-
ORAL COLONIZATION BY CANDIDA ALBICANSR.D. Cannonl*W.L.
Chaffin2'Department of Oral Sciences and Orthodontics, Faculty of
Dentistry, University of Otago, PO Box 647, Dunedin, New Zealand;
2Department of Microbiology and Immunology, Texas TechUniversity
Health Sciences Center, Lubbock, Texas, USA; *corresponding
author
ABSTRACT: Candida albicans is a commensal yeast normally present
in small numbers in the oral flora of a large proportion ofhumans.
Colonization of the oral cavity by C. albicans involves the
acquisition and maintenance of a stable yeast
population.Micro-organisms are continually being removed from the
oral cavity by host clearance mechanisms, and so, in order to
surviveand inhabit this eco-system, C. albicans cells have to
adhere and replicate. The oral cavity presents many niches for C.
albicanscolonization, and the yeast is able to adhere to a plethora
of ligands. These include epithelial and bacterial cell-surface
mole-cules, extracellular matrix proteins, and dental acrylic. In
addition, saliva molecules, including basic proline-rich
proteins,adsorbed to many oral surfaces promote C. albicans
adherence. Several adhesins present in the C. albicans cell wall
have nowbeen partially characterized. Adherence involves lectin,
protein-protein, and hydrophobic interactions. As C. albicans cells
evadehost defenses and colonize new environments by penetrating
tissues, they are exposed to new adherence receptors andrespond by
expressing alternative adhesins. The relatively small number of
commensal Candida cells in the oral flora raises thepossibility
that strategies can be devised to prevent oral colonization and
infection. However, the variety of oral niches and thecomplex
adherence mechanisms of the yeast mean that such a goal will remain
elusive until more is known about the contri-bution of each
mechanism to colonization.
Key words. Candida albicans, colonization, adherence,
candidiasis.
(1) IntroductionThe presence of Candida albicans in the oral
cavity is not
indicative of disease. In many individuals, C. albicans isa
minor component of their oral flora, and they have noclinical
symptoms. In certain sections of the population,however, oral
candidiasis occurs frequently and necessi-tates antifungal therapy.
Oral presentations of candidia-sis vary from the large white
plaques of pseudomembra-neous candidiasis on the tongue and buccal
mucosa tothe palatal erythematous lesions of chronic atrophic
can-didiasis, and to angular cheilitis on the labial commis-sures
(Samaranayake, 1990; Scully et al., 1994; Shay et al.,1997). The
primary etiological agent of oral candidiasis isthe yeast C.
albicans; however, other species that causedisease less commonly
include C. tropicalis, C. glabrata, C.krusei, C parapsilosis, C.
guilliermondii, and C. dubliniensis(Odds, 1988; Fridkin and Jarvis,
1996; Sullivan andColeman, 1998). Sequelae of mucosal colonization,
par-ticularly of the gastrointestinal tract, may include
pene-tration of the vascular system by Candida cells
andhematogenous dissemination (Cole et al., 1996). Thesecells can
then infect a variety of organs in immunocom-promised individuals
and cause disseminated or sys-temic disease.
It is difficult to give a precise oral carriage rate for
C.albicans, since this depends on the age and health of the
population studied. A compilation of data from a num-ber of
reports showed that the mean carriage rate forhealthy individuals
(no known underlying disease) was17.7% (range, 1.9-62.3%), whereas
mean carriage in hos-pitalized individuals (without clinical
candidiasis) was40.6% (range, 6.0-69.6%) (Odds, 1988). These data
indi-cate that the health of an individual is a predisposingfactor
for C. albicans colonization. A large number of sitesin the oral
cavity can be colonized; in healthy individuals,C. albicans is most
commonly isolated from the mid-lineof the middle and posterior
thirds of the tongue, thecheek, or the palatal mucosa (Arendorf and
Walker, 1979,1980; Borromeo et al., 1992).
It is of interest that only a proportion of the popula-tion is
colonized by C. albicans, and only a subset of theseindividuals
develops candidiasis. Few longitudinal stud-ies have been carried
out on healthy individuals to see ifCandida colonization is
continuous. However, daily sam-pling has shown that C. albicans
carriage persisted in aproportion of healthy people and that
colonizationrecurred in a majority of the remaining subjects
(Gergelyand Uri, 1966; Williamson, 1972). In a study of 163neonates
in an intensive care and surgical unit, 21 of theneonates initially
carried C. albicans in their mouths, butonly five yielded 6 or more
yeast-positive cultures overthe 17-week study period (Sharp et al.,
1992). Theseneonates were colonized for periods of between 7 and
63
1013) 359 383(1999)Crit Rev Oral 35910(3)-359-383 (1999) 359Crit
Rev Oral Biol Med
by guest on June 1, 2015 For personal use only. No other uses
without permission.cro.sagepub.comDownloaded from
-
(2) AcquisitionColonization Acquisition (a) Candida species
inhabit a variety of
environments (Odds, 1988). C. albi-
b)cans has been isolated from primates,domesticated and other
mammals,
GWt C. albicans Systemic marsupials, and birds. In contrast,\
(d) Disease other Candida species have been iso-\i' X lated from a
much narrower range ofMucosl hosts (Odds, 1988). In humans, C.
Removal (c) disease Oral cavity Ialbicans preferentially
colonizess Oa mucosal surfaces, and the intestinaltract is believed
to be a major reser-voir for infection (Odds, 1988; Cole etFigure
1. A model showing the interrelationship of factors involved in
colonization of the oral al., 1996). C. albicans can colonize
cavity by C. albicans: (a) acquisition, (b) growth, (c) removal,
and (d) tissue damage and practically any site in the
gastroin-penetration. testinal tract (Cole et al., 1996), from
the oral cavity to the rectum andperi-anal tissues, allowing
anal-oralinoculation to occur (Soll et al., 1991).
days. The C. albicans strains were biotyped, and there was The
vulvovaginal regions of approximately 40% ofunequivocal evidence
for more than one infecting bio- healthy women are colonized by
Candida species (Soll ettype in only 8.1% of colonized neonates. In
immunocom- al., 1991), and the genito-urinary tract presents
anotherpromised hosts, candidiasis is often caused by a resident
reservoir for oral inoculation. C. albicans survives betterstrain
(Powderly et al., 1993; Voss et al., 1994), and the on moist
surfaces than dry inanimate objects, but if thesame strain can
cause recurrent infections (Miyasaki et degree of contamination is
high enough, viable cells willal., 1992). Some of the factors
involved in the balance remain on dry surfaces for at least 24
hours (Rangel-among clearance of C. albicans, colonization, and the
Frausto et al., 1994). Many studies of nosocomial can-development
of candidiasis have been reviewed previ- didiasis in clinical
settings have been carried out toously (Cannon et al., 1995a). The
objective of this review determine how patients acquire infections
(Hunter et al.,is to focus on the initial, critical, step of
colonization, 1990; Vazquez et al., 1993, 1998; Fridkin and Jarvis,
1996).and to discuss the factors involved in colonization and It is
evident that the most common means of transfer ishow current
research might lead to therapeutic interven- contact with carriers,
often the hands of hospital staff,tions that could prevent
colonization and, thus, preclude although various Candida species
can be cultured fromcandidiasis. inanimate objects (Hunter et al.,
1990; Vazquez et al., 1993,
Colonization of the oral cavity by C. albicans can be 1998;
Strausbaugh et al., 1994; Jarvis, 1996; Pfaller, 1996).defined as
the acquisition and maintenance of a stable A worrying finding in
one of these studies was that C.population of C. albicans cells
which does not give rise to albicans could be cultured from the
food given to twoclinical disease. A model based on this definition
is patients in a bone marrow transplant unit (Vazquez et al.,shown
in Fig. 1. Colonization depends on the rate of 1993). Indeed,
yeasts, including Candida species, are rela-acquisition-that is,
the rate at which yeast cells enter tively common contaminants of
both processed andthe oral cavity-growth, and removal of cells from
the unprocessed foods (Buck et al., 1977; Viljoen andmouth by
swallowing and oral hygiene. In a simplified Greyling, 1995). In a
dental setting, the internal surfacesmodel, if the rate of removal
is greater than that of acqui- of dental unit water lines can
become coated with bacte-sition and growth, clearance will take
place. If the rate of ria-rich biofilms (Tippett et al., 1988;
Peters and McGaw,removal is the same as that of acquisition and
growth, 1996). Although there are no reports of yeasts being
pres-then there will be colonization. If the rate is lower and ent
in these biofilms, this may reflect the culturingthere is tissue
damage, it will lead to candidiasis. The methods used. If C.
albicans were a component of suchpresentation of candidiasis will
depend on the tissue biofilms, contaminated water lines could
constitute acolonized, the virulence factors expressed by the
Candida significant risk of inoculating oral cavities with yeast.
Incells, and the host response. So, colonization depends people
whose mouths are colonized with C. albicans, theon several factors:
the acquisition or entry of cells into yeast can be found in saliva
at an average concentrationthe oral cavity, the attachment and
growth of those cells, of 300 to 500 cells per mL (Arendorf and
Walker, 1980).the penetration of tissues, and the removal of cells
from This will allow for transfer during kissing and other
directthe oral cavity. Each of these factors will be examined.
saliva-saliva contact. There are ample opportunities,
360 ritRevOralBio Me lO3):39 33 (999360 Crit Rev Oral Biol Med
10(3):359-383 (1999)
by guest on June 1, 2015 For personal use only. No other uses
without permission.cro.sagepub.comDownloaded from
-
therefore, for the entry of Candidaspecies into the oral cavity
by manualinoculation, saliva transfer, or contami-nated food and
drink.
(3) Maintaining an OralCandida Population
The entry of Candida cells into the oralcavity is not sufficient
for colonization;they must be stably maintained. Sincethe oral
cavity is a continuous-flowenvironment, yeast cells will be
washedout by saliva and swallowed unlessthey adhere and replicate.
Growth con-ditions in the oral cavity are so poor(there is
practically no growth in salivaunless it is supplemented with
glucoselSamaranayake et al., 19861) that cellshave to adhere to be
maintained.Adhesion is therefore of critical impor-tance in
colonization. Adherence ismediated between moieties of theCandida
cell wall and host surfaces, andso an understanding of
colonizationrelies upon knowledge of these sur-faces.
(A) THE C. ALBICANS CELL WALLThe cell wall is essential both to
the biology of C. albicansand to its interactions with the human
host in health anddisease. Although frequently called a dimorphic
fungus,the organism is, in fact, polymorphic and may adoptgrowth
not only in yeast or hyphal modes but also aspseudohyphae and may
produce chlamydospores in cer-tain growth conditions (Odds, 1988).
The initial emer-gence of hyphae from yeast cells is often referred
to asgerm tube formation. While both yeast and hyphae canbe found
in lesions, and different adhesins are expressedon hyphae as
discussed below, hyphal cells clump exten-sively and have been
less-well-studied in adherenceassays than yeast. The cell wall is
the structure responsi-ble for supplying the rigidity that
maintains the uniqueshapes that characterize fungal growth. The
surface ofthe organism is the site of the physical
interactionsbetween the fungus and host proteins and tissues
thatlead to adherence, and between the fungus and theimmune system
that lead to clearance.
The cell wall is composed primarily of carbohydrate(80-90%),
1-glucan, chitin, and mannan (Fig. 2a; for moreextensive
discussion, see reviews by Shepherd, 1987;Cassone, 1989; Fleet,
1991; Fukazawa and Kagaya, 1997;Chaffin et al., 1998). The
components of the cell wallsfrom yeast and hyphal forms are
similar, although thereis some quantitative variation. Chitin (an
unbranched
(a)
(b)
Figure 2. Molecular interactions between the cell wall of C.
albicans and oral surfaces. (a)Schematic representation of the
architecture and composition of the C. albicans cell wall:41M
chitin, 41i49 3(1,3)-glucan, "`..i 13(1,6)-glucan, .L,
mannoprotein, (&phosphodiester linkage, and plasma membrane.
(b) Interactions of C. albicans withmolecules and surfaces in the
oral cavity that may contribute to colonization.
polymer of N-acetylglucosamine) is a minor constituentthat is
variously reported to contribute from 1 to 10% ofthe cell wall's
dry weight. The higher levels are associat-ed with hyphal cells
which are reported to containapproximately three times more chitin
than yeast cells.However, a recent study reports that chitin
measure-ments depend greatly on the method used (Munro et
al.,1998). (-glucan (a branched polymer containing 3-1,3and 3-1,6
linkages) is the main constituent, accountingfor 47 to 60% of the
cell wall's dry weight. These twomicrofibrillar polysaccharides,
while found throughoutthe cell wall, are more concentrated in the
inner portionnear the plasma membrane and provide a rigid
skeleton.The other main component is mannan, also sometimescalled
phosphomannoprotein or phosphopeptidoman-nan complex, which
accounts for about 40% of the cellwall. Mannans are composed of
mannose polymers cova-lently linked to a protein moiety mostly by
N-glycosidiclinkages through di-N-acetylchitobiose to
asparagineresidues. The mannose component consists of a back-bone
of (x-1,6-linked mannose molecules to which areattached
oligosaccharide side-chains containing man-nose residues with
ax-1,2, ox-1,3, 3-1,2, 13-1,6 linkages andsome a-1,6 branches. Some
of these side-chains alsocontain a phosphodiester linkage to short
1-1,2 manno-oligosaccharides. The N-glycosyl moieties of
high-molec-
1013) 359 383 1999) Crit Rev Oral 36110(3) -359-383 (I 999) Crit
Rev Oral Biol Med 361 by guest on June 1, 2015 For personal use
only. No other uses without permission.cro.sagepub.comDownloaded
from
-
TABLE 1C. albicans Adhesins and BEC Ligands
Adhesin Ligand ReferenceCarbohydrate
Chitin Unknown Segal, 1996Factor 6 oligomannosaccharide Unknown
Miyakawa et al., 1992
Protein66-kDa fimbrial protein Glycosphingolipid Yu et al.,
1994aFibronectin binding protein
(multiple candidates) Fibronectin Klotz and Smith, 1992; Gozalbo
et al., 1998;Yan et al., 1998b
iC3b binding protein(multiple candidates) iC3b Eigentler et al.,
1989; Hostetter et al., 1990;
Alaei et al., 1993Fucose binding protein Fucose-containing
oligosaccharide
(1 5-kDa fragment) (blood group antigen?) Cameron and Douglqs,
1996GIcNAc or glucosamine Host oligosaccharide-containing
binding protein (190 kDa) GIcNAc or glucosamine Enache et al.,
1996SAP (secreted aspartyl Unknown (proteolytic modification
proteinase) of host or fungus?) Watts et al., 1 998ALS gene
familyALSI Unknown Fu et al., 1998ALA1 Unknown Gaur and Klotz,
1997Other proteins (38-kDa, 54-kDa
candidate species) Unknown Imbert-Bernard et al., 1995
ular-weight yeast cell mannoproteins average more than
components, 3-1,3-glucan, 3-1,6-glucan, chitin, and600 mannose
residues and those from germ tubes more mannoprotein (Kollar et
al., 1997). The analysis of thisthan 300 residues. In addition,
single mannose residues material suggested that 3-1,6-glucan with
some f-I,3-and short, unbranched manno-oligosaccharides may be
glucan branches may be linked to the reducing end of0-linked to
protein through serine and threonine. chitin. Covalent attachment
of mannoprotein to 3-1,6-Mannoproteins are found throughout the
cell wall and glucan is through a remnant of the mannoprotein
GPIappear to be the dominant component at the cell sur- (glycosyl
phosphatidyl inositol) anchor. However, eachface. Electron
microscopic analysis shows a variable complex may not contain all
four components, and thenumber of cell wall layers (from 3 to 8)
that seems to be proportion of cell wall polysaccharide involved in
thisrelated to the technique used, the strain, and growth type of
structure is unclear.conditions of the fungus (Cassone et al.,
1973; Rico et al.,1991). This layering appears to be the result of
quantita- (B) C. ALBICANS ADHESINStive differences in the
individual components in different Adhesins are the fungal surface
moieties that mediateregions of the wall. Fimbriae, which may
extend 110 to binding of C. albicans to other cells (host or
microbial),300 nm, radiate from the surface (Fig. 2a; Yu et al.,
1994a). inert polymers, or proteins. Different experimentalThe
fimbrial subunit appears to be a highly glycosylated approaches and
reagents have been used to identify C.glycoprotein with an apparent
molecular mass of 66 kDa albicans adhesins (also called binding
proteins or recep-(Yu et al., 1994a). tors) and host ligands
(sometimes also called receptors).
The architecture of the yeast wall has been studied There is
disagreement among some of these studies asmore extensively in
Saccharomyces cerevisiae, and a number to the identity, number of
candidal receptors for variousof observations suggest that the
candidal cell wall will fit ligands, and the inhibitors of
adherence (reviewedthe same model. In a recent study, material was
isolated recently in Fukazawa and Kagaya, 1997; Sturtevant andfrom
a cell wall digest that contained all of the major wall Calderone,
1997; Chaffin et al., 1998). Our incomplete
362 Crit Rev Oral Biol Med 10(3):359-383 (1999) by guest on June
1, 2015 For personal use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
understanding of the factors that influence the adher-ence
interactions is a likely source of apparently conflict-ing
observations. Following are examples of five factorswith the
potential to affect observations:
(1) Some C. albicans strains are more adherent thanothers
(Schmid et al., 1995b).
(2) Some strains possess adhesins with differentspecificities
(Critchley and Douglas, 1987a,b).
(3) In vitro growth conditions of the fungus-such astemperature
(Lee and King, 1983), medium composition(Alloush et al, 1996; Yan
et al., 1998b), carbon source(McCourtie and Douglas, 1985;
Gustafson et al., 1991), orthe presence of a specific inducer (Yan
et al., 1998a)-may alter the expression of an adhesin.
(4) Fungal cell viability may affect the extent of bind-ing
(Gorman et al., 1986).
(5) The binding capacity of exfoliated human cellsused in
adherence assays may differ among donors andfrom the same donor on
different days (Sandin et al.,1987b) and with hormonal status
(Theaker et al., 1993).
Despite differences between some studies, the gen-eral
conclusions are that C. albicans possesses multipleadhesins and
that there may be more than one adhesinthat recognizes a host
ligand or cell. Most adhesins iden-tified to date are mannoprotein,
and, for individualadhesins, both the protein and/or carbohydrate
portionshave been implicated in adherence.
(C) ADHERENCE TO SKINSkin is a site of normal colonization as
well as infectionssuch as diaper rash and intertriginous
candidiasis.Infections usually occur in individuals with some loss
ofnormal skin defenses such as abrasion and maceration,and yeast
growth is promoted by a warm, moist environ-ment (Samaranayake,
1990; Scully et al., 1994). Systemicconditions such as diabetes,
obesity, and various medicaltreatments may also contribute to
susceptibility to skinand other mucocutaneous infections (Odds,
1988;Samaranayake, 1990). C. albicans cells bind in vitro to
cor-neocytes (keratinized cells of stratum corneum) fromindividuals
in these susceptible groups at twice the fre-quency with which they
bind to corneocytes from healthyindividuals (Srebrnik and Segal,
1990). Amino sugars,mannosamine, glucosamine, and galactosamine
inhibit-ed binding of C. albicans to human corneocytes and tobuccal
epithelial cells (BECs) (Collins-Lech et al., 1984).
Achitin-soluble extract (CSE) also inhibited binding of C.albicans
to human corneocytes (Kahana et al., 1988). C. albi-cans cells
exposed to nikkomycin, a chitin synthetaseinhibitor, had decreased
chitin content and showed a cor-responding decrease in adherence to
BECs (Segal et al.,1997). The site of adherence between the yeast
cells andepithelial cells labeled intensely with wheat germ
agglu-tinin, a lectin-recognizing chitin. In a porcine stratum
corneum model, three isolates from oral infectionsadhered more
than a commensal isolate (Law et al., 1997).Removal of lipid from
the stratum corneum led to dou-bling of the number of adhered
organisms. Specificepithelial lipids can modulate fungal adherence,
sincebinding was inhibited by fatty acids, sterols, andceramides
and was unaffected by squalene, steryl esters,cholesterol esters,
and triglycerides. In a murine stratumcorneum model, yeast cells of
C. albicans and C. stellatoideaadhered in greater numbers than
those of C. tropicalis,while C. guilliermondii, C. krusei, and C.
parapsilosis cellsshowed little or no adherence (Ray and Payne,
1988). Thishierarchy of adherence was similar to that observed
withhuman epidermal corneocytes and BECs (Ray et al., 1984).In the
murine stratum corneum model, the adherent cellsacquired fibrils
and strands of an amorphous materialbetween the yeast and
corneocyte cell surface, formedcavitations at the site, and
produced hyphae that invadedcorneocytes distal to the yeast
attachment (Ray andPayne, 1988). Depletion of lipids had no effect
on adher-ence in this study, but pepstatin, an inhibitor of the
fun-gal secreted aspartyl proteinase, inhibited the formationsof
cavities around the adherent cells. Epidermolytic pro-teases,
likely including the secreted aspartyl proteinase,have been
isolated from strains recovered from patientswith cutaneous disease
(El-Maghrabi et al., 1990).Pepstatin, bovine brain gangliosides,
and convalescenthuman serum all reduced binding of yeast cells to
cor-neocytes. Although there may be differences amongadhesins for
corneocytes and BECs and vaginal epithelialcells (VECs), it is
likely that there are at least some com-mon adhesins. It appears,
therefore, that C. albicans chitinand proteinase may be important
in skin colonization.
(4) Adherence to Oral SurfacesThe oral cavity presents a number
of surfaces for C. albi-cans adhesion. These include BECs, the
inert polymers ofdental prostheses, teeth, and other oral
micro-orga-nisms. Adherence to each of these surfaces and the
mod-ulating effect of saliva on adhesion will be discussed.
(A) ADHERENCE TO BECsExfoliated BECs are probably the
best-investigatedhuman cell type in C. albicans adherence studies,
and sev-eral adhesin/ligand interactions have been proposed(Table
1). In interpreting the results of BEC adhesionassays, one should
note the following features: Fungalstrain and source of epithelial
cells affect adherence(Sandin et al., 1987b), the number of C.
albicans cells thatbind to individual BECs is variable (Sandin et
al., 1987a),there are both binding and non-binding BECs (Gorman
etal., 1986; Polacheck et al., 1995), and both viable and
non-viable C. albicans cells bind to BECs, with the non-viablecells
having a greater adherence (Gorman et al., 1986).
10(3) 359-383 (1999) Crit Rev Oral Biol 36310(3)-.359-383 (1999)
Crit Rev Oral Biol Med 363 by guest on June 1, 2015 For personal
use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
The binding capacity of BECs from newborn full-terminfants for
Candida is less than that of BECs from prema-ture infants,
school-age children, and adults, but increas-es over the infants'
first few days (Davidson et al., 1984;Cox, 1986; Polacheck et al.,
1995). Adherence was greaterto BECs from children with oral
infection or colonizationthan to BECs from uninfected controls but
increasedduring a course of antibiotic therapy in previously
unin-fected children (Cox, 1983). Menstrual cycle affectsepithelial
binding capacity, since BECs collected on day5 showed a binding
capacity higher than that of cells col-lected on day 15, 22, or 28
(Theaker et al., 1993). In onestudy, VECs from the first and fourth
weeks had a bind-ing capacity higher than that of VECs from the
second orthird week (Segal et al., 1984), while another study
sug-gested that binding capacity peaked between the thirdand fourth
weeks (Bibel et al., 1987). VECs collected frompregnant or diabetic
women also bound more C. albicanscells than those from non-pregnant
or non-diabetic con-trols, and 'infection isolates' adhered better
than 'colo-nizing isolates' (Segal et al., 1984). However, there
was nodifference in the binding capacity of VECs from womenwith
recurrent vaginitis compared with healthy controls(Trumbore and
Sobel, 1986). Palatal epithelial cells fromacrylic-denture-wearers
with non-insulin-dependent dia-betes bound more fungal cells than
did epithelial cells ofnon-diabetic individuals (Dorocka-Bobkowska
et al.,1996). In another study, adherence to BECs from diabet-ic
individuals was similar to adherence to BECs from nor-mal
individuals (Polacheck et al., 1995). Thus, adherenceto BECs is
affected by many host factors, and hormonaleffects on adherence
could be mediated by altering theexpression of adhesins on C.
albicans cells or ligands onhost cells.
There is also variability in binding to BECs from dif-ferent
donors (Sandin et al., 1987b). Adherence differedwhen epithelial
cells were collected on different dates,but gender was not a
factor. C. albicans adhered in greaternumbers to BECs from AIDS
patients than to BECs fromhealthy individuals or transplant
patients (Schwab et al.,1997). C. albicans isolates from patients
in the early stagesof AIDS adhered to BECs less well than did those
fromhealthy individuals. However, adherence of isolatesincreased
with the progression of AIDS until it exceededthat of control
isolates (Pereiro et al., 1997). Isolates fromimmunocompetent
patients with esophageal candidiasisadhered better than isolates
from patients who wereheavily colonized but not symptomatic
(Wellmer andBernhardt, 1997). Although there were differences
amongstrains, isolates from candidiasis patients were moreadherent
and formed germ tubes more rapidly than theother isolates. Analysis
of these results is complicated bythe fact that the adherence of
strains from AIDS patientshas mostly been measured in vitro and may
not correlate
to adherence in vivo for patients who are on courses
ofantifungal drugs.
The effect of treating BECs and/or C. albicans cellswith
antimicrobial agents on subsequent adherence hasbeen studied
extensively. The consensus of opinion isthat treatment of BECs or
yeast cells with any of a varietyof agents-including chlorhexidine,
hexetidine,dequalinium chloride, cetrimide, cetylpyridinium
chlo-ride, octenidine, pirtenidine, taurolidine,
propamidineisethionate, noxythiolin, and aqueous garlic
extract-reduced adherence (Gorman et al., 1986, 1987a,b; Tobgi
etal., 1987; Ghannoum, 1990; Ghannoum et al., 1990; Jonesand
Fowler, 1994). Also, treatment of C. albicans withpropamidine
isethionate, octenidine, pirtenidine, andaqueous garlic extract
reduced germ tube formation(Jones and Fowler, 1994; Jones et al.,
1997). Exposure toantifungal drugs may also reduce adherence to
epithelialcells. Subinhibitory concentrations of amphotericin
B,nystatin, miconazole nitrate, and 5-fluorocytosinereduced binding
of C. albicans, C. tropicalis, and C. kefyr toBECs, and the effect
of amphotericin B and 5-fluorocyto-sine combined was greater than
that of either alone(Abu-el Teen et al., 1990). A one-week course
of flucona-zole also reduced the adherence of C. albicans to
BECs(Darwazeh et al., 1991). Drug treatment could be
affectingcell-surface charge, or wall and membrane biosynthesisand
structure.
Binding of C. albicans to exfoliated epithelial cells isaffected
by growth conditions of the fungus and can beinhibited by several
reagents. Although there are somedifferences among studies that may
reflect the complex-ity of growth conditions and adhesin
expression, there isprogress in characterizing the interactions and
identify-ing the fungal adhesins and host receptors. Growth of
C.albicans in media containing glucose, sucrose, galactose,xylitol,
or maltose enhanced binding to BECs and HeLacells; maltose was the
most effective and glucose theleast effective sugar (Samaranayake
and MacFarlane,1982). Growth in the presence of glucocorticoids,
dexa-methasone, or triamcinolone acetonide also increasedadherence
to BECs (Ghannoum and Abu Elteen, 1987).In one study, organisms
grown at 250C were more adher-ent than those grown at 37C (Lee and
King, 1983). Cell-surface hydrophobicity, which is increased at the
lowergrowth temperature, is suggested to contribute to, butnot be
the predominant mechanism of, adherence toBECs (Hazen, 1989). A
decrease in hydrophobicity maycontribute partially to the decrease
in binding followingtreatment of C. albicans with cetylpyridium
chloride, tau-rolidine, chlorhexidine acetate, or
providone-iodine(Jones et al., 1991, 1995). Another study
demonstratedthat an increase in temperature during growth
promotedadherence, and, as with growth on different carbonsources,
this may be due to increased expression of an
364 Grit Rev Oral Biol Med 10(3):359-383 (1999)364 Crit Rev Oral
Biol Med 10(3):359-383 (1999) by guest on June 1, 2015 For personal
use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
adhesin for a high copy number receptor (Staddon et
al.,1990).
Concanavalin A, a lectin-recognizing mannan, is aninhibitor of
adherence to BECs (Sandin and Rogers,1982; Sandin, 1987; Macura and
Tondyra, 1989). Glucose,galactose, sucrose, or mannose enhanced
adherence toBECs, while xylose, ribose, fructose, maltose, lactose,
orraffinose had no effect on adherence (Macura andTondyra, 1989).
Another study found no effect of galac-tose, N-acetylglucosamine
(GIcNAc), ribose, or xylose(Sandin, 1987). Thus, certain sugar
residues may beinvolved in a lectin-like adherence interaction or
maycross-bridge between adhesin and ligand. Lipids extract-ed from
C. albicans or C. tropicalis inhibited binding toBECs and involved
individual phospholipids, sterols,and steryl ester but not
triacylglycerols or free fatty acids(Ghannoum et al., 1986).
Several C. albicans cell wall proteins have been iden-tified as
adhesin candidates for epithelial cells. In onestudy, a yeast cell
wall extract fractionated by con-canavalin A-affinity
chromatography followed by ion-exchange chromatography yielded a
fraction that sub-stantially inhibited yeast cell binding to BECs
(Imbert-Bernard et al, 1995). This fraction contained four
moi-eties, of which the 38- and 54-kDa proteins were sug-gested as
adhesins. 0-linked mannoproteins may also beinvolved in C. albicans
adherence to epithelial cells. The C.albicans CaMNTI gene encodes a
mannosyl transferaseinvolved in 0-linked mannosylation, and a
Camntl null-mutant showed reduced adherence to BECs (Buurman etal.,
1998).
Fibronectin was one of the first molecules to be sug-gested as a
ligand recognized by a C. albicans adhesin(Skerl et al, 1984). Both
BECs and VECs stained with anti-fibronectin antibody, and yeast
cells pre-treated withfibronectin showed reduced binding to BECs
and VECscompared with untreated cells (Skerl et al., 1984; Kalo
etal., 1988). The complement fragment iG3b has also beenimplicated
as a ligand involved in epithelial andendothelial cell adherence
(Gustafson et al., 1991; Bendeland Hostetter, 1993; Bendel et al.,
1995; also see reviewsby Hostetter, 1994; Chaffin et al., 1998).
Glucose-growncells express more iC3b receptor than
glutamate-growncells and show increased binding to human
umbilicalvein cells (HUVCs) (Gustafson et al., 1991). Antibody
tothe human iC3b integrin receptor, iC3b, and several
RGD(arginine-glycine-aspartic acid)-containing peptidesfrom iC3b
reduced binding of C. albicans to HUVCs orHeLa cells (Bendel and
Hostetter, 1993). After growth ofHeLa cells in serum-free medium,
iC3b and fibronectinwere detected on the cell surface, and
treatment withanti-C3 antibody, but not anti-fibronectin
antibody,reduced adherence of C. albicans. although the
reverseeffect was observed with C. tropicalis. The candidates
for
fibronectin and iC3b adhesin(s) are described in moredetail
below.
The secreted aspartyl proteinases (SAPs) also appearto
contribute to C. albicans adherence to BECs and othersubstrates
(Ghannoum and Abu Elteen, 1986; El-Maghrabi et al., 1990; Watts et
at., 1998). The SAP genefamily consists of at least seven members
encoding 42-to 45-kDa aspartyl proteinases (Hube et al., 1994;
Monodet al., 1994). The expression of proteinase isozymesdepends on
the strain, cellular morphology, and environ-mental factors (White
and Agabian, 1995). Strains defi-cient in one or more of these
genes have been con-structed. Deletions in SAPI, SAP2, or SAP3
reducedadherence of the organism to poly-L-lysine, an
extracel-lular matrix (ECM) preparation, or (slightly) to
BECs(Watts et al., 1998). However, a triple Asap 4-6 mutantshowed
decreased adherence to the first two substratesbut increased
adherence to BECs. Pepstatin inhibitedbinding of the parental
strain to all three substrates. Inaddition to any direct effect on
adhesion, proteinasesmay act on the yeast surface to modify
adhesins or hostsurfaces to expose ligands.
BEC glycosphingolipid is also an adherence targetfor C.
atbicans. Several pathogenic yeasts, including C.albicans, bind to
lactosylceramide ltGal( 1 -4)3Glc(l - )Cerl(limenez-Lucho et al.,
1990). C. albicans fimbriae bound toBECs and reduced the binding of
C. albicans yeast cells toBECs (Yu et al., 1994b). Purified
fimbriae bind to an asialo-GM, Igangliotetraosylceramide: 3Gal(1
-3)f3GalNAc( 1-4)3Gal( l-4)f3Glc( 1-I )Cerl immobilized on
microtiterplates. The binding of fimbriae to BECs was inhibited
upto 80% by asialo-GMP. Pseudomonas aeruginosa also bindsto this
glycosphingolipid through pili, and the adhesinsfrom P. aeruginosa
and C. albicans appear to share a com-mon binding domain
(adhesintope) (Yu et al., 1994c,1996). Antibodies to this domain in
the P. aeruginosa pilusprotein inhibit binding of both organisms to
BECs, and apeptide derived from this region is also inhibitory.
The presence of candidal lectin-like epithelialadhesins that
recognize L-fucose or GlcNAc has beenreported (Critchley and
Douglas, 1987a,b). Fucoseinhibits binding of some strains to BECs,
and glu-cosamine or GlcNAc inhibits the binding of other
strains,suggesting strain-specific receptors. Synthesis of
thelectin-like material increased when organisms weregrown on
galactose (McCourtie and Douglas, 1985).Extracellular material
recognizing L-fucose inhibitedbinding of the homologous strain.
Fucose has beenshown to bind to yeast and hyphal cells with
approxi-mately 2 x 107 binding sites per hyphal cell, mostly
locat-ed adjacent to the hyphal tip (Vardar-Unlu et al., 1998).
Afragment of an L-fucose-binding protein was purified byaffinity
chromatography with the blood group H trisac-charide antigen that
terminates in fucose, and it was sug-
365103 5 8 (9)Gie rlBo e
Crit Rev Oral Biol Med10(3)-.359-383 (1999) by guest on June 1,
2015 For personal use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
gested that blood group antigens may act as epithelialcell
receptors for C. albicans (Cameron and Douglas,1996). The purified
fragment inhibited binding to BECsby up to 80%. Binding to an
esophageal cell line (Het-l)is partially mediated by a lectin-like
interaction (Enacheet al., 1996). Yeast cells grown on galactose
adhere betterthan those grown on glucose, and GlcNAc or
glu-cosamine reduce binding by about 40%. A 190-kDa gly-coprotein
detected in cell wall extracts of galactose-grown cells was
postulated to be responsible for theincreased adherence (Enache et
al., 1996).
The serotype A determinant, factor 6, of C. albicansmannan has
also been implicated in epithelial adher-ence (Miyakawa et al.,
1992). A mutant strain deficient inthe factor 6 determinant, or
serotype B strains, showedreduced adherence to BECs compared with
non-mutantserotype A strains. Mannan from the wild-type strain
andanti-factor 6 antibody inhibited adherence to BECs.
A genetic approach has yielded two candidate C. albi-cans
adhesins for epithelial cells. The fact that S. cerevisiaecells
adhere to a variety of surfaces significantly less wellthan do C.
albicans cells has been used by several groupsto screen C. albicans
genomic libraries for sequences thatconfer adherence on the
non-adherent yeast. Separatescreens have identified two members of
a family of relat-ed genes. Members of the C. albicans ALS
(agglutinin-likesequence) family are related to S. cerevisiae
agglutiningenes that mediate cell-cell interactions during matingof
haploid cells (Hoyer et al., 1995). Als proteins have acentral
domain of a tandemly repeated motif that is richin serine,
threonine, and proline. The sequence of ALSIcarries a signal for a
GPI (glycosyl phosphatidyl inositol)anchor. Another member of the
family, ALAl, was isolat-ed by the screening of a library for
sequences that con-ferred adherence to ECM (Gaur and Klotz,
1997).Transformed yeast cells bound to fibronectin, laminin,and
collagen IV. In addition, adherence to BECs wasincreased,
suggesting that the adhesin may be multi-functional, recognizing
multiple ligands, and mediatingadherence to different tissues. More
recently, ALS1 hasagain been isolated in a screen of a C. albicans
genomiclibrary for sequences conferring increased adherence
toendothelial cells (Fu et at., 1998). Expression of ALS1
alsosubstantially increased binding of S. cerevisiae to the
FaDuoropharyngeal epithelial cell line. More definitive evi-dence
for the role of these proteins in candidal adher-ence awaits
further analysis in that organism.Nonetheless, members of the Als
protein family are cer-tainly candidates for adhesins that mediate
adherence ofC. albicans in the oral cavity.
In addition to physical immobilization, adherence ofCandida
cells to BECs may lead to alterations in fungalgene expression
(Bailey et al., 1995). Analysis of proteinssynthesized by C.
albicans three hours following adhesion
to BECs showed that proteins of 52-56 kDa differed in theextract
of attached yeast cells compared with those fromunattached yeast or
from BECs alone. Furthermore, anti-phosphotyrosine antibodies
recognized 54-kDa and 60-kDa species from the attached cells but
not from cells incontrol cultures. These results suggest that
contact of C.albicans with a surface may activate signaling
pathwaysthat result in the expression of adhesins. Some C.
albicansstrains demonstrate the phenomenon of phenotypicswitching
(Soll, 1997). It is postulated that a 'masterswitch' is responsible
for turning off one set of genes andswitching on another set, some
of which may be involvedin virulence. It is possible that contact
with a particularsurface activates a set of genes involved in
adherence to,and penetration of, that surface.
(B) ADHERENCE TO INERT POLYMERSC. albicans adheres to a variety
of materials found in med-ical devices, such as catheters and oral
prostheses. Thisadherence may promote colonization and infection.
C.albicans is able to form biofilms on the surfaces of
thesematerials (reviewed in Chaffin et al., 1998). In
addition,colonization may contribute to the deterioration of
thedevices (Marcuard et al., 1993; Gottlieb and Mobarhan,1994;
Busscher et al., 1997; van Weissenbruch et al., 1997),and adherent
organisms may be less susceptible to anti-fungal drugs (Kayla and
Ahearn, 1995; Hawser, 1996).Most studies have focused on oral
devices which maycontain multiple materials. Since these studies
used dif-ferent fungal growth conditions, different
adherenceassays, and different methods of analysis, the
resultscannot be compared directly.
Hydrophobicity has been frequently, but not univer-sally,
implicated as a major factor in the adherence ofCandida species to
inert polymers. The more hydrophobicspecies C. tropicalis, C.
glabrata, and C. krusei adhered moreto these polymers, including
those found in dentureresin materials, than the less hydrophobic C.
albicans, C.stellatoidea, and C. parapsilosis (Klotz et al., 1985;
Minagi etal., 1985, 1986; Miyake et al., 1986). Isolates of C.
krusei, anemerging pathogen, showed variable but
greaterhydrophobicity than C. atbicans isolates, and there was
acorrelation between hydrophobicity and adherence toHeLa cells but
not to denture acrylic (Samaranayake etal., 1995). This suggests
that factors other thanhydrophobicity might contribute to the
hierarchy of viru-lence among Candida species. In an earlier study,
isolatesof C. albicans showed greater adherence to acrylic
thanisolates of other species (Segal et al., 1988). Adherence
isincreased on rough acrylic and silicone rubber surfacescompared
with smooth surfaces (Verran and Maryan,1997). The acrylic base for
dentures supported lessadherence of C. albicans than tissue
conditioners and asoft liner (Okita et al., 1991).
366 Crit Rev Oral Biol Med lO(3):359-383 (1999)366 Crit Rev Oral
Biol Med 10(3):359-383 (1999) by guest on June 1, 2015 For personal
use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
As dental prostheses are exposed to saliva and oralbacteria, a
complex biofilm develops to which C. albicanscells can adhere. The
fungus is present in biofilms in var-ious morphological forms, and
extracellular materialmay also be present (Hawser and Douglas,
1994). Theextent of biofilm formation is dependent on the nature
ofthe inert material; the greatest biofilm formation wasfound on
latex, which is frequently used in urinarycatheters, followed by
silicone elastomer and polyvinylchloride, often found in central
venous catheters.Formation of a biofilm was least on polyurethane
and100% silicone. In vitro, gentle liquid flow increased
theformation of the extracellular matrix material, in whichthe
organism was embedded, compared with static con-ditions (Hawser et
al., 1998).
As with adhesion to BECs, treatment of either dentalacrylic or
C. albicans cells with antimicrobial agentsaffects adhesion.
Incubating acrylic with chlorhexidinegluconate, amphotericin B,
nystatin, but not a histidinepolypeptide, reduced binding to the
polymer (McCourtieet al., 1985, 1986a,b; Spiechowicz et al., 1990).
Exposure ofstationary-phase cells to chlorhexidine for a short
peri-od, or growth of C. albicans in a sublethal concentration
ofchlorhexidine, reduced the adherence of the cells com-pared with
unexposed cells, and the treated cells weremore susceptible to the
action of :3-glucanase. This sug-gests an effect of chlorhexidine
on the fungal cell wall.When C. albicans was grown in subinhibitory
concentra-tions of antifungals, exposure to azalomycin F
andaculeacin A increased subsequent adherence to acrylic,while
exposure to miconazole, ketoconazole, andamphotericin B did not
alter adherence (Miyake et al.,1990). Exposure to drugs did not
change cell-surfacehydrophobicity, while the negative charge of the
cell sur-face decreased in the more adherent cells, suggestingthat
a decrease in electric repulsive force enhanced bind-ing. Growth of
C. albicans, C. krusei, C. kefyr, C. tropicalis, C.parapsilosis,
and C. guilliermondii in subinhibitory concen-trations of sodium
hypochlorite resulted in subsequentreduction in adherence of all C.
albicans strains and mostother species to polystyrene and BECs
(Webb et al.,1995). Growth in hypochlorite appeared to increase
thenumbers and amounts of certain proteins in cell wallextracts
from C. albicans and C. parapsilosis, again indicat-ing alterations
in the cell wall composition.
Several surface mannoproteins, among themhydrophobic proteins,
have been suggested as adhesincandidates for plastics (reviewed by
Fukazawa andKagaya, 1997; Chaffin et al., 1998). Yeast cells grown
ingalactose were more adherent to acrylic than thosegrown in medium
containing glucose, sucrose, fructose,or maltose (McCourtie and
Douglas, 1981). Materialfound in the growth medium, when used to
pre-treatacrylic or BECs, promoted adherence of C. albicans cells
to
acrylic but reduced adherence to BECs (McCourtie andDouglas,
1985). When germ tubes that adhered to poly-styrene were physically
removed, several mannoproteinswere subsequently solubilized from
the plastic (Tronchinet al., 1988). Two major constituents of 60
and 68 kDa andtwo minor constituents of high molecular mass (.
200kDa) were obtained. While the relationship of the small-er
species to similar-sized proteins described below asrecognizing
other ligands is unknown, the size similarityhas supported
conjecture that there may be multi-func-tional adhesins recognizing
a variety of ligands. A 58-kDaand a 37-kDa protein which bind
fibrinogen and laminin,respectively, also bind to plastic and have
been suggest-ed to possess hydrophobic domains (Lopez-Ribot et
al.,1991, 1995). Among extracted cell wall proteins, there aremany
that have hydrophobic domains. Analysis of pro-teins adsorbed to
latex beads showed a spectrum of pro-teins in the 20- to 67-kDa
range that may be more abun-dant in extracts from germ tubes
(Lopez-Ribot et al.,1991). Hydrophobic interaction chromatography
ofextracted proteins suggested that the hydrophobic pro-teins were
usually smaller (< 50 kDa) than thehydrophilic proteins (> 90
kDa), perhaps reflecting theextent of glycosylation (Hazen and
Hazen, 1992, 1993;Hazen and Glee, 1994). Hydrophilic cells exhibit
a denselayer of fibrils not observed on hydrophobic cells, and
ithas been proposed that this layer masks the hydropho-bic species.
In keeping with this suggestion, the abun-dance of the acid-labile
phosphodiester-linked manno-oligosaccharides was less in mannan
from hydrophobiccells than in that from hydrophilic cells (Masuoka
andHazen, 1997).
(C) ADHESION TO TEETHThe mouth is a unique part of the body in
that it containsexposed mineralized tissues, in the form of teeth.
Beadsof crystalline hydroxyapatite (HA) have been used inadhesion
assays as a model for studying microbial adhe-sion to tooth
surfaces (Clark et al., 1978). C. albicans cellsdo not bind well to
hydrated HA beads, but adherence isstimulated greatly by
pre-incubation of the beads witheither whole or parotid saliva (
Cannon et al., 1995b;O'Sullivan et al., 1997). Adherence to
saliva-coatedhydroxyapatite (SHA) beads is strain-specific
(O'Sullivanet al., 1997), and strains more frequently associated
withcandidiasis adhere significantly better to SHA beadsthan do
less pathogenic strains (Schmid et al., 1995b).
(D) CO-ADHERENCEC. albicans cells co-adhere with several species
of oralbacteria, including Streptococcus gordonii, S. mutans, S.
oralis,S. sanguis, S. salivarius, and Actinomyces species
(Richardsand Russell, 1987; Branting et al., 1989; Jenkinson et
al.,1990; Holmes et al., 1995b; Millsap et al., 1998). The
10(3359383(199) Cit Rv Orl Bd Me 36367Crit Rev Oral Biot
Med10(3)-359-383 (1999) by guest on June 1, 2015 For personal use
only. No other uses without permission.cro.sagepub.comDownloaded
from
-
growth conditions for the bacteria, however, can
affectco-adherence (Richards and Russell, 1987; Millsap et
al.,1998), and some assays do not take into account thekinetics
associated with the larger size of yeast cells(Millsap et al.,
1998). Colonizing acrylic with oral strepto-cocci in the presence,
but not in the absence, of sucroseenhanced binding of C. albicans
(Richards and Russell,1987; Branting et al., 1989). C. albicans
bound in greaternumbers to acrylic pieces coated with S. sanguis,
S.mutans, or S. sobrinus than to uncoated acrylic, but in thiscase,
pre-incubation of the bacteria with sucrose toinduce synthesis of
extracellular polymers did notincrease binding (Vasilas et al.,
1992). Protein-protein andlectin interactions have been proposed
for the adhesiveinteractions between Candida and bacteria,
althoughhydrophobic and electrostatic interactions may also
takepart (Millsap et al., 1998). Both carbohydrate (Holmes etal.,
1995b) and protein molecules (Holmes et al., 1996)that act as C.
albicans receptors have been identified onthe surface of S.
gordonii. A carbohydrate containingrhamnose, glucose, GlcNAc, and
galactose, isolated fromthe cell walls of S. gordonii cells, acted
as a receptor for C.albicans adherence in an in vitro assay (Holmes
et al.,1995b). Gene disruption experiments have shown that
C.albicans adherence to bacteria is multifactorial, and
inter-actions involving the S. gordonii cell-surface
polypeptidesCshA, CshB, SspA, and SspB contribute to
co-adherence(Holmes et al., 1996). The co-adherence of C. albicans
withoral bacteria is species-specific. Pre-treating BECs or
den-ture acrylic with S. salivarius, Escherichia coli, or
Porphyromonasgingivalis reduced subsequent adherence of C. albicans
cells(Nair and Samaranayake, 1996a,b). Also, in one report,
abiofilm of S. gordonii reduced adherence of most C.
albicansstrains and other species to polystyrene (Webb et
al.,1995). This would suggest that C. albicans recognizes spe-cific
receptors on certain oral bacteria which are expressedunder
particular growth conditions.
(E) ADHERENCE TO SALIVA MOLECULESIn the oral cavity, proteins
from saliva selectively adsorbto surfaces to form acquired
pellicles. The acquiredenamel pellicle has been particularly
well-studied, andafter two hours' formation it has been found to
containimmunoglobulins, mucin, at-amylase, cystatins, proline-rich
proteins, lysozyme, glucosyltransferases, albumin,fibrinogen, and
serum components (Kraus et al., 1973;Rolla et al., 1983; Al-Hashimi
and Levine, 1989; Jensen etal., 1992; Edgerton et al., 1996). The
composition of thepellicle depends on the underlying surface
(Edgerton etal., 1996) and the composition of the saliva (Jensen et
al.,1992; Edgerton et al., 1996). The intra-oral composition
ofsaliva varies (Sas and Dawes, 1997), and this affects
thepellicles formed at different sites, and hence the patternof
microbial colonization. Since all surfaces are coated
with a salivary pellicle, it is reasonable to suppose
thatmicrobial adherence interactions involve adsorbed
salivamolecules. Saliva pellicles increase the adherence of
C.albicans cells to HA beads (Cannon et al., 1995b),
poly-methylmethacrylate (Edgerton et al., 1993), and to S.
gor-donii cells (Holmes et al., 1995a) (Fig. 2b). Adherence of
C.albicans was greater to dental acrylic coated with wholesaliva
than to uncoated acrylic, and a coating of parotidsaliva stimulated
adherence more than a coating of sub-mandibular-sublingual saliva
(Vasilas et al., 1992). In anearlier study, however, adhesion to
acrylic was reducedby an 18-hour whole-saliva pellicle
(Samaranayake et al.,1980). Also, coating of acrylic surfaces with
whole salivareduced the contact angle and decreased the binding
ofhydrophobic Candida strains, while the adherence ofmore
hydrophilic C. albicans was unaffected (Miyake et al.,1986).
Adherence to two experimental silicone soft-lin-ing materials was
less than to a commercial product orthe acrylic base, varied with
the strains, and was reducedwhen the materials were coated with
saliva (Waters et al.,1997). In another study, however, coating
soft liners withsaliva or serum increased adherence of C. albicans
andbiofilm formation, although the effect varied with thematerial
and protein source (Nikawa et al., 1997). Coatingalso increased
firm colonization and hyphal invasion,although the plasticizer used
affected cavitation.Incubating polymethylmethacrylate beads with
sub-mandibular-sublingual saliva enhanced C. albicans bind-ing
compared with coating them with parotid saliva(Edgerton et al.,
1993). Binding was reduced by treatmentof yeast cells with protease
or glycosidase or incubationwith mannose or galactose.
Interestingly, C. albicans cellsdo not detectably bind proteins
from saliva in the fluidphase, apart from small amounts of mucin
MG1 and MG2(Edgerton et al., 1993; Newman et al., 1996), which
wouldexplain why added saliva did not inhibit adherence of
C.albicans to SHA beads (Cannon et al., 1995b). This indi-cates
that C. albicans may have specific adhesins that rec-ognize
cryptitopes on saliva molecules that are exposedwhen the molecules
adsorb to surfaces (Fig. 2b). Suchadhesins would promote
colonization and prevent sali-va-mediated aggregation and clearance
from the oralcavity.
In order to identify the saliva proteins to which C.albicans
cells adhere, investigators have developed blotoverlay assays
(Newman et al., 1996; O'Sullivan et al.,1997), in which saliva
proteins are separated by SDSpolyacrylamide gel electrophoresis,
electroblotted ontonitrocellulose membranes, and incubated with
eitherradiolabeled (O'Sullivan et al., 1997) or
fluorescentlylabeled (Newman et al., 1996) yeast
cells.Autoradiography or photography, respectively, revealsthe
protein bands to which the cells bind. These studieshave identified
basic proline-rich proteins, including IB-
368~~~~~~~ ~ ~ ~ ~GrtRvOa ilMdl()3933(99368 Crit Rev Oral Biol
Med 10(3):359-383 (1999)
by guest on June 1, 2015 For personal use only. No other uses
without permission.cro.sagepub.comDownloaded from
-
6 (O'Sullivan et al., 1997) and Psi (Newman et al., 1996),as
receptors for C. albicans adhesion.
(5) GrowthIn order to maintain Candida populations in the oral
cav-ity, cells must grow and multiply at a rate at least equalto
that of clearance. The growth rate of C. albicans in sali-va is too
low to be measured accurately, due to carbonsource limitation
(Samaranayake et al., 1986), which ispresumably caused by the large
number of bacteria insaliva. Therefore, any metabolic activity that
helps C. albi-cans acquire carbon or nitrogen will aid its growth
andsurvival in the oral cavity. C. albicans secretes aspartyl
pro-teinases, which are believed to contribute to the orga-nism's
virulence in several ways (Hoegl et al., 1996).Tissue destruction
may aid fungal penetration, but thisprocess could also release
peptides as a source of nitro-gen or carbon. The proteinase Sap2,
for example,degrades gastrointestinal mucin, and mucin can act as
asole nitrogen source for C. albicans (Colina et al., 1996).
Inaddition, C. albicans secretes the hydrolytic enzyme
N-acetylglucosaminidase (also called hexosaminidase),which cleaves
chitobiose, the dimer of GIcNAc, into twomolecules of GIcNAc
(Sullivan et al., 1984; Niimi et al.,1997a). C. albicans can use
GlcNAc as either a carbon ornitrogen source.
N-acetylglucosaminidase activity isshown by C. albicans and C.
dubliniensis and to a lesserextent by C. tropicalis cells (Niimi
and Cannon, unpub-lished observation), species found relatively
frequently inthe oral cavity. It is tempting to speculate that a
functionof this enzyme may be to cleave terminal GIcNAc
residuesfrom host glycoproteins, and that this scavenging
activitygives these Candida species a selective growth
advantage.
Competition with other oral micro-organisms fornutrients, such
as glucose, affects the growth rate ofCandida cells. It is
recognized that antibiotic treatment,which reduces the number of
oral bacteria, is a predis-posing factor for oral candidiasis
(Samaranayake, 1990).Oral bacteria are present in most oral sites
at concentra-tions much higher than C. albicans, and so the
Candidacells must compete with them for adhesion sites
andnutrients, and be exposed to bacterial toxins and
by-products.
(6) Evading Host Clearance MechanismsA major factor influencing
the balance among clearance,colonization, and candidiasis is the
interaction betweenC albicans cells and the host defenses (Cannon
et al.,1 995a). Immune system defects are a major risk factor
forcandidiasis. Innate defenses include the epithelial barri-er and
anti-candidal compounds in saliva such aslysozyme (Tobgi et al.,
1988), histatins (Xu et al., 1991),lactoferrin (Nikawa et al.,
1993), and calprotectin (Mulleret al., 1993; Challacombe, 1994).
Acquired immunity
includes the production of immunoglobulins and, if tis-sues are
penetrated, the involvement of macrophagesand polymorphonuclear
leukocytes (Challacombe, 1994).The major immunoglobulin in saliva
is secretory IgA(SIgA); serum immunoglobulins enter the saliva via
thegingival crevicular fluid, but are present at low
concen-trations. SIgA does not fix complement efficiently; itsmajor
role is the agglutination of micro-organisms,which are then
swallowed more easily. Anti-Candida SIgAcan be detected in saliva,
and its concentration isincreased in whole or parotid saliva from
HIV-positiveindividuals, but reduced in AIDS patients,
suggestingthat a compensatory response is overcome with
progres-sive immunodeficiency (Challacombe and Sweet, 1997).
C. albicans can be ingested by neutrophils andmononuclear
phagocytic cells (for a review of interac-tions with macrophages,
see Vazquez-Torres and Balish,1997). The interactions between these
cells and C. albi-cans appear to involve both opsonic and
non-opsonic fac-tors. Components from the classic and alternative
com-plement pathways can also enhance phagocytosis bymacrophages
and neutrophils (Solomkin et al., 1978;Marodi et al., 1991). C.
albicans activates the alternatepathway of complement, and both
iC3b and C3d frag-ments can bind to C. albicans (adhesins for these
frag-ments are discussed below). A candidal-protective mech-anism
has been proposed in which fungal binding ofiC3b blocks neutrophil
CR3 recognition of iC3b andphagocytosis of iC3b-coated C. albicans
is reduced. Yeastcells coated with an anti-human CR3 antibody,
thatblocked the candidal binding protein, were phagocy-tosed by
neutrophils to a greater extent than uncoatedcells (Gilmore et al.,
1988). The clumping of C. albicanscells coated with C3 fragments
has also been proposedas a candidal-protective effect, since these
aggregatesare too large to be phagocytosed (Heidenreich andDierich,
1985). On the other hand, host protection ispostulated for the
binding of serum vitronectin, sinceCandida cells coated with
vitronectin show enhancedbinding to macrophages and phagocytosis
(Limper andStanding, 1994).
Macrophage mannose receptors also mediate theadherence of C.
albicans (reviewed by Vazquez-Torres andBalish, 1997). Binding of
C. albicans to murine spleen andlymph node tissue is primarily to
macrophages (Kanbe etal., 1993). A (31,2-linked mannotetraose in
the acid-labileC. albicans mannan as well as an acid-stable
structurewere identified as adhesins (Li and Cutler, 1993; Kanbeand
Cutler, 1994). A monoclonal antibody to
(1,2-linkedoligomannosaccharide, but not a monoclonal antibodyto
the acid-stable mannan epitope, in the presence ofcomplement,
enhanced phagocytosis of yeast cells byneutrophils (Caesar-TonThat
and Cutler, 1997). Solublemannan can inhibit phagocytosis of
complement-C3-
lO(3:3593831999 Crf Re Ora Bil Me 36369Crit Rev Oral Biot
Med10(3):359-383 (1999) by guest on June 1, 2015 For personal use
only. No other uses without permission.cro.sagepub.comDownloaded
from
-
coated C. albicans by macrophage (Kolotila et al., 1987).Thus,
C. albicans phagocytosis may be mediated by eitherthe
complement-dependent or mannose receptor path-ways. The effect of
mannan on the complement-mediat-ed pathway, however, suggests that
the pathways mayinteract physiologically or physically. C. albicans
hasdeveloped other ways of evading the innate and theacquired
immune system. The C. albicans cell wall manno-proteins proposed to
be receptors for SIgA binding arereleased into the medium as yeast
cells produce hyphae(Ponton et al., 1996). The shedding of these
receptors mayenable the hyphae to escape clearance. In addition,
theimmunoglobulins present in saliva, including SIgA, IgG,and IgM,
are substrates of C. albicans secreted aspartylproteinase (Ruchel,
1986; Reinholdt et al., 1987).
(A) FINDING BETTER ENVIRONMENTSC. albicans will grow more
quickly and reach higher cellconcentrations in environments with
less effective clear-ance mechanisms and better nutritional supply.
Candidacells have an array of mechanisms (virulence factors)which
enable them to colonize new environments, aprocess often involving
tissue penetration. Polarizedhyphal growth facilitates directional
growth toward a dif-ferent environment, and thigmotropism or
contact sen-sing (Sherwood et al., 1992) could allow the hyphae
toinvade tissues. C. albicans cells also secrete a number
ofhydrolytic enzymes which may play a role in tissuedestruction and
penetration. Phospholipase or N-acetyl-glucosaminidase production
by C. albicans strains, forexample, correlated with virulence in
mouse infectionmodels (Jenkinson and Shepherd, 1987; Ibrahim et
al.,1995). Many workers have investigated the role of secret-ed
aspartyl proteinase isozymes in the pathogenesis ofcandidiasis, and
demonstrated their expression in vivo(Borg and Ruchel, 1988; Ray
and Payne, 1988; El-Maghrabi et al., 1990; De Bernardis et al.,
1995; Hoegl et al.,1996; Borg-von Zepelin et al., 1998). In
addition to playinga role in adherence, as discussed above,
secreted pro-teinases may aid in tissue destruction and
penetration.
Penetration of tissues brings C. albicans into intimatecontact
with other cellular structures and host moleculeswhich could act as
adhesion receptors. Tissue penetra-tion also often involves
breaching endothelial barrierswith consequential endothelial cell
injury (Filler et al.,1995). Depletion of endothelial cell iron
reduces phago-cytosis of the fungus and results in less cell injury
(Frattiet al., 1998). Phagocytosis and subsequent injury is
alsoreduced by treatment of endothelial cells with gammainterferon
(Fratti et al., 1996). Migration across anendothelial layer is
facilitated by hyphal formation (Zinketal., 1996).
Several potential adhesins and ligands that mediatebinding of
the fungus to endothelial cells have been
identified and may share identity with moleculesinvolved in
epithelial adherence (reviewed by Hostetter,1994; Fukazawa and
Kagaya, 1997; Chaffin et al., 1998).The iC3b binding protein(s)
discussed below contributedto adherence to umbilical vein
endothelium. Two mono-clonal antibodies that recognize human
integrin subunittm partially inhibited adherence (Gustafson et al.,
1991).Adherence to cultured human dermal microvascularendothelial
cells was also reduced by treatment with ananti-human CR3
monoclonal antibody (Lee et al., 1997).Fibronectin has been
observed, by indirect immunofluo-rescence, on endothelial cells as
well as on epithelialcells (see above). Fibronectin was detected on
rabbit aor-tic valves in a model of non-bacterial thrombotic
endo-carditis (Scheld et al., 1985). C. albicans and C.
tropicalis,which are often isolated from infections, bound
signifi-cantly better to fibronectin in vitro than the
infrequentlyisolated C. krusei. During the analysis of adherence
toendothelial monolayers, it was noted that C. albicansadhered
preferentially to subendothelial ECM (Klotz,1987; Klotz and Maca,
1988). The candidal binding pro-teins for iC3b and fibronectin are
discussed below.
(B) ADHERENCE TO ECM AND SERUM PROTEINSC. albicans binds to
several host proteins found in serumand in ECM (Table 2; reviewed
by Hostetter, 1994;Fukazawa and Kagaya, 1997; Sturtevant and
Calderone,1997; Chaffin et al., 1998). Serum proteins that bind to
thefungus include serum albumin, transferrin, fibrinogen,and the
complement C3 fragments, C3d and iC3b(Heidenreich and Dierich,
1985; Bouali et al., 1986; Pageand Odds, 1988). ECM components that
bind to the fun-gus include fibronectin, laminin, entactin,
collagen typeI and type IV, and vitronectin (Skerl et al., 1984;
Boucharaet al., 1990; Klotz, 1990; Jakab et al., 1993; Lopez-Ribot
andChaffin, 1994). While the host ligands have been identi-fied,
less progress has been made with the identificationof the fungal
adhesins, many of which appear to beexpressed more abundantly on
the surfaces of germtubes than on yeast cells (Heidenreich and
Dierich, 1985;Bouali et al., 1986; Bouchara et al., 1990;
Lopez-Ribot andChaffin, 1994). Some adhesins may recognize
multipleligands, since similar-sized proteins have been
identifiedas potential adhesins for several ligands. Differences
inadhesin identification among various studies furthercomplicate
the issue.
The interactions between human serum albumin andtransferrin and
C. albicans have not been studied exten-sively. The two proteins
were observed by indirectimmunofluorescence to bind preferentially
to germtubes (Page and Odds, 1988). More recently, binding of
C.albicans cells to bovine serum albumin, immobilized onmicrotiter
plates, was reported, and binding was affectedby pre-incubation of
the yeast with alanine, proline, and
370 Crit Rev Oral Blo! Med 10(3):359-383370 Crit Rev Oral Biol
Med 10(3):359-383 (1999) by guest on June 1, 2015 For personal use
only. No other uses without permission.cro.sagepub.comDownloaded
from
-
TABLE 2Oral Ligands for C. albicans Adhesion, and Possible
Interventions to Prevent Colonization
Adhesion Ligand Type of Interaction Reference Possible
Interventionsa
BEC Protein-protein, lectin, hydrophobic, Hazen, 1989; Macura
CHOb (e.g., CSEC), peptidefimbrial glycoprotein-BEC and Tondyra,
1989;glycosphingolipid Jimenez-Lucho et al., 1990;
Yu et al., 1 994a,b,c; Imbert-Bernardet al., 1995; Cameron
andDouglas, 1996
Dental acrylic Hydrophobic, glycoprotein-acrylic Klotz et al.,
1985; McCourtie Surface modificationand Douglas, 1985;Minagi et
al., 1985
Oral bacteria Lectin, protein-protein Holmes et al., 1995b,
1996; ?CHOMillsap et al., 1998
Adsorbed salivary proteins Lectin, ?protein-protein Newman et
al., 1996; ?CHOO'Sullivan et al., 1997
iC3b Protein-protein Hosteffer et al., 1990; Alaei Peptideet
al., 1993
ECM proteins: Fibronectin Protein-protein, ?lectin Skerl et al.,
1984; Klotz et al., Peptide1993; Gozalbo et al., 1998;Yan et al.,
1998a
Laminin Protein-protein Bouchara et al., 1990; Gozalbo Peptideet
al., 1998; Yan et al., 1 998a
Entactin Protein-protein L6pez-Ribot and Chaffin, 1994
PeptideCollagen Protein-protein Klotz et al., 1993; Chaffin et al.,
Peptide
1998Vitronectin Protein-protein, lectin Limper and Standing,
1994; ?CHO
Olson et al., 1996a Non-specific interventions and those
involving multiple interactions include antimicrobial mouthwashes.b
Carbohydrate (monosaccharides or oligosaccharides).c Chitin-soluble
extract.
leucine but not with other amino acids (Hawser and (Bouali et
al., 1987; Casanova et al., 1992; Martinez et al.,Islam, 1998).
Human fibrinogen binds extensively to 1994). Binding of C. albicans
to platelets appears to begerm tubes, while binding to yeast cells
appears to mediated via the interaction with fibrinogen (Robert et
al.,depend on growth conditions (Bouali et al., 1987; Page 1996).
Since fibrinogen can be a component of theand Odds, 1988). The
major protein that interacted with enamel pellicle (Kraus et al.,
1973), it may also be a recep-fibrinogen had a molecular mass of 68
kDa and also tor for C. albicans adherence in the oral cavity.bound
to plastic, laminin, and C3d (Tronchin et al., 1988; C. albicans
germ tubes have the ability to rosette anti-Annaix et al., 1990;
Bouchara et al., 1990). A 58-kDa pro- body-sensitized erythrocytes
coated with C3d or iC3btein, encoded by FBP1, is also a
fibrinogen-binding pro- (Heidenreich and Dierich, 1985). Functional
and anti-tein, but it does not appear to recognize the other lig-
genic similarities between the fungal and mammalianands (Casanova
et al., 1992; Lopez-Ribot et al., 1997). The proteins that bind
these moieties have led to frequent58-kDa fibrinogen-binding
protein is modified by cova- use of terminology adopted from
mammalian systems.lent attachment of the ubiquitin polypeptide
(Seputlveda The fungal binding proteins are sometimes referred to
aset al., 1996), contains epitopes or sequences from a type an
integrin analog or CR3 (iC3b binding proteins) andIV collagen
molecule (Seputlveda et al., 1995), and is CR2 (C3d binding
protein). The receptor for iC3b mayexpressed in vivo (L6pez-Ribot
et al., 1996). The fibrino- play a role in adherence to epithelial
and endothelialgen-binding species are distributed heterogeneously
on cells, as discussed above. Antibodies that recognize sub-the
cell surface, as determined by transmission, scan- units of human
integrins which bind iC3b react with thening electron, and confocal
fluorescence microscopy fungal cell surface. Using an anti- aXm
antibody and fluo-
37110(3):359-383 (1999) Crit Rev Oral Biol Med by guest on June
1, 2015 For personal use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
rescence flow cytometry, Gilmore et al. (1988) observed alow
level of receptor expression on yeast cells that wasaffected by the
concentration of glucose in the growthmedium and was
strain-specific. There is no consensuson the identity of the iC3b
binding protein(s), since vari-ous studies implicate different
components, including165-, 130-, 66-, 55-, and 42-kDa proteins
(Eigentler et al.,1989; Hostetter et al., 1990; Alaei et al.,
1993).Oligonucleotide probes based on the postulated similar-ity to
integrins were used to isolate a candidal integrinhomologue Intlp
(Gale et al., 1996). Although it has lim-ited similarity to
integrins, the deduced Intlp sequencecontains a membrane spanning
region and divalentcation-binding motifs. Expression of the gene in
S. cere-visiae resulted in aberrant cell morphology with
formationof germ-tube-like structures (Gale et al., 1998). A
mutantstrain deficient in expression of this gene exhibitedreduced
binding to HeLa cells, reduced hyphal formationunder some
conditions, and reduced virulence in mice.Several hypotheses on the
role of iC3b- and C3d-bindingin the pathogenesis of candidiasis
have been proposed."Bystander" deposition of complement fragments
onerythrocytes following activation of the alternate com-plement
pathway by the fungus may promote fungalrosetting of coated
erythrocytes. This could provide fun-gal access to erythrocyte iron
through candidal surfacehemolysin-mediated erythrocyte lysis (Manns
et al.,1994). Also, binding iC3b may mask the cells and
preventphagocytosis, as discussed above.
Two major C3d-binding components have been iden-tified, a 60-kDa
moiety expressed on the surfaces of germtubes and a 50-kDa moiety
in the yeast plasma mem-brane (Calderone et al., 1988; Linehan et
al., 1988). Inother studies, antibodies to the major C3d-binding
pro-tein (60-kDa species) reacted with several components.These
molecules included a major 50- to 60-kDa moietyand minor species of
94, 67-68, 60, 50, 40, 31, and 20 kDa(Kanbe et al., 1991; Franzke
et al., 1993; Lopez-Ribot et al.,1995). The C3d receptor is
expressed, in vivo, on orga-nisms recovered from peritoneal lavage
and in kidneysfrom infected mice (Kanbe et al., 1991). A role for
C3d-binding in pathogenesis is mostly speculative.Hypotheses
include: that aggregation of opsonized andunopsonized cells could
protect the fungus from phago-cytosis; a role in iron acquisition,
as suggested above foriC3b binding; or binding to any host cell on
which C3dwas deposited.
The ability of C. albicans to bind to the ECM
ligandsfibronectin, laminin, entactin (nidogen), types I and
IVcollagen, and vitronectin has been the focus of numer-ous studies
in the last decade (for recent reviews, seeFukazawa and Kagaya,
1997; Sturtevant and Calderone,1997; Chaffin et al., 1998). Some of
these ECM compo-nents are able to form complexes among
themselves,
and, thus, ECM presents a host target with multiplebinding sites
for the fungus. The identity of the adhesinsthat recognize ECM
components is unresolved and thesubject of some disagreement
between studies.Fibronectin is found in both plasma and ECM and
wasone of the first host proteins identified as a ligand pro-moting
C. albicans adherence (Skerl et al., 1984). Species of62 kDa and 72
kDa have been isolated by affinity chro-matography as adhesin
candidates for fibronectin andcollagen (Klotz et al., 1993). The
expression of fibronectin-binding capacity is regulated in part by
environmentalconditions. Growth medium and temperature can alterthe
extent of binding (Jakab et al., 1993; Negre et al.,
1994).Recently, growth in the presence of hemoglobin has
beenreported to induce enhanced expression of a 55-kDapromiscuous
adhesin that recognized fibronectin,laminin, and fibrinogen (Yan et
aci., 1998a).Glyceraldehyde-3-phosphate dehydrogenase, a
33-kDaprotein present predominantly on the yeast cell surface,is
also a fibronectin- and laminin-binding protein(Gozalbo et al.,
1998). Laminin adhesins of 68 kDa and 62kDa have been identified by
ligand affinity blotting(Bouchara et al., 1990). These receptors
may be the sameas the proteins binding to plastic (Tronchin et al.,
1988)and fibrinogen (Annaix et al., 1990) and, indeed,
lamininbinding was reduced competitively by fibrinogen(Bouchara et
al., 1990). A 37-kDa receptor for lamininthat appeared not to
recognize other ligands was firstdetected with an antibody produced
to a human high-affinity laminin receptor (Lopez-Ribot et al.,
1994). Threeproteins (65 kDa, 44 kDa, 25 kDa) were detected, by
lig-and affinity blotting, as candidates for entactin
bindingproteins (Lopez-Ribot and Chaffin, 1994). Binding of
cellwall protein to immobilized entactin was reduced byfibronectin
and laminin. Vitronectin appears to bindboth to cell wall protein
(30-kDa moiety; Limper andStanding, 1994) and to 13-glucan (Olson
et al., 1996).Fibronectin inhibited binding of vitronectin to the
fungus(Jakab et al., 1993).
Several studies have reported that peptides with theRGD
(arginine-glycine-aspartic acid) or related motifwere inhibitors of
binding between fungal adhesins andECM components (Klotz et al.,
1992; lakab et al., 1993;Lopez-Ribot and Chaffin, 1994), while
other studies havesuggested that the RGD motif is not a contributor
to theinteractions (Negre et al., 1994; Yan et al., 1998b).
Parallelsbetween fungal ECM binding proteins and mammalianintegrins
that recognize ECM ligands and the RGD motifraise the possibility
of structural and antigenic related-ness. Antibody to the a531
mammalian integrin andmonoclonal antibodies to each subunit reacted
with C.albicans, and the reactivity increased as yeast cells
formedgerm tubes (Santoni et al., 1994). Both high- and
low-affinity receptors have been detected, and the number of
372 Grit Med 10(3):359-383 (1999)372 Crit Rev Oral Biol Med by
guest on June 1, 2015 For personal use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
binding sites per C. albicans cell ranges from 5000 to98,000
(Bouchara et al., 1990; Limper and Standing, 1994;Negre et al.,
1994). In vivo evidence for a role of adhesionto fibronectin or any
other ECM ligand in pathogenesis isscant. In a rabbit model of
infection, administration of anRGD-containing peptide reduced
fungal counts in sever-al organs (Klotz et al., 1992). The
promiscuity of adhesins,the involvement of multiple adhesins in
adherence, andthe presence of some host ligands, such as
fibronectin,in both superficial and deep sites suggest that there
isnot a strict separation between the adhesins and mech-anisms the
fungus uses to colonize cutaneous andmucocutaneous surfaces and
that those that areinvolved in deep tissue invasion.
(7) Clinical Significance of Colonizationand Its Prevention
(A) INCIDENCE OF CANDIDIASISThe balance between colonization and
mucosal candidia-sis (Fig. 1) depends on the effectiveness of the
hostdefenses. A corollary of this is that candidiasis is often
anindication of an underlying immune deficiency. Oral can-didiasis
affects a large proportion of HIV-positive individ-uals and those
with AIDS (Greenspan and Greenspan,1996; Darouiche, 1998), and
approximately 90% of AIDSpatients have suffered from oropharyngeal
or esophagealcandidiasis at some stage of their illness (Alexander
andPerfect, 1997) Oropharyngeal candidiasis is also oftenseen in
the elderly, and merits investigation of potentialpre-disposing
factors (Shay et al., 1997). A comprehensiveanalysis of the
literature concerning the treatment ofimmunocompromised patients
with oropharyngeal oresophageal candidiasis indicates that optimal
results areachieved with the triazoles fluconazole and
itraconazole(Darouiche, 1998). They offer clinical efficacy at
least com-parable with that of the polyenes and imidazoles,
togeth-er with a highly favorable mycological cure rate.
During the 1980s and 1990s, the frequency of nosoco-mial
candidiasis has increased dramatically (Beck-Sagueand Jarvis, 1993;
Pfaller, 1995, 1996). In a study of datafrom the USA National
Nosocomial InfectionsSurveillance System, C. albicans was the most
frequentlyisolated fungal pathogen (59.7%) in hospital
environ-ments (Beck-Sague and Jarvis, 1993). Transfer of
Candidabetween individuals often occurs via the hands of healthcare
workers (Strausbaugh et al., 1994), and nosocomialtransmission can
occur without candidiasis outbreaks(Schmid et al., 1995a). This
reinforces the need for appro-priate cross-infection control in the
dental surgery.
(B) RECURRENCEA common feature of candidiasis is recurrence.
Factorswhich favor recurrence include: re-inoculation from
colo-
nized individuals and the environment; underlyingimmune
suppression; and endogenous reservoirs forreinfection. We have
discussed the ease with which C.albicans can enter the oral cavity
and the central impor-tance of immune suppression in the
development of can-didiasis. There can also be reservoirs for
re-infectionwithin the oral cavity. C. albicans can be isolated
fromplaque (Arendorf and Walker, 1980), oral biofilms (Kaylaand
Ahearn, 1995; Hawser, 1996), and occluded acrylicdenture surfaces,
where it will be relatively well-protect-ed from topical antifungal
agents but could be releasedinadvertently by routine dental hygiene
procedures andlead to colonization of other oral sites.
(C) CANDIDA SPECIES AND DRUG RESISTANCEAlthough C. albicans
remains the most common cause ofnosocomial candidiasis, infections
due to non-albicansspecies are increasing (Pfaller, 1996). Some of
thesespecies, namely, C. glabrata and C. krusei, are
intrinsicallyless sensitive to azole drugs. Another interesting
trend isthe increased association of a novel Candida species,
C.dubliniensis, with oral infection in AIDS patients (Coleman
etal., 1997; Sullivan and Coleman, 1998). C. dubliniensis isclosely
related to C. albicans (Coleman et al., 1997), and aswith C.
albicans, stable fluconazole resistance can beinduced by exposure
to the drug (Albertson et al., 1996;Moran et al., 1997). Treatment
of AIDS patients with pro-longed courses of azole antifungal agents
appears to haveselected for the development of azole-resistant C.
albicansstrains (White, 1997a; Darouiche, 1998). Resistance can
bedue to mutations in the drug target (White, 1997b;Sanglard et
al., 1998) or over-expression of drug effluxpumps (Sanglard et al.,
1995; Albertson et al., 1996). Over-expression of drug pumps from
the ATP binding cassette(ABC) family of efflux pump can lead to
cross-resistance toseveral azole antifungals (Albertson et al.,
1996; Niimi et al.,1997b). Most azole-resistant strains, however,
retain sen-sitivity to amphotericin B (Albertson et al., 1996).
(D) PROSPECTS FOR STRATEGIESTO PREVENT COLONIZATION
An attractive alternative to treating patients with
can-didiasis, which is often recurrent, is to prevent the
infec-tion from occurring. This could be achieved by increasingthe
concentration of natural anti-candidal compounds inthe mouth or by
preventing adherence. Denture resin hasbeen modified to enhance
adsorption of the anti-candi-dal salivary protein histatin 5
(Edgerton et al., 1995).Although adsorbed histatin 5 did not have
an anti-candi-dal effect, desorbed histatin did, and modified
dentureacrylic loaded with histatin 5 could provide a
localizedcontrolled release of the protein. There is also
theprospect of using gene therapy to over-express histatinsin
saliva (O'Connell et al., 1996).
10(3) 359 383 (1999) Crit Rev Oral Biol Med 37310(3):359-383
(1999) Crit Rev Oral Biol Med 373 by guest on June 1, 2015 For
personal use only. No other uses without
permission.cro.sagepub.comDownloaded from
-
Another approach is the prevention of colonizationby inhibiting
C. albicans adherence. This could beachieved by immunizing the host
or by physical interfer-ence with adherence mechanisms (Table
2).Theoretically, the application of soluble receptors, lig-ands,
or the domains of these molecules involved inadherence could be
used to prevent microbial coloniza-tion (Mandel, 1996). There is
evidence that chitin isinvolved in C. albicans adherence, and a
chitin-solubleextract (CSE) has been used to inhibit yeast adhesion
toa variety of cells (Segal, 1996). Although most
adhesioninhibition experiments have been carried out in vitro,
ini-tial in vivo results appear to be promising (Segal, 1996).The
oral environment is rich in proteinases, and any useof peptides to
inhibit adhesion may necessitate frequentapplication. Other
approaches to prevent microbial colo-nization include immunization
(lenkinson and Lamont,1997). Passive immunization with recombinant
plantmonoclonal secretory antibodies to an S. mutans adhesinhas
been used to inhibit specific microbial colonizationin humans (Ma
et al., 1998). As significant adhesins in C.albicans are
identified, this approach could also be usedto preclude
candidiasis. Salivary IgA antibodies havebeen shown to reduce the
adherence of C. albicans cells toBECs (Epstein et al., 1982;
Challacombe, 1994), and thestimulation of a mucosal immune response
with a C. albi-cans adhesin, possibly expressed by another resident
oralmicrobe, may prevent colonization.
The inhibition of S. mutans colonization is being con-sidered
for the prevention of caries (Ma et al., 1998), aprocess which will
make a significant, and possibly dele-terious, change to the oral
flora. C. albicans cells, howev-er, are present in low numbers, and
so removal is notlikely to have adverse effects on the remaining
flora.From this review, it is evident that C. albicans can utilize
anumber of adherence mechanisms, and it is far from cer-tain that
inhibiting any one will prevent colonization. Inaddition, C.
albicans embedded in plaque may be protect-ed from anti-adhesive
compounds. However, the prophy-lactic treatment of susceptible
individuals with thesecompounds after aggressive dental hygiene
might pre-vent the incorporation of C. albicans into plaque and
thedevelopment of a protected reservoir.
(8) ConclusionC. albicans is truly an opportunistic organism. It
is themost frequent cause of candidiasis because it is the
mostsuccessful yeast at colonizing the oral cavity and so isoften
in a position to take advantage of immune sup-pression in the host.
C. albicans is the most adherentCandida species, which is probably
due to its ability toadhere to many different ligands (Fig. 2b,
Table 2). It pos-sesses other virulence factors, such as the
secretion ofhydrolytic enzymes and the ability to evade the
immune
system, which give it a growth advantage over otheryeast.
Subsets of these virulence factors may be activat-ed by changes in
the environment or contact with sur-faces. In colonized
individuals, however, C. albicans is usu-ally present in the oral
cavity only in relatively low num-bers. This indicates that there
is a fine balance betweencolonization and clearance, and there is
the prospect,with further analysis of major adherence interactions,
fortipping this balance in favor of clearance.
AcknowledgmentsRDC acknowledges financial support from the New
Zealand LotteriesBoard and the University of Otago. WLC
acknowledges support fromUS Public Health Service grants A123416
and A140675. We aregrateful to Dr. Ann Holmes for helpful
discussions.
REFERENCESAbu-el Teen K, Ghannoum M, Stretton Rl (1989).
Effects
of sub-inhibitory concentrations of antifungal agentson
adherence of Candida spp. to buccal epithelial cellsin vitro.
Mycoses 32:55 1-562.
Al-Hashimi I, Levine Ml (1989). Characterization of in
vivosalivary-derived enamel pellicle. Arch Oral Biol
34:289-295.
Alaei S, Larcher C, Ebenbichler C, Prodinger WM,Janatova J,
Dierich MP (1993). Isolation and biochem-ical characterization of
the iC3b receptor of Candidaalbicans. Infect Immun
61:1395-1399.
Albertson GD, Niimi M, Cannon RD, Jenkinson HF (1996).Multiple
efflux mechanisms are involved in Candidaalbicans fluconazole
resistance. Antimicrob AgentsChemother 40:2835-2841.
Alexander BD, Perfect JR (1997). Antifungal resistancetrends
towards the year 2000. Drugs 54:657-678.
Alloush HM, Lopez-Ribot JL, Chaffin WL (1996). Dynamicexpression
of cell wall proteins of Candida albicansrevealed by probes from
cDNA clones. J Med Vet Mycol34:91-97.
Annaix V, Bouchara I-P, Tronchin G, Senet J-M, Robert R(1990).
Structures involved in the binding of humanfibrinogen to Candida
albicans germ tubes. FEMSMicrobiol Immunol 2:147-154.
Arendorf TM, Walker DM (1979). Oral candidal popula-tions in
health and disease. Br Dent J 147:267-272.
Arendorf TM, Walker DM (1980). The prevalence andintra-oral
distribution of Candida albicans in man. ArchOral Biol 25:1-10.
Bailey A, Wadsworth E, Calderone RA (1995). Adherenceof Candida
albicans to human buccal epithelial cells:host-induced protein
synthesis and signaling events.Infect Immun 63:569-572.
Beck-Sague CM, Jarvis WR (1993). Secular trends in the
374 Crit Rev Oral Biol Med 1O(3):359-383374 Crit Rev Oral Biol
Med 10(3):359-383 (1999) by guest on June 1, 2015 For personal use
only. No other uses without permission.cro.sagepub.comDownloaded
from
-
epidemiology of nosocomial fungal infections in theUnited
States, 1980-1990. 1 Infect Dis 167:1247-1251.
Bendel CM, Hostetter MK (1993). Distinct mechanisms ofepithelial
adhesion for Candida albicans and Candidatropicalis. I Clin Invest
92:1840-1849.
Bendel CM, St Sauver J, Carlson S, Hostetter MK
(1995).Epithelial adhesion in yeast species: correlation
withsurface expression of the integrin analogue. I Infect
Dis171:1660-1663.
Bibel DJ, Aly R, Lahti L, Shinefield HR, Maibach HI(1987).
Microbial adherence to vulvar epithelial cells.I Med Microbiol
23:75-82.
Borg M, Ruchel R (1988). Expression of extracellular
acidproteinase by proteolytic Candida spp. during experi-mental
infection of oral mucosa. Infect Immun 56:626-631.
Borg-von Zepelin M, Beggah S, Boggian K, Sanglard D,Monod M
(1998). The expression of the secretedaspartyl proteinases Sap4 to
Sap6 from C. albicans inmurine macrophages. Mol Microbiol
28:543-554.
Borromeo GL, McCullough MI, Reade PC (1992).Quantitation and
morphotyping of Candida albicansfrom healthy mouths and from mouths
affected byerythematous candidosis. J Med Vet Mycol 30:477-480.
Bouali A, Robert R, Tronchin G, Senet J-M (1986). Bindingof
human fibrinogen to Candida albicans in vitro: a pre-liminary
study. I Med Vet Mycol 24:345-348.
Bouali A, Robert R, Tronchin G, Senet IM (1987).Characterization
of binding of human fibrinogen tothe surface of germ-tubes and
mycelium of Candidaalbicans. J Gen Microbiol 133:545-551.
Bouchara JP, Tronchin G, Annaix V, Robert R, Senet IM(1990).
Laminin receptors on Candida albicans germtubes. Infect Immun
58:48-54.
Branting C, Sund ML, Linder LE (1989). The influence
ofStreptococcus mutans on adhesion of Candida albicans toacrylic
surfaces in vitro. Arch Oral Biol 34:347-353.
Buck ID, Bubucis PM, Combs Ti (1977). Occurrence
ofhuman-associated yeasts in bivalve shellfish fromLong Island
Sound. Appl Environ Microbiol 33:370-378.
Busscher HI, Geertsema-Doornbusch GI, van der Mei HC(1997).
Adhesion to silicone rubber of yeast and bac-teria isolated from
voice prostheses: influence of sali-vary conditioning films. J
Biomed Mat Res 34:201-209.
Buurman ET, Westwater C, Hube B, Brown AJP, Odds FC,Gow NAR
(1998). Molecular analysis of CaMntlp, amannosyl transferase
important for adhesion and vir-ulence of Candida albicans. Proc
Natl Acad Sci USA95 7670 7675.
Caesar-TonThat T-C, Cutler JE (1997). A monoclonal anti-body to
Candida albicans enhances mouse neutrophilcandidacidal activity.
Infect Immun 65:5354-5357.
Calderone RA, Linehan L, Wadsworth E, Sandberg AL(1988).
Identification of C3d receptors on Candida albi-cans. Infect Immun
56:252-258.
Cameron Bl, Douglas LI (1996). Blood group glycolipidsas
epithelial cell receptors for Candida albicans. InfectImmun
64:891-896.
Cannon RD, Holmes AR, Mason AB, Monk BC (1995a).Oral Candida:
clearance, colonization, or candidiasis?I Dent Res 74:
1152-1161.
Cannon RD, Nand AK, lenkinson HF (1995b). Adhere