Journal of Medical and Veterinary Mycology (1992), 30,219-231
Fatal systemic candidiasis of gastrointestinal origin: an experimental model in mice compromised by anti-cancer treatment H. S A N D O V S K Y - L O S I C A
t, L. B A R R - N E A 2 AND E. SEGAL t
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
1Department of Human Microbiology, and 2Department of Histology and Cell Biology, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel (Accepted 3 February 1992)
An experimental model of fatal systemic candidiasis originating from the gastrointestinal (GI) tract of compromised mice is presented. ICR female mice were compromised by a single anti-cancer treatment: irradiation (4 or 6 Greys i.e. 400~500 rads), methotrexate (MTX) (3 mg per mouse, intraperitoneally) or 5-fluorouracil (SFU) (200 mg kg -t, intravenously). Three days later, compromised and non-treated control mice were exposed to Candida albicans administered orally. Morbidity and mortality due to candidiasis were monitored for 30 days post-candidal inoculation. Increased and longer GI colonization was noted among the MTX and 5FU treated mice, or 6 Greys irradiated mice (up to 92.3% for over 30 days in anti-cancer treated mice). The stomach was found to be the major part of the GI tract involved in fungal colonization. A significant number (53.8-83.3%) of the anti-cancer treated mice developed systemic candidiasis originating from the GI tract, which was fatal in 30-80% of the infected animals. In systemically infected animals, candidal antigen was demonstrated in the serum, and fungal abscesses containing C. albicans were observed in the liver, kidneys and spleen. C. albicans was isolated from the infected organs. The severity of the infection, as reflected by the number of fungi in visceral organs, and by mortality during the 30 days post-candidal inoculation, indicated differences in the course and nature of the infection among the three treatment groups (i.e. MTX, 5FU, 6 Greys). The increased fungal association with the GI mucosa appears to be one of the factors leading to proliferation of the fungus in the GI tract, which may possibly be followed by dissemination to visceral organs, causing systemic candidiasis and mortality.
Candida albicans is considered to be a commensal of the gastrointestinal (GI) tract [5, 29, 36] of the normal population, however in compromised individuals, such as cancer patients, fungal Gi colonization increases significantly [5, 29, 35]. Thus, it is generally accepted that in such individuals, the GI tract may be a possible portal of entry for the organism into the viscera, leading to systemic candidiasis [13, 28, 38]. The magnitude of the problem is illustrated by the vast number of reports [1, 2, 18, 22, 24, 27, 38], describing human systemic candidiasis and involvement of the GI tract in severely compromised patients, particularly those with haematologic malignancies. As it is generally believed that microbial attachment to host tissues is essential for the development of infection [30], we focused our investigations on the importance of this phenomenon in compromised hosts. In recent studies we have investigated the in vitro attachment of C. albicans to GI mucosal surfaces [32, 33]. It was noted that fungal adherence to GI mucosa from compromised mice was significantly increased when compared with that to tissues from naive animals. Consequently, we initiated the present Correspondence address: Prof. E. Segal, Department of Human Microbiology, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel. 219
220
SANDOVSKY-LOSICA E T A L .
in vivo study, investigating systemic candidiasis originating from the GI tract of mice
treated by irradiation, methotrexate (MTX) or 5-fluorouracil (5FU). METHODS
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Growth and preparation of C. albicans for animal inoculation C. albicans strain CBS 562 (Centraalbureau voor Schimmelcultures, Delft, The Netherlands) was used throughout this study, as in previous experiments [32-34]. The fungus was grown on a gyrotory shaker in yeast extract broth (0.5 % Bacto yeast extract; 3% glucose) at 28°C for 18 h. Yeast cells were harvested by centrifugation (3500 rev. min -1) washed three times in sterilized distilled water and resuspended to the desired concentration.
Anti-cancer treatments
Six-week-old, ICR female mice (mean weight 25-28 g), maintained under conventional conditions, five mice per cage, were subjected to one of the following anti-cancer treatments. Irradiation. Mice were exposed to a nonqethal, single dose (whole body) of 4 Greys (48
mice) or 6 Greys (99 mice) of Cobalt irradiation. M T X treatment. MTX (Abic Ltd. Israel) was injected intraperioneally (131 mice) at a
final concentration of 3 mg per mouse. (Six-percent of the mice died due to the treatment on days 1-2 post-treatment; these animals were excluded from the experiment.) 5FU treatment. 5FU (Abic Ltd. Israel) was injected intravenously (121 mice) into the
tail vein with a non-lethal dose of 200 mg kg- 1body weight. Immunosuppression of the treated animals was monitored by the decrease in white blood cell counts and spleen weights, as described in detail in previous studies [32, 33]. Non-treated mice served as controls. Challenge of the animals with C. albicans
Drinking water (sterile distilled water) supplemented with 5% sucrose, containing 5 × 107 C. albicans m1-1 was administered to the compromised mice for 24 h, on the 4th day post-treatment; control mice (with no pre-treatment) were also challenged orally with the yeast. The mean calculated inoculum per mouse was 2-5 × 108 organisms (based on an average water consumption of 5 ml per mouse). An additional control group of mice treated with the various anti-cancer treatments and not challenged with C. albicans was also included. Following candidal challenge, animals were surveyed for morbidity and mortality for 30 days. On different days post-candidal challenge, animals were sacrificed (by cervical translocation) and morbidity was assessed by: (i) enumeration of C. albicans in the GI tract and visceral organs, and (ii) histological examination of GI and visceral organs. Biood samples from several animals were tested for the presence of candidal antigen using a commercial latex agglutination kit (Difco, USA).
FATAL SYSTEMIC CANDIDIASIS
221
Mortality was monitored for 30 days post-candidal challenge. Post-mortem studies were carried out in some of the experiments on mice that had died during the 30 day monitoring period. Liver and kidneys were examined for the presence of fungal abscesses and were cultured to detect C. albicans.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Enumeration of C. albicans in the GI tract and visceral organs The GI tract of the sacrificed animals was removed, rinsed with sterile distilled water and segmented into the esophagus, stomach, small intestine and cecum. Tissue disks (0.7 cm in diameter) were cut with a sterile metal punch from each segment (the stomach tissue segments always included parts of secreting, non-secreting and cardial atrium line tissue). The esophagus was homogenized as a whole organ (due to its size). Part (upper left lobe) of the liver (approximately 300 mg) and the whole kidneys (approximately 300 mg x 2) were also removed. GI tissue disks or visceral organs were homegenized separately in sterile distilled water (1 ml) using a tissue homogenizer (Polytron-Kinematica, GmbH, Littau, Germany). Diluted samples of the homogenates were plated on Sabouraud glucose agar containing chloramphenicol and incubated at 28°C for 24-48 h. Colonies were identified by microscopy and other conventional techniques for the identification of C. albicans [8]. Growth of more than 10 colony forming units c.f.u, per 0-1 ml of tissue homogenate (or organ, i.e. esophagus) from at least one part of the GI tract was indicative for fungal GI colonization. This strict criterion was chosen to avoid consideration of random growth of a few colonies. The same strict criterion as for GI colonization was chosen for systemic infection. Growth of >10 colonies per 0-1 ml of liver or kidney homogenate was considered an indication of systemic infection. Histological examination Segments of the GI tissues, liver and kidneys were fixed, embedded in paraffin and sections (6/urn) cut and stained with Periodic Acid Schiff (PAS) for light microscopy. Unstained or calcofluor (0-01%, Polyscience Inc. USA) stained [26] tissue homogenates were also used for direct examination under light and fluorescent microscopes (UV light 340-460 nm range), respectively, for detection of fungal elements. Statistical analyses Colonization or systemic infection values were analysed by the Fisher exact test for differences between treated and non-treated groups [10]. Mortality slopes were analysed using two statistical tests: Breslow [3] and Mantel [23]. Confidence levels of at least P < 0.05 were considered significant. RESULTS Candida GI colonization in compromised mice The percentage GI colonization among a total of 70 MTX, 60 5FU and 94 irradiated (4 Greys - 48, or 6 Greys - 46) mice compared with that of 103 non-treated animals was assayed on various days (days 1-2; 3-4, 6-7; 13-14; 21; 30) post-candidal challenge. A higher percentage (up to 92%) of compromised mice were colonized with C. albicans,
222
SANDOVSKY-LOSICA E T A L . Mice 9"9% Control
Ii i i
20.2%
MTX- treoted
21.2%
5FU-treoted
11.3%
57"7%
57-7%
195%
22.8%
36-5%
20-0%
22"3%
72.4%
20-7%
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
21,1%
5-5% 5"5%
Irradiated 4Greys 19.6% Irradiated 6 Greys
51,8%
10.7%
N
17.9%
i i
Esophog l---1Stomach ~ l n t e s t , n e
I~Cecum
FIG. 1. Distribution of C. albicans isolations from the GI tract of irradiated, MTX, 5FU treated and nontreated mice. The data represent the % of C. albicans isolation from various parts of the GI tract of treated and non-treated mice, calculated as proportions of the total number of candidal isolations from the GI tract of each treatment group. The data were analysed by the Fisher exact test; P < 0.05 was considered significant. The differences between fungal isolations from the stomach, small intestine or esophagus of MTX or 5FU treated animals and isolations from the corresponding organs of non-treated mice were significant.
40
2O (a) Nmve m,ce
(b) Irradloted mLce o
30 20
Io
o
7 g ~n g z-
u~
I0
tl
o
g ,
I
i
O
.11 0
0
-
Dqys post d'~nd/do inoculahon
I0 1~0 50 DQys post Cond/dG ~noculutlon
4Greys 0 6Greys g co co
80
co co
c) 5-Fu freqted mtce
--. 400 " (d) MTX lreoted rnlce
60
6 500
40
200
2C
tO0
+ +
0
I0 20 Dqys post L?Lz'd,'do inoculohon
~0
0
'
I0 20 ' Days pos~ C~Tnd/do inoculo'~lon
3'0
FIG. 2. Enumeration of C, albicans c.f.u, in stomach tissue of irradiated, MTX, 5FU treated and non-treated mice. The data represent the number of c.f.u, per tissue at various time intervals post-candidal challenge. Each point indicates an individual animal.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
223
although colonization was also noticed among naive mice (up to 75%). Furthermore, while in compromised mice colonization lasted for at least 21 days post-candidal challenge and for some treatments (5FU or irradiation) over the survey period of 30 days, in normal mice candidal colonization lasted for only 2 weeks. Cultures from non-challenged mice were negative for C. albicans. Figure 1 shows the distribution of Candida in the GI tract. The data represent the relative frequency (in %) of fungal colonization in each of four parts of the GI tract: esophagus, stomach, small intestine and cecum, calculated from pooled data from each group of mice during the 30 days post-candidal challenge. It can be seen that the stomach is the main site of colonization by Candida in normal as well as in compromised mice. However, the frequency of fungal isolation from the stomach was lower from MTX or 5FU treated mice than from control animals. Furthermore, in MTX or 5FU treated mice the relative frequencies of candidal colonization in the esophagus or the small intestine were higher then in normal animals (the differences were statistically significant, P < 0.05, Fisher exact test). Following the observation that the stomach serves as the main part of the GI tract for C. albicans colonization, we compared the degree of colonization in this organ among animals of the various groups quantitatively, by enumerating the candidal c.f.u. in the stomach tissue. As shown in Fig. 2 (a-d), it can be seen that the level of c.f.u, in the stomach tissue of MTX or 5FU treated or 6 Greys irradiated mice, was higher than in non-compromised animals, or in those irradiated with a low dose (4 Greys). The data in this figure show the distribution of c.f.u, among the individual mice of the different groups. Calculated maximal means (not shown) reveal even more pronounced differences between compromised and non-compromised animals. Specifically, the maximal mean c.f.u, per stomach tissue disk of control and 4 Greys irradiated mice was 1.8 x 103 and 1.7 x 103, respectively, while that of MTX and 5FU treated or 6 Greys irradiated mice was 4.3 X 10 4, 1.2 × 10 4 or 3"6 × 10 3, respectively.
Course of systemic candidiasis and mortality in compromised mice The percentage of mice with systemic candidiasis during the 30 day follow-up period, based on isolation of Candida from viscera is presented in Table 1. Among the compromised mice, a large proportion (maximal value, 83%) of animals developed systemic infection, while among the non-compromised mice there was a very low percentage of systemic involvement. It was also noted that in the MTX treated mice, candidal infection was acute while in 5FU treated, or irradiated mice (6 Greys) it apparently had a longer course. It should be emphasized that in about 14 days most of the MTX treated mice succumbed to the infection (Fig. 3), therefore the small animal numbers appearing in Table 1 on day 21 and beyond represent only those animals which did not, apparently, develop systemic infection. Irradiation with 4 Greys resulted in a low percentage of systemically infected animals (maximal value, 25 %), but a higher level of irradiation (6 Greys), resulted, as indicated before, in GI colonization with chronic systemic infection. In some of the compromised animals Candida antigenemia was detected in undiluted animal sera using a commercial latex agglutination kit. Table 2 summarizes the mean number of fungi isolated from visceral organs (liver or kidneys) and the number of fungal isolations out of the total number of mice screened. It can be seen that, among the non-treated mice or those irradiated at the
224
SANDOVSKY-LOSICA E T A L . TABLE 1. Systemic candidal infection in compromised mice following oral challenge with C. albicans %
Day a
Non-treated
MTX
5FU
Irr 4 Greys
Irr 6 Greys
0 (0/10) 0 (0/20) 10.3 (3/29) 12.1 (4/33) 0 (0/6) 0 (0/5)
41.7 c (5/12) 61.9¢ (13/21) 8.3 (1/12) 23-1 (3/13) 0 (0/8) 0 (0/4)
61-6c (8/13) 61-9~ (13/21) 68-7c (11/16) 75" (6/8) 83.3 (5/6) 16.7 (1/6)
25 (1/4) 7.7 (1/13) 0 (0/13) 0 (0/12) n.d.
0 (0/4) 23.1 (3/13) 53-8 ¢ (7/13) 33.3 (4/12) n.d.
0 (0/6)
25 (1/4)
1-2 3-4 6-7 Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
with systemic infection (No. infected mice/totalb)
13-14 21 30
a Days post-candidal challenge. b Number of mice with systemic infection (based on isolation of Candida from viscera) per total number of dissected animals on a given day. c Significant differences, P < 0.05, Fisher exact test. n.d. = Not done.
level of 4 Greys, the number of fungal isolations was not significant (6/103 and 2/48, respectively). In addition, even in the animals with infected viscera, the mean number of candidal c.f.u, isolated per organ was low (0.2-1 × 103). In contrast, in compromised mice C. albicans was isolated from visceral organs from a significant number of animals (14/70; 41/70; 13/46). Moreover, the number of candidal c.f.u, isolated from visceral
~-
~
I00
-
80
-
n
O-
X~-~X~=61
=
61
4o' m~Fi~4~li-I-ill-i-n-Ig-lg-i
/
I
I
,mm ~=~s -i-II-II-II
0
I0 20 Doys post E. o/b/cans chollenge
ZO
- i - Irrod~oted 6Greys -0- MTX -X- 5FU
FIG. 3. Mortality of irradiated, MTX, 5FU treated and non-treated mice. Data are pooled values from several 30 day follow-up surveys, n = total number of animals (at the beginning of the follow-up). Mortality data of the three groups were analysed by the Breslow [3] and Mantel [23] tests, and the mortality plots of the three groups were found to be significantly different (P < 0.001). The calculated mean survival times were: irradiated mice 23.9 days: MTX treated mice 9-4 days; 5FU treated mice, 12-8 days.
225
FATAL SYSTEMIC CANDIDIASIS TABLE 2. Candida isolation from visceral organs Mean Candida c.f.u. × 103 per organ a (No. of mice with Candida/total) h
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Organ
Non-treated
MTX
5FU
Irr 4 Greys
Irr 6 Greys
Liver (300 mg)
0-4 (6/103)
2.0 (19/70)
7-6 (37/70)
0.2 (2/48)
67-1 (12/46)
Kidneys (300 mg)
1.0 (1/103)
2-7 (14/70)
3900 (41/70)
0 (0/48)
2500 (13/46)
a Mean values. b Total number of mice from which Candida was isolated from the kidneys or liver per total number of animals screened.
organs of these animals was significantly higher (from 2-7-4000 times higher than the value of the controls). Furthermore, the data in Table 2 point to differences in the severity of infection among animals of the various groups. Specifically, in MTX-treated mice the level of systemic infection, as judged by the mean of candidal c.f.u., as well as by the number of infected animals, was lower than in 5FU treated or 6 Greys irradiated mice. Figure 3 illustrates the mortality data of compromised mice post-candidal challenge. No animals among the normal mice or those irradiated by low dose (4 Greys) died during the monitoring period. In mice irradiated with the high dose (6 Greys), or in MTX or 5FU treated animals the systemic candidal infection was fatal, and mortality was associated with high morbidity as judged by c.f.u, counts from visceral organs (Table 2). About 80% of the mice compromised by MTX or 5FU, and about 30% of the irradiated (6 Greys) animals succumbed to systemic candidiasis during the 30 days observation period. According to statistical analysis the mean survival time (MST) (for MTX treated mice, 9-4 days; for 5FU treated mice, 12-8 days; and for 6 Greys irradiated mice, 23-9 days) as well as the mortality plots of the three groups of compromised mice were statistically different. Histological examination of infected tissues Gross examination of the GI tract of the compromised mice post-challenge revealed macroscopic changes in the mucosa of the stomach, especially in the cardial atrium line (Fig. 4a, b - 6 Grey irradiated mouse). Examination of viscera revealed macroabscesses in the kidneys, liver and spleen of these animals (Fig. 5a - liver of 5FU treated mouse). Direct examination of calcoftuor stained or unstained tissue homogenates from the stomach or visceral organs of compromised mice revealed fungal elements, including hyphae and budding yeasts (Fig. 6a, b). Tangential sections of stomach tissues (stained by PAS), from compromised mice (Fig. 4c - 6 Greys irradiated mouse) showed large amounts of hyphal elements associated with the squamous epithelium in the region of the cardial atrium line. Deeper sections of the stomach tissues (in the sub-mucosa) revealed invasion of fungal elements Sections of liver and kidneys (stained by PAS) from infected, compromised mice showed large abscesses composed mainly of hyphal elements (Fig. 5b, c-liver of 5FU treated mouse).
SANDOVSKY-LOSICA E T A L .
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
226
FIG. 4. Gross pathology and histopathology of stomach mucosa (inner part) of 6 Greys irradiated mouse inoculated with C albicans. (a) Stomach mucosa from an irradiated (6 Greys) mouse inoculated with C. albicans. Arrow indicates cardial atrium line, note pathological changes. (b) Stomach mucosa from a naive mouse. Arrow points to cardial atrium line. (c) Light microscopy of transverse section through the cardial atrium line of stomach tissue (PAS staining x 160), showing fungal lesion comprising blastospores and hyphal elements (see arrow).
227
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
FIG. 5. Gross pathology and histopathology of liver from a 5FU treated mouse inoculated with C. albicans. (a) Livers, arrows point to fungal abscesses. (b) Light microscopy of liver tissue section (PAS staining × 16). Arrow points to fungal lesion. (c) Enlargement of (b) (× 250), note fungal invasion into blood vessels (arrow).
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
228
SANDOVSK¥-LOSICA E T A L .
FIG. 6. Calcofluor-stained C. albicans in tissue homogenates. (a) Blastospores and hyphal elements in stomach tissue homogenate from MTX treated mouse inoculated with C. albicans (fluorescent microscope, Zeiss, x 400). (b) Blastospores and hyphal elements in renal tissue homogenate from an irradiated (6 Greys) mouse inoculated with C. albicans (fluorescent microscope, Zeiss, x 400).
DISCUSSION We have presented in this study an animal model of systemic candidiasis originating from the GI tract in mice treated with anti-cancer therapy. The development of the animal model was a continuation of our previous in vitro studies investigating the interaction of C. albicans with the GI mucosa from anti-cancer treated mice [32, 33]. Those in vitro studies indicated that the interaction of the fungi with the GI tissues from compromised animals was enhanced. An increased interaction was also noted in the present in vivo investigation, as expressed by the number of Candida c.f.u, associated with the GI mucosa. In this study, we present evidence that oral administration of C. albicans to mice compromised by either irradiation (6 Greys), or treatment with MTX or 5FU, resulted in GI candidiasis that disseminated into visceral organs. The candidal sys-
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
229
temic infection led to death of a large proportion of these animals. The data of the 5FU model are of particular interest, since candidiasis in such a model has not been described previously. Furthermore, compared with our study, in which animals were compromised by a single treatment administered once, most studies by other investigators have involved animals immunosuppressed by a combination of several treatments, administered more than once [7, 11, 14, 21]. Moreover, some of those treatments [9, 12, 16, 21, 27] led only to increased candidal colonization in the GI tract with no systemic involvement, while in our model the GI colonization led to fatal systemic candidiasis. It is possible that the differences between those models [9, 12, 16, 20, 21, 27] and ours could be related either to the C. albicans strain or the mouse strain used, or the maintenance conditions of the mice. The C albicans strain (CBS 562) used by us apparently possesses a high pathogenic potential, as was also demonstrated previously in another study from our laboratory comparing the pathogenicity of several isolates of C albicans [35]. It was noted in that study that the optimal infective dose for induction of systemic infection by strain CBS 562 was 100 times lower than the infective dose of five other C. albicans strains tested. Difference in the pathogenic capacity among C albicans strains, as well as variations in the susceptibility to Candida of genetically distinct mouse strains, were also noted by others [6, 19, 31]. The data presented here revealed that the duration of GI colonization was longer (over 30 days for some treatments) in anti-cancer treated mice, and C. albicans was isolated most frequently from the stomach tissues. The number of fungal c.f.u, attached to stomach tissues increased during the follow-up period and was significantly higher in compromised mice. Thus, the increased fungal proliferation in the stomach, as demonstrated by c.f.u, counts and histological examination, may possibly lead to fungal dissemination into the blood (as judged by detection of candidal antigen in the serum) and visceral organs, such as the liver, spleen and kidneys. As to the portal of entry of Candida from the GI tract into the viscera there are different views [15, 21], indicating among others persorption through the upper small intestine or passage through the stomach tissue. In mice pre-treated with anti-cancer therapy the frequency of fungal isolation from the oesophagus was doubled compared with that from naive animals. This finding may possibly be analogous to the situation observed in debilitated patients, in whom oropharyngeal and oesophageal candidiasis have been identified with increased frequency [13, 38]. Candidiasis of the viscera was noted by gross pathology revealing fungal abscesses in the spleen, liver and kidneys. The presence of the fungal particles in the liver and kidneys was demonstrated histologically and by quantitative enumeration in culture. These organs are considered targets in systemic candidiasis [18, 28, 29]. As judged by the number of candidal c.f.u, in kidneys or liver, the infection was more severe in 6 Greys irradiated, or 5FU treated mice, than in MTX treated animals. The nature of the systemic candidiasis in the three treatment groups was different. In MTXtrated mice the infection seemed to be acute, whereas in irradiated or 5FU treated animals it had a more chronic course. It may be assumed that the differences could be associated with differences in the cytotoxic action of these treatments [17, 25, 37, 39]. The systemic infection in the compromised mice was fatal. As shown by mortality plots, the systemic infection caused death in a large proportion of animals, but there were differences in the mean survival time among the three groups of treated mice. The control animals subjected to anti-cancer treatment and not inoculated with C. albicans,
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
230
SANDOVSKY-LOSICA ET AL.
did not develop infection and no mortality was observed among these animals. It should be emphasized, that those MTX treated animals which succumbed due to the anti-cancer treatment (before challenge), were excluded from the experiment, as the follow-up started from the challenge day. Thus, the death of the compromised mice inoculated with C. albicans may be attributed to systemic candidiasis. In most studies of other investigators exploring GI candidiasis, death from systemic candidiasis was recorded primarily in an infant mouse model [31], or in genetically immunodeficient germ-free adult mice [4]. In summary, this in vivo study indicates that anti-cancer treatment can result in an increased interaction of C. albicans with the GI mucosa, which may be followed by massive fungal proliferation in the GI tract, leading eventually to dissemination and systemic infection which is fatal to some of the animals. ACKNOWLEDGEMENTS We thank Mrs M. Kidron for excellent technical assistance in photography and Mr P. Vennor from Abic Ltd. for supplying the methotrexate and 5-fluorouracil. We also thank Mrs I. Glarneter from the Department of Statistics, Tel-Aviv University, for assistance in the statistical analyses.
REFERENCES 1. BODEY, G. P. 1984. Candidiasis in cancer patients. American Journal of Medicine, 77(4D), 13-19. 2. BODEY, G. P. 1985. Fungal infections and fever of unknown origin in neutropenic patients. American Journal of Medicine, 80 (suppl. 5C), 112-119. 3. BRESLOW,N. 1970. A generalized Kruskal Wallis test for comparing k samples subject to unequal patterns of sensorship. Biometrica, 57, 579-594. 4. CANTORNA, M. T. & BALISH, E. 1990. Mucosal and systemic candidiasis in congenitally immunodeficient mice. Infection and Immunity, 58, 1093-1100. 5. COHEN, R., ROTH, F. J., DELGADO, E., AHEARN, D. G. & KALSER, M. H. 1969. Fungal flora of the normal human small and large intestine. New England Journal of Medicine, 280, 638-641. 6. COLE, G. T., LYNN, K. T. & SESHAN, K. R. 1990. An animal model for oropharyngeal, esophageal and gastric candidosis. Mycoses, 33, 7-19. 7. COLE, G. Z., LYNN, K. T., SESHAN, K. R. t~; POPE, L. M. 1989. Gastrointestinal and systemic candidosis in immunocompromised mice. Journal of Medical and Veterinary Mycology, 27, 363-380. 8. COOPER, B. H. tf¢.HUNTER, M. S. 1985. Yeasts of medical importance In: E. H. LENNETrE (Ed.) Manual of Clinical Microbiology, 4th edn., pp. 526-541. American Society for Microbiology, Washington, D.C. 9. DEMARIA, A., BUCKLEY, H. & VON LICHTENBERG, F. 1976. Gastrointestinal candidiasis in rats treated with antibiotics cortison and azathioprine. Infection and Immunity, 13, 1761-1770. 10. DIXON, W. & MASSEY, F. J. JR. 1957. Introduction to Statistical Analysis. 2nd edn., McGraw-Hill Book Company, Inc. New York. 11. DOMER, J. E. & HECTOR, R. F. 1987. Enhanced immune responses in mice treated with penicillin-tetracycline or trimethoprim-sulfamethoxazole when colonized intragastrically with Candida albicans. Antimicrobial Agents and Chemotherapy, 31, 6914597. 12. EKENNA, O. & SHERERTZ, R. J. 1987. Factors affecting colonization and dissemination of Candida albicans from the gastrointestinal tract of mice. Infection and Immunity, 55, 1558-1563. 13. ERAS, P., GOLDSTEIN, M. J. & SHERLOCK, P. 1972. Candida infections of the gastrointestinal tract.
Medicine (Baltimore), 51,367-379. 14. GREENFIELD, R. A., TROUTT, D. L., RICKARD, R. C. • ALTMILLER, D. H. 1988. Comparison of antibody, antigen, and metabolic assays in rat models of systemic and gastrointestinal candidiasis. Journal of Clinical Microbiology, 26, 409-417. 15. GUENTZEL,M. N., COLE, G. T. & POPE, L. M. 1985. Animal models for candidiasis, In: M. R. McGINNlS (Ed.) Current Topics in Medical Mycology. Vol. 1, pp. 57-116. Springer-Verlag, New York. 16. GUENTZEL,M. N. & HERRERA, C. 1982. Effects of compromising agents on candidosis in mice with persistent infections initiated in infancy. Infection and Immunity, 35, 22-28.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
231
17. GWAVAVA,N. J. T., PINKERTON, C. R., GLASGOW,J. F. T., SLOAN,J. M. & BRIDGES,J. M. 1981. Small bowel enterocyte abnormalities caused by methotrexate treatment in acute lymphoblastic leukemia of childhood. Journal of Clinical Pathology, 34, 790--795. 18. HARON, E., FELD, R., TUFFNELL, P., PATYERSON,B., HASSELBACK,R. & MATLOW,A. 1987. Hepatic candidiasis: an increasing problem in immunocompromised patients~ American Journal of Medicine, 83,17-26. 19. HECTOR, R., DOMER, J. & CARROW, E. W. 1982. Immune responses to Candida albicans in genetically distinct mice. Infection and Immunity, 38, 1020-1028. 20. HERRERA, C. & GUENTZEL, N. 1982. Mice with persistent gastrointestinal Candida albicans as a model for antifungal therapy. Antimicrobial Agents and Chemotherapy, 21, 51-53. 21. KENNEDY, M. J. 1989. Regulation of Candida albicans populations in the gastrointestinal tract: Mechanisms and significance in GI and systemic candidiasis. In: M. R. McGINNIS & M. BORGERS (Eds) Current Topics in Medical Mycology. Vol. 3, pp. 315~,02. Springer-Verlag, New York. 22. MAKSYMIUK,A. W., THONGPROESERT, S., HOPFER, R., LUNA, M., FAINSTEIN, V. & BODEY, G. P. 1984. Systemic candidiasis in cancer patients. American Journal of Medicine, 77(4D), 20-27. 23. MANTEL,N. 1966. Evaluation of survival data and two new rank order statistics arriving in its consideration. Cancer Chemotherapy Reports, 50, 163-170. 24. MEUNIER-CARPENTIER,F., KIEHN, F. E. & ARMSTRONG, D. 1981. Fungemia in the immunocompromised host. American Journal of Medicine, 71,363-370. 25. MILLES, S. S., MUGGIA, A. & SPIRO, H. M. 1962. Colonic histologic changes induced by 5-fluorouracil. Gastroenterology, 43, 391-399. 26. MONHEIT,J. E., COWAN,D. F. & MOORE, D. G. 1984. Rapid detection of fungi in tissues using calcofluor white and fluorescence microscopy. Archives of Pathology and Laboratory Medicine, 108, 616-618. 27. MYEROWITZ,R. L. 1981. Gastrointestinal and disseminated candidiasis. An experimental model in the immunosuppressed rat. Archives of Pathology and Laboratory Medicine, 105, 138-143. 28. MYEROWITZ,R. L., PAZIN, G. L. & ALLEN, C. M. 1977. Disseminated candidiasis: changes in incidence, underlying diseases, and pathology. American Journal of Clinical Pathology, 68, 29-38. 29. ODDS, F. C. 1988. Candida and Candidosis, 2nd edn. Baillibre Tindal, London. 30. OFEK, I. • BEACHFY, E. G. 1980. General concepts and principles of bacterial adherence in animals and man. In: E. G. BEACHEY (Ed.) Bacterial Adherence (Receptors and Recognition), Series B:6, pp. 3-29. Chapman and Hall, London. 31. POPE, L. M., COLE, G. T., GUENTZEL, M. N. & BERRY, L. J. 1979. Systemic and gastrointestinal candidiasis of infant mice after intragastric challenge. Infection and Immunity, 25, 702-707. 32. SANDOVSKY-LosICA,H. & SEGAL, E. 1989. Interaction of Candida albicans with murine gastrointestinal mucosa from irradiated animals: in vitro adhesion and prevention. Journal of Medical and Veterinary Mycology, 27, 345-352. 33. SANDOVSKY-LOS1CA,H. t~ SEGAL, E. 1990. Interaction of Candida albicans with murine gastrointestinal mucosa from methotrexate and 5-fluorouracil treated animals: in vitro adhesion and prevention. Journal of Medical and Veterinary Mycology, 28, 279-287. 34. SEGAL,E., LEHRER,N. & OFEK,I. 1982. Adherence of Candida albicans to human vaginal epithelial cells: inhibition by amino sugars. Experimental Cell Biology, 50, 13-17. 35. SEGAL,E., NUSSBAUM, S. & BARR-NEA, L. 1985. Protection against systemic infections with various Candida species elicited by vaccination with Candida albicans ribosomes. Sabouraudia: Journal of Medical and Veterinary Mycology, 23, 275-285. 36. STONE, H. H., GEHEBER, C. E., KOLB, L. D. 8z KITCHENS, W. R. 1973. Alimentary tract colonization by Candida albicans. Journal of Surgical Research, 14, 273-276. 37. TRIER, J. S. 1962. Morphologic alterations induced by methotrexate in the mucosa of human proximal intestine. I. Serial observations by light microscopy. Gastroenterology, 42, 295-305. 38. TRIER J. S. & BJORKMAN, D. J. 1984. Esophageal, gastric and intestinal candidiasis. American Journal of Medicine, 77(4D), 3943. 39. TRIER, J. S. & BROWNING, T. H. 1966. Morphologic response of the mucosa of human small intestine to X-ray exposure. Journal of Clinical Investigation, 45, 194-204.
Fatal systemic candidiasis of gastrointestinal origin: an experimental model in mice compromised by anti-cancer treatment H. S A N D O V S K Y - L O S I C A
t, L. B A R R - N E A 2 AND E. SEGAL t
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
1Department of Human Microbiology, and 2Department of Histology and Cell Biology, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel (Accepted 3 February 1992)
An experimental model of fatal systemic candidiasis originating from the gastrointestinal (GI) tract of compromised mice is presented. ICR female mice were compromised by a single anti-cancer treatment: irradiation (4 or 6 Greys i.e. 400~500 rads), methotrexate (MTX) (3 mg per mouse, intraperitoneally) or 5-fluorouracil (SFU) (200 mg kg -t, intravenously). Three days later, compromised and non-treated control mice were exposed to Candida albicans administered orally. Morbidity and mortality due to candidiasis were monitored for 30 days post-candidal inoculation. Increased and longer GI colonization was noted among the MTX and 5FU treated mice, or 6 Greys irradiated mice (up to 92.3% for over 30 days in anti-cancer treated mice). The stomach was found to be the major part of the GI tract involved in fungal colonization. A significant number (53.8-83.3%) of the anti-cancer treated mice developed systemic candidiasis originating from the GI tract, which was fatal in 30-80% of the infected animals. In systemically infected animals, candidal antigen was demonstrated in the serum, and fungal abscesses containing C. albicans were observed in the liver, kidneys and spleen. C. albicans was isolated from the infected organs. The severity of the infection, as reflected by the number of fungi in visceral organs, and by mortality during the 30 days post-candidal inoculation, indicated differences in the course and nature of the infection among the three treatment groups (i.e. MTX, 5FU, 6 Greys). The increased fungal association with the GI mucosa appears to be one of the factors leading to proliferation of the fungus in the GI tract, which may possibly be followed by dissemination to visceral organs, causing systemic candidiasis and mortality.
Candida albicans is considered to be a commensal of the gastrointestinal (GI) tract [5, 29, 36] of the normal population, however in compromised individuals, such as cancer patients, fungal Gi colonization increases significantly [5, 29, 35]. Thus, it is generally accepted that in such individuals, the GI tract may be a possible portal of entry for the organism into the viscera, leading to systemic candidiasis [13, 28, 38]. The magnitude of the problem is illustrated by the vast number of reports [1, 2, 18, 22, 24, 27, 38], describing human systemic candidiasis and involvement of the GI tract in severely compromised patients, particularly those with haematologic malignancies. As it is generally believed that microbial attachment to host tissues is essential for the development of infection [30], we focused our investigations on the importance of this phenomenon in compromised hosts. In recent studies we have investigated the in vitro attachment of C. albicans to GI mucosal surfaces [32, 33]. It was noted that fungal adherence to GI mucosa from compromised mice was significantly increased when compared with that to tissues from naive animals. Consequently, we initiated the present Correspondence address: Prof. E. Segal, Department of Human Microbiology, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel. 219
220
SANDOVSKY-LOSICA E T A L .
in vivo study, investigating systemic candidiasis originating from the GI tract of mice
treated by irradiation, methotrexate (MTX) or 5-fluorouracil (5FU). METHODS
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Growth and preparation of C. albicans for animal inoculation C. albicans strain CBS 562 (Centraalbureau voor Schimmelcultures, Delft, The Netherlands) was used throughout this study, as in previous experiments [32-34]. The fungus was grown on a gyrotory shaker in yeast extract broth (0.5 % Bacto yeast extract; 3% glucose) at 28°C for 18 h. Yeast cells were harvested by centrifugation (3500 rev. min -1) washed three times in sterilized distilled water and resuspended to the desired concentration.
Anti-cancer treatments
Six-week-old, ICR female mice (mean weight 25-28 g), maintained under conventional conditions, five mice per cage, were subjected to one of the following anti-cancer treatments. Irradiation. Mice were exposed to a nonqethal, single dose (whole body) of 4 Greys (48
mice) or 6 Greys (99 mice) of Cobalt irradiation. M T X treatment. MTX (Abic Ltd. Israel) was injected intraperioneally (131 mice) at a
final concentration of 3 mg per mouse. (Six-percent of the mice died due to the treatment on days 1-2 post-treatment; these animals were excluded from the experiment.) 5FU treatment. 5FU (Abic Ltd. Israel) was injected intravenously (121 mice) into the
tail vein with a non-lethal dose of 200 mg kg- 1body weight. Immunosuppression of the treated animals was monitored by the decrease in white blood cell counts and spleen weights, as described in detail in previous studies [32, 33]. Non-treated mice served as controls. Challenge of the animals with C. albicans
Drinking water (sterile distilled water) supplemented with 5% sucrose, containing 5 × 107 C. albicans m1-1 was administered to the compromised mice for 24 h, on the 4th day post-treatment; control mice (with no pre-treatment) were also challenged orally with the yeast. The mean calculated inoculum per mouse was 2-5 × 108 organisms (based on an average water consumption of 5 ml per mouse). An additional control group of mice treated with the various anti-cancer treatments and not challenged with C. albicans was also included. Following candidal challenge, animals were surveyed for morbidity and mortality for 30 days. On different days post-candidal challenge, animals were sacrificed (by cervical translocation) and morbidity was assessed by: (i) enumeration of C. albicans in the GI tract and visceral organs, and (ii) histological examination of GI and visceral organs. Biood samples from several animals were tested for the presence of candidal antigen using a commercial latex agglutination kit (Difco, USA).
FATAL SYSTEMIC CANDIDIASIS
221
Mortality was monitored for 30 days post-candidal challenge. Post-mortem studies were carried out in some of the experiments on mice that had died during the 30 day monitoring period. Liver and kidneys were examined for the presence of fungal abscesses and were cultured to detect C. albicans.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Enumeration of C. albicans in the GI tract and visceral organs The GI tract of the sacrificed animals was removed, rinsed with sterile distilled water and segmented into the esophagus, stomach, small intestine and cecum. Tissue disks (0.7 cm in diameter) were cut with a sterile metal punch from each segment (the stomach tissue segments always included parts of secreting, non-secreting and cardial atrium line tissue). The esophagus was homogenized as a whole organ (due to its size). Part (upper left lobe) of the liver (approximately 300 mg) and the whole kidneys (approximately 300 mg x 2) were also removed. GI tissue disks or visceral organs were homegenized separately in sterile distilled water (1 ml) using a tissue homogenizer (Polytron-Kinematica, GmbH, Littau, Germany). Diluted samples of the homogenates were plated on Sabouraud glucose agar containing chloramphenicol and incubated at 28°C for 24-48 h. Colonies were identified by microscopy and other conventional techniques for the identification of C. albicans [8]. Growth of more than 10 colony forming units c.f.u, per 0-1 ml of tissue homogenate (or organ, i.e. esophagus) from at least one part of the GI tract was indicative for fungal GI colonization. This strict criterion was chosen to avoid consideration of random growth of a few colonies. The same strict criterion as for GI colonization was chosen for systemic infection. Growth of >10 colonies per 0-1 ml of liver or kidney homogenate was considered an indication of systemic infection. Histological examination Segments of the GI tissues, liver and kidneys were fixed, embedded in paraffin and sections (6/urn) cut and stained with Periodic Acid Schiff (PAS) for light microscopy. Unstained or calcofluor (0-01%, Polyscience Inc. USA) stained [26] tissue homogenates were also used for direct examination under light and fluorescent microscopes (UV light 340-460 nm range), respectively, for detection of fungal elements. Statistical analyses Colonization or systemic infection values were analysed by the Fisher exact test for differences between treated and non-treated groups [10]. Mortality slopes were analysed using two statistical tests: Breslow [3] and Mantel [23]. Confidence levels of at least P < 0.05 were considered significant. RESULTS Candida GI colonization in compromised mice The percentage GI colonization among a total of 70 MTX, 60 5FU and 94 irradiated (4 Greys - 48, or 6 Greys - 46) mice compared with that of 103 non-treated animals was assayed on various days (days 1-2; 3-4, 6-7; 13-14; 21; 30) post-candidal challenge. A higher percentage (up to 92%) of compromised mice were colonized with C. albicans,
222
SANDOVSKY-LOSICA E T A L . Mice 9"9% Control
Ii i i
20.2%
MTX- treoted
21.2%
5FU-treoted
11.3%
57"7%
57-7%
195%
22.8%
36-5%
20-0%
22"3%
72.4%
20-7%
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
21,1%
5-5% 5"5%
Irradiated 4Greys 19.6% Irradiated 6 Greys
51,8%
10.7%
N
17.9%
i i
Esophog l---1Stomach ~ l n t e s t , n e
I~Cecum
FIG. 1. Distribution of C. albicans isolations from the GI tract of irradiated, MTX, 5FU treated and nontreated mice. The data represent the % of C. albicans isolation from various parts of the GI tract of treated and non-treated mice, calculated as proportions of the total number of candidal isolations from the GI tract of each treatment group. The data were analysed by the Fisher exact test; P < 0.05 was considered significant. The differences between fungal isolations from the stomach, small intestine or esophagus of MTX or 5FU treated animals and isolations from the corresponding organs of non-treated mice were significant.
40
2O (a) Nmve m,ce
(b) Irradloted mLce o
30 20
Io
o
7 g ~n g z-
u~
I0
tl
o
g ,
I
i
O
.11 0
0
-
Dqys post d'~nd/do inoculahon
I0 1~0 50 DQys post Cond/dG ~noculutlon
4Greys 0 6Greys g co co
80
co co
c) 5-Fu freqted mtce
--. 400 " (d) MTX lreoted rnlce
60
6 500
40
200
2C
tO0
+ +
0
I0 20 Dqys post L?Lz'd,'do inoculohon
~0
0
'
I0 20 ' Days pos~ C~Tnd/do inoculo'~lon
3'0
FIG. 2. Enumeration of C, albicans c.f.u, in stomach tissue of irradiated, MTX, 5FU treated and non-treated mice. The data represent the number of c.f.u, per tissue at various time intervals post-candidal challenge. Each point indicates an individual animal.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
223
although colonization was also noticed among naive mice (up to 75%). Furthermore, while in compromised mice colonization lasted for at least 21 days post-candidal challenge and for some treatments (5FU or irradiation) over the survey period of 30 days, in normal mice candidal colonization lasted for only 2 weeks. Cultures from non-challenged mice were negative for C. albicans. Figure 1 shows the distribution of Candida in the GI tract. The data represent the relative frequency (in %) of fungal colonization in each of four parts of the GI tract: esophagus, stomach, small intestine and cecum, calculated from pooled data from each group of mice during the 30 days post-candidal challenge. It can be seen that the stomach is the main site of colonization by Candida in normal as well as in compromised mice. However, the frequency of fungal isolation from the stomach was lower from MTX or 5FU treated mice than from control animals. Furthermore, in MTX or 5FU treated mice the relative frequencies of candidal colonization in the esophagus or the small intestine were higher then in normal animals (the differences were statistically significant, P < 0.05, Fisher exact test). Following the observation that the stomach serves as the main part of the GI tract for C. albicans colonization, we compared the degree of colonization in this organ among animals of the various groups quantitatively, by enumerating the candidal c.f.u. in the stomach tissue. As shown in Fig. 2 (a-d), it can be seen that the level of c.f.u, in the stomach tissue of MTX or 5FU treated or 6 Greys irradiated mice, was higher than in non-compromised animals, or in those irradiated with a low dose (4 Greys). The data in this figure show the distribution of c.f.u, among the individual mice of the different groups. Calculated maximal means (not shown) reveal even more pronounced differences between compromised and non-compromised animals. Specifically, the maximal mean c.f.u, per stomach tissue disk of control and 4 Greys irradiated mice was 1.8 x 103 and 1.7 x 103, respectively, while that of MTX and 5FU treated or 6 Greys irradiated mice was 4.3 X 10 4, 1.2 × 10 4 or 3"6 × 10 3, respectively.
Course of systemic candidiasis and mortality in compromised mice The percentage of mice with systemic candidiasis during the 30 day follow-up period, based on isolation of Candida from viscera is presented in Table 1. Among the compromised mice, a large proportion (maximal value, 83%) of animals developed systemic infection, while among the non-compromised mice there was a very low percentage of systemic involvement. It was also noted that in the MTX treated mice, candidal infection was acute while in 5FU treated, or irradiated mice (6 Greys) it apparently had a longer course. It should be emphasized that in about 14 days most of the MTX treated mice succumbed to the infection (Fig. 3), therefore the small animal numbers appearing in Table 1 on day 21 and beyond represent only those animals which did not, apparently, develop systemic infection. Irradiation with 4 Greys resulted in a low percentage of systemically infected animals (maximal value, 25 %), but a higher level of irradiation (6 Greys), resulted, as indicated before, in GI colonization with chronic systemic infection. In some of the compromised animals Candida antigenemia was detected in undiluted animal sera using a commercial latex agglutination kit. Table 2 summarizes the mean number of fungi isolated from visceral organs (liver or kidneys) and the number of fungal isolations out of the total number of mice screened. It can be seen that, among the non-treated mice or those irradiated at the
224
SANDOVSKY-LOSICA E T A L . TABLE 1. Systemic candidal infection in compromised mice following oral challenge with C. albicans %
Day a
Non-treated
MTX
5FU
Irr 4 Greys
Irr 6 Greys
0 (0/10) 0 (0/20) 10.3 (3/29) 12.1 (4/33) 0 (0/6) 0 (0/5)
41.7 c (5/12) 61.9¢ (13/21) 8.3 (1/12) 23-1 (3/13) 0 (0/8) 0 (0/4)
61-6c (8/13) 61-9~ (13/21) 68-7c (11/16) 75" (6/8) 83.3 (5/6) 16.7 (1/6)
25 (1/4) 7.7 (1/13) 0 (0/13) 0 (0/12) n.d.
0 (0/4) 23.1 (3/13) 53-8 ¢ (7/13) 33.3 (4/12) n.d.
0 (0/6)
25 (1/4)
1-2 3-4 6-7 Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
with systemic infection (No. infected mice/totalb)
13-14 21 30
a Days post-candidal challenge. b Number of mice with systemic infection (based on isolation of Candida from viscera) per total number of dissected animals on a given day. c Significant differences, P < 0.05, Fisher exact test. n.d. = Not done.
level of 4 Greys, the number of fungal isolations was not significant (6/103 and 2/48, respectively). In addition, even in the animals with infected viscera, the mean number of candidal c.f.u, isolated per organ was low (0.2-1 × 103). In contrast, in compromised mice C. albicans was isolated from visceral organs from a significant number of animals (14/70; 41/70; 13/46). Moreover, the number of candidal c.f.u, isolated from visceral
~-
~
I00
-
80
-
n
O-
X~-~X~=61
=
61
4o' m~Fi~4~li-I-ill-i-n-Ig-lg-i
/
I
I
,mm ~=~s -i-II-II-II
0
I0 20 Doys post E. o/b/cans chollenge
ZO
- i - Irrod~oted 6Greys -0- MTX -X- 5FU
FIG. 3. Mortality of irradiated, MTX, 5FU treated and non-treated mice. Data are pooled values from several 30 day follow-up surveys, n = total number of animals (at the beginning of the follow-up). Mortality data of the three groups were analysed by the Breslow [3] and Mantel [23] tests, and the mortality plots of the three groups were found to be significantly different (P < 0.001). The calculated mean survival times were: irradiated mice 23.9 days: MTX treated mice 9-4 days; 5FU treated mice, 12-8 days.
225
FATAL SYSTEMIC CANDIDIASIS TABLE 2. Candida isolation from visceral organs Mean Candida c.f.u. × 103 per organ a (No. of mice with Candida/total) h
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
Organ
Non-treated
MTX
5FU
Irr 4 Greys
Irr 6 Greys
Liver (300 mg)
0-4 (6/103)
2.0 (19/70)
7-6 (37/70)
0.2 (2/48)
67-1 (12/46)
Kidneys (300 mg)
1.0 (1/103)
2-7 (14/70)
3900 (41/70)
0 (0/48)
2500 (13/46)
a Mean values. b Total number of mice from which Candida was isolated from the kidneys or liver per total number of animals screened.
organs of these animals was significantly higher (from 2-7-4000 times higher than the value of the controls). Furthermore, the data in Table 2 point to differences in the severity of infection among animals of the various groups. Specifically, in MTX-treated mice the level of systemic infection, as judged by the mean of candidal c.f.u., as well as by the number of infected animals, was lower than in 5FU treated or 6 Greys irradiated mice. Figure 3 illustrates the mortality data of compromised mice post-candidal challenge. No animals among the normal mice or those irradiated by low dose (4 Greys) died during the monitoring period. In mice irradiated with the high dose (6 Greys), or in MTX or 5FU treated animals the systemic candidal infection was fatal, and mortality was associated with high morbidity as judged by c.f.u, counts from visceral organs (Table 2). About 80% of the mice compromised by MTX or 5FU, and about 30% of the irradiated (6 Greys) animals succumbed to systemic candidiasis during the 30 days observation period. According to statistical analysis the mean survival time (MST) (for MTX treated mice, 9-4 days; for 5FU treated mice, 12-8 days; and for 6 Greys irradiated mice, 23-9 days) as well as the mortality plots of the three groups of compromised mice were statistically different. Histological examination of infected tissues Gross examination of the GI tract of the compromised mice post-challenge revealed macroscopic changes in the mucosa of the stomach, especially in the cardial atrium line (Fig. 4a, b - 6 Grey irradiated mouse). Examination of viscera revealed macroabscesses in the kidneys, liver and spleen of these animals (Fig. 5a - liver of 5FU treated mouse). Direct examination of calcoftuor stained or unstained tissue homogenates from the stomach or visceral organs of compromised mice revealed fungal elements, including hyphae and budding yeasts (Fig. 6a, b). Tangential sections of stomach tissues (stained by PAS), from compromised mice (Fig. 4c - 6 Greys irradiated mouse) showed large amounts of hyphal elements associated with the squamous epithelium in the region of the cardial atrium line. Deeper sections of the stomach tissues (in the sub-mucosa) revealed invasion of fungal elements Sections of liver and kidneys (stained by PAS) from infected, compromised mice showed large abscesses composed mainly of hyphal elements (Fig. 5b, c-liver of 5FU treated mouse).
SANDOVSKY-LOSICA E T A L .
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
226
FIG. 4. Gross pathology and histopathology of stomach mucosa (inner part) of 6 Greys irradiated mouse inoculated with C albicans. (a) Stomach mucosa from an irradiated (6 Greys) mouse inoculated with C. albicans. Arrow indicates cardial atrium line, note pathological changes. (b) Stomach mucosa from a naive mouse. Arrow points to cardial atrium line. (c) Light microscopy of transverse section through the cardial atrium line of stomach tissue (PAS staining x 160), showing fungal lesion comprising blastospores and hyphal elements (see arrow).
227
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
FIG. 5. Gross pathology and histopathology of liver from a 5FU treated mouse inoculated with C. albicans. (a) Livers, arrows point to fungal abscesses. (b) Light microscopy of liver tissue section (PAS staining × 16). Arrow points to fungal lesion. (c) Enlargement of (b) (× 250), note fungal invasion into blood vessels (arrow).
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
228
SANDOVSK¥-LOSICA E T A L .
FIG. 6. Calcofluor-stained C. albicans in tissue homogenates. (a) Blastospores and hyphal elements in stomach tissue homogenate from MTX treated mouse inoculated with C. albicans (fluorescent microscope, Zeiss, x 400). (b) Blastospores and hyphal elements in renal tissue homogenate from an irradiated (6 Greys) mouse inoculated with C. albicans (fluorescent microscope, Zeiss, x 400).
DISCUSSION We have presented in this study an animal model of systemic candidiasis originating from the GI tract in mice treated with anti-cancer therapy. The development of the animal model was a continuation of our previous in vitro studies investigating the interaction of C. albicans with the GI mucosa from anti-cancer treated mice [32, 33]. Those in vitro studies indicated that the interaction of the fungi with the GI tissues from compromised animals was enhanced. An increased interaction was also noted in the present in vivo investigation, as expressed by the number of Candida c.f.u, associated with the GI mucosa. In this study, we present evidence that oral administration of C. albicans to mice compromised by either irradiation (6 Greys), or treatment with MTX or 5FU, resulted in GI candidiasis that disseminated into visceral organs. The candidal sys-
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
229
temic infection led to death of a large proportion of these animals. The data of the 5FU model are of particular interest, since candidiasis in such a model has not been described previously. Furthermore, compared with our study, in which animals were compromised by a single treatment administered once, most studies by other investigators have involved animals immunosuppressed by a combination of several treatments, administered more than once [7, 11, 14, 21]. Moreover, some of those treatments [9, 12, 16, 21, 27] led only to increased candidal colonization in the GI tract with no systemic involvement, while in our model the GI colonization led to fatal systemic candidiasis. It is possible that the differences between those models [9, 12, 16, 20, 21, 27] and ours could be related either to the C. albicans strain or the mouse strain used, or the maintenance conditions of the mice. The C albicans strain (CBS 562) used by us apparently possesses a high pathogenic potential, as was also demonstrated previously in another study from our laboratory comparing the pathogenicity of several isolates of C albicans [35]. It was noted in that study that the optimal infective dose for induction of systemic infection by strain CBS 562 was 100 times lower than the infective dose of five other C. albicans strains tested. Difference in the pathogenic capacity among C albicans strains, as well as variations in the susceptibility to Candida of genetically distinct mouse strains, were also noted by others [6, 19, 31]. The data presented here revealed that the duration of GI colonization was longer (over 30 days for some treatments) in anti-cancer treated mice, and C. albicans was isolated most frequently from the stomach tissues. The number of fungal c.f.u, attached to stomach tissues increased during the follow-up period and was significantly higher in compromised mice. Thus, the increased fungal proliferation in the stomach, as demonstrated by c.f.u, counts and histological examination, may possibly lead to fungal dissemination into the blood (as judged by detection of candidal antigen in the serum) and visceral organs, such as the liver, spleen and kidneys. As to the portal of entry of Candida from the GI tract into the viscera there are different views [15, 21], indicating among others persorption through the upper small intestine or passage through the stomach tissue. In mice pre-treated with anti-cancer therapy the frequency of fungal isolation from the oesophagus was doubled compared with that from naive animals. This finding may possibly be analogous to the situation observed in debilitated patients, in whom oropharyngeal and oesophageal candidiasis have been identified with increased frequency [13, 38]. Candidiasis of the viscera was noted by gross pathology revealing fungal abscesses in the spleen, liver and kidneys. The presence of the fungal particles in the liver and kidneys was demonstrated histologically and by quantitative enumeration in culture. These organs are considered targets in systemic candidiasis [18, 28, 29]. As judged by the number of candidal c.f.u, in kidneys or liver, the infection was more severe in 6 Greys irradiated, or 5FU treated mice, than in MTX treated animals. The nature of the systemic candidiasis in the three treatment groups was different. In MTXtrated mice the infection seemed to be acute, whereas in irradiated or 5FU treated animals it had a more chronic course. It may be assumed that the differences could be associated with differences in the cytotoxic action of these treatments [17, 25, 37, 39]. The systemic infection in the compromised mice was fatal. As shown by mortality plots, the systemic infection caused death in a large proportion of animals, but there were differences in the mean survival time among the three groups of treated mice. The control animals subjected to anti-cancer treatment and not inoculated with C. albicans,
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
230
SANDOVSKY-LOSICA ET AL.
did not develop infection and no mortality was observed among these animals. It should be emphasized, that those MTX treated animals which succumbed due to the anti-cancer treatment (before challenge), were excluded from the experiment, as the follow-up started from the challenge day. Thus, the death of the compromised mice inoculated with C. albicans may be attributed to systemic candidiasis. In most studies of other investigators exploring GI candidiasis, death from systemic candidiasis was recorded primarily in an infant mouse model [31], or in genetically immunodeficient germ-free adult mice [4]. In summary, this in vivo study indicates that anti-cancer treatment can result in an increased interaction of C. albicans with the GI mucosa, which may be followed by massive fungal proliferation in the GI tract, leading eventually to dissemination and systemic infection which is fatal to some of the animals. ACKNOWLEDGEMENTS We thank Mrs M. Kidron for excellent technical assistance in photography and Mr P. Vennor from Abic Ltd. for supplying the methotrexate and 5-fluorouracil. We also thank Mrs I. Glarneter from the Department of Statistics, Tel-Aviv University, for assistance in the statistical analyses.
REFERENCES 1. BODEY, G. P. 1984. Candidiasis in cancer patients. American Journal of Medicine, 77(4D), 13-19. 2. BODEY, G. P. 1985. Fungal infections and fever of unknown origin in neutropenic patients. American Journal of Medicine, 80 (suppl. 5C), 112-119. 3. BRESLOW,N. 1970. A generalized Kruskal Wallis test for comparing k samples subject to unequal patterns of sensorship. Biometrica, 57, 579-594. 4. CANTORNA, M. T. & BALISH, E. 1990. Mucosal and systemic candidiasis in congenitally immunodeficient mice. Infection and Immunity, 58, 1093-1100. 5. COHEN, R., ROTH, F. J., DELGADO, E., AHEARN, D. G. & KALSER, M. H. 1969. Fungal flora of the normal human small and large intestine. New England Journal of Medicine, 280, 638-641. 6. COLE, G. T., LYNN, K. T. & SESHAN, K. R. 1990. An animal model for oropharyngeal, esophageal and gastric candidosis. Mycoses, 33, 7-19. 7. COLE, G. Z., LYNN, K. T., SESHAN, K. R. t~; POPE, L. M. 1989. Gastrointestinal and systemic candidosis in immunocompromised mice. Journal of Medical and Veterinary Mycology, 27, 363-380. 8. COOPER, B. H. tf¢.HUNTER, M. S. 1985. Yeasts of medical importance In: E. H. LENNETrE (Ed.) Manual of Clinical Microbiology, 4th edn., pp. 526-541. American Society for Microbiology, Washington, D.C. 9. DEMARIA, A., BUCKLEY, H. & VON LICHTENBERG, F. 1976. Gastrointestinal candidiasis in rats treated with antibiotics cortison and azathioprine. Infection and Immunity, 13, 1761-1770. 10. DIXON, W. & MASSEY, F. J. JR. 1957. Introduction to Statistical Analysis. 2nd edn., McGraw-Hill Book Company, Inc. New York. 11. DOMER, J. E. & HECTOR, R. F. 1987. Enhanced immune responses in mice treated with penicillin-tetracycline or trimethoprim-sulfamethoxazole when colonized intragastrically with Candida albicans. Antimicrobial Agents and Chemotherapy, 31, 6914597. 12. EKENNA, O. & SHERERTZ, R. J. 1987. Factors affecting colonization and dissemination of Candida albicans from the gastrointestinal tract of mice. Infection and Immunity, 55, 1558-1563. 13. ERAS, P., GOLDSTEIN, M. J. & SHERLOCK, P. 1972. Candida infections of the gastrointestinal tract.
Medicine (Baltimore), 51,367-379. 14. GREENFIELD, R. A., TROUTT, D. L., RICKARD, R. C. • ALTMILLER, D. H. 1988. Comparison of antibody, antigen, and metabolic assays in rat models of systemic and gastrointestinal candidiasis. Journal of Clinical Microbiology, 26, 409-417. 15. GUENTZEL,M. N., COLE, G. T. & POPE, L. M. 1985. Animal models for candidiasis, In: M. R. McGINNlS (Ed.) Current Topics in Medical Mycology. Vol. 1, pp. 57-116. Springer-Verlag, New York. 16. GUENTZEL,M. N. & HERRERA, C. 1982. Effects of compromising agents on candidosis in mice with persistent infections initiated in infancy. Infection and Immunity, 35, 22-28.
Med Mycol 1992.30:219-231. Downloaded from informahealthcare.com by McMaster University on 10/30/14. For personal use only.
FATAL SYSTEMIC CANDIDIASIS
231
17. GWAVAVA,N. J. T., PINKERTON, C. R., GLASGOW,J. F. T., SLOAN,J. M. & BRIDGES,J. M. 1981. Small bowel enterocyte abnormalities caused by methotrexate treatment in acute lymphoblastic leukemia of childhood. Journal of Clinical Pathology, 34, 790--795. 18. HARON, E., FELD, R., TUFFNELL, P., PATYERSON,B., HASSELBACK,R. & MATLOW,A. 1987. Hepatic candidiasis: an increasing problem in immunocompromised patients~ American Journal of Medicine, 83,17-26. 19. HECTOR, R., DOMER, J. & CARROW, E. W. 1982. Immune responses to Candida albicans in genetically distinct mice. Infection and Immunity, 38, 1020-1028. 20. HERRERA, C. & GUENTZEL, N. 1982. Mice with persistent gastrointestinal Candida albicans as a model for antifungal therapy. Antimicrobial Agents and Chemotherapy, 21, 51-53. 21. KENNEDY, M. J. 1989. Regulation of Candida albicans populations in the gastrointestinal tract: Mechanisms and significance in GI and systemic candidiasis. In: M. R. McGINNIS & M. BORGERS (Eds) Current Topics in Medical Mycology. Vol. 3, pp. 315~,02. Springer-Verlag, New York. 22. MAKSYMIUK,A. W., THONGPROESERT, S., HOPFER, R., LUNA, M., FAINSTEIN, V. & BODEY, G. P. 1984. Systemic candidiasis in cancer patients. American Journal of Medicine, 77(4D), 20-27. 23. MANTEL,N. 1966. Evaluation of survival data and two new rank order statistics arriving in its consideration. Cancer Chemotherapy Reports, 50, 163-170. 24. MEUNIER-CARPENTIER,F., KIEHN, F. E. & ARMSTRONG, D. 1981. Fungemia in the immunocompromised host. American Journal of Medicine, 71,363-370. 25. MILLES, S. S., MUGGIA, A. & SPIRO, H. M. 1962. Colonic histologic changes induced by 5-fluorouracil. Gastroenterology, 43, 391-399. 26. MONHEIT,J. E., COWAN,D. F. & MOORE, D. G. 1984. Rapid detection of fungi in tissues using calcofluor white and fluorescence microscopy. Archives of Pathology and Laboratory Medicine, 108, 616-618. 27. MYEROWITZ,R. L. 1981. Gastrointestinal and disseminated candidiasis. An experimental model in the immunosuppressed rat. Archives of Pathology and Laboratory Medicine, 105, 138-143. 28. MYEROWITZ,R. L., PAZIN, G. L. & ALLEN, C. M. 1977. Disseminated candidiasis: changes in incidence, underlying diseases, and pathology. American Journal of Clinical Pathology, 68, 29-38. 29. ODDS, F. C. 1988. Candida and Candidosis, 2nd edn. Baillibre Tindal, London. 30. OFEK, I. • BEACHFY, E. G. 1980. General concepts and principles of bacterial adherence in animals and man. In: E. G. BEACHEY (Ed.) Bacterial Adherence (Receptors and Recognition), Series B:6, pp. 3-29. Chapman and Hall, London. 31. POPE, L. M., COLE, G. T., GUENTZEL, M. N. & BERRY, L. J. 1979. Systemic and gastrointestinal candidiasis of infant mice after intragastric challenge. Infection and Immunity, 25, 702-707. 32. SANDOVSKY-LosICA,H. & SEGAL, E. 1989. Interaction of Candida albicans with murine gastrointestinal mucosa from irradiated animals: in vitro adhesion and prevention. Journal of Medical and Veterinary Mycology, 27, 345-352. 33. SANDOVSKY-LOS1CA,H. t~ SEGAL, E. 1990. Interaction of Candida albicans with murine gastrointestinal mucosa from methotrexate and 5-fluorouracil treated animals: in vitro adhesion and prevention. Journal of Medical and Veterinary Mycology, 28, 279-287. 34. SEGAL,E., LEHRER,N. & OFEK,I. 1982. Adherence of Candida albicans to human vaginal epithelial cells: inhibition by amino sugars. Experimental Cell Biology, 50, 13-17. 35. SEGAL,E., NUSSBAUM, S. & BARR-NEA, L. 1985. Protection against systemic infections with various Candida species elicited by vaccination with Candida albicans ribosomes. Sabouraudia: Journal of Medical and Veterinary Mycology, 23, 275-285. 36. STONE, H. H., GEHEBER, C. E., KOLB, L. D. 8z KITCHENS, W. R. 1973. Alimentary tract colonization by Candida albicans. Journal of Surgical Research, 14, 273-276. 37. TRIER, J. S. 1962. Morphologic alterations induced by methotrexate in the mucosa of human proximal intestine. I. Serial observations by light microscopy. Gastroenterology, 42, 295-305. 38. TRIER J. S. & BJORKMAN, D. J. 1984. Esophageal, gastric and intestinal candidiasis. American Journal of Medicine, 77(4D), 3943. 39. TRIER, J. S. & BROWNING, T. H. 1966. Morphologic response of the mucosa of human small intestine to X-ray exposure. Journal of Clinical Investigation, 45, 194-204.
