866

Binding of Unopsonized Cryptococcus neoformans by Human Bronchoalveolar Macrophages: Inhibition by a Large-Molecular-Size Serum Component Stuart M. Levitz, Abdulmoneim Tabuni, Randall Wagner, Hardy Kornfeld, and Edwin H. Smail

Section of Infectious Diseases and Pulmonary Center. Evans Memorial Department ofClinical Research and Department ofMedicine. University Hospital. Boston University Medical Center. Boston. Massachusetts

Infections due to the yeast-like fungus Cryptococcus neoformans occur with greatly increased frequency in patients with impaired cell-mediated immunity, especially those with AIDS [1]. With rare exceptions, clinical isolates of C. neoformans possess a polysaccharide capsule, which is thought to be a major virulence factor [2, 3]. Compared with encapsulated isolates, acapsular isolates of C. neoformans are relatively avirulent in mouse models of cryptococcosis [4]. The liability of C. neoformans to inhibition and killing by most (but not all) effector cell populations is diminished by the presence of capsule on the surface of the organism [5-9]. Binding by phagocytes of serum-opsonized encapsulated C. neoformans occurs with decreased avidity compared with serum-opsonized acapsular mutants [5, 7, 10, 11]. Moreover, binding by cultured macrophages of serum-opsonized acapsular, but not encapsulated, C. neoformans leads to a phagocytic event with internalization of the organism [10]. The mechanisms by which capsule exerts its antibinding and antiphagocytic effects are incompletely elucidated. The problem does not appear to be a failure of opsonization, as incubation in serum leads to the deposition oflarge amounts of complement components, mostly in the form of iC3b, on

Received 27 January 1992; revised 24 April 1992. Informed consent was obtained from blood donors. Human experimental guidelines of the US Department of Health and Human Services and of Boston University Medical Center were followed in the conduct of the research. Financial support: National Institutes of Health (AI-25780. AI-28408. HL-44846); American Lung Association (career investigator award to H.K.). Reprints or correspondence: Dr. Stuart M. Levitz. Section of Infectious Diseases. University Hospital, 88 E. Newton St., Boston. MA 02118. The Journal oflnfectious Diseases 1992;166:866-73 © 1992 by The University of Chicago. All rights reserved. 0022- [899/92/6604-0026$01.00

the surface of encapsulated C. neoformans [12, 13]. Moreover, monoclonal antibodies directed against complement receptors profoundly inhibit binding of serum-opsonized C. neoformans to human cultured monocytes [10]. While capsular polysaccharide imparts a negative surface charge and hydrophilic surface upon C. neoformans, these physicochemical properties can be dissociated from liability to phagocytosis by mouse peritoneal macrophages [14, 15]. Capsule may inhibit binding by masking ligands, such as mannans and {j-glucans, on the cell wall of the fungus that would otherwise be available to interact with their corresponding phagocytic receptors. Exposure to C. neoformans in humans is thought to occur as a consequence of inhalation ofairborne yeast cells, whose size is ideal for alveolar deposition [2]. Thus, bronchoalveolar macrophages, the major resident phagocyte lining the bronchoalveolar tree, presumably form an important first line host defense mechanism against cryptococcosis. Indeed, antifungal activity ofbronchoalveolar macro phages has been demonstrated, even in the absence of opsonins [7, 16, 17]. However, the conditions under which human bronchoalveolar macrophages recognize inhaled C. neoformans have not been well characterized. Levels ofmany complement components are low or undetectable in the alveolar space [18], and specific anticryptococcal immunoglobulins are not likely to be present during initial exposure to the fungus. Although a functionally active alternative complement pathway sometimes may be present in the alveolar space [19], components ofthe pathway could be depleted early in pulmonary cryptococcal infections as a result of the potent ability of C. neoformans to activate and deplete complement [20]. In the present study, binding of C. neoformans by human bronchoalveolar macrophages in the presence and absence of opsonins was characterized.

Downloaded from http://jid.oxfordjournals.org/ at :: on January 16, 2015

Infection with Cryptococcus neoformans usually begins after inhalation of airborne organisms. Since levels of opsonins in the alveolar space may be low, the ability of human bronchoalveolar macrophages to bind C. neoformans in the presence and absence of opsonins was studied. Bronchoalveolar macrophages bound unopsonized C. neoformans. Surprisingly, component(s) in pooled human serum (PHS) inhibited binding, as evidenced by 26% and 71% inhibition of binding when 20% PHS and heat-inactivated PHS (HI-PHS), respectively, were added to the system. Separation of PHS by molecular size revealed that the inhibitory component had an apparent molecular weight> 106 and was inhibitory at nanomolar concentrations. PHS stimulated and HI-PHS had no effect on binding of acapsular C. neoformans and zymosan particles to bronchoalveolar macrophages. These data demonstrate that bronchoalveolar macrophages can bind unopsonized, encapsulated C. neoformans, but that serum component(s) inhibits binding.

1ID 1992; 166 (October)

Binding of C. neoformans by Macrophages

Materials and Methods

sterile gauze, the cells were centrifuged and rid of contaminating erythrocytes by hypotonic lysis with distilled water. Cells were washed in cold PBS, counted, and resuspended in medium at 5 X 105/ml., Differential cell counts on cytocentrifuged preparations of lavaged cells stained with a modified Wright's stain (Leukostat; Fischer Diagnostics, Orangeburg, NY) demonstrated an average of 91%bronchoalveolar macrophages (range, 83%-99%) and always 97% polymorphonuclear leukocytes after lysis of red blood cells with hypotonic saline. The low density or peripheral blood mononuclear cell fraction contained >90% monocytes after nonadherent cells were washed away, as described below. Cultured monocytes were generated as in prior studies [10] by incubation of monocytes in medium containing 10% human male AB serum for 6 days. C. neoformans. Encapsulated, serotype D strain B350 I, and acapsular isogenic mutant strain CAP67 (ATCC 52817) (obtained from Eric Jacobson, Medical College of Virginia, Richmond) were maintained and harvested as described [5, 7). The encapsulated strain had an average capsular thickness of 1.5 JLm [6). To facilitate rapid, accurate identification of yeasts, most experiments used fluorescein isothiocyanate-Iabeled, heatkilled organisms. Selected experiments yielded similar results with live, unlabeled organisms. Yeasts were heat-killed by immersion in a 60°C water bath for 30 min and then labeled with FITC as described [5, 7). C. neoformans (I X 108) was preopsonized by incubation in 1 mL PHS for 30 min at 37°C, followed by three washes in PBS. Binding assays. In 5 JLL of medium, 2.5 X 103 freshly lavaged cells were added per well of 60-well sterile Terasaki tissue typing trays (Robbins Scientific, Mountain View, CA). Wells were incubated 2 h (except for wells with cultured monocytes, which were cultured 6 days, as described above). Cells were then treated with 2 JLL of inhibitors dissolved in PBS. Control cultures received just PBS. Finally, 3 JLL of C. neoformans or zymosan particles (1.5 X lOS/well in medium) was added. This number of particulate stimuli was just sufficient to completely cover the bottom surface of the well. Phagocytes and particulate stimuli were incubated for 30 min (except where noted), after which unbound organisms and nonadherent leukocytes were washed off by successively dipping the tray three times in each of five flasks containing PBS supplemented with 1.3 mM calcium and 0.8 mM magnesium. This washing procedure effectively removed virtually all unbound organisms. The cells were fixed with 1% formaldehyde in PBS. Using an inverted microscope at 200X, 100-250 phagocytes were randomly selected under bright field and then examined under epifluorescence to determine the number of yeasts associated per phagocyte and the percentage of phagocytes with one or more yeasts. Binding index represents the number of cell-associated yeasts per 100 phagocytes. Due to variability in baseline binding indices between donor phagocytes, some results are expressed as percentage inhibition of binding, which was calcu-

Downloaded from http://jid.oxfordjournals.org/ at :: on January 16, 2015

Materials. All reagents used, except where noted otherwise, were obtained in the highest quality available from Sigma (St. Louis). Medium used in all experiments, unless stated otherwise, was RPMI 1640 (GIBCO, Grand Island, NY) supplemented with L-glutamine, penicillin, and streptomycin. All incubations, except when noted, were in a 37°C, 5% CO 2 , 95%air, humidified environment. Pooled human serum (PHS) was obtained by pooling serum from ~ I0 healthy volunteers under conditions preserving full complement activity. Five different batches of PHS were used. HI-PHS was prepared by incubation of PHS at 56°C for 30 min. For selected experiments, PHS was depleted of >97% of its IgG by incubation with protein A Sepharose 4B (Pharmacia, Piscataway, NJ) exactly as in previous studies [10]. Human serum depleted of the third component of complement (C3) was purchased. Human Cohn fraction II/III contained 3% albumin, 3% alpha 2 globulins, 8% beta globulins, and 86%gamma globulins, while Cohn fraction IV contained 20%albumin, 50%alpha globulins, and 30%beta globulins, as determined by electrophoresis done by the manufacturer (Sigma). Purified human immunoglobulin containing >90% IgG (Gamastan; Cutter Biological, Berkeley, CA) was purchased as a 15%-18% protein solution stabilized with glycine and was diluted in PBS before use. Fractionation ofserum. PHS (2.5 mL) was dialyzed against PBS and then separated on the basis of molecular size using a calibrated 90- X 1.5-cm column packed with 200 mL ofSephacryl S-300 (Pharmacia) in PBS. Sephacryl S-300 has a molecular weight separation range for globular proteins from I X 104 to 1.5 X 106 . Individual 3-mL fractions were concentrated using a microconcentrator (Centricon 30; Amicon, Beverly, MA). The protein concentration of the fractions was determined by shifts in the absorbance of Coomassie brilliant blue using a commercial kit (protein assay; Bio-Rad, Richmond, CA). The average apparent molecular weight of the fractions was calculated using a standard curve generated by plotting the elution volume of globular proteins of known molecular weight. HI-PHS (2 mL) was dialyzed against 10 mMTRIS buffer, pH 8.6, and separated on the basis of charge using a 3- X 2-cm column packed with 7 mL of DEAE Sephacel (Pharmacia). Fractions (2.5 mL) were eluted with increasing concentrations of NaCI using a gradient former (Bethesda Research Laboratories, Gaithersburg, MD) containing 95 mL ofTRIS buffer in one compartment and 95 mL of 1 M NaCl in TRIS buffer in the other compartment. Fractions were dialyzed against PBS, concentrated using a microconcentrator, and adjusted to 15 mg protein/ml, by the addition of PBS. To screen for activity, groups of 10 fractions were pooled. Active fractions were then pooled in groups of three or four and retested. Protein electrophoresis and densitometry readings on the active fractions were done by the Clinical Chemistry Laboratory, University Hospital, using an automated system (Electrophoresis Systems with model 710 densitometer; Ciba Corning Diagnostics, Palo Alto, CA). Bronchoalveolar macrophages. Healthy volunteers underwent flexible fiberoptic bronchoscopy and bronchoalveolar lavage with instillation of 240 mL of normal saline, as in previous studies (unpublished data). About 50%-80%ofthe instilled fluid was recovered using gentle suction. After being strained through

867

Levitz et al.

868

80

250

• X

w

Q

70 200

60 50

150

~

C!J Z

40 100

30

Q

z

iii

50 0

asI:L

0 0

Z

en :z:

I:L I

~

en :z:

I:L

as

as!

f

A:% I:L +

g

-

III C;

z

::)

IL

.-:I: ~ ~

4(

20

In

10

0

0

~

IL

~I:L

w.!.

lated as: [I - (binding index of experimental group/binding index of control group)] X 100 [10]. Most assays were scored while the reader was blinded with respect to the experimental conditions used in each well. Assay to distinguished attached from fully internalized organisms. The above binding assay does not distinguish between organisms bound to the surface of phagocytes versus those that have been fully phagocytosed (internalized). To make this distinction, the Uvitex assay was used exactly as described [10], except the assay was done in Terasaki tissue typing trays. Bound-only organisms were also distinguished from those internalized by examination under phase-contrast microscopy [l0, 21]. Statistics and presentation ofdata. Means and SE were compared using the two-tailed, two-sample (unpaired) t test, except where noted. Unless stated otherwise, all data represent mean ± SE of at least three separate experiments, each done in triplicate or quadruplicate.

Results Initial experiments examined the opsonic requirements for binding of encapsulated C. neoformans to human bronchoalveolar macrophages (figure I). Despite the absence of opsonins, binding of the yeast to bronchoalveolar macrophages was seen. Unexpectedly, the addition of 20% PHS containing an intact complement system not only failed to improve the binding index, but actually inhibited binding by

a small, but significant, degree. However, if the C. neoformans was preopsonized in PHS under conditions allowing for complement activation and deposition and then washed free of serum, the binding index increased. This suggested that component(s) in PHS inhibited binding. This hypothesis was substantiated by experiments demonstrating the capacity of 20% HI-PHS to inhibit cryptococcal binding to bronchoalveolar macrophages by > 70% (figures 1,2). Inhibition of binding (79.2%± 2.6%) was also seen using 20%heatinactivated AB serum. The inhibitory effect of HI-PHS diminished if the concentration was decreased (32% ± 8.3% and 21.1 %± 7.4% inhibition of binding with HI-PHS concentrations of 10%and 2%, respectively). HI-PHS also inhibited the binding of preopsonized C. neoformans to bronchoalveolar macrophages. Increasing the incubation period of bronchoalveolar macrophages and unopsonized C. neoformans from 30 min to 2 h increased the binding index from 71.7% ± 8.9% to 190.0% ± 17.8% (P = .007). However, inhibition ofcryptococcal binding (67.4% ± 2.5%) was still seen when HI-PHS was added to the 2-h incubation mixture. HI-PHS inhibited binding ofa serotype A (ATCC 62070) and a serotype B (ATCC 34877) strain ofC. neoformansby 65.5% ± 14.3% and 67.5% ± 5.1%, respectively. As judged by phase-contrast microscopy (figure 2) and Uvitex staining, 60% at the higher concentration used (figure 3). Not shown on figure 3, purified human IgG (6 mg/mL) and IgM (200 JLg/mL), at concentrations greater than found in 20% PHS, did not significantly inhibit (-2.3% ± 11.6% and -18.3% ± 18.8%, inhibition of binding, respec-

tively). HI-PHS depleted of IgG inhibited binding by 64.4% ± 9.8%. Results similar to those obtained with the Cohn fractions were seen when PHS was separated on the basis of charge on a DEAE Sephacel column (figure 4). However, while activity was seen over a broad range of column fractions, pooled fractions 20-22 were particularly active, with >80% inhibition of binding using 3 rng/ml. protein. When the protein concentration of this fraction was decreased to 600 and 200 JLg/mL, 50.7% ± 3.6% and 30.0% ± 7.2% inhibition of binding, respectively, were observed. Protein electrophoresis of fractions 20-22 demonstrated 30% gamma globulins, 31%

Levitz et al.

870 80