Mycobacterial Glycolipid Cord Factor Trehalose 6,6′-Dimycolate Causes a Decrease in Serum Cortisol during the Granulomatous Response

Abstract

Serum cortisol levels were evaluated in mice following intravenous administration of purified mycobacterial glycolipid trehalose 6,6′-dimycolate (TDM). C57BL/6 mice develop lung granulomas in response to TDM, while A/J mice are deficient in this process. Administration of TDM to C57BL/6 mice led to a rapid reduction in serum cortisol, concurrent with initiation of the granulomatous response and cytokine and chemokine mRNA induction. Cortisol levels were lowest on day 5 after TDM administration, but there was significant production of IL-6, TNF-α and IL-1β messages. Granuloma formation and full immune responsiveness to TDM were only apparent upon a sufficient decrease in levels of systemic cortisol. Treatment of the C57BL/6 mice with hydrocortisone abolished inflammatory responses. Histologically nonresponding A/J mice exhibited higher constitutive serum cortisol and demonstrated different kinetics of cortisol reduction upon administration of TDM. A/J mice demonstrated hyperplastic morphology in the suprarenal gland with a high degree of vacuolization in the medullary region and activation of cells in the zona fasciculata and zona reticularis. The A/J mice were dysregulated with respect to cytokine responses thought to be necessary during granuloma formation. The high constitutive serum cortisol in the A/J mice may therefore contribute to pulmonary immunoresponsiveness and the establishment of an environment counterproductive to the initiation of granulomatous responses. The identification of a mycobacterial glycolipid able to influence serum cortisol levels is unique and is discussed in relation to immunopathology during tuberculosis disease.

Introduction

The role of glucocorticoids in immunomodulation and the granulomatous response to mycobacterial antigens is poorly understood. Experiments demonstrated nearly 35 years ago that glucocorticoids can suppress immune responses to Mycobacterium tuberculosis (MTB) infection [1]. Indeed, cortisone can inhibit granuloma formation in response to both killed mycobacteria and to mycobacteria-derived factors [2, 3]. A correlation between endocrine function and cytokine production during active tuberculosis has been suggested [4–6]; however, the exact relationship between cortisol and immunopathology during active disease remains unknown. To date, no specific mycobacterial factor has been implicated in this process.

Genetically distinct mice differ in their susceptibility to pathogens, including MTB [7–12]. It was observed that serum cortisol decreased dramatically during MTB induction of granuloma formation in mice [6]. Specifically, decreased cortisol was most apparent during the initiation of immunopathology. Mice that did not develop granulomas had greater constitutive levels of cortisol and reduced relative decreases in cortisol during infection. Because lowered serum cortisol correlated positively with histologic and immunologic events consistent with granuloma formation, it was suggested that reduction in cortisol levels in response to mycobacterial factors exerts a direct influence on initiation of the granulomatous response.

When administered in the appropriate physical conformation, purified mycobacterial cord factor glycolipid trehalose 6,6′-dimycolate (TDM) induces granulomatous responses in C57BL/6 mice, but not in A/J mice [13–15]. The histological and cytokine profiles during the response to TDM in these mice mimic in part the immunopathology observed during MTB infection. We hypothesized that granulomatous responses and immunoresponsiveness to TDM in these mice correlate with relative levels of systemic glucocorticoids. Therefore, we were interested in characterizing immune events induced by purified mycobacterial TDM in both A/J and C57BL/6 mice in connection with serum cortisol profiles and granulomatous responses.

Materials and Methods

Generation of TDM-Coated Beads

TDM obtained by petroleum ether extraction from Mycobacterium smegmatis was a gift of Dr. Kuni Takayama, Madison, Wisc., USA. The TDM used in these studies was similar to that described elsewhere [16, 17]. Styrene divinylbenzene beads (Bangs Laboratories Inc., Fishers, Ind., USA) with mean diameters of 5.1 μm were coated with TDM by mixing them with TDM dissolved in hexaneethanol (9:1) and then rapidly evaporating the solvent under a stream of air. The TDM is deposited as an even monolayer (0.37 μg/cm²); the amount of TDM was calculated to produce surface layers of TDM one or two molecules thick [15]. Final suspension was at 10⁶ beads/ml in saline.

Mice and Administration of TDM Beads

Eight-week-old C57BL/6 and A/J mice (Jackson Laboratory, Bar Harbor, Me., USA) were injected intravenously with 10⁵ TDM-coated beads, or uncoated beads, in 0.1 ml of saline. TDM produced lung granulomas in C57BL/6 mice [13] but not in A/J mice [15]. Administered TDM produces a marked inflammatory response, including granuloma-like lesions in the lung [18–20], with lesions that peak in number and size in the first week and resolve over the following 2 weeks. Granulomatous responses are characterized by a rapid influx of monocytic cells and initiation of proinflammatory cytokine production [14, 15, 21].

The mice were bled prior to injection, then at various times in the 14 days after TDM administration. After sacrifice, lung tissue was aseptically removed and processed for histology or for RNA evaluation, and serum was evaluated for changes in cortisol levels. Use of mice was approved by the Animal Welfare Committee of University of Texas-Houston Medical School, AWP #A3413-01.

Evaluation of Serum Cortisol

3–5 mice were bled prior to injection and at subsequent times in the 14 days after injection. Blood was collected at the same time each day in microtubes and allowed to clot overnight at 4°C, and then centrifuged at 13,000 rpm. Serum was diluted 1/10 in phosphate-buffered saline (PBS), pH 7.4. Cortisol was measured in an Abbott TDX/FLX analyzer using a fluorescence polarization immunoassay (Abbott Laboratories, Abbott Park, Ill., USA) with an assay sensitivity of 0.64 μg/dl (95% confidence). Alternatively, 0.1-ml undiluted samples were analyzed at the Memorial Hermann Hospital Chemistry Lab (Houston, Tex., USA) using the Access ImmunoAssay System. Significance was determined using Student’s t test.

Treatment with Hydrocortisone

Procedures for administration of hydrocortisone as anti-inflammatory mediator during infectious disease in mice have been described previously [22]. Briefly, mice received subcutaneous (tail base) hydrocortisone (0.4 mg of 17-hydroxycortisone; 11β, 17α, 21-trihydroxypregn-4-ene-3,20-dione; Sigma Aldrich) or diluent control (PBS) in a 0.1-ml delivery volume on day 0 1 h prior to administration of TDM-coated beads, and again on days 2 and 4 after TDM administration.

Isolation and Purification of mRNA

RNase-free plastic and water were used throughout the experiments. Lung samples (approximately 25 mg) were homogenized in 0.5 ml of RNAzol (Tel-Test, Inc., Friendswood, Tex., USA). 0.05 ml of chloroform was added, then the sample was shaken vigorously for 30 s, chilled on ice for 15 min and centrifuged at 13,000 rpm for 15 min at 4°C. The aqueous phase was transferred to a fresh tube and an equal volume of isopropanol (2-propanol) was added. The samples were stored overnight at −20°C followed by 15 min at 4°C. After centrifugation at 13,000 rpm for 15 min, the pellet was washed with ultrapure 75% ethanol. The RNA was resuspended in diethylpyrocarbonate-treated water containing 1 mM EDTA.

Reverse Transcription

cDNA was synthesized and quantitated from RNA as described elsewhere [23]. Briefly, 1 μg of total RNA together with 1× Superscript II reverse transcriptase (RT) buffer (Gibco, Grand Island, N.Y., USA), 0.1 M dithiothreitol, 2.5 mM deoxynucleotide triphosphates (Boehringer Mannheim, Indianapolis, Ind., USA), 80 U of random hexamer oligonucleotides (Boehringer Mannheim) and 20 U of RNase inhibitor (RNasin, Promega) in a 25-μl volume was heated for 5 min at 70°C, then chilled for 5 min at 4°C. RT was added and the mixture was incubated for 60 min at 37°C. The reaction was heated to 90°C for 5 min. Finally, the sample was diluted 1/8 using distilled water.

PCR Amplification of cDNA

Amplification of the PCR products was accomplished using a sense primer biotinylated on the 5′ terminal nucleotide to facilitate later capture using streptavidin. To the PCR reaction mixture the following components were added: 0.25 mM deoxynucleotide triphosphate mix, 1× PCR buffer (50 mM KCl, 10 mM Tris-HCl, 1.5 mM MgCl₂; Promega), 0.2 μM sense and antisense primers, 5 μl of cDNA, and 1 U of Taq polymerase (Promega). Denaturation, annealing and elongation temperatures for PCR were 94, 54 and 72°C for 30, 30 and 40 s, respectively, using a PTC-100-96V thermal cycler (MJ Research, Waltham, Mass., USA). Numbers of cycles for amplification are listed in table 1. Quantitative comparison of RT-PCR products was done during the linear phase of amplification [24, 25], where detected amplicons were normalized to β-actin mRNA levels from non-bead-administered controls. Representative cDNA samples were also subjected to 40 cycles of amplification and electrophoresis followed by ethidium bromide staining to confirm a single-band product.

Table 1.

Mouse bioluminescent RT-PCR primer sets and probes

Primers			Product size, bp	Number of cycles	Probe
β-actin	S	biotin-ATGGATGACGATATCGCT	185	23	GAGCACCCTGTGCTGCTC-dig
	AS	ATGAGGTAGTCTGTCAGGT
IL-1β	S	biotin-GCAACTGTTCCTGAACT	383	24	GGATGATGATGATAACCTG-dig
	AS	CTCGGAGCCTGTAGTGCA
IL-6	S	biotin-TTGGTCTCTGCAAGAGACT	428	31	AGTAGGGAAGGCCGTGGTTGTCAC-dig
	AS	TGTATCTCTCTGAAGGACT
TNF-α	S	biotin-GTCTACTTTGGAGTCATTGC	304	28	CTCTTCAAGGGACAAGGCTG-dig
	AS	GACATTCGAGGCTCCAGTG
IFN-γ	S	biotin-AACGCTACACACTGCATCT	400	30	TCGCCTTGCTGTTGCTGA-dig
	AS	GAGCTCATTGAATGCTTGG

S = Sense; AS = antisense; biotin = 5′-biotinylated; dig = 3′-digoxigenin; bp = base pairs.

Oligonucleotide Primers and Probes

Sequences were synthesized at Sigma-Genosys (Woodlands, Tex., USA) and are listed in table 1 [11]. The murine primers are RNA specific in that both the 5′ and 3′ primers span 1 intron, thus precluding amplification of genomic DNA. The primer sequences were labeled with biotin on the 5′ end to facilitate later capture on streptavidin plates, and probes were labeled with digoxigenin for detection using bioluminescence [24, 26, 27].

Bioluminescent Hybridization Immunoassay

Bioluminescent detection is optimal for measurement of PCR products derived from the early phase of log amplification [25, 28]. 5 μl of biotinylated PCR product was denatured by the addition of 1/4 volume (1.25 μl) of denaturation buffer (1 M NaOH, 200 mM EDTA) for 5 min at room temperature. To neutralize, an equal volume (6.25 μl) of neutralization buffer (0.15 M Na₂HPO₄, pH 6.0) was added. The samples were added to wells of Streptavidin-Microtiterplates^® (MicroCoat, Penzberg, Germany) which already contained 1–2 ng of digoxigenin-labeled probe in hybridization buffer (62.5 mM Na₂HPO₄, 0.94 M NaCl, 94 mM citric acid, 10 mM MgCl₂, 0.125% Tween-20, 0.0625% BSA, 15 mM NaN₃, pH 6.5). The wells were covered and incubated for 2 h at 42°C in a water bath. After washing 4 times with 150 μl of ChemFLASH wash buffer (20 mM Tris, 5 mM EDTA, 0.15 M NaCl, 0.05% Tween-20, 15 mM NaN₃, pH 7.5; Chemicon International Inc., Temecula, Calif., USA), 5 ng of a conjugate of AquaLite^® and anti-digoxigenin (Chemicon) in assay buffer (25 mM Tris, 10 mM EDTA, 2 mg/ml BSA, 0.15 M KCl, 0.05% Tween-20, 15 mM NaN₃, pH 7.5) was added. After incubation for 30 min with agitation (150 rpm) at room temperature, the wells were washed (4 times) with wash buffer. The conjugate was detected by measuring the flash reaction on a Berthold microplate luminometer LB 96V (EG&G Berthold, Germany), integrating for 2 s after injection of AquaLite Trigger Solution (Chemicon). The apparatus precision was determined using 2.5 × 10⁻¹⁰ moles of AquaLite, resulting in a coefficient variation (CV) of 5.0% [27].

Stimulation of Bone Marrow-Derived Macrophages

Bone marrow-derived macrophages (BMM) were established as described previously [14], with modifications. Briefly, C57BL/6 femurs were flushed with PBS, and 3 × 10⁷ cells were aliquoted into 75-cm² tissue culture flasks (Costar). Cells were grown in EMEM containing 10% FBS, 10 ng/ml recombinant murine GM-CSF (Chemicon), 100 U/ml penicillin and 100 μg/ml streptomycin (complete medium). Cells were incubated at 37°C in 5% CO₂ overnight. Nonadherent cells were collected and incubated for 7 days and fed twice with additional complete medium. Finally, adherent cells were removed, washed, resuspended in DMEM containing 5% FBS and plated at 1 × 10⁶ cells/well in 24-well plates (Costar). TDM-coated styrene divinylbenzene beads were diluted in PBS and incubated with macrophages at a dilution of approximately 3 beads per macrophage.

ELISA Monokine Quantitation

Levels of IL-1β, TNF-α, IL-6 and IFN-γ in cell supernatants and TNF-α in lung mince preparations were measured by sandwich ELISA. Costar 96-well vinyl assay plates were coated with capture antibody overnight (IL-1β at 2.5 μg/ml, TNF-α at 0.5 μg/ml, IL-6 at 1 μg/ml; all from R&D Systems, Minneapolis, Minn., USA). Plates were washed three times with wash buffer (0.05% Tween-20 in PBS). Blocking buffer (1% BSA, 5% sucrose, 0.05% NaN₃ in PBS) was added for 3 h. After three washings, 100 μl of cell supernatants were added. Plates were incubated for 2 h. Biotin-conjugated secondary antibodies were added after washing (IL-1β at 80 ng/ml, TNF-α at 200 ng/ml, IL-6 at 154 ng/ml; all from R&D Systems). Plates were incubated for 2 h, washed and then developed using streptavidin-horseradish peroxidase (Sigma, St. Louis, Mo., USA) and TMB Microwell Peroxidase Substrate (Kirkegaard and Perry, Gaithersburg, Md.). Absorbance was read at 570 and 450 nm on an ELISA plate reader (Molecular Devices). The means of triplicate wells were calculated based on a standard curve constructed for each assay, using recombinant murine IL-1β, TNF-α, IL-6 and IFN-γ (R&D System). The limit of sensitivity was 5 pg/ml.

Statistical Analysis

All data are presented as the mean ± 1 standard error. Normally distributed data were analyzed by the unpaired t test when the difference between two means was compared within groups. The Wilcoxon matched-pairs signed-ranks test was used to compare differences between groups matching means over the course of treatment. A difference between group means was considered significant at a level of p < 0.05.

Histology

Lung and adrenal tissues from mice were fixed with 10% neutral buffered formalin and embedded together in paraffin blocks. Slides were made from each block and stained with hematoxylin and eosin (HE) for histological analysis.

Results

Lung Granulomatous Inflammation Induced by TDM

The induction and resolution of lung granulomas following a single intravenous injection of TDM coated onto styrene divinylbenzene beads was characterized using morphologic parameters of inflammation in both histological responder C57BL/6 and nonresponder A/J mice. A diffuse alveolitis with small patches of inflammatory cells including macrophages, neutrophils and lymphocytes was apparent in responding TDM bead-treated mice by day 3. By day 5, more focal, compact, granulomatous responses were seen throughout the lung (fig. 1A). Higher magnification depicted the accumulation of activated monocytic cells (epithelioid cells) with large amounts of cytoplasm and pale nuclei centered within the inflammatory response (fig. 1C). This response was contained with no evidence of pneumonitis; a minimal thickening of the alveolar walls was restricted to regions immediately surrounding the response. Immunoreactions in the C57BL/6 mice appeared to be histologically resolved by day 14 with recovery of normal lung architecture. In contrast, A/J mice and control mice treated with noncoated beads demonstrated little or no inflammatory reactions histologically over the course of the experiment. The reaction in A/J mice was indistinguishable from that in bead-treated controls (fig. 1B, D).

Fig. 1.

Histopathology following administration of TDM. A single intravenous injection of TDM produced focal development of discrete inflammatory lesions that peaked in size and number on day 5 in C57BL/6 mice, but not in A/J mice. A A granuloma is shown located near a bronchiole in a C57BL/6 mouse. C Higher magnification of a similar response in a C57BL/6 mouse. B, D A/J mice treated similarly demonstrated no evidence of a granulomatous response. Uncoated beads produced minimal inflammatory reaction indistinguishable from that shown for the A/J mice. The lung sections were stained with HE, as described in Materials and Methods, and photographed at a magnification of × 40 (A, B) and × 200 (C, D).

Decreased Serum Cortisol following TDM Bead Administration

Serum cortisol content was monitored following administration of TDM beads in both histologic responder C57BL/6 and nonresponder A/J mice (fig. 2). C57BL/6 mice demonstrated an immediate and marked decline in serum cortisol after exposure to TDM. One day after injection, serum cortisol levels had decreased by nearly 50% to 1.66 ± 0.81 μg/dl. Levels of serum cortisol were significantly reduced on days 3 and 5 (p ≤ 0.002) compared to non-TDM-treated mice, with the reduction continuing through to day 5 (0.85 ± 0.26 μg/dl) and correlating with peak granulomatous responses. Cortisol levels then increased, beginning on day 7 (2.77 ± 1.00 μg/dl) and continuing through to resolution of the response, as measured on day 14 (4.51 ± 0.24 μg/dl). Treatment with uncoated beads led to an acute decrease in serum cortisol 1 day after injection which was not significantly different from that in noninjected control mice. This level increased thereafter, with serum cortisol levels through to day 10 remaining elevated compared to both noninjected and TDM-injected control mice. Similar to that demonstrated in the TDM bead group, administration of the uncoated beads also led to serum cortisol increases at later times, albeit in the absence of pathology.

Fig. 2.

Relative serum cortisol levels following TDM bead administration. C57BL/6 (solid symbols) and A/J (open symbols) mice were bled on days 0, 1, 3, 5, 7, 10 or 14 after intravenous injection with 10⁵ TDM beads (left panel), or through to day 10 using 10⁵ uncoated beads (right panel). Cortisol was measured by fluorescence polarization immunoassay. Assay sensitivity was 0.6 μg/dl (95% confidence). Data are expressed as mean values for 3 or 4 mice with standard errors shown.

The response seen in the histologically nonresponsive A/J mice was markedly distinct from that in the C57BL/6 mice (fig. 2). In the A/J mice, serum cortisol levels remained relatively constant throughout the first 3 days (4.85 ± 1.26 and 4.64 ± 0.65 μg/dl for days 1 and 3, respectively). On day 5, a dramatic reduction was apparent, with serum cortisol levels lowered significantly (1.70 ± 0.24 μg/dl, p ≤ 0.002) to roughly 30% of constitutive control levels. The drop in serum cortisol was short-lived; by day 7 after TDM administration, serum cortisol had increased to levels that were not significantly different from either control levels or those seen on days 1 and 3 after TDM administration. Treatment with uncoated beads led to an acute reduction in serum cortisol only on day 1 after injection, with levels rising thereafter. This initial decrease was insignificant compared to non-bead-injected controls; levels recovered by 3 days after administration of uncoated beads.

An evaluation of serum cortisol levels between mouse strains was also performed (table 2). Constitutive serum cortisol prior to TDM bead injection was significantly higher in A/J mice (5.57 ± 0.99 μg/dl) than in C57BL/6 mice (3.23 ± 0.64 μg/dl) (p < 0.009). Decreases in cortisol occurred following TDM bead administration in both sets of mice; however, absolute levels of serum cortisol in the A/J mice remained statistically higher than the levels in C57BL/6 mice throughout the course of experimentation (Wilcoxon ranks test, p ≤ 0.05). Direct comparison of cortisol levels at each time point indicated significant differences between mice throughout the first 5 days of experimentation (p ≤ 0.02). Variation in the recovery response at the later time points was evident, with average cortisol levels approaching significant differences during the resolution phase of the response. Comparing groups, the differences between patterns of response in the C57BL/6 and A/J mice were significant for both the TDM bead and uncoated bead treatments (p ≤ 0.04).

Table 2.

Serum cortisol (μg/dl) in C57BL/6 and A/J mice after TDM injection

	C57BL/6	A/J	p value
Day 0	3.23 (0.64)	5.57 (0.99)	0.009
Day 1	1.66 (0.81)	4.85 (1.26)	0.008
Day 3	1.45 (0.50)	4.64 (0.65)	0.003
Day 5	0.85 (0.26)	1.70 (0.24)	0.014
Day 7	2.77 (1.00)	3.76 (1.57)	0.158

p values determined by Student’s t test between mouse strains. Data expressed as mean values with standard errors.

Pulmonary Inflammation Was Reduced in Mice Treated with Hydrocortisone

Hydrocortisone treatment had a dramatic effect in reducing inflammatory responses in the TDM bead-treated C57BL/6 mice. Lung sections from mice administered TDM beads were examined on days 3 and 5 after treatment with hydrocortisone. Histologic examination revealed markedly reduced pulmonary pathology, with only minor indications of initiation of a granulomatous response. From all tissue examined on day 3, only one minor inflammatory focus was identified (fig. 3A). High magnification revealed accumulation of intraalveolar monocytes surrounding a nidus of TDM-coated beads (fig. 3B). Examination of tissue on day 5 demonstrated a complete absence of inflammation and no evidence of accumulation of TDM-coated beads (fig. 3C).

Fig. 3.

Lung histopathology in hydrocortisone-treated mice. C57BL/6 mice treated with hydrocortisone failed to develop granulomatous responses to TDM beads. A Three days after intravenous injection of TDM in hydrocortisone-treated mice, there was only a minor presence of inflammatory cells (arrow). B Higher magnification of inflammatory foci reveals the presence of beads surrounded by relatively few monocytes (arrow). C By 5 days after treatment, no inflammation is apparent. The lung sections were stained with HE and photographed at a magnification of × 40 (A, C) and × 400 (B).

Hyperplastic Suprarenal Gland in Histologically Nonresponsive Mice

A possible explanation for high constitutive serum cortisol in the A/J mice was found upon histological examination of the suprarenal gland. There was a high level of vacuolization evident in the medullary region (fig. 4), most likely due to packaging events for transport and release of molecules produced in both the medullary and cortex regions. In contrast, C57BL/6 mice did not show hyperplastic morphology in any region of the suprarenal gland. High magnification of cells within the cortex region revealed activated cells in the zona fasciculata and zona reticularis apparent in the A/J mice (fig. 5), evidenced by increased cytoplasm in cells of these regions. There was no evidence of hyperplasticity apparent in the zona fasciculata and zona reticularis of the C57BL/6 mice. There were no readily apparent histological differences between mouse strains in the zona glomerulosa of the suprarenal gland. Examination of adrenal glands from A/J or C57BL/6 mice on day 5 after TDM administration did not show a discernible reduction in the hyperplastic state in the cortex or medullary regions (not shown).

Fig. 4.

Comparison of suprarenal glands in the C57BL/6 and A/J mice. A comparison of the adrenal glands of C57BL/6 (left panel) and A/J (right panel) mice demonstrates a hyperplastic state within the zona fasciculata and zona reticularis regions (responsible for glucocorticoid production) in the A/J mice. High vascularization within the medulla is also evident in the A/J compared to the C57BL/6 mice. HE. × 100.

Fig. 5.

Hyperplastic suprarenal gland in A/J mice. Expanded view of the adrenal cortex region in C57BL/6 (left panel) and A/J mice (right panel). Zone regions responsible for glucocorticoid production demonstrate a hyperplastic state in the A/J mice. Z.gl. = Zona glomerulosa; Z.fas = zona fasciculata; Z.ret = zona reticularis. HE. × 400.

TDM Increases Inflammatory and Cytokine mRNA

The increased pulmonary inflammation and granulomatous response in the C57BL/6 mice warranted further examination of molecular events. Changes in mRNA for inflammatory mediators within the lung were examined (fig. 6). Messages for IL-6, TNF-α and IL-1β were significantly elevated by 1 day after TDM administration, with peak levels during the initiation phase of the response. IL-6 mRNA was increased nearly 6-fold on days 1 and 3 (5.8 ± 0.65- and 5.4 ± 0.81-fold, respectively), preceding peak granuloma size on day 6 (3.3 ± 0.45-fold). TNF-α mRNA was elevated 9-fold on day 1 and remained elevated on days 3 and 6 (6.7 ± 2.95- and 5.4 ± 0.38-fold, respectively). IL-1β remained elevated between 2.8- and 3.4-fold through to day 6. All three of these mRNAs returned to constitutive levels by the time granulomatous responses had resolved (day 14). In contrast, no apparent change in IFN-γ was detected throughout the course of the granulomatous response and resolution. There was also no significant change in IL-2, IL-4, IL-10 or IL-12(p40) message levels compared to noninjected or bead-alone control mice (not shown).

Fig. 6.

Increased inflammatory message in C57BL/6 lung tissue. Expression of mRNA in the lungs of C57BL/6 mice following intravenous injection with 10⁵ TDM beads. Message was evaluated by quantitative RT-PCR and bioluminescent hybridization immunoassay. Samples of cDNA from lung on days 0, 1, 3, 6 and 14 were amplified individually using primers specific for IL-6, IL-1β, TNF-α and IFN-γ. Product was quantitated as β-actin-normalized relative light unit values. No message was detected for IL-2, IL-4, IL-10 or IL-12p40. The -fold increase was compared with values obtained for noninjected lungs. Data are expressed as mean values for up to 10 mice with standard errors shown. * p ≤ 0.05, ** p ≤ 0.01 within groups (t test).

Histologic nonresponder A/J mice exhibit a distinct mRNA profile in response to TDM. RT-PCR analysis examined changes in IL-6, TNF-α, IL-1β and IFN-γ mRNA levels in the A/J mice following administration of TDM beads (fig. 7). Elevation of IL-6 mRNA levels occurred, but was delayed until days 3 and 5; significant changes (p ≤ 0.02) were seen only on day 3, when a 3.4-fold increase in message was apparent. Marked differences from the responding mice were identified in examination of the other inflammatory cytokines. IL-1β remained unchanged in the A/J mice throughout the time period examined, with no difference compared to control nonchallenged or noncoated bead-treated controls. The A/J mice exhibited significant increases in TNF-α mRNA through to day 5; however, the increase was maintained through to day 14 after TDM bead injection. The largest contrast between strains existed with respect to IFN-γ mRNA, which was significantly elevated in the A/J mice on days 1 and 3 (6.0 ± 2.52- and 14.0 ± 1.62-fold, respectively) after TDM administration. IFN-γ message remained elevated through to day 14; however, the high variability between animals at this time rendered the changes insignificant compared to control mice.

Fig. 7.

Increased inflammatory message in A/J lung tissue. Expression of mRNA in the lungs of A/J mice following intravenous injection with 10⁵ TDM beads. Message was evaluated by quantitative RT-PCR and bioluminescent hybridization immunoassay. Samples of cDNA from lung on days 0, 1, 3, 5 and 14 were amplified individually using primers specific for IL-6, IL-1β, TNF-α and IFN-γ. The -fold increase was compared with values obtained for noninjected lungs. Data are expressed as mean values for up to 10 mice with standard errors shown. * p ≤ 0.05, ** p ≤ 0.02 within groups (t test).

Increased Proinflammatory Response of BMM to TDM

The changes in immunopathology in the C57BL/6 mice were consistent with an influx of mononuclear cells due to chemotactic molecules and elicitation of proinflammatory responses. We wished to further examine the relative role of naïve macrophages from C57BL/6 and A/J mice in inducing proinflammatory cytokines to TDM. BMM were incubated with TDM-coated beads, and culture supernatants were evaluated for TNF-α, IL-1β and IL-6 protein at various intervals after incubation with TDM-coated beads (fig. 8). IL-6, IL-1β and TNF-α were all produced in high quantities in the in vitro cultures by C57BL/6 BMM, with responses significantly higher than responses to uncoated beads by 24 h (p ≤ 0.05). In contrast, A/J BMM were significantly impaired (p ≤ 0.05) in the production of IL-6 and TNF-α through to 72 h after incubation compared to the C57BL/6-derived cells. Both the C57BL/6 and A/J cells produced IL-1β in response to TDM beads; however, the C57BL/6 BMM had increasing responses at 72 h that were significantly elevated over those of the A/J cells. Stimulation of BMM with TDM-coated or uncoated beads did not induce IFN-γ protein expression at any time point examined (not shown).

Fig. 8.

Production of proinflammatory cytokines by BMM in response to TDM. C57BL/6 (solid bars) and A/J (open bars) BMM were treated with TDM-coated beads, and IL-6 (A), IL-1β (B) and TNF-α (C) levels were assessed in culture supernatants through to 72 h. C57BL/6 BMM produced IL-6 and TNF-α at levels significantly higher than the A/J BMM by 4 h. IL-1β levels were significantly higher in C57BL/6 BMM compared to A/J BMM by 72 h after incubation, Background levels for uncoated beads were highest at 24 h for TNF-α (60 pg/ml) and at 72 h for IL-6 and IL-1β (12 and 7 pg/ml, respectively). * p ≤ 0.05.

Discussion

Although the observation of glucocorticoid modulation during tuberculosis infection has been known for many years, there has been insufficient investigation of specific mycobacterial components which may cause this response. The identification of a mycobacterial glycolipid having the property of influencing (decreasing) serum cortisol levels is unique and may aid our understanding of immunopathological changes during MTB-related disease. In these experiments, granulomatous responses and full immune responsiveness to TDM were only seen upon a sufficient decrease in relative levels of systemic glucocorticoids.

Relative amounts of cortisol may be crucial in susceptibility to tuberculosis and may play a defining role in ensuing pathology [29]. We established a model to further study the role of cortisol during tuberculosis, utilizing mouse strains with natural discrepancies in both constitutive serum cortisol and granulomatous responsiveness. Increased cytokine mRNA in the lung during infection has been shown to be correlated with decreased serum cortisol levels [6]. However, it was not clear whether the decreased cortisol was due to (1) the host response to the infectious event, (2) the host response to a mycobacteria-derived factor or (3) the response due to a combination of both. In part, the initial change in serum cortisol may be influenced by the presence of beads in lung tissue, as uncoated beads demonstrated an insignificant, transient drop in cortisol. Of interest, the recovery of serum cortisol levels in uncoated bead-administered groups led to a rebound effect with elevated levels at later times, perhaps reflecting artifactual occurrences due to clearance of beads through other tissues. However, no additional changes were noted in adrenal glands following administration of coated or uncoated beads [unpubl. observation].

A/J and C57BL/6 mice differ naturally with respect to the distribution of endogenous glucocorticoids and serum cortisol levels. We confirm that both phenotypic and histological differences exist in adrenal glands from these mice [30, 31]. Tanaka et al. [31] also observed high vacuolization within the medullary region in A/J mice, consistent with high levels of molecular packaging events for transport of both medullary and cortex region-derived products. The biological significance of 75% greater constitutive serum cortisol in A/J mice compared to C57BL/6 mice remains unknown. Thaete et al. [30] suggested that these differences are consistent with the hypothesis that A/J mice are relatively deficient in the prophylactic activities of endogenous glucocorticoids. Indeed, measurements of serum cortisol may not actively represent levels in the lung; the ratio of cortisol to cortisone may be more revealing [32]. Recently, Rook and Hernandez-Pando [33] and Chomarat et al. [34] suggested that local cytokines may affect the concentration of tissue cortisol via an enzymatic ‘shuttle’ that can deactivate cortisol by converting it to cortisone, or activate cortisone by converting it to cortisol. If true, the local cortisol concentration during granulomatous reactions may be independent of circulating cortisol levels. Yet, it seems likely that elevated systemic cortisol would exhibit an immunomodulatory effect within tissue.

More certain is that the relative level of T helper 1 (Th1) and T helper 2 (Th2) cellular responses during MTB infection is critical with respect to the protective response against MTB disease and related pathology. The balance of cytokine evoked during pathogen-induced inflammation may be affected by cortisol. T cells maturing in the presence of elevated glucocorticoids exhibit a preferential Th2 cytokine profile [35–38]. Experiments using A/J mice are consistent with the notion that elevated constitutive serum cortisol influences the development of Th2 responses, most likely through induced IL-10 in the absence of strong initial IL-12 [6] leading to a change in the ratio of cytokines during infection. Complicating this issue is the observation of development of a strong IFN-γ response in the A/J mice following TDM bead administration. At this time, the source of the IFN-γ remains unidentified; monocytes do not appear to be responsible for IFN-γ production. One possible explanation may be that even a low Th2 response can exacerbate destructive pathology in experimental tuberculosis [33], despite the presence of an adequate Th1 response.

Many properties of mycobacterial cord factor TDM are dependent upon the physical state of the molecule as presented to biological systems [14, 39, 40]. The experiments described here suggest an additional property of TDM in its ability to directly influence local environments in vivo through reduction of serum glucocorticoids. If cortisol levels are sufficiently lowered for the duration of the response, granulomatous responses may ensue, as seen for the C57BL/6 mice. If the decrease in serum cortisol is lessened in duration and magnitude, as in the A/J mice, sufficient pressure remains to limit immunological activation. It would therefore follow that in A/J mice, T cells responding to MTB and cord factor are more likely to mature within the presence of elevated glucocorticoids, and would exhibit a preferential Th2 cytokine profile [35, 36]. This fits the pattern observed during A/J infection.

The present model is limited to the examination of acute responses to TDM in the absence of sensitization. TDM by itself does not induce IL-2 and IL-4 or IL-10 upon acute, primary in vivo stimulation. Secondary responses to TDM are different [41, 42]. For example, Lima et al. [42] recently demonstrated that TDM-coated microparticles with a slow release time (60 days) elicited high levels of TNF-α and IL-6 in bulk lung cell homogenates; production of IL-4, IL-10, IL-12 and IFN-γ was also described, signifying factors present in secondary, and not primary, responses.

In light of the requirement for TNF-α in granuloma formation, it was puzzling to see increased TNF-α in the A/J mice, which did not produce focal granulomatous responses. TNF-α and IL-1β are important inflammatory mediators for granuloma development and maintenance [43], especially during tuberculosis [44]. The high cortisol levels in A/J mice do not completely inhibit the immune response; rather, inhibition is only seen for granuloma formation, which is a very complex event. High TNF-α levels in A/J mice have been reported following TDM administration [15], as well as during infection [11, 12]. TNF-α is clearly required for protection against MTB [44, 45], as well as for granuloma development. Perhaps, in this instance, TNF-α acts more in a destructive manner, contributing to necrotic lesions during infection [29]. However, in terms of TDM-induced granuloma, there may not be sufficient quantities of secondary signals to mount a true deleterious response. Indeed, the TNF-α produced may be inactive or adsorbed by soluble receptors, as is reported for MTB-infected alveolar macrophages that release soluble TNF-α receptor 2 [46].

There is increasing evidence that cortisol can inhibit production of monocytic chemotactic factors and proinflammatory mediators [47, 48]. More than 25 years ago, TDM was identified as an elicitor of chemotactic responses from leukocytes [49]. Recent investigations have defined a role for TDM in recruitment of cells and modulation of production of cytokines in tuberculosis [42]. TDM microparticles demonstrated increased attraction of mononuclear cells to the lung with little increase in polymorphonuclear cells. This is consistent with our personal observations that the changes in immunopathology in C57BL/6 mice were consistent with an influx of mononuclear cells due to chemotactic molecules. Indeed, MCP-1, MIP-1α, MIP-2 and IP-10 mRNA levels increased significantly (2- to 4-fold) by day 1 after injection, with chemokine mRNA remaining elevated through initiation of granuloma development [unpubl. observation]. The in vitro experiments using BMM in the absence of cortisol indicate that innate components moderate responses to TDM. Significant differences in production of IL-6, IL-1β and TNF-α were identified between the C57BL/6 and A/J BMM after incubation with TDM beads. Specifically, monocytes can be activated by IL-6, leading to the secretion of CC chemokines, which can be inhibited in vitro using hydrocortisone [47]. One hypothesis is that a TDM-induced decrease of cortisol in vivo would allow responding monocytes to upregulate production of chemokines, thus initiating an influx of cells necessary for maintaining granulomatous responses.

The underlying basis for A/J genetic histologic nonresponsiveness to TDM is likely to be due only in part to glucocorticoids, as other factors in these mice play a role in responses during infection [21, 50]. The findings reported here indicate a need to further explore the influence of cortisol on initial events involved in immunoregulation and immunopathology during acute tuberculosis infection, and the contribution of cord factor to this process. As such, the experiments described here form the basis of a model to study specific factors and their relationship to glucocorticoid involvement in tuberculosis.