Published online before print October 2, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-elicited endothelial activation
* Departments of Medicine, Oncology and Clinical Immunology Unit, H San Raffaele Scientific Institute and Università Vita-Salute San Raffaele, and
IIT Network-Molecular Neuroscience Unit, Milan, Italy;
Department of Internal Medicine III and
IZKF Research Group 2, Nikolaus-Fiebiger-Center of Molecular Medicine, University Hospital Erlangen, Erlangen, Germany
1 Correspondence: Departments of Medicine, Oncology and Clinical Immunology Unit, H San Raffaele University Hospital, DIBIT 3A1, via Olgettina 58, 20132 Milano, Italy. E-mail: manfredi.angelo{at}hsr.it
ABSTRACT
TNF-
plays an important role in the natural history of rheumatoid arthritis (RA), a systemic disease characterized by endothelial activation and synovial involvement with bone erosions. Neuroendocrine signals contribute as well to RA, but their role is poorly understood. We measured in 104 RA patients and in an equal number of sex- and age-matched, healthy controls the blood levels of chromogranin A (CgA), a candidate marker linking the neuroendocrine system to TNF--mediated vascular inflammation. CgA levels were significantly higher in patients with RA and remained stable over time. High levels of CgA were significantly associated with severe extra-articular manifestations, namely pulmonary fibrosis, rheumatoid vasculitis, serositis, and peripheral neuropathy. RA sera curbed the response of human microvascular endothelial cells to TNF-, as assessed by the expression of ICAM-1, the release of MCP-1/CCL2, and the export of nuclear high-mobility group box 1; the effect abated in the presence of anti-CgA antibodies. The efficacy of the blockade was significantly correlated with the CgA concentration in the serum. The recombinant aminoterminal portion of CgA, corresponding to residues 1–78, had similar inhibitory effects on endothelial cells challenged with TNF-. Our results suggest that enhanced levels of CgA identify patients with extra-articular involvement and reveal a negative feedback loop that limits the activation of endothelial cells in RA.
Key Words: arthritis • inflammation • HMGB1 • cell activation • neuropeptides
INTRODUCTION
The failure of homeostatic control mechanisms underlies articular and vascular manifestations of rheumatoid arthritis (RA). The hypothalamic-pituitary-adrenal and -gonadal axes and the sympathetic nervous system are implicated in the regulation of inflammation and the pathogenesis of RA [1
2
3
4
]. Elevated blood levels of neuropeptide Y and chromogranin A (CgA), markers of adrenergic outflow, reveal the activation of the sympathetic nervous system in RA [5
6
7
]. The concentrations of CgA and soluble TNF-RI and -II in the blood of RA patients are correlated significantly [5
], suggesting a cross-talk between neuroendocrine and inflammatory pathways. Conversely, a defective regulation of the parasympathetic nervous system, which suppresses the production of inflammatory cytokines, tissue injury, and cell damage in peripheral organs [8
], has been suggested in RA patients [9
]. High levels of the high-mobility group box 1 (HMGB1) protein, a nuclear molecule with an important extracellular action involved in experimental and human synovitis [10
11
12
13
14
], have also been reported in RA [9
].
CgA is a 49-kDa acidic polypeptide, stored in the dense-core secretory granules of adrenergic neurons, which contributes to the biogenesis, trafficking, storage, and release of catecholamines and other neuropeptides [15 ]. Secreted CgA undergoes proteolytic processing, yielding biologically active fragments [16 ]. The concentration of CgA in the blood of patients with RA is higher than in the synovial tissue, suggesting a systemic release of the molecule [7 ]. Likewise, we described increased blood levels of CgA in patients with giant-cell arteritis, with no evidence of the protein expression in temporal artery samples [17 ].
CgA is a biologically active molecule: Intact CgA and its amino-terminal proteolytic fragment maintain in vitro and in vivo the barrier function of the endothelium challenged with TNF- [18
, 19
], influence the myocardial function, and regulate TNF--mediated vascular inflammation in patients with chronic heart failure [20
21
22
23
]. The endothelium and the vasculature are therefore candidate targets for CgA released in the bloodstream.
The goal of this study was to verify whether CgA is involved, via its action on the vessels, in systemic, extra-articular manifestations of RA, most of which are caused directly or associated strictly with vessel inflammation [24 25 26 ]. To address this issue, we assessed the concentration of the molecule in the blood of RA patients and verified in vitro its biological activity on microvascular endothelial cells.
MATERIALS AND METHODS
Patients
We enrolled 104 patients with RA (Table 1
), classified according to the 1987 American Rheumatism Association/American College of Rheumatology revised criteria [27
], followed in the Rheumatology and Clinical Immunology Unit of the San Raffaele Scientific Institute Hospital (Milan, Italy) and in the Department of Internal Medicine III of the Friedrich-Alexander University (Erlangen, Germany). Patients were not eligible if they had concomitant New York Heart Association Stage III or IV cardiac failure, uncontrolled hypertension, atrophic gastritis, neuroendocrine malignancies, sepsis or severe infections, severe renal failure (<50 ml/min creatinine clearance), and hepatic insufficiency (Child class B or C). All patients provided written informed consent to this study. Table 1
summarizes some characteristics of the patients, including ongoing treatment with diseas-modifying antirheumatic drugs, corticosteroids, and/or nonsteroidal, anti-inflammatory drugs. An equal number (104) of age- and sex-matched, healthy volunteers were studied in parallel as controls. Demographic features and relevant medical data, including present or past extra-articular manifestations, bone erosions, comorbidity, and drug treatments, were registered. Clinical evaluation at the day of venipuncture included the assessment of disease activity by means of the European League Against Rheumatism DAS28, which is determined by a formula that takes into account the number of tender and swollen joints, ESR, and the patient visual analog scale for disease activity. Patients were evaluated for the presence of extra-articular involvement, according to previously defined criteria (Table 2
) [28
, 29
]. Severe extra-articular manifestations comprised rheumatoid vasculitis, interstitial lung disease, serositis and peripheral neuropathy, and mild extra-articular manifestation rheumatoid nodules and sicca syndrome. Blood samples were collected by venipuncture at enrollment from 104 patients and 104 controls. Fourteen RA patients were selected randomly to measure prospectively biological parameters. Samples were obtained 6 months after the baseline and then every 3 months. We collected two to eight samples per patient, depending on the time of enrollment. Clinical evaluation was repeated at each blood collection.
|
View this table: [in a new window] |
Table 1. Characteristics of RA Patients
|
|
View this table: [in a new window] |
Table 2. Characteristics of RA Patients with Extra-Articular Manifestations
|
RI and -II was measured by commercial ELISA kits (R&D Systems, Minneapolis, MN, USA), according to the manufacturer’s instructions. All experiments have been performed at least in triplicate.
Cells
Human microvascular endothelial cell 1 (HMEC-1), kindly provided by the Center for Disease Control and Prevention (Atlanta, GA, USA), was cultured in MCDB-131 medium (Invitrogen, Carlsbad, CA, USA), supplemented with 10% heat-inactivated FCS (Cambrex Bio Science, Milano, Italy), hydrocortisone (1 µg/ml, Sigma Chemical Co., St. Louis, MO, USA), epidermal growth factor (10 ng/ml, Sigma Chemical Co.), L-glutamine 2 mM, penicillin (50 U/ml), and streptomycin (50 µg/ml, Invitrogen).
Cell activation
Adherent HMEC-1 was cultured in 12-well plates, and 90–95% confluent cells were stimulated with TNF- (500 pg/ml, R&D Systems) overnight. Purified vasostatin-1 (VS-1; 1 µM unless otherwise indicated) was added 1 h before TNF-. Supernatants were then harvested and stored at –20°C and analyzed by ELISA [CCL2/MCP-1, CXCL8/IL-8, soluble receptor for advanced glycation end products (RAGE) and CXCL16, R&D Systems], according to the manufacturer’s instructions. Each sample was analyzed in duplicate. Cells were detached with PBS containing EDTA (0.5 mM, Sigma Chemical Co.), incubated in PBS/4% FCS for 30 min, and incubated with unlabeled, anti-human ICAM-1, VCAM-1 (R&D Systems), PECAM-1, CD105 (BD PharMingen, San Diego, CA, USA), and von Willebrand factor (Immunotech Instrumentation Laboratories, Milan, Italy) primary antibodies for 1 h at 4°C (10 µg/ml final concentration). Cells were then washed and incubated with PE-labeled goat anti-mouse antibodies as second-step reagents (10 µg/ml, BD PharMingen) for 45 min and then fixed in 1% paraformaldehyde (Sigma Chemical Co.). Control samples were incubated with isotype-matched irrelevant Igs (BD PharMingen). Data were acquired on a Beckman Coulter FC500 flow cytometer. Results are expressed as relative fluorescence intensity (RFI), calculated as follows: mean fluorescence intensity (MFI) with the specific mAb/MFI in the presence of isotype-matched, irrelevant Igs. To verify the action of sera on endothelial activation, HMEC-1 cells were cultured on 12-well plates. Confluent cells (90–95%) were washed once with PBS with calcium and magnesium (Cambrex Bio Science, Milano, Italy), and culture medium containing 10% serum from RA patients was added. A pool of 10 sera from healthy subjects was used as a control. We tested sera from RA patients who had never been treated with anti-TNF- agents. When indicated, anti-CgA, anti-VS-1, and anti-C-terminal-CgA antisera (1:200, see above) were added in parallel. After 1 h, cells were challenged with TNF- (500 pg/ml final concentration) and incubated overnight. Cells were then stained with anti-ICAM-1 antibodies as described and analyzed by flow cytometry. All experiments have been performed at least in triplicate.
Immunofluorescence and confocal microscopy
Intracellular staining for HMGB1 was performed in HMEC-1, cultured on glass slides, and stimulated overnight with TNF- (500 pg/ml). Purified VS-1 (1 µM final concentration) was added 1 h before TNF-. After incubation, cells were fixed with 4% paraformaldehyde. The membrane was permeabilized with HEPES-Triton buffer (Sigma Chemical Co.) and subsequently blocked with PBS/4% BSA. Cells were then incubated with unlabeled mouse anti-human HMGB1 mAb (1:200; BD PharMingen) for 1 h at 37°C and subsequently, with FITC-labeled goat anti-mouse secondary antibody (1:200; BD PharMingen). Hoechst 33358 was used for nuclei counterstaining (Molecular Probes, Eugene, OR, USA). Control samples were challenged with isotype-matched irrelevant Igs (BD PharMingen). Slides were washed, dried, mounted, and examined under a Leica TCS SP2 confocal microscope (Leica, Heidelberg, Germany). Leica confocal software was used to acquire images (liquid color spectrophotometer) and Adobe Photoshop 6.0 software (Adobe Systems, San Jose, CA, USA) to process them.
Statistical analysis
Data are presented as mean ± SEM. We used the Mann-Whitney U test to compare continuous parameters between RA patients and controls and between different subgroups of patients, according to distribution and dimension of the samples. The correlation between CgA levels and other continuous variables in RA patients (biological markers and DAS28) has been calculated by Spearman correlation analysis and is expressed as r coefficient. Comparison of discrete variables (sex, gender, and the presence of ANA and RF) was performed using the 
2 test. The receiver operating characteristic (ROC) curve was used to identify the threshold value of circulating CgA with optimal sensitivity and specificity to distinguish patients with extra-articular manifestations. The correlation between CgA levels and continuous parameters is expressed as Spearman’s r coefficient. Comparisons between data of the in vitro experiments were made using the Mann-Whitney U test. A P value below 0.05 was considered significant.
RESULTS
High levels of CgA identify extra-articular RA involvement
CgA mean concentration in the sera of 104 consecutive RA patients, whose characteristics are described in Table 1
, was 283.6 ± 97.9 ng/ml (range 11.2–10010 ng/ml). This value was significantly higher than in 104 sex- and age-matched, healthy controls (55.5±3.9 ng/ml, range 9–122.2 ng/ml; P<0.005; Fig. 1A
). There was no correlation between CgA and ESR values (r=0.14), CRP levels (r=0.03), or disease activity, assessed using the validated DAS28 score (r=0.08). Circulating RF was detected in 82 of 104 RA patients (78.8%). There was no correlation between CgA and RF (r=–0.07). Moreover, we failed to detect significant differences in the concentration of CgA in patients with or without bone erosions (55 of 104). Twenty-three of 104 RA patients had extra-articular manifestations (Table 2)
, and in five patients, extra-articular manifestations coexisted. CgA levels were significantly higher in patients with extra-articular manifestations than in patients with exclusive articular involvement (787.4±354 vs. 134.3±2.3 ng/ml; P<0.005; Fig. 1B
). Twelve patients had severe extra-articular manifestations, which are associated to enhanced mortality: pulmonary fibrosis, peripheral neuropathy, systemic rheumatoid vasculitis, and serositis. CgA levels were even higher in these patients (1562.6±261.1 ng/ml; P=0.0006 compared with patients with no extra-articular involvement; Fig. 1B
). ROC curve analysis showed that CgA concentrations exceeding 300 and 700 ng/ml identified extra-articular manifestations with 90% and 97% specificity, respectively. As expected [28
, 32
], we observed a significantly lower (P=0.005) female:male ratio among patients with extra-articular manifestations (1.3:1) as compared with patients with articular disease only (5.6:1). ANA or RF did not differ significantly in patients with or without extra-articular manifestations.
 View larger version (12K): [in a new window] |
Figure 1. Levels of CgA in RA patients are high and associated with extra-articular manifestations. (A) We evaluated the concentrations of CgA in the serum of 104 RA patients and 104 healthy controls. CgA levels were significantly higher in patients (283.6±97.9 ng/ml) as compared with sex- and age-matched, healthy controls (55.5±3.9 ng/ml). Data are shown as box plots, and first, second (median), and third quartiles define box area and vertical lines extending to fifth and 95th percentiles. ***, P < 0.005 (Mann Whitney test). (B) Significantly higher levels of CgA were observed in RA patients with severe extra-articular manifestations (EAM) as compared with patients with selective articular involvement. Data are shown as box plots, and first, second (median), and third quartiles define box area and vertical lines extending to fifth and 95th percentiles. ***, P < 0.005 (Mann Whitney test).
|
 View larger version (15K): [in a new window] |
Figure 2. CgA levels are stable over time in RA patients. (A) We prospectively collected serial samples (two to eight samples) from 14 randomly selected patients (Pt.) over a period of several months (6–24 months). Arrows indicate the timing of blood sampling in each patient. (B) CgA (CHGA) serum concentration was remarkably stable over the time. Each symbol represents a single determination at 3-month intervals.
|
-elicited endothelial cell activation, substantially up-regulated the expression of ICAM-1, a transmembrane glycoprotein of the Ig superfamily involved in cell adhesion via integrin receptors. CgA-containing RA sera (10%), but not control sera, inhibited TNF--induced ICAM-1 expression on endothelial cells (Fig. 3
). To limit possible biases, we excluded sera of patients treated with anti-TNF- agents, namely infliximab, etanercept, and adalimumab. To assess whether CgA is involved in the inhibitory effect of the sera, we used antisera raised against the N- and C-terminal portion of CgA. The treatment rescued the ability of TNF- to induce ICAM-1 up-regulation on microvascular endothelial cells: The rescue was significant, only in sera that contained the highest concentrations of CgA, and the efficacy of the blockade was significantly correlated with the concentration of the molecule (r=0.799; Fig. 3
).
 View larger version (34K): [in a new window] |
Figure 3. CgA contained in RA sera inhibits ICAM-1 up-regulation by TNF-. We evaluated by flow cytometry the expression of surface ICAM-1 in cultured microvascular endothelial cells. (A) Histogram plots from a representative experiment depicting ICAM-1 expression in resting cells (third panel from top) or in cells challenged overnight with TNF- (500 pg/ml; bottom three panels) in the absence or presence of CgA-containing serum, which prevented TNF--elicited ICAM-1 up-regulation (bottom two panels). ICAM-1 up-regulation was rescued by adding blocking anti-CgA antiserum (bottom panel). (B) ICAM-1 expression was evaluated by flow cytometry in resting cells (filled column), cells challenged overnight with TNF- (500 pg/ml) in the presence of FCS (striped column), or sera from RA patients (Patients 1–9, open columns) or normal human donors [normal human serum (NHS; dotted column)]. (C) Open columns represent the percent of rescue of TNF--elicited ICAM-1 up-regulation obtained by adding anti-CgA to the serum of patients (1–9) or to NHS. (D) The rescue was correlated with concentration of CgA in the serum of the patient (r=0.799; P<0.005; Spearman’s test).
|
-elicited up-regulation of ICAM-1 (P<0.005; Fig. 4
). The effect was specific, as the expression of CD105 and von Willebrand factor was not influenced (Fig. 4)
and dose-dependent, with a plateau at approximately 1 µM (Fig. 4
, inset).
 View larger version (28K): [in a new window] |
Figure 4. CgA1–78 reproduces the effects of CgA-containing sera on microvascular endothelial cell activation. (A) Overnight challenge with TNF- (500 pg/ml) induced ICAM-1 up-regulation, as assessed by flow cytometry, which abated in the presence of CgA1–78 (1 µM). (B) Overnight challenge with TNF- (500 pg/ml) induced the release of MCP-1/CCL2 and (C) IL-8/CXCL8 by endothelial cells. MCP-1/CCL2 but not IL-8/CXCL8 release abated in the presence of CgA1–78 (1 µM). (Inset) ICAM-1 up-regulation induced by TNF- was inhibited by increasing concentrations of CgA1–78 (x-axis). ***, P < 0.001 (Mann-Whitney test).
|
, and it had no effect on the secretion of CXCL8/IL-8 (Fig. 4)
. Soluble RAGE and CXCL16 were undetectable (not shown). As a further marker of endothelial activation, we traced the intracellular compartmentalization of HMGB1. Upon stimulation with TNF-, HMGB1 expression increased in the nucleus and in the cytoplasm (Fig. 5
), an event that is required for its eventual, atypical secretion by endothelial cells [33
, 34
]. The action of TNF- was abrogated in the presence of CgA1–78 (Fig. 5)
.
 View larger version (22K): [in a new window] |
Figure 5. CgA1–78 inhibits the nuclear export of HMGB1 from activated endothelial cells. We evaluated by indirect immunofluorescence the expression of HMGB1 (green) in cultured endothelial cells. Nuclei were counterstained with Hoechst (blue). Resting cells (A) and cells challenged with CgA1–78 alone (1 µM; C) did not express detectable, extranuclear HMGB1. Upon overnight challenge with 500 pg/ml TNF- (B), a substantial increase in HMGB1 expression was detectable in the nucleus and in the cytoplasm. The shift was abrogated in the presence of CgA1–78 (1 µM; D).
|
A sustained, vascular damage underlies clinically heterogeneous, severe extra-articular manifestations in RA, including peripheral neuropathy, rheumatoid vasculitis, interstitial lung disease, and serositis [24 ]. Inflammatory vascular lesions are fairly common in RA, and often, they are apparently not associated with overt clinical manifestations [24 ]. However, clinically silent inflammatory lesions could explain the increased atherosclerosis of coronary arteries in the absence of a higher frequency of common cardiovascular risk factors, which is responsible for the increased mortality related to cardiovascular events among RA patients [28 , 35 ], in particular, those with severe extra-articular manifestations [36 , 37 ]. In our study, CgA concentrations exceeding 300 and 700 ng/ml identify extra-articular manifestations with 90% and 97% specificity, respectively. This result has important implications, as markers associating to or predicting systemic involvement in RA would be valuable in the clinical setting [28 , 29 , 32 ]. Extra-articular manifestations of RA, once established, often represent a challenge for clinicians and escape early recognition, as they are asymptomatic for a long time.
This observation provides clues to entangle the vascular dysfunction in RA patients: CgA and its N-terminal portion contribute to the regulation of the myocardial function, play a role in cardiovascular diseases in which TNF- is involved, such as heart failure [20
21
22
23
], and protect vessels against TNF--induced vascular leakage (reviewed in ref. [16
]). We found that high concentrations of CgA correlate with extra-articular manifestations in RA, which in turn depend on persistent vascular inflammation.
CgA blood levels were relatively stable over time, irrespective of disease activity and acute-phase reactant fluctuations. Such a pattern is typical of other moieties relevant to the natural history of RA. This is the case of autoantibodies, such as those recognizing epitopes derived from proteoglycans, collagens, chondrocyte or synovial antigens, citrullinated residues, and RF. However, this is the first observation of the stable elevation of a bona fide neuroendocrine marker present in RA patients’ sera.
Other inflammatory molecules have been well-characterized, which are produced as a response to persistent triggers and play a protective role, limiting tissue damage: this is, for example, the case of the IL-1βR antagonist, whose blood levels and ex vivo production represent independent predictors, inversely correlated with joint damage evolution (see ref. [38 ]). This notion has actually been exploited to develop a widely used treatment for RA and related conditions, including juvenile idiopathic arthritis [39 ].
On the basis of this evidence, we investigated in vitro the effects of the N-teminal fragment of CgA on human endothelial cells. We found that CgA curbs the effects of TNF- on ICAM-1 expression and of MCP-1/CCL2 release by endothelial cells. ICAM-1 and MCP-1/CCL2 represent crucial targets of TNF- action involved in the recruitment of inflammatory leukocytes at the site of inflammation, in their firm adhesion to endothelia, and eventually controlling their extravasation. In addition, data from experimental models and in vitro studies of human arteries support the notion that these molecules play important roles in atherosclerosis development [40
41
42
].
The CgA N-teminal fragment also influenced the nuclear export of HMGB1, a nuclear DNA-binding protein secreted by a variety of cell types upon activation, which exert proinflammatory effects [13
, 33
, 34
, 43
44
45
]. High levels of HMGB1 are found in the synoviae and in the blood of RA patients, and synovial macrophages challenged with HMGB1 produce high amounts of TNF-, IL-1β, and IL-6 [10
, 11
]; HMGB1 injected once in the joint of experimental mice induces a destructive polyarthritis, and it plays an important role in endothelial cell activation [46
47
48
].
Recently, an elegant study indicated that gold compounds, a treatment used for several decades in the treatment of RA, act, at least partially, by inhibiting the extracellular release of HMGB1 from activated macrophages and by promoting the nuclear retention of this protein [14 ]. It is fascinating that the endogenous CgA molecule, which is present in the blood of some RA patients at levels that are comparable with those found in the blood of patients with neuroendocrine malignancies, mediates an identical effect on HMGB1 transport.
Altogether, these results suggest that CgA acts at a crucial point in the natural history of RA and that it could be involved in the regulation of events that are pivotal to the establishment of vascular inflammation. Prospective clinical studies will allow us to verify the cardiovascular outcome of patients with high circulating levels of CgA.
ACKNOWLEDGEMENTS
This work was supported by the Ministero della Salute, the Ministero dell’Università e della Ricerca (MIUR), and the Associazione Italiana per la Ricerca sul Cancro (AIRC). The Center for Disease Control and Prevention kindly provided HMEC-1.
Received June 11, 2008; revised July 17, 2008; accepted August 31, 2008.
REFERENCES
in rheumatoid arthritis Ann. N. Y. Acad. Sci. 1069,428-437[CrossRef][Medline]-induced vascular leakage FASEB J. 18,554-556 (TNF) in chronic heart failure Adv. Exp. Med. Biol. 482,351-359[Medline]This article has been cited by other articles:
 |
|
 |
|
  K. B. Helle The chromogranin A-derived peptides vasostatin-I and catestatin as regulatory peptides for cardiovascular functions Cardiovasc Res, January 1, 2010; 85(1): 9 - 16. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
