Measurement of blood clearance time by Limulus G test ofCandida-water soluble polysaccharide fraction, CAWS, in mice

Kiyoshi Kurihara a, Noriko N. Miura a, Michiharu Uchiyama a, Naohito Ohno a;*,Yoshiyuki Adachi a, Maki Aizawa b, Hiroshi Tamura b, Shigenori Tanaka b,

Toshiro Yadomae a

a Tokyo University of Pharmacy and Life Science, School of Pharmacy, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japanb Seikagaku Corporation, 3-1253 Tateno, Higashiyamato, Tokyo 207-0021, Japan

Received 25 February 2000; received in revised form 19 June 2000; accepted 29 June 2000

Abstract

The Limulus G test, responsive to L-1,3-D-glucan, is a well-established method for the detection of invasive fungal infection. We haverecently found that Candida albicans released a water-soluble polysaccharide fraction (CAWS) into synthetic medium (Uchiyama et al.,FEMS Immunol. Med. Microbiol. 24 (1999) 411^420). CAWS was composed of a mannoprotein-L-glucan complex and activated Limulusfactor G, and thus would be similar to the Limulus active substance in patient's blood. In a preliminary investigation, we have found thatCAWS is lethal when administered intravenously in a murine system. In this study, we examined the toxicity and then the fate of CAWS inmice. The lethal toxicity was strain-dependent and strain DBA/2 was the most resistant. The toxicity was, at least in part, reduced bysalbutamol sulfate and prednisolone treatment in the sensitive strains. On intravenous administration, the half clearance time (t1=2) wasapproximately 40 min in mice (DBA/2). On intraperitoneal administration, CAWS appeared in the blood with a peak concentration at 1 h.In order to establish a treatment plan, it is important to demonstrate the onset and the termination of deep-seated mycosis. The Limulus Gtest is suitable for the above purpose; however, it is necessary to fully understand the fate of L-1,3-D-glucan in patients' blood ß 2000Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Limulus test; Mannan^glucan complex; Candida-water soluble polysaccharide fraction; Lethal toxicity; Half clearance time

1. Introduction

Invasive deep-mycoses have been becoming increasinglycommon due to the growing number of immunocomprom-ised hosts caused by progressed chemotherapy or immuno-suppressants, such as steroids. These infections occurin severely ill patients with hematological malignancies,leukemia, and acquired immune de¢ciency syndrome(AIDS), and in those undergoing myelosuppressive cyto-toxic chemotherapy [1]. From the stand point of diagnosis,symptoms of infection by bacteria and fungi are similarbecause of systemic in£ammation. In addition, isolation ofthe causative microbe is time consuming. Misinterpreta-tion resulting in the wrong choice antibacterial or antifun-gal chemotherapy can be fatal. Metabolites of microbesare often used for early detection.

Candida albicans is a medically important fungus, whichinduces a disseminated candidiasis and candidemia in im-munocompromised hosts [2], and releases a polysaccharidefraction into the blood. The cell wall of Candida is mainlycomposed of two polysaccharides, mannan and L-glucan[3^8]. In a previous study, we used a chemically de¢nedliquid medium to culture Candida spp. and collected thewater-soluble fraction named CAWS [9]. We have re-ported that CAWS is mainly composed of mannoproteinand L-glucan. At least in part, the mannan moiety wasfound to be covalently linked to the L-glucan moiety.

It is generally accepted that the blood concentration ofmedicines is an important parameter in maintaining phar-macological actions and reducing side e¡ects. We havelong been working on analyzing the immunopharmacolog-ical activity of L-1,3-D-glucan preparations. We found thatthe half clearance time of an antitumor L-1,3-D-glucanpreparation, SPG (soni¢lan), was approximately 6 h inmice. However, on repeated administration, L-1,3-D-glu-can was saturated in the metabolizing organs, liver and

0928-8244 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 9 2 8 - 8 2 4 4 ( 0 0 ) 0 0 1 9 0 - 5

* Corresponding author. Tel. : +81 (426) 765570;Fax: +81 (426) 765570; E-mail : ohnonao@ps.toyaku.ac.jp

FEMSIM 1244 21-8-00

FEMS Immunology and Medical Microbiology 29 (2000) 69^76

www.fems-microbiology.org

spleen, resulting in a delay in the clearance from theblood, and ¢nally, a high concentration in the blood[10]. This characteristic feature is strongly related to theabsence of a speci¢c metabolizing enzyme for L-1,3-D-glu-can in mice.

The Limulus test is a well-established method for thedetection of both Gram-negative sepsis and invasive fun-gal infection. At present, a L-1,3-D-glucan-speci¢c Limulustest, the G test, is clinically used to diagnose invasive deep-mycosis [11]. However, for the above reason, it is neces-sary to determine the fate of L-glucan. The goal of ourproject is to clarify the blood clearance of CAWS in ex-perimental animals to establish a better therapeutic proto-col for fungal infection by using the G test. However, ithas been known for some sorts of yeast mannan to exhibitlethal toxicity in mice [12]. Mannan is the most abundantcomponent in CAWS, and in the preliminary investiga-tion, CAWS was lethal to ICR mice within 30 min similarto anaphylactic shock.

In the present study, we: (1) screened inbred strains lesssensitive to mannan mediated lethal toxicity; (2) testedseveral medicines to reduce the sensitivity ; and (3) mea-sured blood clearance of CAWS in the insensitive mice.We found that the half clearance time of CAWS in bloodwas 35 min. It was therefore cleared signi¢cantly fasterthan the clinically used L-1,3-D-glucan, Soni¢lan (SPG)[13].

2. Materials and methods

2.1. Materials

All strains of C. albicans, Candida parapsilosis and Sac-charomyces cerevisiae were purchased from the Institutefor Fermentation, Osaka, Japan, maintained on Sabour-aud agar (Difco, USA) at 25³C and transferred once every3 months. SPG (Soni¢lan prepared from Schizophyllan[14,15]) was kindly provided by Kaken Pharmaceutical(Tokyo, Japan). Fungitic G test MK was from Seikagaku(Tokyo, Japan). Distilled water (DIW) and physiologicalsaline were from Otsuka (Tokyo, Japan). The visking tube(MW cuto¡: 3500) was from Spectrum Medical, CA,USA. Salbutamol, prednisolone, gadolinium chloride hex-ahydrate, D-galactosamine hydrochloride, cyclophospha-mide monohydrate and D-mannose were from Wako (Osa-ka, Japan). Carrageenan was from Sigma, USA. Mannan(from yeast) was from Nakarai Chemicals, (Kyoto, Ja-pan).

2.2. Mice

Male 5^10-week-old closed colonies (ICR and ddY),inbred strains (A/J, AKR/N, BALB/c, CBA/N, C3H/He,C3H/HeJ, C3H/HeN, C57BL/6 and DBA/2) mutant strain(KSN nu-nu) and hybrids (B6C3F1; C57BL/6UC3H/He,

BDF1; C57BL/6UDBA/2, CBF1; BALB/cUC57BL/6,CDF1; BALB/cUDBA/2, DC3F1; DBA/2UC3H/Heand WBB6F1-w/wv ; WB-W/+UC57BL/6-Wv/+) were ob-tained from Japan SLC, Shizuoka, Japan.

2.3. Media

C-limiting medium originally described by Shepherdand Sullivan [16] was used to grow all strains of yeast.C-limiting medium contained (per liter) : sucrose 10 g,(NH4)2SO4 2 g, KH2PO4 2 g, CaCl2W2H2O 0.05 g,MgSO4W7 H2O 0.05 g, ZnSO4W7 H2O 1 mg, CuSO4W5H2O 1 mg, FeSO4W7 H2O 0.01 g, biotin 25 Wg, ¢nal pH5.2. Five liters of medium was placed in a glass jar of amicroferm fermentor (New Braswick), and cultured at27³C with 5 l min31 of aeration and with 400 rpm ofstirring.

2.4. Preparation of water soluble fraction (WS)

Water soluble fraction (WS) was prepared as previouslydescribed [9]. Brie£y, to collect the CAWS, an equal vol-ume of ethanol was added to the culture and the precip-itate was collected, which included cells and secreted mac-romolecules. The precipitate was then suspended in analiquot of distilled water and the solubilized part was col-lected and acetone dried. Similar methods were used tocollect a water soluble fraction derived from C. parapsilo-sis (CPWS) and S. cerevisiae (SCWS). A chemical charac-terization of WS was done previously [9]. Brie£y, eventhough there were strain di¡erences in the yeast, the per-centage of carbohydrate and protein content was78 þ 6.6% and 15 þ 6.3%, respectively. And the M (man-nan)/G (glucan) ratio was 6.3 þ 1.3 in the case of C. albi-cans IFO 1385

2.5. Preparation of plasma for L-D-glucan assay

CAWS was diluted with physiological saline and auto-claved. The optimal dose of CAWS was administered tomice by intravenous or intraperitoneal injection. An ali-quot of blood was collected from the tail vein at appro-priate intervals using heparinized capillaries. After centri-fugation, plasma samples were stored at 4³C until the Gtest (Fungitec G test MK) was performed.

2.6. Measurement of blood concentration of CAWS usingthe Fungitec G test MK

Activation of factor G (Limulus reactivity) by L-1,3-D-glucan was measured by a chromogenic method using aL-1,3-D-glucan-speci¢c reagent (Fungitec G test MK),which eliminates factor C [11^13]. Plasma containingCAWS was mixed with 0.5 N NaOH to degrade any pro-tease in plasma. The resulting mixture was serially dilutedwith 0.01 N NaOH. Usually, dilutions were made with

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^7670

0.01 N NaOH and a sample solution was used directly forthe Limulus reaction without neutralization. DilutedNaOH was con¢rmed to be usable for the Limulus reac-tion by the high bu¡er action of the reagent. Reactionswere performed in a £at-bottomed 96-well Toxipet plate96F (Seikagaku) and a microplate reader (WellreaderSK603, Seikagaku). The CAWS concentration in bloodwas calculated from the relative reactivity to CAWS as astandard. Disposable plastic materials for tissue culture orclinical use were employed and all glassware was sterilizedat 260³C for 3 h. The operations were performed in du-plicate under aseptic conditions. The data are shown asthe average of two experiments.

2.7. Calculation of half-life by cricket version 1.3

In order to compare the half-life of CAWS with that ofSPG under various conditions in mice, curve ¢tting wasperformed using the program Cricket version 1.3. Thehalf-life (t1=2) was calculated using the one-compartmentmodel.

3. Results

3.1. Lethal toxicity of CAWS in various strains of mice

In the preliminary experiment, on intravenous adminis-tration CAWS (from C. albicans IFO 1385, 1 mg) waslethal to ICR mice. In contrast, on intraperitoneal admin-istration CAWS did not show any toxicity. As described inSection 1, some of the yeast mannan fractions are known

to be lethal to mice when administered intravenously [12].Expression of lethal toxicity presumably depends on themannan moiety in CAWS. To con¢rm the lethal toxicity,the WS from various strains of yeast were tested. Asshown in Table 1, WS from all strains of yeast used inthis study exhibited lethality. And so we screened CAWS-mediated lethal toxicity in various strains of mice. A doserange of 10 Wg to 2 mg mouse31 was used. As shown inTable 2, DBA/2 mice were the least sensitive to CAWS.BALB/c, C3H strains, WBB6F1-W/Wv, KSN nu-nu andddY exhibited lethality as did ICR mice. Initially, thesemice were motionless and had a decreased body temper-ature. Later, they died with respiratory distress and con-vulsion within 30 min. Although the A/J, AKR/N, CBA/Nand C57BL/6 mice survived, they were distressed and mo-tionless. Among the hybrids, B6C3F1 and DC3F1 exhib-ited lethality.

Table 1Lethal toxicity of WS from various strains of yeast

Strains Dead/total

C. albicans 0579 1/6C. albicans 0583 3/5C. albicans 1061 3/5C. albicans 1385 3/3C. albicans 1594 3/5C. parapsilosis 0708 1/5C. parapsilosis 1068 4/5S. cerevisiae 1136 2/3

WS, prepared from various strains of yeast, was dissolved in saline andautoclaved. Each WS (1 mg mouse31) was administered to ICR mice byintravenous injection. Lethality was monitored 1 h later.

Table 2Lethal toxicity of CAWS in various strains of mice

Dead/total

2 mg mouse31 1 mg mouse31 0.25 mg mouse31 0.1 mg mouse31 0.01 mg mouse31

ICR ^ 3/3 2/3 0/3 ^ddY ^ 3/3 2/3 1/3 ^A/J 0/3 0/3 ^ ^ ^AKR/N 0/3 0/5 ^ ^ ^BALB/c 0/3 2/9 ^ ^ ^CBA/N 0/3 0/3 ^ ^ ^C3H/He ^ 2/3 ^ 2/3 0/3C3H/HeJ ^ 3/3 ^ 2/3 1/3C3H/HeN ^ 3/3 3/3 3/3 ^C57BL/6 0/3 0/5 ^ ^ ^DBA/2 0/5 0/8 ^ ^ ^KSN nu-nu ^ 3/3 3/3 2/2 0/2B6C3F1 ^ 3/5 1/3 ^ ^BDF1 0/3 0/5 ^ ^ ^CBF1 0/3 0/5 ^ ^ ^CDF1 0/4 0/4 ^ ^ ^DC3F1 ^ 3/3 3/3 0/2 0/2WBB6F1-w/wv ^ 3/3 3/3 2/2 0/2

^, not done. Each 2, 1, 0.25, 0.1, or 0.01 mg of CAWS was intravenously administered to various strains of mice. Lethality was monitored 1 h later.Closed colonies: ICR and ddY. Inbred strains: A/J, AKR/N, BALB/c, CBA/N, C3H/HeJ, C3H/HeN, C57BL/6, DBA/2. Mutant strain: KSN nu-nu.Hybrids: B6C3F1, BDF1, CBF1, CDF1, DC3F1, WBB6F1-w/wv. CAWS used in this experiment is derived from C. albicans 1385.

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^76 71

In addition, we tested whether aging a¡ected the lethal-ity. As a consequence, aging (ICR mice over 6 months old)did not have an e¡ect on lethality (dead/total : 1 mgmouse31, 2/3; 0.25 mg mouse31, 0/3; 0.1 mg mouse31,1/3; CAWS used in this experiment was derived from1385).

3.2. E¡ect of various treatments on lethal toxicity inducedby CAWS

It is generally accepted that some sorts of activities are

modi¢ed by treatment, and such results provide informa-tion about the mechanisms at work. Various reagents havebeen found to act as macrophage blockers. Carrageenan(sulfated polygalactose) primes leukocytes to enhance lipo-polysaccharide (LPS)-induced tumor necrosis factor(TNF) production and lethality, and destroys macro-phages; it is used to inhibit macrophage function [17,18].Gadolinium chloride (GdCl3) is also well known as a se-lective blocker of macrophages, especially Kup¡er cells[19^22]. Colloidal aggregates of GdCl3 at a pH above6.0 in the circulation after intravenous injection, might

Table 3E¡ect of various treatments on lethal toxicity induced by CAWS

Pretreatment Dose mouse31 Dead/total

CAWS dose

2 mg mouse31 1 mg mouse31 0.25 mg mouse31 0.1 mg mouse31

Untreated ^ ^ 3/3 2/3 0/3Carrageenan 5 mg ^ 4/6 0/3 0/7Gadolinium 0.2 mg 2/3 0/3 ^ ^Cyclophosphamide 6 mg ^ 2/3 2/3 0/3Galactosamine 20 mg 1/4 1/3 0/2 ^Prednisolone 90 Wg 3/3 1/3 ^ ^

9 Wg 2/3 0/3 ^ ^Salbutamol 300 Wg 0/3 0/3 ^ ^

30 Wg 0/3 1/3 ^ ^Mannose 22.5 mg ^ 3/3 ^ ^

15 mg ^ ^ 3/3 ^Yeast mannan 1 mg ^ 0/3 ^ ^

0.5 mg ^ 0/3 ^ ^0.1 mg ^ 1/3 ^ ^

^, not done. Carrageenan, cyclophosphamide, galactosamine, prednisolone and salbutamol sulfate were administered intraperitoneally, 12 h, 3 days,1 min, 2 h and 30 min before CAWS, respectively. Gadolinium chloride was administered intravenously 1 and 2 days before CAWS. Mannose andyeast mannan were administered intravenously 5 min before CAWS. ICR mice and CAWS derived from C. albicans 1385 were used in this experiment.

Fig. 1. Comparison of Limulus factor G reactivity of WS from various strains of yeast and SPG treated with distilled water or NaOH. Each samplewas dissolved in saline (5 mg ml31), and sterilized by autoclaving. An aliquot of each sample (5 Wl) was mixed with distilled water (45 Wl) or 0.5 NNaOH (45 Wl). The resulting solutions were diluted with distilled water or 0.01 N NaOH (¢nal concentration of CAWS or SPG was 1000 or 100 ngml31, respectively). Milliabsorbance min31 was measured with the Fungitec G test MK using a Wellreader SK603 (kinetic assay).

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^7672

be taken up by Kup¡er cells and kill them by destroyingtheir phagosomes [20]. As shown in Table 3, these sub-stances reduced the lethality slightly, but not completely.In the case of cyclophosphamide-treated mice, no reduc-tion in toxicity was observed. Galactosamine enhances thesensitivity to TNF and lethal toxicity induced by LPS [23^25]. However, the treatment did not modulate the CAWS-induced toxicity.

Platelet-activating factor presumably plays an importantrole in the lethality of yeast mannan. On this account, wehad used salbutamol sulfate and prednisolone to reducethe e¡ect. As shown in Table 3, salbutamol sulfate andprednisolone, at least in part, reduced the lethality.

The, lethality seems to depend on the mannan moiety inCAWS. Consequently, we used mannose and mannan(from yeast; Nakarai) to inhibit the CAWS-mediated le-thality. We found a commercially available mannan prep-aration (from yeast ; Nakarai) that did not exhibit lethalityby itself (lethal toxicity of yeast mannan in ICR mice:dead/total : 1 mg mouse31, 0/3; 0.25 mg mouse31, 0/3;0.1 mg mouse31, 0/3), and used it to block the mannosereceptors. Yeast mannan, but not mannose, inhibited thelethality (Table 3).

3.3. Conformation of WS assessed by Limulus G test

It is generally accepted that a high molecular massL-1,3-D-glucan forms two conforms, a single helix and atriple helix. The helix structure was interchangeable ontreatment with NaOH or autoclaving. In contrast, thesmaller L-1,3-D-glucan exists as a random coiled conform-er. Conformation is an important contributing factor inthe biological as well as physicochemical properties ofL-1,3-D-glucan [26], thus the conformation of WS was ex-amined before testing blood clearance. We have previously

established that the single helix and the random coiledconformer reacted strongly with Limulus G, but the triplehelix conformer did not [27]. For example, the single helixform of SPG treated with NaOH showed high reactivity toLimulus factor G compare to SPG treated with distilledwater. In addition, the single helix conformer could bechanged to the triple form by heat treatment.

In this study, the Limulus reactivity of WS from variousstrains of yeast was tested after autoclaving and/or NaOHtreatment. As shown in Fig. 1, the Limulus reactivity ofWS treated with NaOH was higher than that of WStreated with distilled water. It is strongly suggested, there-fore, L-1,3-D-glucan in WS is, for the most part, present inthe random coiled form rather than the triple helix form.For this reason, we used NaOH treatment before theLimulus G test in this study.

Fig. 2. Blood concentration of CAWS on intravenous administration athigh dose in DBA/2 mice. CAWS (1385:1000 Wg) was intravenously ad-ministered to DBA/2 mice. An aliquot of blood was collected from thetail vein at appropriate intervals using heparinized capillaries. Plasmawas obtained and each sample was dissolved in 0.5 N NaOH and seri-ally diluted with 0.01 N NaOH. CAWS concentrations were measuredwith Fungitec G test MK using a Well-reader SK603 (kinetic assay).

Fig. 3. Blood concentration of CAWS on intravenous administration atlow dose in DBA/2 mice. CAWS (1385:100 Wg) was intravenously ad-ministered to DBA/2 mice. An aliquot of blood was collected from thetail vein at appropriate intervals using heparinized capillaries. Plasmawas obtained and each sample was dissolved in 0.5 N NaOH and seri-ally diluted with 0.01 N NaOH. CAWS concentrations were measuredwith Fungitec G test MK using a Wellreader SK603 (kinetic assay).

Fig. 4. Blood concentration of SPG on intravenous administration atlow dose in DBA/2 mice. SPG (100 Wg) was intravenously administeredto DBA/2 mice. An aliquot of blood was collected from the tail vein atappropriate intervals using heparinized capillaries. Plasma was obtainedand each sample was dissolved in 0.5 N NaOH and serially diluted with0.01 N NaOH. SPG concentrations were measured with Fungitec G testMK using a Wellreader SK603 (kinetic assay).

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^76 73

3.4. Measurement of blood concentration of CAWS byintravenous or intraperitoneal administration in DBA/2mice

CAWS (from C. albicans IFO 1385: 1000 Wg or 100 Wg)was intravenously administered to DBA/2 mice and bloodwas collected from the tail vein at appropriate intervals.We measured the chronological variation in blood concen-tration of CAWS by Limulus G test. As shown in Figs. 2and 3, the half-life of CAWS at high dose (1000 Wg) andlow dose (100 Wg) was 36 and 19 min, respectively. It wassigni¢cantly faster than that of SPG, which is clinicallyused as immunomodulating L-1,3-D-glucan (Fig. 4).

In the case of intraperitoneal administration, the max-imum blood concentration of CAWS was approximately

300 Wg ml31 at 60 min after CAWS administration (Fig.5). In ICR mice, CAWS was not lethal on intraperitonealadministration, and showed a similar clearance to that inDBA/2 mice.

3.5. Inhibition of the toxicity of CAWS by using salbutamolsulfate or prednisolone, to measure the bloodconcentration of CAWS in ICR mice

We had to investigate the e¡ect of inter-strain di¡erenceof mice on the clearance rate. Thus, we wanted to measurethe blood clearance of CAWS in ICR mice. As describedearlier, the lethality of CAWS in ICR mice could be, atleast in part, inhibited by salbutamol sulfate and predni-solone. Thus, we measured the blood clearance of CAWSin salbutamol or prednisolone treated ICR mice. Asshown in Fig. 6, the half-life of CAWS in salbutamoland prednisolone prophylactically treated ICR mice was57 and 64 min, respectively. The rate was slightly slowerthan that in DBA/2.

4. Discussion

We have reported that CAWS was mainly composed ofa complex of mannan, L-1,6-D-glucan and L-1,3-D-glucan[9]. Because of the lethal toxicity of CAWS on intravenousadministration in ICR mice, it was di¤cult to measure theblood clearance. It is generally accepted that certain activ-ities are modulated by genetic background. For example:(1) C3H/HeN mice are more sensitive to endotoxin shockprovoked by LPS than C3H/HeJ mice [28]. It has alreadybeen veri¢ed that the disparate sensitivity depends on apoint mutation of amino acid in Toll-like-receptor 4 whichis the signaling receptor of LPS; (2) delayed type-acutephase response (DT-APR) and vascular dilation and hem-orrhage (VDH) by antitumor L-1,6;1,3-D-glucan, lentinan,were controlled by two genes [29,30]; and (3) there weregenetic di¡erences between BALB/c and B10.D2 T-cellsfor T-helper phenotype development [31]. In this study,we found that CAWS-mediated lethal toxicity was straindependent and DBA/2 mice were the least sensitive. Le-thality was observed in KSN nu-nu mice. This result sug-gested that the T-cell might not be participated in thelethality. In addition, we investigated the role of mastcells, which play a central role in anaphylaxis, in the tox-icity of CAWS by using mast-cell de¢cient WBB6F1-w/wv

mice. Lethal toxicity was also observed in this strain, sug-gesting weak contribution to the lethality. The ¢ndingsthat marked alterations in toxicity were not observed incarrageenan, gadolinium chloride or cyclophosphamidetreated mice, suggest that the macrophage and lymphaticsystem do not play a central role in lethality. Carrageenanand galactosamine, typically used to enhance the sensitiv-ity and lethality of LPS, did not modulate the toxicity,suggesting independence to LPS-mediated toxicity.

Fig. 6. Blood concentration of CAWS on intravenous administration inICR mice prophylactically treated with salbutamol or prednisolone. Sal-butamol (300 Wg) or prednisolone (90 Wg) was administered 30 or120 min before CAWS, respectively. An aliquot of blood was collectedfrom the tail vein at appropriate intervals using heparinized capillaries.Plasma was obtained and each sample was dissolved in 0.5 N NaOHand serially diluted with 0.01 N NaOH. CAWS concentrations weremeasured with Fungitec G test MK using a Wellreader SK603 (kineticassay).

Fig. 5. Blood concentration of CAWS on intraperitoneal administrationin ICR and DBA/2 mice. CAWS (1385: 1 mg) was intraperitoneally ad-ministered to DBA/2 and ICR mice. An aliquot of blood was collectedfrom the tail vein at appropriate intervals using heparinized capillaries.Plasma was obtained and each sample was dissolved in 0.5 N NaOHand serially diluted with 0.01 N NaOH. CAWS concentrations weremeasured with Fungitec G test MK using a Wellreader SK603 (kineticassay).

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^7674

We measured the blood clearance of CAWS in DBA/2mice. The half-life of CAWS (V40 min) was faster thanthat of SPG (V6 h). These results indicated that CAWSmight disappear via a di¡erent pathway like to SPG. Asmentioned above, the major component of CAWS is man-nan, thus, like mannan in blood, CAWS may be trappedby speci¢c receptors, such as the mannose receptor.

In addition, we had to consider the inter-strain di¡er-ence in clearance among mice. By using salbutamol andprednisolone, we measured the blood concentration inICR mice, and found the half-life to be as fast as that inDBA/2.

We also measured the clearance of intraperitoneally ad-ministered CAWS. The blood concentration reached amaximum at around 1 h. We have previously reportedthe value of SPG as 2^4 h in ICR mice. The faster attain-ment of the maximum level with CAWS would be due tothe faster clearance of CAWS from the blood stream.

Conformation of L-1,3-D-glucan is an important factorcontributing to biological activity, such as Limulus factorG activation, nitric oxide synthesis by macrophages, andso on. The clearance of SPG was also a¡ected by its con-formation. The single helical conformer cleared faster thanthe triple helix. In addition to the triple and the singlehelix conformation, L-1,3-D-glucan has a random coiledconformation, especially at lower molecular mass andwhen charged. These facts strongly suggested that theclearance of CAWS is in£uenced by its conformation. Ingeneral, the conformation of the triple helix is stable com-pared with that of the single helix. Autoclaving of L-1,3-D-glucan preparations induced a conformation change fromthe single to the triple form. As shown in Fig. 1, the triplehelix of SPG was changed to a single helix by sodiumhydroxide treatment, resulting in a form highly reactiveto Limulus factor G. Sodium hydroxide treatment is alsoimportant to negate the in£uence of conformation on theLimulus G test. In the WS preparation, the autoclavedpreparation showed signi¢cant activity and this activitywas slightly increased after sodium hydroxide treatment.The WS preparation would mainly be of the random coilconformation. In a previous study, we analyzed the struc-ture of WS by NMR spectroscopy and found the ratio ofL-1,3-D-glucan moiety to be low. This ¢nding also sup-ports the conformation of the L-1,3-D-glucan moiety ofWS as a random coil.

Mikami et al. used salbutamol, a L2-adrenoceptor ago-nist, to reduce the lethality of yeast mannan [12]. Accord-ing to their results, PAF- and yeast mannan-induced le-thality was blocked by such a treatment. They hadconcluded that yeast mannan was able to induce PAF inmice, and the factor that causes the lethality was PAF.Similarly, we have been demonstrated that the CAWS-in-duced lethality was blocked by treatment with salbutamol.These results indicated that CAWS-induced lethality de-pends on the mannan moiety, and PAF plays a centralrole in the toxicity.

In conclusion, our results strongly suggested that theLimulus factor G-reactive substance in patients' bloodcame from the infecting fungi and would rapidly disappearwith the killing of fungi in infected tissue.

Acknowledgements

We express our thanks to Daichi Kakuta for his excel-lent technical assistance.

References

[1] Garner, R.E. and Hudson, J.A. (1996) Intravenous injection of Can-dida-derived mannan results in elevated tumor necrosis factor alphalevels in serum. Infect. Immun. 64, 4561^4566.

[2] Branco, L., Pitta, M.L., Bernardes, L., Galrinho, A., Agapito, A.F.,Ramos, J.M., Quininha, J., Figueiredo, L., Bento, R. and Mata, A.(1997) A review of infectious endocarditis due to Candida. Rev. Port.Cardiol. 16, 967^974.

[3] Osumi, M. (1998) The ultrastructure of yeast cell wall structure andformation. Micron 29, 207^233.

[4] Bishop, C.T., Blank, F. and Gardner, P.E. (1960) The cell wall poly-saccharides of Candida albicans : glucan, mannan and chitin. Can. J.Chem. 38, 869^881.

[5] Shepherd, M.G. (1991) The structure and function of Candida albi-cans cell wall. Jpn. J. Med. Mycol. 32, 63^73.

[6] Kapteyn, J.C., Montijn, R.C., Dijkgraaf, G.J., Van den Ende, H. andKlis, F.M. (1995) Covalent association of beta-1,3-glucan with beta-1,6-glucosylated mannoprotein in cell walls of Candida albicans.J. Bacteriol. 177, 3788^3792.

[7] Jiang, B., Sheraton, J., Ram, A.F., Dijkgraaf, G.J., Klis, F.M. andBussey, H. (1996) CWH41 encodes a novel endoplasmic reticulummembrane N-glycoprotein involved in beta 1,6-glucan assembly.J. Bacteriol. 178, 1162^1171.

[8] Montijin, R.C., Rinsum, J., Schagen, F.A. and Klis, F.M. (1994)Glucomannoproteins in the cell wall of Saccharomyces cerevisiae con-tain a novel type of carbohydrate side chain. J. Biol. Chem. 269,19338^19342.

[9] Uchiyama, M., Ohno, N., Miura, N.N., Adachi, Y., Aizawa, M.W.,Tamura, H., Tanaka, S. and Yadomae, T. (1999) Chemical and im-munochemical characterization of limulus factor G-activating sub-stance of Candida spp. FEMS Immunol. Med. Microbiol. 24, 411^420.

[10] Miura, N.N., Ohno, N., Aketagawa, J., Tamura, H., Tanaka, S. andYadomae, T. (1996) Blood clearance of (1C3)-L-D-glucan in MRLlpr/lpr mice. FEMS Immunol. Med. Microbiol. 13, 51^57.

[11] Obayashi, T., Yoshida, M., Mori, T., Goto, H., Yasuoka, A., Iwa-saki, H., Teshima, H., Kohno, S., Horiuchi, A., Ito, A., Yamaguchi,H., Shimada, K. and Kawai, T. (1995) Plasma (1C3)-L-D-glucanmeasurement in diagnosis of invasive deep mycosis and fungal febrileepisodes. Lancet 345, 17^20.

[12] Mikami, T., Fukushi, K., Ishitani, M., Ishitani, K., Suzuki, S. andSuzuki, M. (1991) Induction of platelet-activating factor in mice byintravenous administration of a neutral fraction bakers' yeast man-nan. Lipids 26, 1404^1407.

[13] Miura, N.N., Ohno, N., Adachi, Y., Aketagawa, J., Tamura, H.,Tanaka, S. and Yadomae, T. (1995) Comparison of the blood clear-ance of triple- and single-helical Schizophyllan in mice. Biol. Pharm.Bull. 18, 185^189.

[14] Norisue, T. (1989) Triple-stranded helical structure of Schizophyllanand its antitumor activity. Seibutu-buturi. 29, 35^38.

[15] Norisue, T., Yanaki, T. and Fujita, H. (1980) Triple helix of a Schiz-

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^76 75

ophyllum Commune polysaccharide in aqueous solution. J. Polym.Sci. Polym. Phys. Ed. 18, 547^558.

[16] Shepherd, M.G. and Sullivan, P.A. (1976) The production andgrowth characteristics of yeast and mycelial forms of Candida albi-cans in continuous culture. J. Gen. Microbiol. 93, 361^370.

[17] Ogata, M., Yoshida, S., Kamochi, M., Shigematsu, A. and Mizugu-chi, Y. (1991) Enhancement of lipopolysaccharide-induced tumor ne-crosis factor production in mice by carrageenan pretreatment. Infect.Immun. 59, 679^683.

[18] Ogata, M., Matsumoto, T., Koga, K., Takenaka, I., Kamochi, M.,Sata, T., Yoshida, S. and Shigematsu, A. (1993) An antagonist ofplatelet-activating factor suppresses endotoxin-induced tumor ne-crosis factor and mortality in mice pretreated with carrageenan. In-fect. Immun. 61, 699^704.

[19] Squiers, E.C., Brunson, M.E. and Salomon, D.R. (1993) Kup¡er cellscan present alloantigen in vitro: an e¡ect abrogate by gadolinium.J. Surg. Res. 55, 571^574.

[20] Iimuro, Y., Yamamoto, M., Kohno, H., Itakura, J., Fujii, H. andMatsumoto, Y. (1994) Blockade of liver macrophages by gadoliniumchloride reduces lethality in endotoxemic rats-analysis of mechanismslethality in endotoxemia. J. Leukocyte Biol. 55, 723^728.

[21] Lazar Jr., G., Lazar, G., Kaszaki, J., Olah, J., Kiss, I. and Husztik,E. (1994) Inhibition of anaphylactic shock by gadolinium chloride-induced Kup¡er cell brockade. Agents Actions 41, C97^C98.

[22] Hardonk, M.J., Dijkhuis, F.W., Hulstaert, C.E. and Koudstaal, J.(1992) Heterogeneity of rat liver and spleen macrophages in gadoli-nium chloride-induced elimination and repopulation. J. LeukocyteBiol. 52, 296^302.

[23] Galanos, C., Freudenberg, M.A. and Reutter, W. (1979) Galactos-amine-induced sensitization to the lethal e¡ects of endotoxin. Proc.Natl. Acad. Sci. USA 76, 5939^5943.

[24] Freudenberg, M.A., Keppler, D. and Galanos, C. (1986) Require-ment for lipopolysaccharide-responsive macrophages in galactos-amine-induced sensitization to endotoxin. Infect. Immun. 51, 891^895.

[25] Lehmann, V., Freudenberg, M.A. and Galanos, C. (1987) Lethaltoxicity of lipopolysaccharide and tumor necrosis factor in normaland D-galactosamine-treated mice. J. Exp. Med. 165, 657^663.

[26] Ohno, N., Emori, Y., Yadomae, T., Saito, K., Masuda, A. and Oi-kawa, S. (1990) Reactivity of limulus amoebocyte lysate toward(1C3)-L-D-glucans. Carbohydr. Res. 207, 311^318.

[27] Nagi, N., Ohno, N., Adachi, Y., Aketagawa, J., Tamura, H., Shibata,Y., Tanaka, S. and Yadomae, T. (1993) Application of limulus test(G pathway) for the detection of di¡erent conformers of (1C3)-L-D-glucans. Biol. Pharm. Bull. 16, 822^828.

[28] Poltorak, A., He, X., Smirnova, I., Liu, M.Y., Hu¡el, C.V., Du, X.,Birdwell, D., Alejos, E., Silva, M., Galanos, C., Freudenberg, M.,Ricciardi-Castagnoli, P., Layton, B. and Beutler, B. (1998) DefectiveLPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations inTlr4 gene. Science 282, 2085^2088.

[29] Maeda, Y.Y., Takahama, S., Kohara, Y. and Yonekawa, H. (1996)Two genes controlling acute phase responses by the antitumor poly-saccharide, lentinan. Immunogenetics 43, 215^219.

[30] Maeda, Y.Y., Takahama, S., Kohara, Y. and Yonekawa, H. (1997)Polygenic control of the expression of biological activities of an anti-tumor polysaccharide, lentinan. Int. J. Immunopharmacol. 19, 469^472.

[31] Guler, M.L., Jacobson, N.G., Gubler, U. and Murphy, K.M. (1997)T cell genetic background determines maintenance of IL-12 signaling:e¡ects on BALB/c and B10.D2 T helper cell type 1 phenotype devel-opment. J. Immunol. 159, 1767^1774.

FEMSIM 1244 21-8-00

K. Kurihara et al. / FEMS Immunology and Medical Microbiology 29 (2000) 69^7676