Requisite Role for Complement C5 and the C5a Receptor in Granulomatous Response to Mycobacterial Glycolipid Trehalose 6,6′-Dimycolate

Authors


Dr Jeffrey K. Actor, PhD, Department of Pathology and Laboratory Medicine, MSB 2.214, University of Texas-Houston Medical School, 6431 Fannin, Houston, TX 77030, USA. E-mail: jeffrey.k.actor@uth.tmc.edu

Abstract

The development of pulmonary granulomatous lesions during mycobacterial infection is a complex phenomenon, in part caused by responses elicited towards the surface glycolipid trehalose 6,6′-dimycolate (TDM; cord factor). The molecular mechanisms underlying granuloma formation following challenge with TDM are not yet completely understood. The present study defines pathologic differences in acute response to Mycobacterium tuberculosis TDM in C57BL/6 mice and mice lacking the C5a receptor (C5aR–/–). Mice were intravenously injected with TDM prepared in water-in-oil-in-water emulsion and examined for histologic response and changes in proinflammatory cytokines and chemokines in lung tissue. Control C5a receptor-sufficient mice demonstrated a granulomatous response that peaked between days 4 and 7. Increased production of macrophage inflammatory protein-1 alpha (MIP-1α), interleukin-1β (IL-1β) and CXC chemokine KC (CXCL1) correlated with development of granulomas, along with modest change in tumor necrosis factor-alpha (TNF-α). In contrast, the C5aR–/– mice revealed markedly exacerbated inflammatory response. The receptor-deficient mice also demonstrated a lack of coherent granulomatous response, with severe oedema present and instances of lymphocytic cuffing around pulmonary vessels. Lung weight index was increased in the C5aR–/– mice, correlating with increased MIP-1α, KC, IL-1β and TNF-α over that identified in the congenic C5aR-sufficient controls. Correlate experiments performed in C5-deficient (B10.D2-H2d H2-T18c Hco/oSnJ) mice revealed similar results, leading to the conclusion that C5 plays a significant role in mediation of chemotactic and activation events that are the basis for maturation of granulomatous responses to TDM.

Introduction

Approximately, 3 million deaths worldwide each year are attributed to tuberculosis, the causative agent of which is Mycobacterium tuberculosis (MTB) [1]. Manifestations of disease in humans range from a classic granulomatous response with containment of organisms, to a more rapid bacterial dissemination with accompanying tissue necrosis [2]. The mechanisms underlying protective granuloma formation in the lung and control of bacterial growth are not yet completely understood, however, recent studies demonstrate an increasing importance for complement in this process. Specifically, C5-deficient A/J mice demonstrated severe pulmonary pathology, disseminated infection and early mortality as compared with C5-sufficient C57BL/6 mice [3].

C5 deficiency has been correlated to enhanced susceptibility to a variety of infectious diseases [4–6]. C5 deficiency as a possible explanation for the attenuated pathology following MTB infection was further examined in congenic C5-deficient mice [7]. C5 deficiency increased the susceptibility of mice to MTB infection with evidence for C5 function in granuloma formation, organism containment and regulation of cellular distribution within the lung, upon organism challenge [7, 8]. While A/J mice demonstrated differences in both chemokines and cytokines, the primary discrepancy identified in the C5-deficient OSN mice was attributed to defects in chemotactic factors [8].

Trehalose-6,6′-dimycolate (TDM), or ‘cord factor’, is a mycobacterial glycolipid cell wall component that is implicated in pathogenesis of tuberculosis infection. TDM is a potent immune adjuvant with inflammatory and immunoregulatory properties that directly affect macrophage response [9, 10], and has been used in models of granulomatous responses which mimic pathology occurring during initiation of acute MTB infection [11–13].

It is unknown whether C5 plays a significant role in mediation of chemotactic and activation events that are the basis for initiation and maintenance of granulomatous response to isolated TDM, or in the regulation of the inflammatory response. Therefore, mice deficient in C5a receptor (C5aR–/– mice) were examined with respect to histological disturbances and manifestations of proinflammatory mediators and chemotactic factors following TDM administration. Results were compared with mice deficient in C5a to determine the role of anaphylactic factors relative to direct mediation via receptor ligand interaction.

Methods

Animals.  C5a receptor-deficient (C5aR–/–) mice were generated by targeted deletion of the murine C5aR gene and determined to be completely C5aR deficient by polymerase chain reaction, Northern blot and immunohistochemistry analyses [14]. C5aR-deficient animals were backcrossed with C57BL/6J mice. Heterozygous C5aR+/– backcrossed mice were interbred, and the resulting F2 C5aR+/+ and C5aR–/– littermates were used for studies [15]. Congenic complement C5-sufficient strain (B10.D2-H2d H2-T18c Hcl/nSnJ; NSN) and complement C5-deficient strain (B10.D2-H2d H2-T18c Hco/oSnJ; OSN) were obtained from Jackson Laboratories (Bar Harbor, ME, USA). The C5-deficient inbred strains produce no C5, because they carry a null hemolytic component allele Hc [16–18]. The two-base pair deletion which causes C5 deficiency [19] and restriction fragment length variants associated with the null Hc allele have been localized within the exon-intron structure of the gene [18, 20]. A DNA polymorphism for C5 segregates in complete concordance with Hc variants. Animal studies were conducted under approval of the UTHHSC Institutional Review Board, document AWP A3413-01. Three to six mice were used per time point, per group. Experiments were repeated up to three times per time point.

TDM challenge and processing.  A 100 µl sample of TDM emulsion was prepared for injection into each mouse as detailed, with modifications as described [13, 21, 22]. About 100 µg of purified TDM was dissolved in 3.2 µl of Freund's incomplete adjuvant (FIA) in a Teflon grinder. After addition of 3.2 µl of 0.1 m phosphate-buffered saline (PBS), a water-in-oil emulsion was made. Then, 93.6 µl of saline containing 0.2% Tween 80 was added to the final concentration (3.2%) of FIA, and a water-in-oil-in-water emulsion was made by mixing. Each 100 µl sample was injected into the tail vein of the mice. All mice were sacrificed at indicated times post TDM challenge. Lung tissue was aseptically removed and processed. Approximately, 30 µg of processed lung was placed in 2 ml media (Dulbecco's modified eagle medium, 5% fetal calf serum heat inactivated). Supernatants were aliquoted after 4 h incubation at 37 °C, and samples frozen until assessed for proinflammatory mediators IL-1β and TNF-α, or chemokines MIP-1α and CXC chemokine KC (CXCL1), by standard ELISA. Alternatively, lung samples were fixed with 10% neutral buffered formalin and embedded in paraffin blocks for further histological examination.

Histological sections and staining.  Organs were collected from replicate mice in 10% buffered formalin, embedded in paraffin and sections were stained using hematoxylin and eosin (H&E). Visual examination of lung tissue from noninfected C5aR–/– mice was identical in all histologic respects to noninfected congenic controls, with no evidence of pulmonary interstitial inflammation in either group prior to infection. The same held true for OSN and NSN mice.

ELISA analysis of lung chemokines and cytokines.  Paired antibodies for capture of cytokines and Quantikine kits for chemokines were from R&D systems (San Diego, CA, USA). Costar 96-well vinyl flat-bottom well plates were coated with capture antibody overnight in PBS as recommended by the manufacturer. Plates were washed three times with wash buffer (0.05% Tween-20 in PBS) and treated with blocking buffer (1% BSA, 5% sucrose, 0.05% NaN3 in PBS) for 1 h. After three washings, 50 µl cell supernatant plus 50 µl PBS were added per well (accomplished in triplicate). Plates were incubated overnight at 4–8 °C. Biotin-conjugated secondary antibodies were added after washing. 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, USA). Absorbance was read at 570 nm and 450 nm on an ELISA plate reader (Molecular Devices, Sunnyvale, CA, USA). The mean of triplicate wells was calculated based on a standard curve constructed for each assay, using recombinant protein from kit manufacturer (R&D Systems). Constitutive expression for noninfected controls was less than 30 pg/mg for chemokines and cytokines.

Statistics.  Results were calculated for presentation at level of cytokine per mg of lung tissue processed. Groups were determined as statistically significant using the student's t-test, with values for significance set at P < 0.05, with up to six mice per time point examined. Data shown is representative of results obtained from individual experiments; experiments were repeated three times with similar results obtained.

Results

Dysregulated inflammatory mediation during initiation of granulomatous response in C5aR–/– mice

C5AR–/– mice demonstrate exacerbated inflammation to TDM challenge

Lung weight indices (LWI) for C5aR–/– mice and C5aR+/+ congenic controls were determined after administration of TDM as an indirect measure of inflammatory response (Fig. 1). Control C5aR-sufficient mice demonstrated increased LWI by 4 days post TDM administration, with significantly increased lung weights over noninjected control mice at 7 and 14 days post injection (P < 0.05). The C5aR–/– mice demonstrated exacerbated inflammation, with significantly increased LWI apparent by 4 days post injection relative to the injected sufficient controls (1.51 ± 0.2; P < 0.05). The C5aR–/– mice maintained high-LWI values (1.80 ± 0.2 on day 7, 1.72 ± 0.3 on day 14) throughout the course of experimentation. Overall, the inflammatory response was significantly greater in the C5aR–/– mice compared to the sufficient control mice at days 4 and 7 post TDM challenge (P < 0.05), with the inflammatory response similar only at the latest time examined.

Figure 1.

Enhanced trehalose 6,6′-dimycolate (TDM)-induced lung inflammation in C5aR-deficient mice. Mice were i.v. challenged with TDM and killed at 4, 7 and 14 days later. Pulmonary inflammation was calculated as a lung weight index (LWI) for both C5aR–/– (□) and C5aR+/+ littermate controls (▪). C5aR–/– mice demonstrated significantly increased LWI at days 4 and 7 compared with C5aR+/+ mice (*, P < 0.05); error bars represent standard deviation from mean values.

Non-focal monocytic infiltration and histological abnormalities in lungs of C5aR–/– mice after TDM challenge

Manifestation of increased lung weights was examined histologically post TDM administration (Fig. 2). In general, the pathology observed correlated well with the LWI data described. Small, focal histiocytic clusters comprised primarily of monocytes were evident by 4 days post TDM challenge in the C5aR-sufficient control mice. By 7 days post TDM administration, the clusters were more complex, being greater in frequency and size. Lymphocytic infiltration ensued; at 14 days post administration, there were numerous well defined granulomas with abundant activated (foamy) macrophages ringed by lymphocytes. These granulomas were situated within the parenchyma and located away from endothelial-comprised vessels (blood vessels or lymphatics). In contrast, the C5aR–/– mice demonstrated marked nonfocal monocytic response by 4 days post TDM administration. By 7 days, there was little evidence for cohesive cellular aggregates; rather monocytic and lymphocytic infiltrates into surrounding parenchyma were increased along with evidence for oedema. Day 14 revealed increased inflammatory response, with little unaffected parenchyma remaining. At this time, widely distributed foamy monocytes were discernible. Multiple incidents of lymphocytic cuffing circumscribing occluded vesicles were evident, with an overall failure in lymphocytic distribution and migration to remaining internal parenchyma.

Figure 2.

Nonproductive trehalose 6,6′-dimycolate (TDM)-induced granulomatous response in C5aR-deficient mice. Histopathology of C5aR–/– and C5aR+/+ mice challenged with TDM. Lungs of C5aR+/+ mice (top) show relatively well preserved pulmonary parenchyma with TDM-induced non-nectrotizing granulomas apparent by 4 days post administration. Granulomas increase in size and number through the course of experimentation. In contrast, C5aR–/– mice (bottom) demonstrate an aggressive early inflammation, extensive oedema and lymphocytic infiltration. Perivascular cuffing is present by day 7, and is prominent by day 14, with relatively little intact parenchyma remaining. (H&E staining, 40×, bar indicates 600 µm).

Elevated chemokines MIP-1α and KC in response to TDM challenge in C5aR–/– mice

A significant increase in chemokines MIP-1α and KC produced within lungs was demonstrated for both C5aR+/+ and C5aR–/– strains relative to nontreated controls following administration of TDM (Fig. 3). The elevated chemokines were maintained throughout the first week examined. Comparison between strains revealed that the C5aR–/– mice demonstrated significantly increased MIP-1α over that produced by the C5aR+/+ mice. At 4 days post treatment, the C5aR–/– mice produced 36.5 ± 5.8 pg/mg tissue, whereas the C5aR+/+ mice produced 9.6 ± 2.6 pg/mg. Similar differences were identified at day 7. The levels of KC produced were also significantly elevated for both strains relative to nontreated controls, however, the variability between mice was larger. The levels of KC in C5aR–/– mice increased to 70.2 ± 36.9 pg/ml on day 4, with similarly elevated levels present on day 7. The C5aR+/+ mice also produced KC, with levels generally decreased compared to that identified in the knockout mice.

Figure 3.

Increased chemokine production in C5aR-deficient mice in response to trehalose 6,6′-dimycolate (TDM) challenge. C5aR–/– mice (□) exhibit enhanced production of macrophage inflammatory protein-1 alpha (MIP-1α) (left) and KC (right) compared to C5aR+/+ mice (▪) at 4 and 7 days post TDM challenge. Results represented in pg/mg lung tissue are representative of data obtained from three separate experiments; differences of significance between groups indicated (*P < 0.05); error bars represent standard deviation from mean values.

Exacerbated proinflammatory response to TDM challenge in C5aR–/– mice

Mediators TNF-α and IL-1β involved in granulomatous inflammatory response were monitored post challenge (Fig. 4). TNF-α exhibited relatively minor change in the C5aR+/+ control mice with limited but detectable production in lungs following TDM challenge. In contrast, the C5aR–/– mice demonstrated marked and significant TNF-α production by 4 days post TDM administration (57.4 ± 10.4 pg/mg), with maintained production through day 7 post challenge (P < 0.05). Both groups of mice demonstrated strong IL-1β response to TDM at both 4 and 7 days post administration. IL-1β produced from lungs of C5aR–/– mice was nearly 115 pg/mg on day 4 and 7; the C5aR+/+ mice also showed substantial increase in IL-1β, however, the amounts produced were significantly (P < 0.05) below the knockout mice at days 4 and 7.

Figure 4.

Increased proinflammatory response in C5aR-deficient mice in response to trehalose 6,6′-dimycolate (TDM) challenge. C5aR–/– mice (□) exhibit enhanced production of tumour necrosis factor-α (TNF-α) (left) and interleukin-1β (IL-1β) (right) compared to C5aR+/+mice (▪) at 4 and 7 days post TDM challenge. Results represented in pg/mg lung tissue are representative of data obtained from three separate experiments; differences of significance between groups indicated (*P < 0.05); error bars represent standard deviation from mean values.

Requirement for complement C5 in response to TDM confirmed in C5-deficient mice

To determine whether C5 has a role in mediation of TDM granulomatous response distinct from C5a/C5a-receptor-ligand interactions, C5-sufficient mice (NSN; B10.D2-H2d H2-T18c Hcl/nSnJ) and complement C5-deficient (OSN; B10.D2-H2d H2-T18c Hco/oSnJ) mice were injected with TDM emulsion and killed 1 week later. Lung weights, pathology, inflammatory cytokines and chemokines were measured. Direct comparisons were made to receptor-sufficient and -deficient mice.

The LWI for the C5-sufficient NSN mice were significantly (P < 0.05) elevated (1.234 ± 0.25) compared to the noninjected controls (0.724 ± 0.03) (Table 1), similar to that previously observed for the C5aR+/+ mice. In contrast, both the OSN and the C5aR–/– mice exhibited increased inflammatory responses compared to the NSN noninjected control group. The OSN mice were significantly elevated (P < 0.05) compared with the NSN TDM challenged group. The C5-deficient OSN mice demonstrated responses slightly higher than that identified in the C5aR–/– group.

Table 1.  Lung weight index (LWI) comparison for C5 sufficient (NSN), C5 deficient (OSN), and C5-receptor knockout (C5aR–/–) mice following administration of trehalose 6,6′-dimycolate (TDM)
Mouse strainTreatmentLWI (average ± StDev)
  • StDev, standard deviation.

  • *

    P < 0.05 versus noninjected mice.

  • P < 0.05 versus NSN injected mice at day 7.

  • n, three to six mice per group.

NSNNone0.724 ± 0.034
C5aR–/–None0.879 ± 0.048
NSNTDM, day 71.234 ± 0.279*
OSNTDM, day 71.737 ± 0.065*
C5aR–/–TDM, day 71.546 ± 0.214*

C5 deficient and C5aR–/– exhibit identical lung pathology in response to TDM

C5-sufficient NSN and C5-deficient OSN mice were examined histologically for development of lung granulomatous response. The NSN control mice developed well contained granulomas similar to that identified in the C5aR+/+ mice, with focal areas of monocytic inflammation surrounded by relatively normal parenchyma. However, elimination of C5 or C5a receptor led to markedly different pathology. The histological pattern in the OSN mice was indistinguishable from the C5aR–/– mice, with high incidence of pneumonitis on day 7 characterized by heavy mononuclear infiltrates, oedema and vascular occlusion (Fig. 5). The inflammatory response in the OSN and C5aR–/– mice was in concert with the increased lung weights. Both deficient strains also demonstrated high levels of lymphocytic cuffing of vasculature with heavy accumulation of small mononuclear cells around blood vessels.

Figure 5.

Comparative histopathology of C5aR–/–, OSN (C5 deficient), C5aR+/+ and NSN (C5 sufficient) mice challenged with trehalose 6,6′-dimycolate (TDM). Seven days post challenge, C5aR+/+ mice (top left) show relatively well preserved pulmonary parenchyma and evidence of focal, non-necrotizing granulomas. Nearly identical histopathology is apparent in NSN mice (top right). The C5aR–/– mice demonstrate an aggressive inflammation, extensive oedema and lymphocytic infiltration (middle left). Similar pathology is seen in the C5-deficient OSN mice (middle right). High-power magnification reveals characteristic perivascular cuffing present in both the C5aR–/– and OSN mice (bottom left, right, respectively). This inability of lymphocytes to disseminate was unique to the C5a and C5a-receptor deficient strains. (H&E staining, 40× and 100×).

OSN mice demonstrate exacerbated cytokine and chemokine response to TDM

Seven days post TDM administration, the lungs of all TDM treated mice demonstrated significant production of chemokines and proinflammatory mediators compared to nontreated controls (Table 2). However, the deficient mice produced markedly elevated amounts of IL-1β and MIP-1α which was significantly (P < 0.05) higher than that produced by the treated control animals. Analysis of the cytokine IL-1β and the chemokine MIP-1α show similar increase for the OSN mice as identified in the C5aR–/– mice; there were no significant differences between C5-deficient group or C5a receptor-deficient group.

Table 2.  Interleukin-1β (IL-1β) and macrophage inflammatory protein-1 alpha (MIP-1α) response (pg/mg) produced in C5 sufficient (NSN), C5 deficient (OSN), and C5a-receptor knockout (C5aR–/–) mice following administration of trehalose 6,6′-dimycolate (TDM)
Mouse strainTreatmentIL-1β (pg/mg) (average ± Stdev)MIP-1α (pg/mg) (average ± Stdev)
  • StDev, standard deviation.

  • *

    P < 0.05 versus control noninjected mice.

  • P < 0.05 versus NSN injected mice on day 7.

  • n, three to six mice per group.

NSNNone1.28 ± 0.340.05 ± 0.01
C5aR+/+ controlNone0.16 ± 0.012.03 ± 0.1.97
NSNTDM, day 786.89 ± 53.5946.79 ± 16.42*
OSNTDM, day 7169.10 ± 74.6774.65 ± 28.36*
C5aR+/+ controlTDM, day 785.66 ± 79.7847.33 ± 14.87*
C5aR–/–TDM, day 7154.10 ± 25.4367.77 ± 8.35*

Discussion

In this study, a role for complement component C5 and the C5a receptor in the initiation of the granulomatous response to TDM challenge was examined. The observations in these studies would indicate that C5 and the C5a receptor are not required for initiation of TDM-induced inflammatory response, but are critical to the development of cohesive granuloma formation. In the absence of C5, or C5aR, granuloma cohesiveness does not occur resulting in a nonfocal pneumonitis with components of oedema and activated monocytic cellular infiltration. This suggests that C5a is critical in maintenance, and perhaps even resolution, of the response.

The histology identified in both the C5-deficient mice and C5aR-deficient mice challenged with TDM correlated well with initial pathology occurring after M. tuberculosis infection of C5-deficient mice [7]. MTB infection of OSN mice demonstrated disseminated infection without organism containment, while control NSN mice were unimpaired in their granulomatous response [8]. In those studies, a dysregulation in chemokine response correlated with bacterial dissemination, lack of granuloma development and overall increased bacterial loads within the lung. The similarities identified in this study support the role of TDM as a useful model to mimic aspects of MTB-induced pathology. In addition, the results of this study argue against a requisite role for C5 and the C5aR in mounting an inflammatory response. Overall, the inflammatory response, demonstrated by histology and LWI of challenged mice, was not affected by the loss of C5 or the loss of the C5a receptor. Rather, exacerbated production of proinflammatory mediators and chemotactic factors occurred. Therefore, these results suggest an additional role for C5 and C5a receptor in the regulation required for maintenance of the granulomatous response.

TDM induces production of proinflammatory cytokines, procoagulant activity and fibrinolysis in MTB susceptible mice, with a primary inflammatory response characterized by the activation of undifferentiated T cells, culminating in a pulmonary granulomatous response [11, 12]. One theory holds that a delayed-type hypersensitive (DTH) response to TDM is necessary for the maintenance of the granulomatous inflammation during mycobacterial infection [21, 22]. In fact, Seggev and colleagues [23, 24] suggested that production of pulmonary lesions by TDM was dependent only upon T lymphocytes, concluding that lymphocytes were required while antibodies and complement were not. In hindsight, those studies were focused towards observing the absence of granulomatous response, not exacerbation as identified here.

C5a and C5b upregulate adhesion molecules on macrophages, neutrophils and other inflammatory cells suggesting that expression of adhesion receptors may be a critical event in maintaining the structure of the granuloma [25–27]. Indeed, C5 may also be necessary for development of cell-mediated immunity postulated to occur via IL-12 [28, 29], although this was not directly examined in this study. Development of granulomatous response is not required for T-cell-mediated response to occur, as evidenced by studies of mice deficient in intercellular adhesion molecule-1 (ICAM-1) challenged with MTB [30]. Perivasuclar cuffing of infiltrating lymphocytes was also observed in the ICAM (interrupted) mice, similar to that shown here using TDM in OSN and C5aR–/– mice. The pathology observed in this study in both the C5aR–/– and the OSN mice does not illuminate the role of C5 or the C5a receptor in the development of T cell chemotactic events. However, the failure of C5aR–/– mice to develop granulomas or direct the infiltration of lymphocytes into the affected parenchyma suggests a key role for adhesion receptors in lymphocyte chemotactic events during granulomatous response. This could work in concert with C5's role in activation of macrophages to initiate events to mediate lymphocytes responding to released chemotactic factors [31].

The increased level of macrophage response in the C5-deficient mice and C5aR-deficient mice treated with TDM is enigmatic. One explanation could involve regulation of macrophage infiltration by neutrophils. C5a induces neutrophil production of macrophage migration inhibitory factor [32], the absence of which would fit results described here. Although C5a has been ascribed a negative regulatory role in neutrophil proinflammatory mediator production during sepsis [33], opposite effects are attributed to alveolar macrophages [34]. In the TDM model, the level of infiltrating neutrophils is thought to be low [35], with primary response ascribed to macrophages. We initially hypothesized that TNF-α would be reduced in both C5aR–/– and C5a-deficient mice, however, this did not occur and there is no clear explanation for increased TNF-α other than a parallel increases in numbers of macrophages present. Overall, this suggests that the role of neutrophils in the TDM model needs revisiting.

In conclusion, the C5a receptor-deficient mice are defective in complement component C5-mediated events that are critical for maturation of granulomatous response to TDM. During tuberculosis infection, C5 deficiency results in improper activation and immune surveillance-related events. The end result is that mycobacterial infections are not effectively contained, resulting in both exacerbation of deleterious pathology and poor control of disease. The TDM-glycolipid model may be ideal for examination of the role of complement components in primary mycobacterial infection.

Acknowledgments

We thank Chinnaswamy Jagannath, PhD for helpful discussions and data review and his laboratory personnel for technical assistance. This work was supported by NIH grants R01HL68537-01, R01HL68520-01, R01AI025011, R01HL074333 and R21HL080313–10.

Ancillary