candida albicans can stimulate stromal cells resulting in enhanced granulopoiesis
TRANSCRIPT
STEM CELLS AND DEVELOPMENT 13:39–50 (2004)© Mary Ann Liebert, Inc.
Candida albicans Can Stimulate Stromal Cells Resulting in Enhanced Granulopoiesis
SUNANDA BASU,1 HUI-HUA ZHANG, CATHY QUILICI, and ASHLEY R. DUNN
ABSTRACT
Previously, we have reported that although unperturbed granulocyte colony-stimulating factor (G-CSF)-deficient (G-CSF2/2) mice are neutropenic, when challenged with Candida albicans, they de-velop a profound neutrophilia. In an attempt to understand the basis of Candida-induced neu-trophilia in G-CSF-deficient mice, we have modified the Dexter bone marrow culture system toproduce an in vitro model that mimics emergency granulopoiesis in vivo. In this model, stromal cul-tures are overlaid with bone marrow cells in the presence or absence of heat-inactivated (HI) Can-dida. Irrespective of the genotype of mice used as a source of bone marrow-derived stromal cells,stimulation of these cultures with HI Candida led to a significantly greater recovery of cells com-pared to unstimulated stromal cultures. In addition, there was a marked increase in the number ofcolony-forming units granulocyte-macrophage (CFU-GM), as well as in the percentage of granulo-cytes in the population of nonadherent cells recovered from HI Candida-stimulated cultures. Theconditioned medium generated from stromal cultures derived from either wild-type or G-CSF2/2
mice exposed to HI Candida, when applied to bone marrow cells in a soft agar clonogenic assaystimulated M-, GM-, and G- type colonies. Interleukin-3 (IL-3) and GM-CSF could not be detectedin the conditioned medium from either HI Candida stimulated or unstimulated stromal cultures.However, IL-6 was detected in the conditioned media from both wild-type and G-CSF2/2 stromalcultures. Addition of anti-IL-6 antibody significantly impaired granulopoiesis in unstimulated andHI Candida-stimulated, wild type, and G-CSF2/2 stromal cultures. Conditioned medium generatedfrom G-CSF/IL-6-deficient stromal cells had the capacity to stimulate bone marrow cells to formcolonies comprised of granulocytes and macrophages in soft agar clonogenic assay. This study dem-onstrates that stromal cells can be stimulated with HI Candida and gives an insight into Candida-mediated granulopoiesis.
39
INTRODUCTION
IT HAS BEEN RECOGNIZED FOR SOME TIME that bone mar-row is the major hematopoietic tissue in the adult mam-
mal. This tissue is the main source of multipotential stemcells as well as the mature myeloid cells that are neededdaily to replace the short-lived circulating blood cells (1).The development of clonal semisolid culture assays has
provided valuable information about the growth and dif-ferentiation of hematopoietic progenitor cells in responseto colony-stimulating factors and has led to the identifi-cation, purificationn and molecular cloning of variouscolony-stimulating factors (2–4). The development ofDexter cultures has demonstrated that long-term mainte-nance of stem cell growth and differentiation in vitro isdependent on bone marrow-derived adherent cells (5–8).
Ludwig Institute for Cancer Research, Melbourne Tumor Biology Branch Post Office Box 2008, Royal Melbourne Hospital,Victoria 3050, Australia.
1Present address: Walther Oncology Center, R2 302, 950 West Walnut Street, Indianapolis, IN 46202.
Alterations in steady-state hematopoiesis are com-monly associated with infection. Yet, only a few reportshave addressed the issue of whether these alterations aredue to the action of pathogens directly on the stromalcells or progenitor cells (9–11) or whether they are theresult of secondary effects following the release of cyto-kines produced in response to the pathogen by cells(12,13). It has been demonstrated previously that duringCandida albicans infection in the mouse, there is an in-crease in granulopoiesis (14,15). Although not specifi-cally demonstrated for Candida infection, neutrophiliaduring infection has often been associated with a rise inserum granulocyte colony-stimulating factor (G-CSF)levels (16). Surprisingly, we have found that G-CSF-de-ficient mice when challenged with Candida mount a pro-found and sustained neutrophilia (14). Because the Can-dida-induced neutrophilia in G-CSF-deficient mice issustained and is also accompanied by an increase in bothprecursor and mature neutrophils in bone marrow, thisclearly indicates that this phenomenon is not merely dueto mobilization of a neutrophil ‘reservoir pool’, but in-volves de novo neutrophil production.
In this study, we exploited a modified Dexter culturesystem (5,17) to study the basis of the observed Candida-mediated neutrophilia. In this model, the established bonemarrow stromal cultures are overlaid with bone marrowcells in the presence or absence of heat-inactivated (HI)Candida. The hematopoiesis seen in HI Candida-stimu-lated stromal cultures mimics in vitro the Candida-in-duced granulopoiesis in mice (14). Our results suggestthat C. albicans can directly stimulate the stromal cellsto support increased granulopoiesis. Moreover, G-CSF,granulocyte-macrophage (GM)-CSF and interleukin-6(IL-6) are dispensable for enhanced granulopoiesis byCandida-stimulated stromal cultures.
MATERIALS AND METHODS
Animals
Eight- to 10-week-old G-CSF-deficient (G-CSF2/2) (18)and wild-type mice were housed in micro-isolator cages.Mice of mixed C57BL/6 and 129/OLA background wereused for all studies. All experiments were conducted in ac-cordance with the Guidelines of the National Health andMedical Research Council, Australia, and the experimen-tal protocols were approved by the Animal Ethics Com-mittee of the Royal Melbourne Hospital Campus and Lud-wig Institute for Cancer Research (Melbourne, Australia).
Preparation of C. albicans
Subcultures of C. albicans (ATCC 18804) were main-tained at 280°C in tryptone broth containing 10% glyc-
erol. Cultures were plated on Sabouraud dextrose agarplates and grown at 37°C for 48 h. The plates were storedat 4°C for a maximum of 4 weeks. Before each experi-ment, a colony was picked from a plate and grown in 5ml of Sabouraud agar broth (SAB) broth at 37°C for 24h in a shaker. The cells were centrifuged, and the pelletwas washed twice with pyrogen-free phosphate-bufferedsaline (PBS) and resuspended in PBS. The cell suspen-sion was briefly sonicated and then autoclaved. The cellswere counted in a hemocytometer and then adjusted tothe desired concentration. These cells will be referred toas HI Candida.
Establishment of bone marrow stromal culturesand their stimulation by HI Candida
Bone marrow stromal cultures were established usingmodification of method described by Dexter et al. (5,17).Bone marrow cells from both femurs of a mouse wereplated in a T 75-cm2 flask and grown in stromal mediumcontaining Iscove’s modified Dulbecco’s medium (IMDM)/RPMI-1640 (50:50); 13 nonessential amino acids, 13 vi-tamins, and 10% HI fetal calf serum (FCS) at 37°C in anatmosphere of 5% O2:10% CO2:85% N2, and allowed tobecome confluent. After 3 weeks, the adherent cells weretrypsinized, washed and subcultured into T 25-cm2 flasksat a density of 1 3 106 cells per flask. After 3 days, thecells were irradiated at 1100 rad (gamma irradiation) toprevent them from multiplying but remain metabolicallyactive and able to produce growth factors. The irradiatedstromal cultures were then seeded with bone marrow cellsin the presence or absence of 2.5 3 106 HI Candida. Thenumber of nonadherent cells was enumerated after 7 daysof incubation by gently tapping the flasks and collectingthe supernatant. The cells were pelleted and the super-natant was filtered, aliquoted, and stored at 270°C to as-say subsequently for the cytokines and colony-stimulat-ing activity (CSA). The pellet was resuspended, and thecells were counted using hemocytometer. The nonadher-ent cells were examined for their morphology and phe-notype and assayed for the presence of committed orprimitive progenitors.
FACScan analysis of nonadherent cell population
Nonadherent cells were harvested by gently tappingthe stromal cell cultures and aspirating the medium. Thecells were centrifuged and the supernatant was discarded.The pellet was washed once with FACScan buffer (PBScontaining 5% heat-inactivated fetal calf serum). Thecells were preincubated with 2.4G2 antibody (Pharmin-gen, San Diego, CA) for 10 min at room temperature toblock Fc receptors (FcgR III/II), and the cells were sub-sequently treated with various antibodies at 4°C for 30min in FACScan buffer. The following antibodies were
BASU ET AL.
40
used: fluorescein isothiocyanate (FITC)-conjugated ratanti-mouse F4/80 and biotinylated rat anti-mouse Gr-1(Ly-6G). Those stained with the biotinylated antibodywere developed with phycoerythrin (PE)-conjugatedstreptavidin (Pharmingen, San Diego, CA). All cells wereanalyzed on a FACScan flow cytometer (Becton Dickin-son, Mountain View, CA) using Cell Quest software. Pos-itive populations were gated on the basis of staining pro-files with an isotype-matched control antibody.
In vitro colony assays
The number of myeloid progenitor cells in the nonad-herent cells recovered from stromal cultures, unstimu-lated or stimulated with HI Candida, was determined us-ing the soft agar clonal culture method described byMetcalf (19). Briefly, 2500 nonadherent cells were platedin a clonogenic soft agar assay in the presence of GM-CSF (10 ng/ml; Peprotech Inc., NJ) and IL-3 (10 ng/ml;Peprotech Inc., NJ). Cultures were incubated at 37°C ina 5% CO2 incubator, and the colonies were scored after14 days of incubation.
CSA of conditioned medium
The conditioned medium (CM) was harvested at var-ious times from established irradiated stomal cultures un-stimulated or stimulated with HI Candida, as describedearlier. In some cases the CM was collected from un-stimulated or stimulated stromal cultures overlaid withfresh bone marrow cells, and in other cases the bone mar-row cells were not added. The colony-stimulating activ-ity in the supernatant of the HI Candida-stimulated bonemarrow stromal cultures was measured in the soft agarclonogenic assay by assessing colony numbers when2.5 3 104 fresh murine bone marrow cells were culturedin the presence of 20% of CM from HI Candida-stimu-lated or unstimulated irradiated bone marrow stromal cul-tures. The colonies were scored after 7 days of incuba-tion. Colony typing was performed by morphologicalexamination of individual colonies after fixing and stain-ing the cultures (19).
IL-6 assay
IL-6 levels in the CM were assayed by the capacity ofthe CM to induce proliferation of the IL-6-dependentmouse hybridoma cell line 7TD1 (20) (obtained from Dr.J. Van Snick, LICR, Brussels Belgium) using a methodas described earlier (17). In assays using neutralizing an-tibody, samples were mixed with a 1:1 volume of a 1:100dilution of polyvalent rabbit anti-mouse IL-6 antibody for2 h before analysis in the 7TD1 bioassay. These assayswere performed in parallel with samples treated with con-trol rabbit serum.
Statistical analysis
Data are presented as mean 6 SD unless otherwisestated. Comparisons were made using two-tailed pairedStudent’s t-test, and a p value of less than 0.05 was con-sidered significant.
RESULTS
Effect of HI C. albicans on recovery ofnonadherent cells
To investigate whether Candida-mediated neu-trophilia observed in vivo (14) could be reproduced invitro, we established adherent stromal cultures from thebone marrow of wild-type and G-CSF2/2 mice as de-scribed in Materials and Methods. After 4 weeks of cul-ture, the established stromal cultures were irradiated,stimulated with HI Candida, and incubated for 24 h. Mi-croscopic examination of these cultures revealed thatsome of the stromal cells had phagocytosed the Candida.We then proceeded to study whether the addition of HICandida to irradiated stromal cells could enhancemyelopoiesis in overlaid bone marrow cells. To this end,the stromal cell cultures from both wild type and G-CSF2/2 were divided into two groups: one group wasstimulated with HI Candida and to the other group anequivalent amount of medium was added. After 7 daysof co-culture, microscopic examination of these culturesrevealed that a greater production of cells in HI Can-dida-stimulated cultures than in corresponding unstimu-lated controls. In cultures stimulated with HI Candida,there was a striking increase in the size and the densityof hematopoietic foci growing in association with theadherent stromal cells (Fig. 1). Without disturbing theadherent layer, cells in the culture supernatant were har-vested and counted. The numerical quantification of thenonadherent cells showed that on day 7 of co-culture,there was a significant increase in the number of non-adherent cells recovered from cultures stimulated withHI Candida as compared to unstimulated cultures (Fig.2). To investigate the identity of the nonadherent cells,the cells were stained with Gr-1 and F4/80 antibodies-cell specific markers for granulocytes and monocytes,respectively. In cultures stimulated with HI Candida, ap-proximately 56% of nonadherent cells were Gr-11ve
compared to 39 % in unstimulated cultures (Fig. 2A,B).Also, the absolute number of Gr-11ve cells in the non-adherent fraction was higher in HI Candida-stimulatedstromal cultures compared to unstimulated cultures (Fig.2 C,D). This result implies that addition of HI Candidanot only results in increase in an overall production ofcells but also led to preferential amplification of cellswithin the granulocytic lineage.
STIMULATION OF STROMAL CELLS BY C. albicans
41
CFC production in response to HI Candida
Because addition of HI Candida to the co-culture ofbone marrow cells and irradiated stromal cells led to anincrease in the myeloid cells in the nonadherent fraction,we further investigated the effect of Candida infectionon the functional capacity of the stromal cells to supporthematopoiesis. To address this question, the productionof progenitor cells was evaluated on day 7 of co-culture,using a conventional soft agar assay in the presence ofvarious growth factors. As shown in Fig. 3, A and B, pro-genitor cells capable of responding to GM-CSF and IL-3 were significantly elevated in the nonadherent popula-tion harvested from cultures stimulated with HI Candida.Surprisingly, the number of myeloid progenitors recov-ered both in the presence or absence of HI Candida washigher in G-CSF-deficient stromal cultures compared towild-type stromal cultures.
Cytokines production in response to HI Candida
To determine whether enhanced myelopoiesis in thepresence of HI Candida was due to enhanced productionof certain growth factors known to promote myelopoiesis,the CM generated from HI Candida-stimulated or un-stimulated bone marrow stromal cultures were analyzedfor various growth factors. Neither IL-3 nor GM-CSF wasdetected in the bioassays set up to measure these factors.
However, low levels of IL-6 were detected in the CMgenerated from both the G-CSF-deficient and wild-typestromal cultures (see Fig. 4A,B). The levels of IL-6 inthe conditioned media from both HI Candida-stimulatedor unstimulated cultures were similar (Fig. 4A,B). Al-though IL-6 could be detected even in CM harvested onday 7, the level was lower compared to that detected inday 2 CM (Fig. 4C).
Effect of anti-IL-6 antibody
Because IL-6 could be detected in the supernatant ofstromal cultures and IL-6 is known to promote granu-lopoiesis (21,22), we explored the possibility that the ac-tion of IL-6 might support myelopoiesis by both wild-type and G-CSF-deficient stromal cells in the presenceand absence of HI Candida. To this end, neutralizing anti-IL-6 antibody was added to the irradiated stromal cul-tures. We had established that 50 mg of anti-IL-6 anti-body could neutralize completely the bioactivity of 50 ngIL-6 as tested in 7TD1 proliferation assay. Because themaximum amount of IL-6 detected in the supernatant ofstromal cultures was never found to be more than 175ng, we added 196 mg of anti-IL-6 Ab to the irradiated es-tablished stromal cultures. These cultures were then over-laid with bone marrow cells in the presence or absenceof HI Candida. The nonadherent cells were recovered af-ter 7 days of co-culture and the proportion of Gr-1- andF4/80-positive cells in the nonadherent cells was deter-mined. As shown in Table 1, addition of anti-IL-6 anti-body clearly partially inhibited myelopoiesis in all cases.However, the inhibition appeared to be selective, prefer-entially inhibiting granulocyte rather than macrophageproduction. To assess further the role of GM-CSF andIL-6 in granulopoiesis in the absence of G-CSF, we stud-ied the effect of anti-IL-6 antibody on granulopoiesis sup-ported by irradiated G-CSF2/2/GM-CSF2/2 stromal cul-tures. Again, we found that anti-IL-6 antibody inhibitedmyelopoiesis (Table 2) to a similar extent as seen for G-CSF2/2 stromal cultures.
CSA of the conditioned medium
Since the above experiments demonstrated that IL-6was not solely responsible for the observed myelopoiesis,we were interested in investigating whether the CM fromthese cultures had CSA. The CSA of the stromal-derivedCM was assayed at different time points following stim-ulation with HI Candida. A significant level of colony-stimulating activity was detected 24 h post HI Candidastimulation of G-CSF-deficient stromal cells and pro-gressively declined over time and no CSA was detectedin day 7 CM (Fig. 5A). Similar profiles of CSA were ob-tained when CM from wild-type stromal cultures was
BASU ET AL.
42
FIG. 1. Increased myelopoiesis in stromal cultures stimulatedwith HI Candida. Established G-CSF2/2 bone marrow stromalcultures, unstimulated or stimulated with HI Candida, were over-laid with G-CSF2/2 bone marrow cells as described in Materi-als and Methods. After 7 days of co-culture, it was evident un-der phase-contrast microscope that compared to unstimulatedstromal cultures (A), there was a striking increase in the size andthe density of hematopoietic foci growing in association with theadherent stromal cells cultures stimulated with HI Candida (B).
STIMULATION OF STROMAL CELLS BY C. albicans
43
FIG. 2. Analysis of nonadherent cells recovered from co-culture of irradiated stromal cells and bone marrow cells. Irradiatedestablished wild-type and G-CSF2/2 stromal cultures were stimulated with HI Candida (represented as C in the axes labels) orunstimulated, and then 2.5 3 106 bone marrow cells from wild-type and G-CSF2/2 mice were overlaid on these cultures as de-scribed in Materials and Methods. After 7 days of co-culture, the nonadherent cells were recovered, counted, and stained withGr-1 and F4/80 antibodies. The total number of nonadherent recovered (A) and the proportion of Gr-11ve cells (B) in the non-adherent cells recovered on day 7 of co-culture from wild-type (blank bars) and G-CSF2 /2 (filled bars) stromal cell cultures. Thetotal number of nonadherent, Gr-11ve and F4/801ve cells recovered on day 7 from wild-type (C) and G-CSF2 /2 (D) stromal cul-tures. Nonadherent and Gr-11ve cells recovered from cultures stimulated with HI Candida were significantly greater than that re-covered from unstimulated culture of same genotype (p , 0.05). Data is represented as mean 6 SD (n 5 6).
A
B
C D
tested. To evaluate whether the decrease in CSA in thesupernatants as a function of time was due to consump-tion of the CSA by the overlaid bone marrow cells, wecollected CM from stromal cultures stimulated with HICandida at different times in the absence of bone mar-row target cells. CSA was not detected at 2 h post HICandida stimulation but could be detected at 8 h postCandida stimulation and remained similarly elevated at24 h. However, even in these cultures, a decrease in ac-tivity was noted over the time course of the experiment(Fig. 5B). It is possible that the decrease in activity inthese cultures is likely to be result of loss of biologicalpotency due to thermolability.
Types of colonies induced by CM
The colonies generated by 24 h unstimulated or stim-ulated stromal CM were further classified into differentsubtypes—namely G, GM, and M, depending on theircellular composition. As shown in Table 3, most of thecolonies generated from unstimulated CM consistedmainly of macrophages. However, when CM from G-CSF2/2 stroma stimulated with HI Candida was testedin the soft agar assay, it gave rise not only to increasednumbers of colonies compared to unstimulated G CSF2/2
stromal CM, but also to G- and GM colonies in additionto the M colonies. To assess the contribution of GM-CSFand IL-6 in stimulating myelopoiesis by G-CSF2/2 stro-mal CM, we also tested CM generated from G-CSF2/2/GM-CSF2/2 and G-CSF2/2/IL-62/2 stromalcultures in soft agar assay. CM generated from HI Can-dida-stimulated G-CSF2/2/GM-CSF2/2 or G-CSF2/2/IL-62/2 stromal cultures also had CSA, and HI Candida-stimulated CM from these stromal cultures gave moreGM-type colonies compared to respective unstimulatedstromal CM. In all situations, HI Candida stimulated stro-mal CM gave two types of GM colonies—predominantlycompact colonies with occasional dispersed colonies.
DISCUSSION
We and others have documented the observation thatCandida infection of experimental mice leads to neu-trophilia (14,15). In view of the pivotal role played by G-CSF in promoting neutrophil production in steady-stateand emergency hematopoiesis, it was perhaps surprisingthat mice deficient in G-CSF develop a profound neu-trophilia in the face of an overt infection with C. albicans.However, it is not known whether during Candida infec-
BASU ET AL.
44
FIG. 3. Colony-forming cells present in nonadherent cells recovered from co-culture of bone marrow and irradiated stromalcells. (A) Wild-type (unfilled bars) and G-CSF2/2 (filled bars) irradiated stromal cultures unstimulated/stimulated with HI Can-dida and were overlaid with either wild-type or G-CSF2/2 bone marrow cells. After 7 days nonadherent cells were harvestedfrom these cultures and counted. The proportion of CFCs present in the nonadherent cells recovered was assayed by plating 2500cells in soft agar clonogenic assay in the presence of IL-3 (10 ng/ml) and GM-CSF (10 ng/ml). The colonies were scored on day14 and the data are presented as mean 6 SD of six cultures and is representative plot of one of three independent experiments.(B) Total number of CFCs present in nonadherent cell fraction recovered on day 7 from wild-type (unfilled) and G-CSF2 /2
(filled) stromal cultures. Total number of CFCs in nonadherent fraction were significantly greater when wild-type bone marrowcells were co-cultured with HI Candida-stimulated irradiated G-CSF2 /2 stromal cell compared to HI-stimulated irradiated wild-type stromal cells (p , 0.05).
A B
tion the fungal antigens directly activate bone marrowcells to produce growth factors that ultimately promotegranulopoiesis. We have addressed this issue by exploit-ing the Dexter stromal culture system (5–8,17) to gain in-sight into the mechanisms by which the bone marrow mi-croenvironment regulates granulopoiesis in response to aninfection, particularly in the absence of G-CSF. We haveshown that established stromal cultures could be stimu-lated by HI Candida leading to enhanced granulopoiesis.
Since the use of live Candida would have resulted inthe death of stromal cell cultures, we challenged stromalcultures with heat-inactivated Candida because this strat-egy had previously been exploited successfully by oth-ers (23,24). Indeed, heat-inactivated Candida has beenshown to stimulate splenocytes to secrete cytokines(23,25,26), and, thus, while the identity of cell-surfaceantigens that mediate these effects has not been firmlyestablished (25,27), their capacity to stimulate target cellsis not compromised by heat inactivation. These effectson cultured cells are consistent with our in vivo studieswhere mice challenged with HI Candida develop neu-trophilia (S.B. and A.R.D., unpublished observation).
HI Candida was taken up by both wild-type and G-CSF-deficient stromal cells and boosted myelopoiesis.However, addition of HI Candida to bone marrow targetcells, in the absence of bone marrow stromal cells, didnot enhance myelopoiesis (data not shown). This sug-gests that the enhanced granulopoiesis seen in the pres-ence of HI Candida is not due to direct stimulation ofprogenitor cells but rather is mediated through activationof stromal cells. Cottrell et al. (10) have also shown thatLeishmania donovani can infect stromal macrophagesleading to selective enhancement of myelopoiesis. Weconsistently observed that the recovery of non-adherentGr-11ve cells as well as CFCs was greater when wild-type marrow was overlaid on G-CSF-deficient bone mar-row stromal cells compared to when they were overlaidon wild-type stromal cells. Although the reason for thisremains unclear, it is possible that the overlaid wild-typebone marrow cells produce G-CSF and this has an addi-tive effect with the compensatory cytokine(s) producedby unstimulated/HI Candida-stimulated G-CSF2/2 stro-mal cells.
The capacity of HI Candida-stimulated G-CSF-defi-
STIMULATION OF STROMAL CELLS BY C. albicans
45
FIG. 4. IL-6 levels in CM from stromal culture. Day 7 CM harvested from co-cultures of unstimulated and HI Candida (rep-resented as C in the axis label) stimulated wild-type (A) and G-CSF2 /2 (B) irradiated stromal cells overlaid with either wild-type and G-CSF2/2 bone marrow cells were tested for IL-6 level using the 7TD1 bioassay at 1:100 dilution. (C) Irradiated G-CSF2/2 stromal culture was overlaid with either wild-type or G-CSF2/2 bone marrow cells and stimulated with HI Candida.IL-6 level in the day 2 and day 7 CM was measured using 7TD1 bioassay at 1:500 dilutions. Data are represented as mean 6
SD (n 5 3) and is representative plot of three independent experiments.
A B
C
BASU ET AL.
46
TA
BL
E1.
EF
FE
CT
OF
AN
TI-
IL-6
AN
TIB
OD
YO
NR
EC
OV
ER
YO
FN
ON
AD
HE
RE
NT
CE
LL
S,
GR
-11
veA
ND
F4/8
01ve
CE
LL
SIN
HI
CA
ND
IDA-S
TIM
UL
AT
ED
AN
DU
NS
TIM
UL
AT
ED
ST
RO
MA
LC
UL
TU
RE
S
WT
str
oma
(tot
al n
umbe
r of
cel
ls r
ecov
ered
310
24 )
G-C
SF2
/2st
rom
a (t
otal
num
ber
of c
ells
rec
over
ed 3
102
4)
Non
adhe
rent
Non
adhe
rent
IL-6
Abb
cell
sG
r-1
F4/
80IL
-6 A
bbce
lls
Gr-
1F
4/80
WT
BM
270
.56
17.
633
.46
11.
46.
46
1.3
WT
BM
293
.06
10.
042
.06
3.0
10.0
6 2
.0W
T B
M1
29.6
6 3
.311
.86
1.3
3.5
6 0
.4W
T B
M1
48.5
6 3
.514
.06
0.3
a5.
06
0.4
GK
O B
M2
44.0
6 7
.113
.06
2.9
5.5
6 1
.8G
KO
BM
265
.56
13.
419
.66
8.2
13.0
6 0
.3G
KO
BM
125
.26
0.3
6.9
6 0
.42.
36
0.3
GK
O B
M1
39.2
6 1
1.1
3.1
6 1
.0a
8.0
6 0
.2W
TB
M2
117.
56
24.
775
.96
14.
26.
76
1.0
WT
BM
220
1.0
6 1
1.0
102.
06
5.2
17.0
6 0
.21
Can
dida
1C
andi
daW
TB
M1
53.0
6 4
.227
.36
4.3
5.6
6 1
.6W
TB
M1
73.0
6 1
4.0
12.4
6 1
.0a
9.5
6 1
.21
Can
dida
1C
andi
daG
KO
BM
211
56
22.
657
.56
13.
012
.46
1.9
GK
OB
M2
100.
06
14.
136
.06
1.0
12.2
6 1
.81
Can
dida
1C
andi
daG
KO
BM
148
.56
7.0
15.6
6 1
.04.
96
0.1
GK
OB
M1
67.5
6 2
3.0
7.0
6 1
.2a
13.0
6 1
.21
Can
dida
1C
andi
da
Ant
i-IL
-6 A
b (1
96 m
g/fl
ask)
was
add
ed t
o th
e ir
radi
ated
est
abls
hed
stro
mal
cul
ture
s. T
hese
cul
ture
s w
ere
then
ove
rlai
d w
ith
bone
mar
row
cel
ls i
n th
e pr
esen
ce o
r ab
senc
e of
HI
Can
dida
. T
he n
onad
here
nt c
ells
wer
e re
cove
red
afte
r 7
days
of
co-c
ultu
re a
nd t
he p
ropo
rtio
n of
Gr-
1- a
nd F
4/80
-pos
itiv
e ce
lls
in t
he n
onad
here
nt c
ells
was
det
erm
ined
. D
ata
are
repr
esen
ted
as m
ean
6S
D o
f ce
lls
reco
vere
d fr
om t
hree
ind
epen
dent
fla
sks.
a p,
0.05
com
pare
d to
tha
t re
cove
red
in a
bsen
ce o
f an
ti-I
L-6
Ab.
b 50
mg
of a
nti-
IL-6
ant
ibod
y co
uld
neut
rali
ze c
ompl
etel
y th
e bi
oact
ivity
of
50 n
g IL
-6 a
s te
sted
in
7TD
1 pr
olif
erat
ion
assa
y.
cient stromal CM to generate a few G colonies, in addi-tion to GM and M colonies, could be due to the presenceof low levels of GM-CSF in HI Candida-stimulated G-CSF2/2 stromal CM. GM-CSF alone or in combinationwith other cytokines produced by stromal cells can stim-ulate CFU-G (26–29). Although we could not detect GM-CSF in the CM harvested from G-CSF2/2 stromal cul-ture, it has been shown previously (30–33) that irradiatedbone marrow stromal cells can produce GM-CSF. How-ever, our data show that G-CSF2/2/GM-CSF2/2 stromalcells could also be stimulated with HI Candida leadingto enhanced myelopoiesis in overlaid bone marrow cells.Moreover, the CM from HI Candida-stimulated G-CSF2/2/GM-CSF2/2 stromal cells can stimulate GM andM types of colonies in a soft agar assay. Therefore, GM-CSF is dispensable for Candida-mediated myelopoiesisby bone marrow stromal cells. This is in agreement withprevious findings in vivo, where GM-CSF was found dis-pensable for sustaining steady-state (34) as well as ‘emer-gency’ hematopoiesis (14).
IL-6 levels were elevated in both wild-type and G-CSF2 /2 stromal cultures, but stimulation of stromalcultures with HI Candida had little effect on IL-6 lev-els. In line with an earlier report (35), we have foundthat addition of anti-IL-6 antibody to stromal culturespartially inhibited myelopoiesis supported by bothwild-type and G-CSF2 /2 stromal cultures. Although,this suggests that IL-6 is partially responsible for dri-ving myelopoiesis in these in vitro cultures, our otherfindings suggest that IL-6 is dispensable for the Can-dida-mediated myelopoiesis. This includes our obser-vation that G-CSF2 /2/IL-62/2 stroma could supportmyelopoiesis (data not shown); moreover, CM har-vested from HI Candida stimulated G-CSF2/2/IL-62 /2
stromal cultures could support myeloid colony forma-tion in the in vitro soft agar assay.
What drives myelopoiesis in the absence of G-CSF?We sought to address this by testing CM harvested fromCandida-stimulated/unstimulated bone marrow stromalcultures from various gene-targeted mice for their abil-ity to generate colonies from bone marrow cells in the invitro colony assay. It has been shown previously that stro-mal cultures can produce various factors including G-CSF, GM-CSF, M-CSF, IL-1, IL-3, IL-6, IL-7, stem cellfactor (SCF), tumor necrosis factor-a (TNF-a), and trans-forming growth factor-b (TGF-b) (36,37). Although onlya subset of these factors can stimulate granulopoiesis di-rectly, others can be co-stimulatory. Interestingly, themost dramatic effect of stimulation of stromal cultureswith HI Candida was observed by the capacity of CMharvested from such cultures to generate significantlyhigher numbers of CFU-GM compared to unstimulatedstromal culture CM. Cotterel et al. (10) have also shownthat L. donovani infection of stromal cells enhanced theirability to support hematopoiesis. This was due to the en-hanced ability of L. donovani-infected stromal cells tosupport colony formation, specifically CFU-GM, and thisinvolved induction of GM-CSF and TNF-a. Similar toour findings in vitro, we had also observed in vivo thatthe peripheral neutrophilia and monocytosis was pre-ceded by an increase in GM-type progenitors in the bonemarrow of growth factor(s)-deficient mice. These obser-vations had led us to speculate that in the absence of G-CSF, a compensatory factor/s is produced that drives pro-liferation of bone marrow cells at the GM-progenitorstage (14). Taken together, our in vivo and in vitro ob-servations suggest that the cytokines(s) involved and thestage at which progenitor cells are expanded during
STIMULATION OF STROMAL CELLS BY C. albicans
47
TABLE 2. EFFECT OF ANTI-IL-6 ANTIBODY ON RECOVERY OF NONADHERENT
CELLS FROM G-CSF/GM-CSF-DEFICIENT STROMAL CELLS
Total number of cells recovered 3 1014
Overlaid Number of nonadherent Gr-11 F4/801
BM cell IL-6 Abb cells cells cells
G-CSF2/2/GM-CSF2/2 2 63.0 6 6.4 17.8 6 3.9 10.8 6 4.3G-CSF2/2/GM-CSF2/2 1 35.0 6 7.1a 2.3 6 1.3a 9.0 6 5.6
1 CandidaG-CSF2/2/GM-CSF2/2 2 94.0 6 11.2 34.1 6 7.7 9.5 6 2.9G-CSF2/2/GM-CSF2/2 1 65.5 6 12.4a 5.8 6 2.5a 9.3 6 3.4
1 Candida
Anti-IL-6 Ab (196 mg/flask) was added to the irradiated establshed G-CSF2/2/GM-CSF2/2 stromal cultures. These cultureswere then overlaid with G-CSF2/2/GM-CSF2/2 bone marrow cells in the presence or absence of HI Candida. The nonadherentcells were recovered after 7 days of co-culture, and the proportion of Gr-1- and F4/80-positive cells in the nonadherent cells wasdetermined. Data are represented as mean 6 SD of cells recovered from three independent flasks.
ap , 0.05 compared to that recovered in absence of anti-IL-6 Ab.b50 mg of anti-IL-6 antibody could neutralize completely the bioactivity of 50 ng IL-6 as tested in 7TD1 proliferation assay.
myelopoiesis are similar in the absence of G-CSF/GM-CSF, G-CSF/IL-6, and G-CSF alone. The production ofneutrophils in the absence of G-CSF involves amplifica-tion of progenitors that are at least bipotential, appear lessdifferentiated than the CFU-G stage, and could possiblybe at the CFU-GM-stage or earlier. Moreover, the find-ing that both G-CSF2/2/GM-CSF2/2 and G-CSF2/2/IL-
62/2-deficient stromal CM could generate GM-typecolonies in the soft agar assay clearly suggests factor(s)other than G-CSF, GM-CSF, and IL-6 contribute to thegeneration of granulocytes.
Our model clearly demonstrates that bone marrow stro-mal cells can be directly activated by HI Candida to drivemyelopoeisis. Current studies are underway to identify
BASU ET AL.
48
FIG. 5. Colony-stimulating activity of conditioned medium harvested from stromal culture. (A) Irradiated G-CSF-deficient stro-mal cells, unstimulated (e ) or stimulated with HI Candida ( u ) were overlaid with G-CSF-deficient bone marrow cells. CM washarvested at the indicated times and assayed for colony-stimulating activity using soft agar clonogenic assay as described in Ma-terials and Methods. (B) CM from irradiated G-CSF2 /2 stromal cultures stimulated with HI Candida was harvested at indicatedtimes and assayed for CSA. In all assays, IL-3 (10 ng/ml) (O) was used as a standard. Data are represented as mean 6 SD. Oneof two separate experiments with similar results is shown.
TABLE 3. COLONY-STIMULATING ACTIVITY INDUCED BY VARIOUS SOURCES OF CM
Total number ofcolonies per Number of Number of Number of2.5 3 104 G-type M-type GM-type
Source of CM BM cellsa coloniesa coloniesa coloniesa
G-CSF2/2 stroma 34 6 1.0 0 6 0.0 30 6 4.5 4 6 3.5G-CSF2/2 stroma 1 HI Candida 51 6 7.0b 8 6 0.1b 20 6 8.0 23 6 1.0b
G-CSF2/2/GM-CSF2/2 stroma 27 6 5.0 0 6 0.0 26 6 5.0 1 6 1.0G-CSF2/2/GM-CSF2/2 stroma 1 HI Candida 58 6 3.0b 0 6 0.0 46 6 2.5b 12 6 3.0b
G-CSF2/2/IL-62/2 stroma 32 6 3.2 0 6 0.0 30 6 2.0 2 6 2.0G-CSF2/2/IL-62/2 stroma 1 HI Candida 66 6 6.4b 0 6 0.0 39 6 7.2 27 6 3.0b
Bone marrow cells from C57/BL6 mice were treated with 24 h CM from various sources at 1:5 dilution in soft agar assay.The colonies were counted on day 7 and thereafter fixed and stained and typed into following types depending on the cellularcomposition of the colonies. G, Granulocytic; M, macrophage; and GM, granulocyte-macrophage.
aValues are represented as mean 6 SD for three different batches of CM of each type.bComparisons significant at p , 0.05 using two-tailed t-test. Comparisons were made between HI Candida-stimulated and
unstimulated CM from same genotype of irradiated stromal culture.
A B
the factor(s) that drive granulopoiesis in the absence ofG-CSF, GM-CSF, and IL-6.
ACKNOWLEDGMENTS
This study was supported in part by a research grantfunded by NHMRC. We are thankful to T. Helman, B.Morrow, and E. Richardson for assistance in the animalhouse; J. Stickland for assistance with the artwork; andDr. Alister Ward and Prof. A.W. Burgess for criticallyreviewing the manuscript.
REFERENCES
1. Metcalf D and MAS Moore, eds. Haemopoietic Cells.(1971). Elsevier North Holland, Amsterdam.
2. Johnson GR and D Metcalf. (1977). Pure and mixed eryth-roid colony formation in vitro stimulated by spleen-condi-tioned medium with no detectable erythropoietin. Proc NatlAcad Sci USA 74:3879–3882
3. Morstyn G and AW Burgess. (1988). Hemopoietic growthfactor. Cancer Res 48:5624–5637
4. Metcalf D. (1989). Hemopoietic growth factors 1. Lancet15:825–827
5. Dexter TM. (1982). Stromal cell associated hematopoiesis.J Cell Physiol 1:87–94 (suppl).
6. Dexter TM, LH Coutinho, E Spooncer, CM Heyworth, CPDaniel, R Schiro, J Chang and TD Allen. (1990). Stromalcells in hematopoiesis. In: Molecular Control of Hemato-poiesis. G Black and J Marsh, eds. Ciba Foundation Sym-posium 1990;148:76–95.
7. Weilbaecher K, I Weissman, K Blume and S Heimfeld.(1991). Culture of phenotypically defined hematopoieticstem cells and other progenitors at limiting dilution on Dex-ter monolayers. Blood 78:945–952
8. Reimann J and H Burger. (1979). In vitro proliferation ofhaemopoietic cells in the presence of adherent cell layers.II. Differential effect of adherent cell layers derived fromvarious organs. Exp Hematol 7:52–58
9. Wang S-Y, C-Y Su, M-L Hsu, L-Y Chen, C-H Tzeng andC-K Ho.(1991). Effect of lipopolysaccharide on the pro-duction of colony-stimulating factors by the stromal cells inlong-term bone marrow culture. Exp Hematol 19:122–127.
10. Cotterell SEJ, CR Engwerda and PM Kaye. (2000). Leish-mania donovani infection of bone marrow stromal macro-phages selectively enhances myelopoiesis, by a mechanisminvolving GM-CSF and TNF-a. Blood 95:1642–1651.
11. Maürer AK, M Wilhelm, F Weissinger, EB Brocker andW Goebel. (2002). Infection of human hematopoietic stemcells with bacterial pathogens. Blood 100: 3703–3709.
12. Murray PJ, RA Young and GQ Daley. (1988) Hema-topoieetic remodeling in interferon-g-deficient mice in-fected with mycobacteria. Blood 91:2914–2924.
13. Lelichuk R, R Graveley and FY Liew. (1988). Suscepti-bility to murine cutaneous leishmaniasis correlates with thecapacity to generate interleukin-3 in response to leishma-nia antigen in vitro. Cell Immunol 111:66–76.
14. Basu S, G Hodgson, H-H Zhang, M Katz, C Quilici andAR Dunn (2000). “Emergency” granulopoiesis in G-CSF-deficient mice in response to Candida albicans infection.Blood 95:3725–3733.
15. Bistoni F, A Vecchiarelli, E Cenci, P Puccetti, P Marconiand A Cassone. (1986). Evidence for macrophage-medi-ated protection against lethal Candida albicans infection.Infect Immun 51:668–674.
16. Kawakami M, H Tsutsumi and T Kumukawa. (1990) Lev-els of serum granulocyte-colony-stimulating factor in pa-tients with infections. Blood 76:1962–1964.
17. Dexter TM, TD Allen and LG Lajtha. (1977). Conditionscontrolling the proliferation of haemopoietic stem cells invitro. J Cell Physiol 91:335–334.
18. Lieschke GJ, D Grail, G Hodgson, D Metcalf, E Stanley,E Cheers, KJ Fowler, S Basu, YF Zhan and AR Dunn.(1994) Mice lacking granulocyte colony-stimulating factorhave chronic neutropenia, granulocyte and macrophageprogenitor cell deficiency, and impaired neutrophil mobi-lization. Blood 84:1737–1746.
19. Metcalf D. The hemopoietic colony stimulating factors. El-sevier, Amsterdam.
20. Van Snick J, S Cayphas, A Vink, C Uyttenhove, PG Coulie,MR Rubira and RJ Simpson. (1986). Purification and NH2-terminal amino acid sequence of a T-cell-derived lym-phokine with growth factor activity for B-cell hybridomas.Proc Natl Acad Sci USA 83:9679–9683.
21. Dalarymple SA, LA Lucian, R Slattery, T McNeil, DMAbud, S Fuchino, F Lee and R Murray. (1998). Interleukin-6 deficient mice are highly susceptible to Listeria mono-cytogenes infection: correlation with inefficient neu-trophilia. Infect Immun 63:2262–2268.
22. Ulich TR, J del Castillo and KZ Guo. (1989). In vivo hema-tologic effects of recombinant interleukin-6 on hematopoi-esis and circulating numbers of RBCs and WBCs. Blood73:108–110.
23. Steinshamm S and A Waage (1992). Tumor necrosis factorand interleukin-6 in candida albicans infection in normal andgranulocytopenic mice. Infect Immun 60:4003–4008.
24. Ortega E, J Algarra, MJ Serrano, G Alvarezde Ceinfuegosand JJ Gaforio. (2001).The use of 7-amino-actinomycin Din the analysis of Candida albicans phagocytosis response.J Immunol Methods 253:189–193.
25. Vecchiarelli A, M Puliti, A Torosantucci, A Cassonee andF Bistoni. (1991). In vitro production of tumor necrosis fac-tor by murine splenic macrophages stimulated with manno-protein constituents of candida albicans cell wall. Cell Im-munol 134:65–76
26. Casalinuovo IA, R Gaziano and P Di Francesco. (2000).Cytokine pattern secretion by murine spleen cells after in-activated candida albicans immunization. Effect of cocaineand morphine treatment. Immunopharmacol Immunotoxi-col 22:35–48.
27. Bromuro C, R La Valle, S Sandini, F Urbani, CM Ausiello,L Morelli, S Fed’ Ostiani, L Romani and A Cassone.(1998). A 70-kilodalton recombinant heat shock protein ofcandida albicans is highly immunogenic and enhances sys-temic murine candiasis. Infect Immun 66:2154–2162.
28. Tsuji T, K Sugimoto, T Yanai, E Takashita and KJ Mori.(1994). Induction of granulocyte-macrophage colony-stim-
STIMULATION OF STROMAL CELLS BY C. albicans
49
ulating factor (GM-CSF) and granulocyte colony-stimulat-ing factor (G-CSF) expression in bone marrow and frac-tionated marrow cell populations by interleukin 3 (IL-3):IL-3 mediated positive feedback mechanisms of granu-lopoiesis. Growth Factors 11:71–79.
29. Hestdal K, SE Jacobsen, FW Ruscetti, DK Longo, TCBoone and JR Keller. (1993). Increased granulopoiesis af-ter sequential administration of transforming growth fac-tor-beta 1 and granulocyte colony-stimulating factor. ExpHematol 21:799–805.
30. Hestdal K, SE Jacobsen, FW Ruscetti, CM Dubois, DLLongo, R Chizzonite, JJ Oppenheim and JR Keller. (1992).In vivo effect of interleukin-1 alpha on hematopoiesis: roleof colony stimulating factor receptor modulation. Blood80:2486–2494.
31. Sutherland HJ, DE Hogge and CJ Eaves (1993). Growthfactor regulation of the maintenance and differentiation oflong-term culture-initiating cells (LTC-IC). Leukemia.Suppl 2:S122–125.
32. Quesenberry PJ, E Kittler, P Lowry and R Crittenden.(1993) Stromal requlation of haemopoiesis and the role ofsubliminal growth factor production. Leukemia Suppl 2:S3–4.
33. Kittler E, H McGrath, D Temeles, R Crittenden, VK Kister,and PJ Quesenberry. (1992). Biologic significance of con-stitutive and subliminal growth factor production by bonemarrow. Blood 79:3168–3178.
34. Seymour JF, GJ Lieschke, D Grail, C Quilici, G Hodgson
and AR Dunn. (1997). Mice lacking both granulocytecolony-stimulating factor (CSF) and granulocyte-macro-phage CSF have impaired reproductive capacity, perturbedneonatal granulopoiesis, lung disease, amyloidisis and re-duced long-term survival. Blood 90:3037–3049
35. Hudak S, LA Thompson-Snipes, C Rocco, J Jackson, MPearce and D Rennick. (1992) Anti-IL-6 antibodies suppressmyeloid cell production and the generation of CFU-c inlong-term bone marrow cultures. Exp Hematol 20:412–417.
36. Quesenberry PJ, RB Crittenden, P Lowry, EW Kittler, SRao, S Peters and FM Stewart. (1994). In vitro and in-vivostudies of stromal niches. Blood Cells 20:97–104
37. Sensebe L, M Deschaseaux, J Li, P Herve and P Charbord.(1997). The broad spectrum of cytokine gene expressionby myeloid cells from the human marrow microenviron-ment. Stem Cells 15:133–143.
Address reprint requests to:Dr. Sunanda Basu
Walther Oncology CenterR2 302
950 West Walnut StreetIndianapolis, IN 46202
E-mail: [email protected]
Received August 15, 2003; accepted January 10, 2004.
BASU ET AL.
50