosome 7 trisomies from multiple C. albicans genotypes passaged through mouse GI tracts in the presence of antibiotics. Direct competition assays showed that the trisomic strain exhibited increased fitness in antibiotic-treated mice [54]. Notably, passage of C. albicans through either mouse kidneys or the gut resulted in different suites of genomic changes, suggesting that different host niches may favor different evolved strategies and/or that different host environments could bias towards certain mutational possibilities. Consistent with this hypothesis, strains passaged through an oropharyngeal candidiasis model acquired trisomies on Chromosome 5 and 6 [53]. One striking observation is that global isolates of C. albicans collected from humans show less frequent aneuploidy than the experimentally-evolved lines [55], suggesting that selection over longer timescales may refine large effect mutations.

Microscopy-based approaches

Systems biology does not necessarily mean ‘omics approaches; instead, the goal is to use techniques where a range of properties can be examined to determine which will be the most informative. One such approach is detailed microscopy, which can reveal new aspects of the C. albicans-host interactions. A benefit of microscopy approaches is the single-cell resolution of features that are lost during bulk culture approaches, including confounding factors such as variation in phagocytosis rates, changes in morphology, or stochasticity in responses.

Recent work demonstrated that C. albicans interactions with host macrophages can result in a pyroptotic host cell death [56–59]. Time-lapse microscopy revealed that there are two stages of host cell death after phagocytosis with macrophages [57,60]; the first is NLRP3-dependent, and the second phase results from competition for glucose [35]. These microscopy-based approaches also revealed that hyphal formation is not sufficient for the induction of pyroptosis [58,60–62]. Without analyzing both the fungus and the host at single-interaction levels, it would not be possible to decouple the hyphal transition from the induction of host death. However, the transition from yeast to filamentous form within macrophages can drive additional processes that are linked with host cell death, such as the production of the cytolytic toxic peptide Candidalyin [63] or physical pressure driven by the elongating hyphae causing phagosomal membrane damage [64]. Microscopy approaches can also be coupled with functional genomic screens, such as with screens for mutants with defects in pyroptosis. Large-scale screens of a filamentous S. cerevisiae library or a tetracycline-repressible C. albicans library demonstrated that components of the fungal cell surface, such as ergosterol or fungal cell wall proteins, are required to drive the host cell death response [62,65,66]. The exposure of these components is regulated by interaction with host cells [66], highlighting the dynamic interactions between microbe and macrophage.

Recent work has used tissue clearing and light-sheet microscopy approaches for examining in vivo fungal infections and the influence of specific immune defects on fungal infections [67,68]. In Aspergillus fumigatus infections of murine lungs, these techniques showed that most fungal conidia did not reach the alveoli and were instead concentrated in the bronchioles [68], altering our understanding of which host cells and processes are important to these infections. Using this approach to examine C. albicans colonization and infection in situ at the whole-organ level will reveal new aspects of fungal interactions with the host.

Conclusions

Although it can cause life-threatening systemic infections, C. albicans is also an abundant fungal species colonizing mammalian mucosal surfaces. As part of this colonization, C. albicans may play a major role in continuously shaping immune surveillance and the host microbiome throughout the lifetime of the host. There is not a singular host-pathogen conflict, but multiple interfaces and interactions between these organisms at multiple levels. As the studies discussed above show, the development of technological advances allows for a deeper analysis of the eco-evolutionary feedback loops between Candida albicans and its host ecosystems. By studying both organisms in parallel during their interactions, it is possible to identify how they shape each other, both in the short term and in their evolutionary trajectories. Further studies to build more complex and representational models are an exciting direction for future research.

Highlights.

• Interactions between Candida albicans and human hosts shape each other’s biology.

• Systems biology provides tools to interrogate these interactions at multiple scales.

• Microscopy, transcriptomics, and proteomics illustrate individual cell responses.

• Functional genomics screens and experimental evolution reveal genotypes that impact fitness.

• Microbial community interactions can modify each other and host health.