OUP user menu

Candida glabrata and Candida albicans; dissimilar tissue tropism and infectivity in a gnotobiotic model of mucosal candidiasis

Caroline Westwater, David A. Schofield, Peter J. Nicholas, Emily E. Paulling, Edward Balish
DOI: http://dx.doi.org/10.1111/j.1574-695X.2007.00287.x 134-139 First published online: 1 October 2007


Germ-free transgenic epsilon 26 (Tgɛ26) mice, deficient in both natural killer (NK)- and T-cells, were inoculated (orally) with each of two Candida glabrata (BG2 or BG1003) or Candida albicans (CAF2-1 or SC5314) strains. Candida glabrata- or C. albicans-colonized mice exhibited similar numbers of viable Candida in the alimentary tract. Neither C. glabrata nor C. albicans caused systemic candidiasis of endogenous (alimentary tract) origin. Candida albicans invaded oroesophageal (tongue, palate, esophagus) and keratinized gastric tissues, evoked hyperkeratosis and a prominent, chronic, granulocyte-dominated, inflammatory response in all infected tissues, stimulated the production of splenic granulocytes and was lethal for the mice within 3–5 weeks after oral colonization. The two C. glabrata strains colonized the alimentary tract and penetrated into the keratinized (cardia-antrum) gastric tissues, but in contrast to C. albicans, were unable to infect oroesophageal tissues. Furthermore, C. glabrata strains were not lethal for the Tgɛ26 mice, and did not evoke an inflammatory response in colonized gastric tissues or stimulate the production of splenic granulocytes. This ‘stealth-like’ behavior could explain the ability of C. glabrata to persist in infected tissues and survive as a commensal in the alimentary tract.

  • candidiasis
  • animal models
  • immunodeficient
  • Candida albicans
  • Candida glabrata
  • mucosal infection


Candida species are the fourth most common cause of hospital-acquired infections in the United States (Edmond et al., 1999). Although Candida albicans is responsible for the majority of infections, Candida glabrata is ranked as the most common non-albicans Candida species (NACS) (Vazquez et al., 1999; Tortorano et al., 2004). C. glabrata has emerged as an important etiologic agent of mucosal, bloodstream and urinary tract infections, due in large part to its intrinsic and acquired resistance to azoles and other commonly used antifungal agents (Pappas et al., 2004; Pfaller & Diekema, 2004). There is an urgent need for animal models that not only mimic the clinical conditions associated with these opportunistic infections but also facilitate genomic and proteomic approaches to understanding the pathogenesis of C. glabrata infections and facilitate the development and evaluation of innovative therapeutic options.

Except for vaginal (Fidel et al., 1996) or bladder (Domergue et al., 2005) infections in rodents, animal models of mucosal C. glabrata infections are few in number. Furthermore, studies examining C. glabrata-mucosal interactions in the alimentary tract are especially lacking. To date most studies have focused on animal models of acute systemic infection in immunosuppressed (cyclophosphamide or 5-fluorouracil) mice (Kamran et al., 2004; Olson et al., 2005). Such animal models bypass the alimentary tract, the natural portal of entry for most Candida infections (Nucci & Anaissie, 2001). Furthermore, immunosuppressive treatments cause a number of poorly defined innate and acquired immune defects and mucosal ulcerations that enhance the hosts' susceptibility to opportunistic bacterial and fungal infections (Walsh et al., 1994). To date, very little is known about the specific virulence mechanism(s) employed by NACS and the specific defense components needed to protect the host from C. glabrata-associated infections. Herein we contrast the capacity of C. glabrata and C. albicans to adhere to, colonize and infect mucosal tissues in the alimentary tract of gnotobiotic mice that have combined deficiencies in both NK-cells and T-cells. To the best of our knowledge this is the first study to delineate the host-C. glabrata interaction at mucosal surfaces in the alimentary tract.

Materials and methods


Candida glabrata strains were kindly provided by Dr Brendan P. Cormack (Johns Hopkins University, Baltimore, MD). Candida glabrata strain BG2 is a wild-type clinical isolate. Candida glabrata BG1003 (EPA6::GFP sir3Δ) is deleted for SIR3 and has the EPA6 ORF replaced with the gene-encoding green fluorescent protein (Domergue et al., 2005). Candida glabrata sir3 mutants exhibit increased expression of EPA adhesin genes, have an increased ability to form a biofilm, and are hyperadherent to mammalian cells in vitro and in vivo (Castano et al., 2005; Domergue et al., 2005; Iraqui et al., 2005). Candida albicans strains SC5314 and CAF2-1 were obtained from Dr Joseph Dolan (Nashville State Community College, Nashville, TN). Candida albicans SC5314 (URA3/URA3) is the wild-type parent of the genetically marked strain CAF2-1 (URA3/ura3Δ::imm434) (Fonzi & Irwin, 1993). Candida strains were grown on Sabouraud Dextrose Agar or Broth (SDA or SDB, respectively).


Transgenic epsilon 26 (Tgɛ26) mice were derived into the germ-free (GF) state at the University of Wisconsin Gnotobiotic Laboratory (Madison, WI) and bred at the Medical University of South Carolina Gnotobiotic Facility (http://www.musc.edu/gnotobiotic). The Tgɛ26 mice were originally generated by overexpressing the full-length human CD3ɛ gene in C57BL/6 × CBA/J mice and are defective in both NK-cell and T-cell functions (Wang et al., 1994). All animal procedures were approved by the Institutional Animal Care and Use Committee and followed the guidelines of the American Veterinary Medical Association.

Oroesophageal and gastric candidiasis

The alimentary tracts of GF Tgɛ26 mice were colonized with a pure culture of Candida by oral inoculation (Balish et al., 2001). The cell inoculum was prepared from a 24-h culture (SDB at 37°C), which had been washed twice in phosphate buffered saline, and counted with a hemocytometer. GF mice were provided with a drinking bottle containing a standardized Candida cell suspension (40 mL, 106 cells mL−1 in phosphate buffered saline) for 4 h. The inoculum source was subsequently removed and the animals were given sterile drinking water for the remainder of the study. Microscopic examination and culturing of fecal pellets showed that the GF mice became heavily colonized with yeasts (C. glabrata) or with yeast and hyphal forms (C. albicans) within 24 h after oral inoculation. The mice remained chronically colonized with a pure culture of each Candida species for the duration of the 8-week study. The number of CFU in the stomach contents, ceca and internal organs (kidney, spleen, liver) of euthanized mice was determined according to Balish (2001). Data are presented as the number of viable Candida CFU (g dry wt tissue or alimentary tract contents)−1.


To assess for the presence of candidiasis, tissues were also fixed, paraffin embedded, sectioned (5 µm), and stained with periodic acid-Schiff reagent to visualize fungi, and counterstained with hematoxylin for characterization of host cells (Balish et al., 2005). Three sections for each tissue harvested from at least three mice were assessed for candidiasis.

Spleen cell isolation and phenotype analysis

Spleens were excised aseptically from GF or Candida-colonized mice and placed into 5 mL of RPMI+ (RPMI 1640 medium supplemented with 10% heat inactivated fetal calf serum, 250 U penicillin, 250 µg streptomycin, and 200 mM l-glutamine). The spleens were disrupted with a Stomacher II homogenizer for 120 s on a setting of ‘high’. After allowing the debris to settle, the tissue was carefully layered over 3 mL Histopaque #1083 (Gibco, Carlsbad, CA), centrifuged (20 min, 450 g), and the buffy coats collected. Spleen cells were washed two times (10 min, 350 g) with RPMI+ and adjusted to 1 × 106 cells mL−1. Phenotyping of GR1 granulocytes was assessed by fluorescent activated cell sorting analysis (>5000 events) using a FITC-labeled GR1 monoclonal antibody according to the manufacturer's instructions (BD Pharmingen, San Diego, CA).

Statistical analysis

All data were subjected to statistical analysis using Sigma Stat version 2.0 (SPSS Science, Chicago, IL). P-values were calculated by the Student' t-test and the Mann-Whitney Rank Sum test. P-values of <0.05 were considered significant.


Alimentary tract colonization

Pure cultures of C. glabrata (BG2, a clinical isolate or BG1003, an epa6 sir3 mutant) and C. albicans (CAF2-1 or SC5314) were able to chronically colonize the alimentary tract of the Tgɛ26 mice equally well. Viable colony counts in specimens harvested weekly from the alimentary tract and feces of C. glabrata- or C. albicans-colonized mice ranged between 105–106 for stomach contents, 107–108 for cecal contents, and 106–107 for feces (CFU [g dry wt tissue or contents]−1).

Histopathology and lethality

Candida albicans caused severe infections of the oral cavity (palate, tongue and esophagus) and keratinized gastric tissues in the NK- and T-cell deficient Tgɛ26 mice (Table 1). Mice apparently succumbed to oroesophageal and gastric candidiasis as no progressive systemic candidiasis was detected (culture and histology) in spleen, liver, kidney or lungs (data not shown). The C. albicans-colonized Tgɛ26 mice became moribund and had to be euthanized within 3 to 5 weeks after oral inoculation (Fig. 1). Both yeast and hyphae were evident in oroesophageal (tongue, palate, esophagus) and keratinized gastric tissues (Fig. 2a, data not shown); however, C. albicans infection (hyphal or yeast invasion) of the glandular portion of the murine stomachs, large and small intestines, or ceca was not observed (data not shown). A heavy infiltration of granulocytes was apparent in all C. albicans infected tissues (Fig. 2a).

View this table:
Table 1

Susceptibility of Tg 26 mice to oroesophageal and gastric candidiasis

Tissues assessed for candidiasis
Candida strainTimeStomachEsophagusTonguePalate
GF560/6 (0)0/6 (0)0/6 (0)0/6 (0)
SC531414–2810/19 (53)9/18 (50)11/18 (61)10/18 (56)
CAF2-114–2113/14 (93)12/14 (86)14/14 (100)8/11 (73)
BG214–4210/10 (100)0/10 (0)0/10 (0)0/10 (0)
BG100314–429/9 (100)0/9 (0)0/9 (0)0/9 (0)
  • Candida albicans strains SC5314 and CAF2-1 and C. glabrata strains BG2 and BG1003.

  • Days after oral inoculation; germ-free (GF) mice were 8 weeks of age when assayed.

  • Number of tissues infected/number of tissues examined by histopathology (percentage of tissues with candidiasis, yeast and/or hyphae).

  • Mice were moribund 3–5 weeks after oral inoculation.

  • Mice were moribund 3 weeks after oral inoculation.

  • Mice survived the duration of the 8-week study and appeared healthy and normal.

  • Majority of tissues were negative for the presence of C. glabrata. A few scattered yeast cells were detected in ≤2 tissue sections; however, yeast cells were loosely associated with the tissue surface.

Figure 1

Survival of C. albicans- or C. glabrata-colonized Tgɛ26 mice. Tgɛ26 mice (n≥9) were orally inoculated with C. albicans (CAF2-1, circle or SC5314, filled square) or C. glabrata (BG2 or BG1003, unfilled square). The survival time for mice colonized with C. albicans was 3 and 5 weeks post-colonization for CAF2-1 and SC5314, respectively. All mice challenged with C. glabrata (BG2 or BG1003) survived the 8-week study. One group of mice (solid triangle, n=9) were colonized via oral inoculation with C. glabrata BG1003 for 1 week and then challenged (orally) with C. albicans CAF2-1. Prior colonization with C. glabrata did not protect the Tgɛ26 mice from lethality after oral challenge with CAF2-1.

Figure 2

Histopathology of Candida-infected stomachs harvested from Tgɛ26 mice 2 weeks after oral association. Mice were challenged with C. albicans (CAF2-1; a) or C. glabrata (BG2 or BG1003; b and c, respectively). The representative tissues shown were fixed in buffered formalin and stained using a standard periodic acid-Schiff reagent and counterstained with hematoxylin. Candida yeast cells (arrows), C. albicans hyphal cells (black arrowheads), and inflammatory cell infiltrates (yellow arrowheads) are indicated. Magnification × 400.

Candida glabrata was able to colonize and invade the keratinized gastric tissue but, in stark contrast to C. albicans, was severely curtailed in its ability to infect oroesophageal tissue (Table 1). Yeast cells were observed layering (resembling a biofilm) and penetrating the keratinized, but not the glandular, tissues of the gastric epithelium (Fig. 2b and c). Neither C. glabrata strain penetrated other mucosal tissues in the alimentary tract such as the tongue, palate, esophagus, small and large intestines or ceca (data not shown). Most interesting is that no inflammatory response was evoked in infected keratinized gastric tissues (Fig. 2b and c) and neither C. glabrata strain was lethal for the Tgɛ26 mice (Fig. 1).

Host response to C. glabrata and C. albicans

In contrast to C. albicans, C. glabrata failed to evoke a granulocyte-dominated inflammatory response in the infected keratinized gastric tissues. We investigated the capacity of C. glabrata to stimulate the hosts' splenic immune cells (GR1+ granulocytes) after alimentary tract colonization. Colonization with BG1003 not only failed to evoke an inflammatory response in keratinized gastric tissues, but also did not increase the percentage of granulocytes (GR1+ cells) in the spleen (Table 2). Conversely, colonization and infection (oroesophageal and keratinized gastric tissues) with C. albicans (SC5314) significantly increased (P≤0.001) the number of GR1+ spleen cells in comparison to GF controls (Table 2).

View this table:
Table 2

Effect of alimentary tract colonization with C. albicans or C. glabrata on GR1+ spleen cells

Microbial statusPercentage GR1+ spleen cells
GF11.4 ± 0.7 (18)
SC531420.7 ± 1.2 (27)
BG10038.3 ± 2.2 (7)
  • Mean ± SEM (number of mice) assayed.

  • Significant increase compared to germ-free (GF) mice (P≤0.001).

  • Significant decrease compared to SC5314-associated mice (P≤0.001).

  • No significant difference compared to GF controls (P=0.380).

Inability of C. glabrata to inhibit lethal oroesophageal candidiasis

Candida glabrata yeast cells colonized the alimentary tract, adhered to, and penetrated, keratinized gastric tissues that are also a favored site for infection by C. albicans (Balish et al., 2001). We assessed whether prior colonization with C. glabrata would interfere with the capacity of C. albicans to infect tissues within the alimentary tract and cause lethal oroesophageal candidiasis. Tgɛ26 mice were colonized with C. glabrata (BG1003) and subsequently challenged with a dose of C. albicans (CAF2-1, oral inoculation 106 cells mL−1) that had previously been shown to produce a lethal infection in GF Tgɛ26 mice. As expected, mice became quickly colonized with C. glabrata (data not shown); however, upon exposure to C. albicans the oroesophageal and keratinized gastric tissues became infected and chronic inflammation became evident (data not shown). Three weeks after oral inoculation with CAF2-1 the mice became moribund and had to be euthanized (Fig. 1). Therefore, prior colonization with C. glabrata did not protect mice from a subsequent challenge with C. albicans.


Our data demonstrated that each of the C. glabrata and C. albicans strains, in the absence of intestinal bacteria, were able to colonize the murine alimentary tract, and adhere to and penetrate keratinized gastric tissues; however, both of the C. glabrata strains had little capacity, compared to C. albicans, to colonize and infect oroesophageal (tongue, palate, esophagus) tissues. Importantly, colonization of the alimentary tract with each of the C. glabrata strains was not lethal (8-week study). In contrast, the C. albicans oroesophageal and gastric infections were lethal for the immunodeficient Tgɛ26 mice within 3–5 weeks after oral inoculation; death was most likely due to occlusion of the esophagus (Balish et al., 2001; Sundstrom et al., 2002). Importantly, progressive systemic candidiasis of endogenous origin was not apparent in mice infected with any C. albicans or C. glabrata strain tested. These data emphasize that additional factors other than a combined deficiency in NK- and T-cells must be required for C. albicans or C. glabrata to cause systemic candidiasis of endogenous origin; functional granulocytes and macrophages are likely protective factors (Vazquez-Torres & Balish, 1997; Balish et al., 2001; Balish et al., 2005). Candida glabrata is haploid, aerobic, monomorphic and auxotrophic for nicotinic acid, pyridoxine and thiamine (Fidel et al., 1999; Domergue et al., 2005). It does not possess many of the putative virulence factors (e.g., dimorphism) associated with C. albicans pathogenesis (Fidel et al., 1999). In spite of their deficits in, or lack of, C. albicans-like virulence factors, the two C. glabrata strains were able to adhere to, colonize and penetrate keratinized gastric tissues, but not the secretory section of the murine stomach. We observed little evidence of oroesophageal (palate, tongue, esophagus) colonization or infection (tissue penetration) in the C. glabrata-colonized mice. The reason for this difference in tissue tropism is unknown at this time, but C. glabrata may require commensal bacteria as receptor sites for adhesion or oroesophageal tissue-specific adhesins may be absent in this Candida species. Alternatively, additional immune deficiencies such as the inability to produce antimicrobial peptides may be required for C. glabrata (but not C. albicans) to infect oroesophageal tissues.

The inability of C. glabrata to evoke an inflammatory response in keratinized gastric tissues is of importance as we also observed that colonization with C. glabrata did not increase the number of GR1+ cells in the spleen. In contrast, colonization and infection with C. albicans evoked a prominent, granulocyte-dominated, inflammatory response in all infected tissues and stimulated the production of GR1+ cells in the spleen. The apparent inability of C. glabrata to stimulate splenic granulocytes could be the basis for the lack of an inflammatory response in colonized and infected tissues. Thus, C. glabrata appears to possess a unique virulence mechanism(s), i.e. non-stimulation of granulocytes in the spleen, and an apparent inability to evoke, or a capacity to suppress, an inflammatory response in colonized and infected keratinized gastric tissues. Such inflammation-deficient activity could explain why C. glabrata is able to colonize and persist in the alimentary tract and cause chronic infections after parenteral (IV or IP) challenge (Fidel et al., 1999; Brieland et al., 2001). Candida glabrata does not evoke a strong inflammatory response in patients (Fidel et al., 1999) and animal models of systemic and vaginal candidiasis have revealed minimal inflammatory cell infiltrate in C. glabrata-infected tissue (Fidel et al., 1996; Brieland et al., 2001; Kamran et al., 2004). The diminished inflammation could be attributed to the suppressed immune responses and/or neutropenia in the debilitated patients, and the administration of immunosuppressive agents, which enhance the susceptibility of animal models to acute systemic and vaginal C. glabrata infections. The lack of an inflammatory response in the C. glabrata-associated gnotobiotic, transgenic Tgɛ26 mice demonstrates that these anti-inflammatory effects can now be attributed to C. glabrata as no antibiotics or immunosuppressive agents are used in this model of infection.

In summary, C. albicans was found to be adherent, invasive, inflammatory and lethal for colonized (alimentary tract) T-cell and NK-cell deficient Tgɛ26 mice. Conversely, C. glabrata, although able to chronically colonize the alimentary tract and penetrate primarily keratinized gastric tissue, did not evoke a host granulocytic inflammatory response. Both C. glabrata strains were less able to colonize and penetrate oroesophageal (tongue, palate, esophagus) tissues and were unable to cause lethal systemic candidiasis of endogenous origin. This model of C. glabrata-host interactions not only provides new opportunities to study the pathogenesis and host response to C. glabrata infections, but will also expedite the evaluation of new and innovative antimycotic therapeutic strategies.


This research was supported by a grant from the National Institutes of Health (DE-13968).


  • Editor: Alex van Belkum


View Abstract