Engagement of Penicillium marneffei conidia with multiple pattern recognition receptors on human monocytes


Sansanee C. Chaiyaroj, Department of Microbiology, Faculty of Science, Mahidol University, Bangkok 10400, Thailand. Tel: +66 2 20 5677; fax: +66 2 644 5411; email: scscy@mahidol.ac.th


P. marneffei is a thermal dimorphic fungus which causes penicilliosis, an opportunistic infection in immunocompromised patients in South and Southeast Asia. Little is known about the innate immune response to P. marneffei infection. Therefore, the initial response of macrophages to P. marneffei conidia was evaluated by us. Adhesion between monocytes from healthy humans and fungal conidia was examined and found to be specifically inhibited by MAbs against PRR, such as MR, (TLR)1, TLR2, TLR4, TLR6, CD14, CD11a, CD11b, and CD18. To study the consequences of these interactions, cytokines were also examined by ELISA. Binding of P. marneffei conidia to monocytes was significantly inhibited, in a dose-dependent manner, by MAbs against MR, TLR1, TLR2, TLR4, TLR6, CD14, CD11b and CD18. When monocytes were co-cultured with the conidia, there was an increase in the amount of surface CD40 and CD86 expression, together with TNF-α and IL-1β production, compared to unstimulated controls. In assays containing anti-TLR4 or anti-CD14 antibody, reduction in the amount of TNF-α released by monocytes stimulated with P. marneffei conidia was detected. In addition, it was found that production of TNF-α and IL-1β from adherent peripheral blood monocytes was partially impaired when heat-inactivated autologous serum, in place of untreated autologous serum, was added to the assay. These results demonstrate that various PRR on human monocytes participate in the initial recognition of P. marneffei conidia, and the engagement of PRR could partly initiate proinflammatory cytokine production.

List of Abbreviations: 


A. fumigatus

Aspergillus fumigatus


alveolar macrophages



C. albicans

Candida albicans


complement receptor three





H. capsulatum

Histoplasma capsulatum


human monocytes


iduronic acid




leukocyte function-associated antigen 1


monoclonal antibodies


macrophage-inflammatory protein-2


multiplicity of infection


mannose receptor


myeloid differentiation primary response gene 88


nuclear factor kappa-light-chain-enhancer of activated B cells


peripheral blood mononuclear cells



P. marneffei

Penicillium marneffei

P. jiroveci

Pneumocystis jiroveci


pattern recognition receptors


room temperature


Toll-like receptor


tumor necrosis factor-α

Penicilliosis marneffei is a disease caused by P. marneffei. The fungus is unique among Penicillium spp. in that it is thermal dimorphic and secretes red pigment in culture medium. At 25 °C, P. marneffei is in filamentous mold form. At 37 °C, it transforms into fission yeast, the pathogenic form which has been found in macrophages of penicilliosis patients. Penicilliosis has been considered to be an important emerging opportunistic fungal infection occurring in immunocompromised patients, predominantly in those with AIDS. The incidence of HIV-related penicilliosis has been increasing in endemic areas including China, Hong Kong (1–3), India (4), and Thailand (1, 5). In Thailand, approximately 100 cases were reported from 1973 to 1990. The number of cases rose to more than 2350 by 1995 and had reached 6808 by March 31, 2006 (Bureau of Epidemiology, Department of Disease Control, Ministry of Public Health; personal communication). Penicilliosis has been defined as one of the AIDS-specific diseases, and can be fatal in immunocompromised patients who receive inappropriate anti-fungal treatment.

A better understanding of the initial interaction of P. marneffei conidia with host target cells, and of host innate immunity to the infection, is required to develop strategies aimed at boosting management with antifungals. Many investigators, including our group, have suggested that infection with P. marneffei originates in the lung following inhalation of airborne conidia (6, 7). When P. marneffei conidia reach the alveoli, the conidia are likely to interact with extracellular matrix components and subsequently be exposed to AM. We have recently demonstrated that laminin, fibronectin and IdoA containing glycosaminoglycans serve as important adhesion molecules on the surface of lung epithelial cells (8). Similarly, macrophages, which possess many cell-surface receptors, can potentially interact either directly with the fungus or via opsonins deposited on the fungal surface.

Several leukocyte surface molecules have previously been identified as receptors for pathogenic fungi. Dong and colleagues have demonstrated that Cryptococcus neoformans utilizes β-integrin (CD18) and CR3 (Mac-1, CD11b/CD18) in its interaction with human neutrophils (9). Mac-1 (CR3) has also been reported to be a receptor for the yeast, Candida albicans (10). With P. marneffei, Rongrungruang and Levitz have demonstrated that the adhesion of unopsonized P. marneffei yeasts to macrophages occurs independently of the mannose and β-glucan receptors, and CD11/CD18; rather, adhesion occurs through interaction with a glycoprotein receptor which contains N-acetyl-β-D-glucosaminyl groups (11).

In addition to integrins, MR and dectin-1 have been demonstrated to bind a wide variety of fungal pathogens, including C. albicans and Aspergillus fumigatus. Generally, MR has been implicated as a phagocytic receptor, and its engagement leads to proinflammatory cytokine production (12). Another C-type lectin receptor, dectin-1, mediates proinflammatory cytokine production on live C. albicans activation (13). Brown and colleagues have reported that, compared to control RAW cells, dectin-transfected RAW cells produce significant concentrations of TNF-α when they are infected with live C. albicans (13). Dectin-1 was also demonstrated to respond to β-glucan on A. fumigatus conidia (14, 15).

In recent years, involvement of TLR in host response to pathogenic fungi has been demonstrated. TLR4, together with CD14, has been reported to function as a receptor for C. neoformans, C. albicans yeasts and A. fumigatus hyphae (16–18). Recognition of a major component of the C. neoformans polysaccharide capsule, GXM, by TLR4 and CD14, but not TLR2, results in signal transduction through MyD88 and NF-κB (16, 19). With C. albicans, an increase in fungal burden has been observed in TLR4-mutant mice in comparison to wild-type mice (20). In contrast, Marr et al. have demonstrated that proinflammatory cytokine secretion by macrophages stimulated with C. albicans yeast results from TLR2 engagement and MyD88 expression, but is not dependent on activation through TLR4 (21, 22). Further work is required to resolve this apparent paradox. An important role for TLR in the recognition of Aspergillus has also been reported by Wang et al. and others (17, 23). Engagement of TLR4 and CD14 with A. fumigatus hyphae has been shown to lead to the production of proinflammatory cytokines, such as TNF-α, IL-1β, and IL-6 (17).

In the present study, it was hypothesized that multiple pattern recognition receptors are involved in the initial interaction of P. marneffei conidia with macrophages. The role of integrins, MR and TLR as receptors for P. marneffei conidia was evaluated by in vitro experiments which employed blocking of monoclonal antibodies against these groups of molecules in order to inhibit binding between P. marneffei conidia and HM. The function of integrins and TLR in the induction of cytokine production from activated macrophages was also investigated.


Fungal pathogen and culture conditions

A standard strain of P. marneffei (ATCC 64102, American Type Culture Collection, Rockville, MD, USA), was cultivated on malt extract agar slant in a flat bottle at 25 °C for 2 weeks. Conidia were harvested as previously described (24). In brief, colonies of P. marneffei were flooded with PBS and gently scraped with a sterile wire loop. The fungal suspension was centrifuged at 300 ×g for 5 min at RT. The conidial-enriched supernatant was collected and enumerated using a hemacytometer. The suspension, contaminated with hyphae of less than 1%, was subjected to storage at 4 °C until use.

P. marneffei conidia labeling

P. marneffei conidia were labeled with FITC (Sigma-Aldrich, St. Louis, MO, USA) using the modified methods of Campbell (25). Briefly, 1.0–1.5 × 109 conidia were suspended in 0.1 mg/ml FITC in NaHCO3 buffer (pH 9.0) and left for 1 h at RT with periodic shaking. FITC-labeled conidia were washed five times with PBS and counted under the fluorescence microscope (Nikon, New York, NY, USA).

MAbs and carbohydrate compounds

Human-specific MAbs (IgG isotype of murine origin) used in this study included antibodies against CD11a (clone 25.3), CD11b (clone BEAR-1), CD18 (clone 7E4), CD14 (clone M5E2), and mannose receptor (clone 3.29B1.10). The MAbs were purchased from Immunotech (Coulter, Marseille, France). The MAbs against human TLR1 (clone GD2.F4) and TLR2 (clone TL2.1) were obtained from Alexis Biochemical (Lausen, Switzerland). The MAb that recognizes the extracellular domain of human TLR6 (clone TLR6.127) was purchased from Hycult biotechnology (Uden, the Netherlands). These MAbs have previously been shown to be blocking MAbs (26–30). Anti-TLR4 MAb (clone HTA125) was a gift from Prof. K. Miyake (University of Tokyo, Tokyo, Japan). An irrelevant mouse IgG isotype (anti-Opisthorchis viverrini antibody) was employed as an antibody control. None of the MAbs was found to affect cell viability as determined by trypan blue exclusion. Mannan and laminarin were purchased from Sigma-Aldrich. For flow cytometric analysis, FITC-labeled anti-human CD14, PE-labeled anti-human CD86, and APC-labeled anti-human CD40 MAbs were obtained from BD Biosciences (San Diego, CA, USA).


Monocytes were prepared from PBMC. First, whole blood was collected from healthy adult volunteers by venipuncture. Ethical clearance was obtained from the Committee on Human Rights to Research Involving Human Subjects of Ramathibodi Hospital, Mahidol University (Protocol Number-ID 05-48-29). Then, PBMC were isolated from whole blood by Ficoll hypaque gradient centrifugation (Lymphopre, Nycomed Pharma, Oslo, Norway). To obtain primary adherent monocytes, PBMC suspended in RPMI-1640 containing 10% fetal bovine serum (complete RPMI-1640, Gibco BRL, Grand Island, NY, USA) were allowed to adhere to a plastic tissue culture plate for 60 min at 37 °C as described by Wang and colleagues (17). After removing non-adherent cells, the adherent cells were subsequently stained with FITC-labeled anti-CD14 MAb, and CD14+ cells were sorted through FACSVantage (Becton Dickenson, Palo Alto, CA, USA). The purity of CD14+ cells isolated by the cell sorter was 98–99%. Approximately 1.0 × 106 macrophages were used per assay, and the experiment was performed in triplicate.

Competitive binding inhibition assay

HM were pretreated for 30 min at 37 °C with various concentrations of MAbs against integrins (1, 5, 10 μg/ml), TLR (1, 5, 10 μg/ml), CD14 (1, 5, 10 μg/ml), MR (10, 50, 100 ng/ml), or IgG isotype control, prior to incubation with FITC-labeled P. marneffei conidia, at a MOI of 100. In order to analyze the combined inhibitory effect of MAbs against TLR2 and TLR6 on conidial adhesion, pretreatment of HM with the two MAbs at equal concentrations (1, 5, 10 μg/ml) was performed. In order to identify fungal cell wall components that are involved in adhesion, either mannan or laminarin (Sigma-Aldrich) at 10, 50, 100 μg/ml, was individually added to the human monocyte culture. FITC-labeled conidia were subsequently allowed to bind to HM for 30 min at 37 °C with gentle rocking every 15 min. Each assay was carried out in a 40 mm-Petri dish (Nalge Nunc International, Rochester, NY, USA) at the final volume of 2 ml. The dishes were then gently washed with 10 ml RPMI-1640 and fixed with 2% glutaraldehyde (Merck, Darmstadt, Germany). The cells were stained and mounted with Giemsa and liquid parafilm, respectively. Each adherence assay was independently performed three times. The number of P. marneffei conidia per 100 HM was determined and expressed as a percentage of inhibition, as follows:


Where “No. bound conidia per 100 host cells in the absence of inhibitor” represents the number of conidia bound in the presence of irrelevant IgG Isotype control.

Activation of HM by P. marneffei conidia

Adherent HM were prepared in tissue culture dishes, as described above. P. marneffei conidia suspended in RPMI-1640 were added to HM culture at MOI of 1, 10, and 100. After co-cultivation at 37 °C for 8 h, HM were harvested by mild flooding with warm culture medium. The harvested cells were stained with FITC-labeled anti-CD14 MAb, PE-labeled anti-CD86 MAb, and APC-labeled anti-CD40 MAb at 4 °C for 30 min. CD14 positive cells were gated, and the degree of CD40 and CD86 expression was determined by FACSCalibur (Becton Dickinson). For each staining, three independent experiments were performed in duplicate and the analysis of 5000 events for each individual experiment was done with the Cell-QUEST program (Becton Dickinson).

Cytokine quantification

CD14+ cells were exposed to P. marneffei conidia at MOI of 100, 37 °C for 8 h. In order to examine their involvement in cytokine production, blocking MAbs against integrins, MR, CD14 or TLR were individually applied to HM cultures for 30 min at 37 °C before the addition of P. marneffei conidia. An irrelevant mouse IgG MAb was utilized as an isotype-specific control. Each assay was performed in a 96-well tissue culture plate (Nalge Nunc) at the final volume of 200 μl. Culture supernatant from each well was harvested, and the concentrations of secreted TNF-α or IL-1β in the supernatant were measured with human TNF-α or IL-1β ELISA kit (Pierce Endogen, Rockford, IL, USA), according to the manufacturer's protocol. To determine the importance of serum components on cytokine production, similar experiments were conducted in media with 0–10% serum supplementation. Autologous sera, unheated or heated at 56 °C for 30 min, were also employed to examine the requirements for heat labile factors for cytokine production from HM.

Statistical analysis

All results are expressed as mean ± standard deviation. Statistical significance of results was determined by a Student's t-test or ANOVA followed by Turkey's post-hoc test, where appropriate. Results were determined to be statistically significant when P values < 0.05 were obtained.


Role of pattern recognition receptors in the recognition of P. marneffei conidia by HM

Recognition of P. marneffei conidia by HM through PRR was examined using MAbs against PRR to inhibit in vitro P. marneffei conidial adhesion to HM. Anti-TLR1, -TLR2, -TLR4 or -TLR6 MAbs demonstrated significant inhibitory effects, in a dose-dependent manner, on the adhesion of P. marneffei conidia to HM (P < 0.05) (Fig. 1a). Among the four antibodies that showed comparable inhibitory effects, anti-TLR2 MAb resulted in the greatest inhibition when used at low concentrations. Although combined treatment with anti-TLR2 and -TLR6 antibodies demonstrated significant inhibitory effects in a dose-dependent manner, an insignificant additional inhibitory effect on conidial adhesion to HM was seen as compared to either single anti-TLR2 or -TLR6 MAb treatment. Additionally, we examined whether CD14 participates in conidial recognition. As shown in Figure 1b, pre-incubation of HM with anti-CD14 MAb at 5 μg/ml diminished conidial adherence to HM by approximately 50%.

Figure 1.

Inhibition of adherence of P. marneffei conidia to human monocytes by various inhibitors. HM were treated with various concentrations of either (a) anti-TLR MAbs, (b) anti-CD14 MAb or anti-MR MAb, (c) mannan or laminarin, or (d) anti-integrin MAbs prior to conidial adhesion. Mean number of conidia bound to HM ± standard deviation in the absence of antibody is 136 ± 34.86 while that in the presence of control IgG is 165 ± 6.48.

To evaluate the ability of the MR to mediate adherence to its ligand on the conidial wall, anti-MR MAb, mannan and laminarin were employed as inhibitors for the binding assay. Interestingly, anti-MR MAb utilized at a concentration 100 times lower than that of anti-CD14 MAb was able to show comparable inhibitory effects. The characteristics of inhibition by these two MAbs were dose-dependent (P < 0.05) (Fig. 1b). Preincubation of HM with 100 μg/ml of a macrophage MR ligand, mannan, resulted in approximately 40% inhibition to conidial binding (Fig. 1c). On the other hand laminarin, which was utilized as an inhibitor for β-glucan receptor, had no direct effect on adhesion.

Numerous studies have demonstrated that integrins are leukocyte receptors for a number of fungi by recognizing mannose and β-glucan carbohydrate structures (9, 31, 32). We therefore assessed the ability of CD11a, CD11b, and CD18, to bind P. marneffei conidia. Significant inhibition of conidial adherence to HM was observed in the assays containing anti-CD11b or anti-CD18 MAb (Fig. 1d). Anti-CD18 MAb at 1 μg/ml conferred nearly 60% inhibition of adhesion. In contrast, MAb against CD11a did not demonstrate significant inhibitory effects.

Monocyte activation and induction of cytokine production by P. marneffei conidia

The above-mentioned experiments demonstrate dependence on several PRR for the initial adhesion of P. marneffei conidia to monocytes. Our next set of experiments assessed whether P. marneffei conidia are able to activate HM and induce its cytokine production. HM were exposed to P. marneffei conidia, and the expression of CD86 and CD40 molecules on the surface of HM was evaluated. When an increasing number of P. marneffei conidia were applied to the HM culture, flow cytometric analysis showed an increase in fluorescent intensity, representing the degree of CD86 and CD40 surface expression. This finding demonstrates that P. marneffei conidia can activate HM and induce monocyte response by increasing the expression of co-stimulatory molecules within 8 h of incubation (Fig. 2a). Notably, treatment of HM with MAbs against CD14, CD18, TLR1, TLR2, or TLR4 prior to HM activation with conidia did not affect the fluorescent intensity of the stained cells (data not shown). In addition, P. marneffei conidia exhibited marked activity of induction of TNF-α and IL-1β production from HM. As shown in Figure 2b, the concentrations of both proinflammatory cytokines released from HM were significantly increased when HM was allowed to interact with P. marneffei conidia at an MOI of 100. On the contrary, when IL-10 concentration was quantified from the conidia-treated culture, there was a 10 fold reduction in secretion as compared to that of the unstimulated control. These results indicate that P. marneffei conidia are not only able to activate HM, but also induce proinflammatory cytokine production thereafter.

Figure 2.

Ability of P. marneffei conidia to activate human monocytes and to induce proinflammatory cytokines production. (a) FACS analysis data show expression of CD86 (left panel) and CD40 (right panel) on the surface of HM exposed to P. marneffei conidia at a MOI of 100. Black line represents the degree of expression of CD86 or CD40 on normal HM. Blue line represents CD86 or CD40 expression on HM activated with intact conidia. (b) Cytokine production by HM after exposure to P. marneffei conidia at the MOI of 100 was measured by ELISA. Data are expressed as mean values ± standard deviation of three independent experiments (*P values < 0.05).

Cytokine production and activation by P. marneffei conidia through PRR

As incubation of HM with conidia of P. marneffei resulted in the stimulation of proinflammatory cytokine production, we further investigated which monocyte surface receptor is involved in signal delivery for such interactions. There was a significant decrease in the concentration of TNF-α in the assay containing monoclonal anti-CD14 and TLR4 antibodies (Fig. 3). The inhibitory effect on TNF-α production by both anti-TLR4 and anti-CD14 antibodies was dose-dependent (Student's t-test; P values < 0.05). Maximal inhibition of 48% was obtained when 10 μg/ml of anti-CD14 MAb were applied. In contrast, pretreatment with MAbs against CD11b, CD18 or MR did not inhibit secretion of TNF-α induced by P. marneffei conidia (Fig. 3). Monoclonal anti-TLR-1, anti-TLR-2 and anti-TLR-6 antibodies also demonstrated no distinct inhibitory effects when compared to the isotype control (data not shown).

Figure 3.

Blockade of proinflammatory cytokine production by anti-CD14 and anti-TLR4 Abs. HM were incubated with various concentrations of MAbs against (•) CD11b, (○) CD18, (▾)MR, (▿) CD14 or (▪) TLR4 prior to exposure to P. marneffei conidia. TNF-α secretion by HM infected with P. marneffei conidia was measured. Data are expressed as mean percentage of induction of TNF-α production ± standard deviation of three independent experiments (*P value < 0.05). Y axis represents percent reduction in the amount of TNF-α released by HM infected with conidia at the MOI of 100 when the culture was treated with MAbs against CD14, integrins, MR, or TLR4. The reported data was normalized with the level of TNF-α released by infected HM pretreated with isotype control antibody.

Taken together, these results demonstrate that, although many PRR serve as binding receptors for P. marneffei conidia, only TLR4 and CD14 are responsible for the induction of TNF-α secretion by human monocytes.

Dependency on serum for the stimulation of HM by P. marneffei conidia

Since we found that CD14 participates in the activation of cytokine production, our next experiment was aimed at determining whether serum factors are involved in proinflammatory cytokine production. In normal physiology, several serum components participate in innate immune responses against pathogen infection. Figure 4a,b demonstrates the requirement for serum in TNF-α and IL-1β production from primary adherent HM. Stimulation of HM by P. marneffei conidia in the presence of various concentrations of autologous serum (2, 4, 8, and 10%) resulted in the induction of TNF-α and IL-1β production in a dose-dependent manner (P < 0.05). Heat treatment of serum at 56 °C for 30 min to inactivate complements and other heat-labile factors significantly reduced the production of both cytokines. No significant change was detected in IL-10 concentrations in the presence of heat-treated serum (data not shown). This finding suggests that serum is important in TNF-α and IL-1β production from HM, and that a heat-labile factor in the serum is also required for proinflammatory cytokine production, particularly in regard to IL-1β.

Figure 4.

Effect of serum factors on proinflammatory cytokine production by human monocytes infected with P. marneffei conidia. HM and P. marneffei conidia were co-cultured in the presence of various concentrations of autologous serum or heat inactivated autologous serum. The amount of (a) TNF-α and (b) IL-1β secretion in culture supernatant was measured by ELISA. Data are expressed as mean values ± standard deviation of three independent experiments (*P value < 0.05 when compared to serum free condition, #P value < 0.05 when compared to 10% HI serum condition).


Initial adherence of P. marneffei conidia to human monocytes

Resident tissue macrophages in alveolar sacs serve as the first line of immune defense by detecting the immediate presence of a pathogen and sending an alarm signal, such as cytokine secretion, to other cells. Alveolar macrophages are abundant, making them the most likely to be the first to encounter air-borne pathogens including P. marneffei conidia. Nevertheless, no study has reported the exact route of infection of P. marneffei, except for showing that the fungus can infect tissue macrophages and monocytes. Alveolar macrophages are a type of tissue macrophage derived from peripheral blood monocytes, and the differences between alveolar macrophages and peripheral blood monocytes in the degree of surface marker expression, such as phagocytic markers (CD36, CD44) and co-stimulatory molecules (CD40, CD80, and CD86), have been stated elsewhere (33). The distinction between the functional characteristics of alveolar macrophages and monocytes is still unclear. Taken together, these considerations made us decide to use adherent HM as the study model.

Adherence capability of P. marneffei conidia to HM was found in the present study, which is the first to demonstrate that several PRR, including TLR1, TLR2, TLR4, and TLR6, are involved in the interaction, and contribute to subsequent innate immune cell activation. CD14 was also shown to be essential for the initial interaction of intact P. marneffei conidia with HM. Our results are consistent with earlier findings that TLR2, TLR4, and CD14 play important roles in host response against fungal conidia (23). In that study, macrophages obtained from TLR2-, CD14-knock out mice or TLR4 defective mice showed impairment of TNF-α production on activation with fungal conidia. Involvement of CD14 was not restricted to mouse innate immune response, but was also required for TNF-α production in human antigen presenting cells (17, 23).

The MR has previously been reported to serve as a recognition site for pathogenic fungi, including C. albicans, Pneumocystis carinii (now named P. jiroveci), and P. marneffei yeast (34, 35). In the present study, a very low concentration of MAb against the MR was able to inhibit adhesion and subsequent phagocytosis of P. marneffei conidia by HM, supporting the hypothesis that MR is a common phagocytic receptor for a wide variety of fungal pathogens. However, the binding mediated by MR could depend on the morphotype of fungal pathogen, as reported by Rongrungruang and Levitz, in that unopsonized yeast of P. marneffei do not bind to human HM via MR (11). Recently, dectin-1, a β-glucan receptor expressed on the surface of monocyte/macrophage lineage cells, has been demonstrated to induce biological activation upon β-glucan engagement (13, 36). Hohl and colleagues have shown that β-glucan restrictively expresses on swelling or metabolically active conidia (14). They have further demonstrated that the production of TNF-α and MIP-2 by murine macrophages is initiated by metabolically active conidia, indicating that the β-glucan receptor plays an important role in the innate immune response against restrictive morphotypes of fungal conidia. Our results support the earlier findings of Hohl and colleagues (14), as no direct inhibitory effect from laminarin on the interaction between resting P. marneffei conidia and HM was seen. Furthermore, mannan significantly inhibited conidial adhesion to the target cells, suggesting that mannose-containing molecule(s) may be responsible for macrophage stimulation. Taken together, P. marneffei conidial adherence to HM is mediated by the MR and the β-glucan receptor plays a lesser role in such interactions.

In addition to TLR and MR, adhesion of pathogenic fungi to human leukocytes via integrins has been reported. LFA-1 (CD11a/CD18), CR3 (CD11b/CD18) and p150,95 (CD11c/CD18) have been shown to mediate binding of H. capsulatum yeasts and microconidia to human macrophages (37, 38). In addition, CR3 has also been shown to mediate macrophage binding to other fungi such as Blastomyces dermatitidis and C. albicans (39). In this study, blockade of P. marneffei conidia to HM with MAbs against integrins suggests that attachment of intact conidia of P. marneffei to HM is mediated through CR3 or Mac-1. Further investigation also indicates that CR3 is required for phagocytosis of intact P. marneffei conidia by macrophages. This result strongly supports the possibility that CD18-associated integrins, especially CD11b/CD18, serve as common receptors for pathogenic fungi on human leukocytes (40, 41).

Host response on initial adherence of P. marneffei conidia to human monocytes

In the present study, the ability of P. marneffei conidia to activate HM was revealed by FACS analysis, which showed increasing CD86 and CD40 expression on the cell surface and a significant increase in TNF-α and IL-1β secretion. Increase in the concentrations of TNF-α, IL-1β, and IL-6 production by monocytes infected with conidia or hyphae of A. fumigatus has been reported elsewhere (17, 42). Furthermore, Sisto and colleagues have demonstrated that the spleen and liver of BALB/c mice initially infected with P. marneffei had local production of high concentrations of TNF-α, IL-12 and IFN-γ, which may be responsible for fungal clearance (43). These findings, as well as ours, strongly suggest that this group of proinflammatory cytokines are commonly secreted in the initial phase of fungal infection. With regards to monocyte production of IL-10, exposure to P. marneffei conidia seems to significantly suppress release of IL-10 from human HM. This result is corroborated by a reduction in IL-10 production, associated with enhanced host resistance against a number of infectious fungi, including C. albicans, A. fumigatus, and C. neoformans (44). Previous work on P. marneffei infection in mice has also demonstrated no difference in the concentrations of IL-10 secreted from the suspension of whole spleen as compared to uninfected controls (43).

In the present study MAbs specific to CD14 or TLR4 diminished TNF-α production by conidial-stimulated HM. Our results are consistent with involvement of CD14 in proinflammatory cytokine production by human monocytes in A. fumigatus infection (17, 23). Previously, Meier et al. also demonstrated the importance of TLR4 on macrophages activated by A. fumigatus conidia in TNF-α production (45). These results suggest that at least TLR4 and CD14 are involved in induction of proinflammatory cytokine production. In addition, TNF-α seems to be a major cytokine responsible for control of P. marneffei conidia infection, and its involvement against other pathologic fungal infections has been reported. Fungal pneumonia occurring after TNF-ablation therapy is evidence which supports the possible pivotal role of TNF-α in control of fungal infection (46).

In the present study, the results of MAb blockade on TNF-α production by HM revealed that anti-TLR4 and CD14 MAbs exerted partial inhibitory effects of 20–50%. We also performed a parallel experiment on blockade of IL-1β production. Here inhibitory effects of MAbs were not as conclusive as in the case of TNF-α production, due to a low and varied degree of IL-1β production from HM of certain donors, leading to lower discriminatory power for that data set (data not shown). Notably, we were surprised to see that treatment of P. marneffei-activated HM with MAbs against CD14, TLR4 or CD18 did not cause a decrease in surface expression of co-stimulatory molecules although there was a reduction in TNF-α production upon treatment with anti-TLR4 and -CD14 MAbs. This might be attributable to the fact that these antibodies do not specifically block at the activation motifs which lead to co-stimulatory molecule expression. In addition, blockade of CD18 (integrin β-subunit) alone might not be sufficient to reduce the Ca2+ transducing signal that is critically regulated by the association of CD18 with the unique α-subunit, CD11b.

With regards to TLR1, TLR2 and TLR6, although the MAbs against these TLR were able to inhibit P. marneffei conidia adherence to HM, they failed to block both TNF-α production and the surface expression of co-stimulatory molecules. It is possible that while these MAbs can inhibit the binding sites, they are not capable of activating blockade. One possible explanation is that the blockade of conidial attachment to TLR1 and TLR2 by the MAbs used in this study might occur at motifs other than the ones directly involved in signal transduction leading to TNF-α production. Secondly, Mab blockade of adhesion to TLR 1 and 2 might be through steric hindrance and not directly related to the activation motif. Additionally, we may have to take into account the distribution and degree of expression of each individual TLR on the surface of HM. It should also be noted, as previously reported by Netea and colleagues, that not every TLR that participates in binding of fungal cells involves induction of cytokine production (47). TLR6, but not TLR1, plays a role in recognition and acts as a modulator for the adaptive immune response in C. albicans infection (47). The inductive role of TLR1, TLR2 and TLR6 in proinflammatory cytokine production should be further elucidated by employing a TLR-transfected cell line as a representative model.

The present study also addressed the importance of serum factors in HM stimulation during P. marneffei infection. Previously, Rongrungruang and Levitz demonstrated that, despite an opsonic requirement for optimal adhesion of P. marneffei yeasts to human macrophages, yeast binding to human macrophages is still present when serum is omitted from the adhesion assay (11). They further reported that stimulation of human macrophages, resulting in phagocytosis and respiratory burst, can occur in the absence of opsonins. Similar results were also observed in the case of P. marneffei conidia in our laboratory (data not shown). Taken together, these data confirm that certain cell wall components of both yeast and conidia, independent of serum factors, interact directly with receptors on the surface of macrophages, leading to cell activation and killing of pathogenic fungi.

With respect to proinflammatory cytokine secretion by HM, it was demonstrated in the present study that serum factors are required for the expression of such cytokines. Heat treatment strongly attenuated cellular activation of serum to induce TNF-α and IL-1β production. Thus serum factors, including heat-labile factors such as complements, are required for proinflammatory cytokine production by HM on infection with P. marneffei conidia. However, inactivation of complements by heat treatment does not result in complete loss of leukocyte activation, implying that other, heat-stable, factors available in serum may be involved in cellular activation by P. marneffei conidia. It is also important to note that HM exhibited similar requirements in response to P. marneffei yeasts (11).

During colonization P. marneffei conidia generally interact with macrophages residing in the alveoli through adhesion to the MR, which may induce internalization of P. marneffei conidia by the macrophages. This interaction is mediated by mannan, one of the major components of the fungal cell wall, and its receptor. It is still not known whether β-glucan exists on the surface of conidia, but the β-glucan receptor on the surface of macrophages seems to play a very minor role in the early phase of the interaction. Similar to MR, CD11b and CD18 participate in conidial attachment, although to a lesser extent. Interestingly, the present study reveals for the first time that TNF-α can be induced by a signal derived from the interaction of P. marneffei conidia with CD14 or TLR4. Assessment of the relative contribution of TLR in proinflammatory cytokine production by host cells in response to P. marneffei is currently ongoing, with the hope of providing insight into the molecular basis of signal transduction.


We appreciate valuable suggestions on statistical analysis made by Montip Tiensuwan (Department of Mathematics, Faculty of Science, Mahidol University). We are grateful to Praveen Chawla for her critical reading and grammatical correction of our manuscript. This work was supported by Thailand Tropical Diseases Research Programme (T-2), National Science and Technology Development Agency (NSDTA) and the Chulabhorn Research Institute. Yuttana Srinoulprasert is supported by the Medical Scholars Program of Mahidol University and the Royal Golden Jubilee Ph.D Scholarship of the Thailand Research Fund. Piyapong Pongtanalert is supported by Mahidol University and BIOTEC, NSTDA, Thailand.