|
|
||||||||
INFLAMMATION/IMMUNITY/MEDIATORS
in macrophage suppresses experimentally induced colitis
Laboratory of Metabolism, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, Maryland
Submitted 16 August 2006 ; accepted in final form 2 November 2006
| ABSTRACT |
|---|
|
|
|---|
(PPAR-
) has been shown to be a protective transcription factor in mouse models of inflammatory bowel disease (IBD). PPAR-
is expressed in several different cell types, and mice with a targeted disruption of the PPAR-
gene in intestinal epithelial cells demonstrated increased susceptibility to dextran sulfate sodium (DSS)-induced IBD. However, the highly selective PPAR-
ligand rosiglitazone decreased the severity of DSS-induced colitis and suppressed cytokine production in both PPAR-
intestinal specific null mice and wild-type littermates. Therefore the role of PPAR-
in different tissues and their contribution to the pathogenesis of IBD still remain unclear. Mice with a targeted disruption of PPAR-
in macrophages (PPAR-
M
) and wild-type littermates (PPAR-
F/F) were administered 2.5% DSS in drinking water to induce IBD. Typical clinical symptoms were evaluated on a daily basis, and proinflammatory cytokine analysis was performed. PPAR-
M
mice displayed an increased susceptibility to DSS-induced colitis compared with wild-type littermates, as defined by body weight loss, diarrhea, rectal bleeding score, colon length, and histology. IL-1
, CCR2, MCP-1, and inducible nitric oxide synthase mRNA levels in colons of PPAR-
M
mice treated with DSS were higher than in similarly treated PPAR-
F/F mice. The present study has identified a novel protective role for macrophage PPAR-
in the DSS-induced IBD model. The data suggest that PPAR-
regulates recruitment of macrophages to inflammatory foci in the colon.
CC chemokine receptor 2; macrophages
(PPAR-
) is a member of the nuclear receptor superfamily of transcription factors. Following activation of PPAR-
by specific ligands, the receptor binds to its obligate heterodimer partner retinoid X receptor. Ligand binding induces a conformational change resulting in the release of corepressors, binding of a distinct set of nuclear coactivators, and activation of gene transcription (47). In addition to endogenous fatty acids and their derivatives (26, 27), PPAR-
binds to a certain class of insulin-sensitizing agents known as the thiazolidinediones (TZDs) (30). Two commonly used TZDs, rosiglitazone and pioglitazone, bind with high affinity to PPAR-
and are in widespread clinical use for anti-type 2 diabetes therapy (34) PPAR-
is highly expressed in brown and white adipose tissue and is shown to be a key transcription factor in adipocyte differentiation (37, 40). Transgenic mice containing a conditional disruption of PPAR-
in adipocytes displayed a severe reduction in brown and white adipocyte cell mass (21, 23, 24).
More recently, PPAR-
was shown to be critical in inflammatory bowel disease (IBD) (1, 5, 17, 25, 31, 38, 42, 44). IBD, which manifests as either ulcerative colitis or Crohns disease, is associated with chronic inflammation of the intestinal tract. PPAR-
ligands can attenuate the severity of mouse models of IBD induced by dextran sulfate sodium (DSS) or 2,4,6-trinitrobenzenesulfonic acid (5, 17, 25, 31, 38, 42, 44). Furthermore, previous work from our laboratory has demonstrated an increase in the severity of IBD and demonstrated increased expression of TNF-
, IL-1
, and IL-6 in a mouse line where PPAR-
is specifically deleted throughout the intestinal epithelium, providing evidence for a direct role of PPAR-
in the colon mucosa (1). However, unexpectedly, the PPAR-
ligand rosiglitazone decreased the severity of DSS-induced colitis and suppressed cytokine production in both colon epithelial-specific PPAR-
null mice and littermate controls, suggesting that PPAR-
expressed in other cell types may also be of importance (1).
Activated macrophages were shown to express high levels of PPAR-
(8), and the PPAR-
agonists TZDs and 15-deoxy-delta 12,14-prostaglandin J2 were found to suppress the inflammatory response by attenuating expression of specific inflammatory mediators via a PPAR-
-dependent pathway (52). Macrophages are critical in the pathogenesis of IBD. Depletion of intestinal macrophages was shown to be protective in the IBD mouse model (33, 50). These results suggest that macrophages in the intestine are critical in the pathogenesis of colitis in animal models for IBD. The present study assessed the role of macrophage PPAR-
in DSS-induced IBD. An increased susceptibility to DSS-induced colitis, as defined by body weight loss, diarrhea, rectal bleeding score, colon length, and histology, was found in mice that contained a macrophage-specific disruption of PPAR-
(PPAR-
M
) compared with wild-type littermate (PPAR-
F/F) mice. Increased chemokine signaling in colons of PPAR-
M
vs. PPAR-
F/F mice suggest a role for macrophage recruitment in the increased susceptibility of PPAR-
M
mice to DSS-induced IBD.
| MATERIALS AND METHODS |
|---|
|
|
|---|
-floxed (PPAR-
F/F) mice containing loxP sites flanking exon 2 (2), were crossed with mice harboring the Cre recombinase under control of the lysozyme M promoter (Lys-cre mice). Mice were interbred for over six generations to produce littermates with the same mixed genetic background (10). Mice, housed in temperature- and light-controlled rooms, were given water and pelleted chow ad libitum. All animal studies were carried out in accordance with Institute of Laboratory Animal Resources guidelines and approved by the National Cancer Institute Animal Care and Use Committee.
Isolation of macrophage, neutrophils, and dendritic cells.
Thioglycollate-induced macrophages were isolated from PPAR-
M
and PPAR-
F/F mice by intraperitoneal injection of 3% thioglycollate medium (Becton Dickinson Microbiology Systems, Cockeyesville, MD). At 72 h postinjection, macrophages were harvested by peritoneal lavage with PBS and plated on plastic noncoated petri dishes (Becton Dickinson Labware, Franklin Lakes, NJ) as previously described (2). Following 2-h incubation, plastic-adhered cells were washed three times with PBS and harvested or were further cultured in RPMI medium (Invitrogen, Grand Island, NY) containing 2% FBS (Gemini Bio-Products, Woodland, CA) with rosiglitazone (1 µM) (LKT Laboratories, St. Paul, MN) or vehicle for 24 h. Purity was shown to be above 95% as assessed by F4/80 immunostaining. Neutrophils were isolated by intraperitoneal injection of 3% thioglycollate medium. At 4 h postinjection, neutrophils were harvested by peritoneal lavage with PBS and purity was 90% as assessed by Wright Giemsa stain. Dendritic cells were isolated from the spleens of PPAR-
M
and PPAR-
F/F mice. The spleens were digested with collagenase D and strained through a 100 µM nylon mesh (BD Biosciences, San Jose, CA) and centrifuged at 400 g for 10 min. The cell pellet was resuspended in 4 ml of RPMI-1640 medium (Invitrogen, Carlsbad, CA) and incubated with Dynabeads (Invitrogen) coated with anti-mouse cd11c antibody (MBL International, Woburn, MA) for 30 min at 4°C with constant agitation. The cell-bound beads were washed five times with RPMI-1640. For cd11c staining, the cells were plated directly on chamber slides (BD Biosciences) and, cd11c was detected using an ABC Mouse Vectastain Elite Kit (Vector Laboratories, Burlingame, CA). The purity was assessed to be >90%. For overnight culture, dendritic cells were placed in RPMI-1640 medium supplemented with 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin/streptomycin, and 50 µM 2-mercaptoethanol.
Induction and assessment of colitis.
PPAR-
M
or PPAR-
F/F mice, 8 to 10 wk old, were administered 2.5% (wt/vol) DSS (MW 35,00044,000) (MP Biomedicals, Aurora, OH) in the drinking water for 7 days. Daily changes in body weight and clinical signs of colitis, such as rectal bleeding, diarrhea, and bloody stool, were assessed and reported as a score from 0 to 4. Hemoccult SENSA (Beckman Coulter, Fullerton, CA) was used for the examination of rectal bleeding. For macroscopic colonic damage, colons were opened longitudinally, flushed with PBS, and fixed in 10% buffered formalin. The colons were Swiss-rolled to examine entire length of the colon and processed in paraffin. Colitis was scored on routine hematoxylin and eosin-stained section, according to previously described morphological criteria (13).
RNA analysis.
RNA was extracted from total colon following DSS administration or from thioglycollate-elicited macrophages using TRIzol reagent (Invitrogen). Northern blot analysis and probes for PPAR-
and acidic ribosomal phosphoprotein (36B4) were previously described (1). Quantitative real-time PCR (qPCR) was performed using cDNA generated from 1 µg total RNA with SuperScript III Reverse Transcriptase kit (Invitrogen). Primers were designed for qPCR using the Primer Express software (Applied Biosystems), and sequences are available upon request. qPCR reactions were carried out by use of SYBR green PCR master mix (Applied Biosystems) in an ABI Prism 7900HT sequence detection system (Applied Biosystems). Values were quantified by the comparative CT method, and samples were normalized to 36B4.
Chemotaxis assay. Transwell inserts with an 8 µM pore size fitted in 96-well plates (Chemicon International, Temecula, CA) were used for chemotaxis assays. Primary macrophages (5 x 105) were loaded into the top well containing 150 µl of RPMI medium. The bottom well contained RPMI medium and a combination of vehicle and MCP-1 (100 ng) (Sigma, St. Louis, MO) or rosiglitazone (1 µM) as indicated in the figure. The plates were incubated at 37°C in a CO2 incubator for 4 h. Following the incubation, the migrated cells were detached, lysed, and fluorescently labeled and analyzed per manufacturers protocol.
Immunohistochemistry. For histological analysis, tissue samples were fixed in 10% neutral buffered formalin overnight and paraffin embedded; 5 µM sections were cut, deparaffinized with xylene, and hydrated in an ethanol gradient. Immunohistochemical analysis was performed with macrophage receptor with collagenous structure (MARCO) antibody (BD Transduction Laboratories, Lexington, KY) by a streptavidin-biotin immunoperoxidase method using the ABC Kit (Vector Laboratories). The signal was visualized by diaminobenzidine staining (DAKO, Carpinteria, CA) and counterstained with hematoxylin.
Western blot analysis.
Thioglycollate-elicited macrophages were isolated from PPAR-
M
and PPAR-
F/F mice, and nuclei were isolated and lysed by use of NE-PER nuclear extraction kit (Pierce). The macrophage nuclear lysate was prepared for Western blotting as previously described (1). The membranes were incubated with an antibody against PPAR-
(Santa Cruz Biotechnology, Santa Cruz, CA), and the signals obtained were normalized to GAPDH (Chemicon International).
Data analysis. Results are expressed as means ± SD. P values were calculated by independent t-test, or for the DSS experiments multifactorial ANOVA test on the basis of genotype and DSS status. P < 0.05 was considered significant.
| RESULTS |
|---|
|
|
|---|
-null mice.
To specifically study the role of macrophage PPAR-
in IBD, PPAR-
F/F mice were crossed with Lys-cre transgenic mice to generate mice lacking expression of PPAR-
in the macrophage. PPAR-
M
mice were born at the expected Mendelian frequencies and exhibited no overt abnormalities compared with PPAR-
F/F littermate mice. To estimate the extent of macrophage-specific disruption of the PPAR-
gene, PCR analysis was used. The null allele amplifies as a 400-bp product; it was detected in genomic DNA of thioglycollate-elicited macrophages from PPAR-
M
mice and was not detected in macrophage DNA isolated from PPAR-
F/F mice (Fig. 1A). The intact floxed allele, which amplifies as a 285-bp product, was only faintly detected in macrophages from PPAR-
M
mice. In contrast, the intact floxed allele was the only band evident in macrophages from PPAR-
F/F mice and from kidney, liver, heart, and brown adipose tissue genomic DNA from PPAR-
M
mice (Fig. 1A). In addition, other myeloid cells were also assessed. In purified spleen dendritic cells (Fig. 1A) or whole spleen (data not shown), no recombination was demonstrated, whereas purified neutrophils demonstrated partial recombination. Consistent with the recombination data, Northern blot analysis and Western blot analysis demonstrated that PPAR-
was nearly completely disrupted in macrophages from PPAR-
M
mice (Fig. 1, B and C). Neutrophils isolated from PPAR-
M
or PPAR-
F/F mice demonstrated no PPAR-
expression as assessed by qPCR using exon 2 specific primers consistent with a recent report (28). To assess the effect of macrophage PPAR-
disruption on gene expression, qPCR was used to analyze mRNAs of PPAR-
target genes. Significant decreases were observed in basal aP2 and CD36 mRNA levels demonstrating a functional consequence of PPAR-
gene disruption in macrophages (Fig. 1D).
|
-null macrophage.
To assess whether disruption of PPAR-
in macrophages has any affect on basal cytokine gene expression levels, qPCR was used to measure cytokines mRNA levels shown to be critical in the pathogenesis of IBD. Thioglycollate-elicited macrophage from PPAR-
M
and PPAR-
F/F mice were incubated with either rosiglitazone (1 µM) or vehicle for 24 h. There were no differences in the expression levels of proinflammatory cytokines between macrophages incubated with vehicle from PPAR-
M
vs. PPAR-
F/F mice (Fig. 2). Rosiglitazone only reduced IL-1
expression in a PPAR-
-independent manner (Fig. 2). Basal expression of inducible nitric oxide synthase (iNOS) mRNA was increased in PPAR-
-null macrophages, and rosiglitazone decreased basal level of expression only in PPAR-
F/F, consistent with previous reports (14, 35) (Fig. 2). In addition, chemokine signaling was also induced in macrophages isolated from PPAR-
M
compared with PPAR-
F/F mice. CC chemokine receptor 2 (CCR2) and its ligand monocyte chemoattractant protein-1 (MCP-1) were both increased in PPAR-
-null macrophages (Fig. 2). This is the first report demonstrating an increase in MCP-1 expression in PPAR-
-null macrophages. Several studies have demonstrated a repressive role for TZDs in MCP-1- and CCR2-mediated signaling pathways (9, 20, 45). Owing to concentration-specific PPAR-
-independent pathways stimulated by TZDs (52), the present study sought to further characterize the PPAR-
dependency. Rosiglitazone repressed CCR2 and MCP-1 gene expression in macrophages isolated from PPAR-
F/F mice (Fig. 2). In contrast, rosiglitazone had no affect on CCR2 and MCP-1 expression in macrophages isolated from PPAR-
M
mice (Fig. 2), demonstrating that the inhibitory effects of rosiglitazone on CCR2/MCP-1 pathway are PPAR-
dependent.
|

M
to DSS-induced IBD.
PPAR-
M
mice (n = 21) showed an increased susceptibility to DSS-induced colitis compared with PPAR-
F/F mice (n = 21). Administration of 2.5% DSS in the drinking water for 7 days induced a significant body weight loss in PPAR-
M
mice compared with PPAR-
F/F mice at 5, 6, and 7 days following DSS treatment (Fig. 2A). PPAR-
M
mice demonstrated relatively poor diarrhea score and bleeding score (Fig. 2, D and E) following 7-day treatment with DSS, two major clinical symptoms associated with IBD. The colon length of PPAR-
M
mice following 7-day treatment with DSS was considerably shortened compared with PPAR-
F/F mice, indicating a greater extent of tissue damage in PPAR-
M
(Fig. 3D). Histological analysis showed significantly increased severity of IBD in PPAR-
M
mice (Fig. 3, E and F). The colons of PPAR-
M
mice following 7-day treatment with DSS demonstrated severe inflammation in the mucosa, muscularis propria, and submucosa with entire loss of the crypts and partial loss of the surface epithelia. The PPAR-
M
mice also displayed massive inflammatory infiltrates, whereas PPAR-
F/F mice displayed only partial loss of the crypts, no loss of the surface epithelia cells, and only mild inflammatory infiltrates. These data demonstrate a novel role for macrophage PPAR-
in protection of DSS-induced colonic injury.
|
, cytokine gene expression profiles were compared in whole colons of PPAR-
M
and PPAR-
F/F mice following 7-day DSS treatment. IFN-
, IL-1
, IL-6, IL-10, TNF-
, CCR2, MCP-1, and iNOS mRNA were induced from colonic tissue following 7-day treatment of DSS in both PPAR-
M
and PPAR-
F/F mice (Fig. 4). There were no differences in IFN-
, IL-6, IL-10, and TNF-
mRNA expression between PPAR-
M
and PPAR-
F/F mice. However, there was significant increase in IL-1
, iNOS, CCR2, and MCP-1 expression between colonic RNA isolated from PPAR-
M
vs. PPAR-
F/F mice following 7-day treatment of DSS. Together, these data suggest an important role for macrophage PPAR-
in the regulation of chemokine signaling during increased inflammation.
|
in monocyte recruitment during DSS-induced IBD.
Owing to the critical role of chemokine signaling in macrophage recruitment (51), in vitro migration assays were performed on macrophages isolated from PPAR-
M
and PPAR-
F/F mice. The data demonstrate that basal level of migration is not affected by PPAR-
disruption. However, upon MCP-1 incubation, macrophages isolated from PPAR-
M
mice exhibited increased migration compared with macrophages isolated from PPAR-
F/F mice. Furthermore, 1 µM of rosiglitazone inhibited MCP-1 induced migration in macrophages isolated from PPAR-
F/F mice but had no effect in PPAR-
disrupted macrophages (Fig. 5A). Macrophage-specific immunostaining for MARCO was performed on colonic sections following a 7-day treatment with DSS (Fig. 5B). A significant increase in macrophage recruitment was observed in PPAR-
M
mice compared with PPAR-
F/F mice. To confirm these results, macrophage-specific CD68 expression was assessed. Interestingly, CD68 expression was significantly increased in colon tissue of untreated PPAR-
M
mice compared with PPARF/F; this was not observed by MARCO staining (data not shown), because of the increased sensitivity of qPCR analysis. CD68 expression was enhanced in both PPAR-
M
and PPAR-
F/F mice following 7-day treatment with DSS compared with untreated mice; however, the increase in CD68 expression was more pronounced in PPAR-
M
mice (Fig. 5C). Taken together, the data demonstrate a novel role for macrophage PPAR-
in the recruitment of macrophages to inflammatory foci in the colon.
|
| DISCUSSION |
|---|
|
|
|---|
expressed in macrophages in the mouse IBD model, macrophage-specific PPAR-
-null mice were generated by using the Cre/loxP strategy with the Lys-cre transgene. PPAR-
M
mice exhibited over 90% loss of PPAR-
expression in macrophage as assessed by Northern and Western blot analysis. In a DSS-induced IBD model, PPAR-
M
mice exhibited more severe colitis as revealed by loss of body weight, shortened colon length, diarrhea, rectal bleeding scores, and macroscopic and histological analysis compared with PPAR-
F/F mice. Cytokine gene expression profiles were compared in macrophages and colonic tissue from PPAR-
M
and PPAR-
F/F mice. iNOS, MCP-1, and CCR2 mRNA levels were significantly increased in macrophage from PPAR-
M
mice. Similarly, in colonic tissue following 7-day treatment with DSS, the mRNAs encoding the chemotactic proteins, MCP-1 and CCR2, and iNOS were induced. In addition there was a significant increase in IL-1
cytokine gene expression in PPAR-
M
vs. PPAR-
F/F mice. This is consistent with an anti-inflammatory role for PPAR-
in macrophage and provides further support for a critical role for macrophages in colitis. Macrophages from PPAR-
M
mice demonstrated an increased expression of iNOS, MCP-1, and CCR2 compared with macrophages isolated from PPAR-
F/F mice. Therefore, the data suggest that endogenous PPAR-
ligands suppress basal levels of iNOS, MCP-1, and CCR2.
Increasing evidence suggests that the MCP-1/CCR2 pathway is important in inflammatory diseases. In an atherosclerotic mouse model, inhibition of CCR2 or its ligand MCP-1 significantly decreased macrophage recruitment and lesion size (7, 15, 19). In addition, a functional polymorphism of the CCR2 gene locus, which reduces CCR2 activity, is associated with decreased risk of coronary atherosclerosis (48). Recently, it was demonstrated that PPAR-
disruption in macrophage increases CCR2 expression and recruitment of macrophages to atherogenic sites (4). In the DSS-induced IBD model, a novel chemokine receptor antagonist, TAK-779, which demonstrates high affinity for CCR2, protected mice following DSS administration. The ameliorative effects of TAK-779 were directly correlated with a decrease in macrophage recruitment into the colonic mucosa (46). In addition, specific disruption of CCR5 and/or CCR2 protected mice from the severe inflammation and mucosal damage induced by DSS. Interestingly, the disruption of CCR2 and CCR5 was not critical for macrophage migration. Instead, both chemokine receptors were able to regulate the migration and differentiation of mucosal T cells (3). This discrepancy may be due to differences in MCP-1 expression levels. MCP-1 levels were not altered in mice with selective deletion of CCR2 or CCR5 (3). Consistent with an earlier report (46), the present study demonstrates increased CCR2 and MCP-1 levels in inflamed colonic tissue from PPAR-
M
mice, which correlates with an increase in macrophage recruitment to the colon.
In addition to chemokine signaling, IL-1
and iNOS expression were increased in colons from PPAR-
M
mice compared with PPAR-
F/F mice. IL-1
, a proinflammatory cytokine, was shown to be critical in gut inflammation by activating numerous immune cell types and appears to be a primary cause of IBD-induced diarrhea (18, 39). IL-1
at high doses induces tissue damage by promoting epithelial cell necrosis and inhibiting the endogenous action of IL-1
ameliorates acute and chronic experimental colitis (11, 12, 39).
iNOS expression is regulated by several proinflammatory cytokines, and it functions to generate high levels of nitric oxide (NO) via oxidative metabolism of L-arginine. NO activates guanylate cyclases, thereby enhancing cyclic GMP synthesis, and is involved in various functions including vasodilatation, inhibition of platelet aggregation, skeletal muscle contractility, and host defense (43). In addition, NO is a highly reactive free radical and can rapidly react with active oxygen species to generate peroxynitrite and cause severe detrimental effects (36, 41). However, IBD models using partially selective iNOS inhibitors or iNOS-deficient mice have led to contradictory results (6, 16, 22, 29, 32, 49, 53). The dual nature of NO is thought to be the reason for conflicting data in determining the role of iNOS in IBD models. The present data demonstrates that local increases in iNOS expression correlate with increased severity of IBD.
Unstimulated macrophages from PPAR-
M
mice displayed an increase in several proinflammatory mediators. In addition, colons from PPAR-
M
mice displayed an increase in macrophage recruitment as assessed by CD68 expression. The present study demonstrates a critical role for endogenous PPAR-
ligands in maintaining homeostasis of inflammatory responses. Therefore, the data suggest that in the absence of macrophage PPAR-
, the anti-inflammatory signal from endogenous PPAR-
ligands are no longer conveyed, making them more susceptible to inflammatory diseases. Although the expression levels of a number of proinflammatory genes are induced in PPAR-
M
mice following administration of DSS, the relative importance of these and possibly other genes in the increased susceptibility of PPAR-
M
mice in DSS-induced IBD model remains unclear. However, the present study has identified a novel and critical role for macrophage PPAR-
in the recruitment and activation of macrophages in the pathogenesis of DSS-induced colitis. These events appear to represent a deleterious cyclical process in which recruited macrophages secrete proinflammatory cytokines that in turn recruit additional macrophages and ultimately cause severe tissue damage. The present study provides rationale to target macrophage PPAR-
in patients diagnosed with IBD. Recently, it was shown that the molecular mechanisms driving ligand-dependent transrepression by PPAR-
in macrophage are distinct from classical PPAR-
-dependent gene transcription (35). It will be of interest to explore how these mechanisms can be used to develop new treatment modalities for IBD.
| GRANTS |
|---|
|
|
|---|
| FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
| REFERENCES |
|---|
|
|
|---|
in colonic epithelial cells protects against experimental inflammatory bowel disease. Gut 55: 10671069, 2006.This article has been cited by other articles:
![]() |
B. M. Necela, W. Su, and E. A. Thompson Peroxisome Proliferator-activated Receptor {gamma} Down-regulates Follistatin in Intestinal Epithelial Cells through SP1 J. Biol. Chem., October 31, 2008; 283(44): 29784 - 29794. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Thorp, G. Kuriakose, Y. M. Shah, F. J. Gonzalez, and I. Tabas Pioglitazone Increases Macrophage Apoptosis and Plaque Necrosis in Advanced Atherosclerotic Lesions of Nondiabetic Low-Density Lipoprotein Receptor Null Mice Circulation, November 6, 2007; 116(19): 2182 - 2190. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||