An abnormal adherence of monocytes to fibronectin in thyroid autoimmunity has consequences for cell polarization and the development of veiled cells

Abstract

Blood monocytes of patients with thyroid autoimmune disease (TAID) display defects in rearranging their cortical actomyosin cytoskeleton (‘polarize’) in response to chemoattractants. Such rearrangements also take place after the adherence of monocytes to the extracellular matrix (ECM). It is therefore not surprising that monocytes are primed after fibronectin (FN) adherence, displaying an enhanced polarization toward chemoattractants.

We investigated the integrin expression and chemoattractant-induced polarization of monocytes of TAID patients before and after FN adherence. Since cytoskeletal rearrangements are also required during the transition of monocytes into veiled antigen-presenting cells (VCs), we investigated such transition of FN-adherent monocytes of TAID patients.

Adherent and nonadherent monocyte populations from TAID patients and healthy controls were subjected to a polarization test with the chemoattractant fMLP (or MCP-1), FACS analyses (FITC-labelled FN, CD29, CD49e, d, b and a) and tested for their capability to develop into veiled APC.

Monocytes of healthy individuals showed an improved chemoattractant-induced cell polarization after FN adherence, not reflected by TAID monocytes, in which chemoattractant-induced polarization worsened. Monocytes of healthy individuals up-regulated CD49e and d integrins and their capability to bind FITC-labelled FN after adherence to a FN-coated plate, as well as enhancing their capability to generate T cell-stimulatory VCs. Monocytes of TAID patients did not.

These data indicate that integrin- (and chemokine-) mediated functions are hampered in monocytes of TAID patients. Because integrin action is pivotal to processes such as monocyte adherence to endothelial cells, uropod formation, migration into tissues and differentiation into APC and macrophages, these defects might underly immune dysbalances important in thyroid autoimmune development.

Introduction

We previously reported in patients with thyroid autoimmune disease [1,2] a hampered response of monocytes to change shape from a round to a triangular motile form when stimulated with the chemoattractant N-formyl-methionyl-leucyl-phenalanine (fMLP). The process of shape change, also called cell polarization, is based on a rearrangement of the interior cytoskeleton of the cell [3]. Chemoattractants and chemokines such as fMLP induce such rearrangement by binding to serpentine G-protein coupled receptors in the cytomembrane of monocytes. This stimulates a complex cascade of 2nd messenger systems including the phosphatidyl inositol (PI) cycle and the Rho-family of GTP-ases, which results – amongst other things – in a coupling of f-actin and other cytoskeletal building blocks to the membrane of the cell [4]. In this way a submembranous myoactin cell cortex is formed, enabling the cell to change shape [5].

A similar formation of submembranous cortical myoactin filaments does occur after contact of monocytes with the extracellular matrix (ECM). Monocytes adhere to the ECM via integrins [6]. The integrins are a family of noncovalent heterodimeric complexes situated in the cell membrane and consisting of an α and a β chain. Integrins do not only act as intermediary molecules linking the ECM fibres to the interior cytoskeleton, they also act as receptors triggering 2nd messenger systems like the PI cycle and the Rho-family of GTP-ases [7]. It is thus not surprising that adherence of monocytes to an ECM matrix synergizes with chemoattractant signalling. Fibronectin (FN)-adhered monocytes have been reported to respond to chemoattractants with an enhanced cell polarization [8,9]. A concerted action between integrin signalling and chemoattractant stimulation is thus of prime importance for the diapedesis of monocytes from the bloodstream into the tissues and their differentiation in the various monocyte-derived cells.

Here we describe an investigation on the chemoattractant-induced cell polarization of monocytes from 20 patients with recently diagnosed thyroid autoimmune disease (10 patients with hyperthyroid Graves' disease; 10 patients with hypothyroid autoimmune thyroiditis). Monocytes had either been or not been adhered to a FN-coated surface prior to fMLP or MCP-1 chemoattractant stimulation.

Several integrins can bind to FN, however, of the α chains CD49d and CD49e are primarily involved [6,7]. Of the β chains CD29 is involved. Because of the molecular heterogeneity of integrin receptors for FN binding, we labelled FN with FITC and studied in addition to CD49 and CD29 expression, the expression of ‘fibronectin-receptors’, i.e. the binding of FITC-labelled FN to patient and control monocytes before and after FN adherence.

Since rearrangements of the cortical actomyosin cytoskeleton are also needed for the development of the cell protrusions characteristic of veiled cells during their differentiation from monocytes [10], we also studied the capability of patient monocytes to develop into accessory veiled cells (VCs) before and after FN adherence. We previously reported on a method to generate considerable numbers of VCs from monocytes [11,12]. Crucial in this method of VC generation is the prevention of monocyte oxidase stimulation via strictly plastic nonadherent culture conditions and the presence of reducing substances like BHA or iodinated compounds (e.g. metrizamide, reverse T3 and T3) to improve the yield of VCs [11,12]. The procedure results in a veiled cell population of which 30–45% of the cells have veiled protrusions, starting from a suspension of round or oval-shaped monocytes. These generated VCs show a strong MHC class II positivity, a decreased expression of the monocyte marker CD14, S100 positivity (a marker of Langerhans dendritic cells), a decreased phagocytic capability, and a clearly enhanced stimulator capability in T cell proliferation assays indicating their improved accessory cell function [11,12]. These in vitro-generated monocyte-derived VC thus resemble in many morphological and functional aspects the cells in vivo known to belong to the heterogenous group of dendritic cells (DC) [13]. We previously showed that the generation of VCs from monocytes and their function is disturbed in patients with an organ-specific autoimmune disease, like autoimmune thyroid disease [1,2,14].

Patients and methods

Patients and controls

Heparinized blood (70 ml) was obtained via venapuncture from the following groups of individuals.

(a) Ten recently diagnosed hyperthyroid Graves' patients (nine females, one male, ages 50 ± 13 (mean ± standard deviation) age-ranges 23–69 years) visiting the Department of Internal Medicine of the Zuiderziekenhuis, Rotterdam. None of the patients received medication at the time of blood collection. All patients had raised serum free thyroxin (fT4) levels and decreased serum thyroid stimulating hormone (TSH) levels. Patients had a diffuse non-nodular appearance of the thyroid on palpation, scan or ultrasonography. TSH-receptor antibodies (Brahms Diagnostics, Berlin, Germany) were positive in 8, thyroid anticytoplasmic antibodies (inhouse indirect immunofluorescence assay) were positive in 7.

(b) Ten recently diagnosed hypothyroid patients (nine females, one male; ages 53 ± 14 years, age ranges 40–75 years) visiting the Department of Internal Medicine of the Zuiderziekenhuis, Rotterdam. None of the patients received medication at the time of blood collection. All patients had lowered serum fT4 levels and raised serum TSH levels. All had positive thyroid anticytoplasmic antibodies. TSH-receptor antibodies were positive in none.

(c) Twenty healthy controls (15 females, 5 males) consisting of laboratory and hospital staff, ages 34 ± 8 years (mean ±sd); age range 21–52 years.

All patients and controls gave informed consent to donate blood for this investigation. The protocol had been approved by the Medical Ethical Committee of the Erasmus University Hospital Dijkzigt.

Monocyte isolation

Peripheral blood mononuclear cell separation was performed using Ficoll‐Paque density gradient centrifugation (density 1·077 g/ml; Pharmacia, Uppsala, Sweden). The cells were washed twice in RPMI 1640 culture fluid enriched with 10% fetal calf serum (FCS; Gibco, Breda, the Netherlands). Aliquots of 40–80 × 10⁶ cells were stored at −150°C in a solution of 10% DMSO (Dimethylsulphoxide) in RPMI 1640 supplemented with 10% FCS.

Monocytes were isolated from the separated and deep-frozen lymphoid cells: an enrichment for the monocytes in the Ficoll-Paque isolated fraction was obtained by Percoll gradient centrifugation: after washing, the Ficoll-isolated pellet containing both monocytes and lymphocytes was resuspended in RPMI 1640 10% FCS and carefully layered on top of an equal volume of Percoll 1·063 (Pharmacia, Diagnostics AC, Uppsala, Sweden). After centrifugation (40 min, 450 g) the cells were collected from the interface, washed twice in the culture fluid (10 min, 500 g) and counted. The suspension almost always contains 80–95% monocyte-specific esterase positive cells, and only suspensions showing this yield of monocytes were used. The monocyte suspension was either directly used for the monocyte polarization assay and for FACS analysis, or for further adherence to fibronectin-coated surfaces.

Preparation of fibronectin-adhered monocytes

Monocytes were suspended at a density of 2 × 10⁶ cells/ml in supplemented RPMI medium; aliquots (2·5 ml) were then dispensed into 50-mm tissue culture dishes (Nunc, Denmark) that had been precoated with varying concentrations of human FN (CLB, Amsterdam, the Netherlands) in a total volume of 2·0 ml of phosphate buffered saline (PBS) for 2 h at 37°C. A coating concentration of 20 μg/ml of FN was found to be optimal, and this concentration was used in subsequent assays. The monocytes were incubated in these FN-coated wells at 37°C for 1 h in a 5% CO₂−95% air atmosphere. Nonadherent monocytes were vigorously washed off with ice-cold PBS (2 × 2 ml).

Firmly FN-adherent monocytes were removed by incubating the cells in 2·0 ml PBS enriched with 3 mm EDTA for 45 min at 4°C, followed by further removal using a rubber policeman. This procedure was repeated twice with RPMI 1640, cells were washed, concentrated by centrifugation and counted using a haemocytometer.

We checked the lymphocytic contamination of the FN-adhered cells by studying the expression of monocyte-esterase and markers such as CD3 (for contaminating T cells), CD19 (for contaminating B cells) and CD56-CD16 (for contaminating NK cells). The lymphocytic contamination of the monocytes after FN adherence was very low: the population contained >90% of cells staining for monocyte-esterase, while contamination with CD3⁺ T cells was as low as 2%, and the B- and NK-cell contamination as low as 5%. The FN nonadherent monocytes were considerably contaminated with other populations of cells (around 15% of T cells and 15% of NK cells) and were not further used in our assays.

The polarization assay

For the polarization assay, aliquots (0·2 ml) of the various suspensions containing 2 × 10⁵ monocytes were added to 12 × 75 mm polypropylene tubes (Falcon Labware Division of Becton Dickinson, Oxford, CA, USA) containing 0·05 ml of either medium, N-formyl-methionyl-leucyl-phenylalanine (fMLP) in medium to reach a final concentration of 10 nm, or MCP-1 (gift Dr A. Montovani). The tubes were incubated at 37°C in a waterbath for 15 min. The incubation was stopped by addition of 0·25 ml ice-cold 10% formaldehyde in 0·05% PBS (pH 7·2). The cell suspensions were kept at 4°C until counting of polarized cells in a haemocytometer using an ordinary light microscope (magnification 250 ×). The test was read blindly by two experienced persons; 200 cells were counted, manually, from each tube. A monocyte was ‘polarized’ if any of the following occurred: a change to an elongated or triangular shape, broadened lamallopodia, and/or membrane ruffling. The responsiveness of a monocyte population was expressed as the percentage of polarized cells in the presence of the chemoattractant minus the percentage of polarized cells in the absence of the chemoattractant. On previous occasions we verified that contaminating lymphocytes (in occasional preparations present in up to 20%) do not polarize under these conditions. The percentage of polarized monocytes was calculated as follows: (% cells polarized due to chemoattractant/% monocyte-specific esterase positive cells) × 100%.

The mean value of the two values recorded blindly by the two experienced observers was taken as the final value. Although this method seems to be a subjective microscopical quantification, it has been proven to be the most sensitive and reliable polarization assay as compared to morphometric and flow cytometric methods [15]. This has also been our experience, with intra- and interassay variation hardly ever exceeding 5% polarization (n > 100). If the data were discrepant in this way (less than 2% of assays), outcomes were not used and the assay was carried out again.

FACS analysis

The various monocyte fractions were analysed using a FACScan flow cytometer (Becton Dickinson) for the following markers using the following monoclonal antibodies: CD49a (VLA1, Serotec, Oxford, England); CD49b (VLA2, Serotec); CD49d (VLA4, Serotec); CD49e (VLA5, Serotec); CD29 (Beckman Coulter, Hialeah, FL, USA) and FITC-labelled FN. FN had been labelled with FITC (Sigma, St. Louis, USA) according to Bergquist et al. [16]. Mononuclear cells were incubated with the FITC-labelled FN and monoclonal antibodies for the detection of the membrane-bound integrins according to earlier described standard FACS analysis techniques [17]. The monoclonal antibodies had been directly conjugated with fluorescein isothiocynate (FITC) or phycoerythrin (PE) by the manufacturer.

Generation of veiled cells (VC)

Monocytes were suspended under plastic-nonadherent conditions (polypropylene tubes, Falcon, Becton Dickinson, USA) at a concentration of 2 × 10⁶/ml in RPMI 1640 (without FCS). T3 (Sigma) was added to the cell suspension at a concentration of 2 × 10⁻¹⁰ m to improve the yield of VCs. This suspension was then incubated at 37°C, 5% CO₂, 100% humidity for 30 min. This was followed by washing in FCS supplemented RPMI 1640 culture fluid (5 min, 500 g), and further culturing in the same culture fluid at 37°C, 5% CO₂, 100% humidity under nonadherent conditions for 16 h (overnight).

Following the 16 h culture period, cells were centrifuged (5 min, 500 g) and resuspended in FCS supplemented RPMI 1640. Cells were then examined under light microscopy at a magnification of 400 ×. Veiled cells were defined as large cells with actively moving cytoplasmic processes or veils, which can easily be identified (see Fig. 1). Measurements were performed blindly, and intra-assay variation remained below 2% (n > 50). Inter-assay variation ranged from 1 to 6%.

Fig. 1.

Time-lapse cinematographic pictures of a living monocyte-derived veiled cell (a–d) and an unchanged monocyte (e–h) in the same suspension (see for generation of these cells Materials and Methods). Pictures are taken 20–30 s apart. Magn. × 1000. Note the constant changes in shape of the veiled cell, extending and withdrawing long, thin cytosplasmic extensions (a–d); this in contrast to the rigid round shape of the unchanged monocyte (e–h).

Allogeneic MLRs were performed in order to measure the accessory capability of the generated VCs. Responder T lymphocytes were obtained from healthy donors and isolated following standard procedures with Ficoll-Isopaque, Percoll density gradient centrifugation, and nylon wool adherence (Leuko-Pak, Fenwall Laboratories, IL, USA). Non-adhering cells recovered were more than 90% CD3 positive. 1·5 × 10⁵ responder cells were cultured in 96-well, flat-bottom microtiter plates (NUNC A/S International, Denmark) with different numbers of irradiated (2000 rad) stimulator cells (VC) to achieve stimulator-to-T cell ratios of 1:5 and 1:10. The culture medium used was RPMI 1640 with 25 mm HEPES and l-glutamine, supplemented with 100 U/ml penicillin G, 0·1 mg/ml streptomycin and 10% human A⁺ serum, in a total volume of 200 μl per well. Controls used were VC alone, and lymphocytes in the presence of 10–50 μg/ml phytohaemagglutinin (PHA) (Wellcome Diagnostics, UK). Cultures were performed in triplicate. On day 5, thymidine incorporation was measured by adding 0·5 μCi 3H-thymidine to each well, then harvesting 16 h later. Scintillation was counted on an LKB 1205 Betaplate liquid scintillation counter (Wallac, Turku, Finland).

Statistical analysis

For statistical testing, the Mann–Whitney test was performed (two-tailed, Instat program). P < 0·05 was considered significant.

Results

The chemoattractant-induced cell polarization of monocytes of thyroid autoimmune patients is defective and does not improve, but worsens after fibronectin adherence

Monocytes of thyroid autoimmune patients were as capable of adhering to FN-coated surfaces as circulating monocytes of healthy control individuals. A mean value of 47% ± 10% (mean ± standard deviation, n = 20) adhering monocytes was found in healthy controls, whereas thyroid autoimmune patients showed a value of 48% ± 10% (n = 20). There was also no difference between the purities of the monocyte suspensions of thyroid autoimmune patients and of healthy controls neither before nor after FN adherence; values ranged from 80 to 95% monocyte-esterase positive cells before and over 95% monocyte-esterase positive cells after FN adherence.

We went on to test the chemoattractant-induced polarization of monocytes of thyroid autoimmune patients and healthy controls before and after FN adherence. As can be seen from Fig. 2, Percoll-purified monocytes of thyroid autoimmune patients were hampered in their polarization response towards fMLP before FN adherence (verifying our earlier data [1]): Values of 27% (mean, s.d. 11, n = 20) versus 35% (s.d. 6, n = 17) of polarized monocytes were found in thyroid autoimmune patients and healthy controls, respectively (P = 0·005). Both hypothyroid and hyperthyroid patients showed this phenomenon.

Fig. 2.

The percentage of monocytes that show a cytoskeletal rearrangement (cell polarization) under the influence of the chemoattractant fMLP. Means ± s.d. are shown of monocytes before (□) and after () fibronectin (FN) adherence of healthy controls (n = 17) and thyroid autoimmune patients (n = 20). *P < 0·01 versus monocytes prior to FN adherence, †P < 0·01 versus monocytes of healthy controls, ‡P < 0·001 versus FN-adhered monocytes of healthy controls.

After adherence to FN, the monocytes of healthy controls had a higher polarization capability as compared to the original monocytes: 45% (s.d. 5%, n = 17) of the cells now responded to the chemoattractant fMLP (Fig. 2). The monocytes of thyroid autoimmune patients, on the contrary, became (after FN adherence) extremely poor in their polarization capability, and only 18% of polarized monocytes (s.d. 11, n = 20) were found when stimulated with fMLP (Fig. 2).

In a few experiments we used the chemokine MCP-1 as a polarization-inducing signal. Monocytes of healthy controls again showed a good polarization response to MCP-1 (40–42%, at optimal MCP-1 dosage 20 ng/ml) after FN adherence. In 3 patients tested (2 hypo, 1 hyper) this response was again very low (11%, 29% and 9%, respectively), reinforcing the observation of a poor capability of the patient monocytes to rearrange their cytoskeleton upon a chemoattractant stimulus, particularly after FN adherence. It must also be noted that we found MIP1α and RANTES to be poor inducers of monocyte polarization both in healthy controls and patients.

The up-regulation of fibronectin receptors on monocytes of thyroid autoimmune patients is hampered after fibronectin adherence

Firstly it must be noted that the expression of ‘fibronectin receptors’, of the various CD49 markers and of CD29 was ‘dull’ on monocytes and at a moderate Mean Fluorescence Intensity (MFI) level of 60–150 as compared to the MFI expression levels we have seen previously on, e.g. dendritic cells (MFIs of 500–2000).

Although we had expected that FN-adherence would have acted as a method to separate ‘fibronectin receptor’ positive monocytes from ‘fibronectin receptor’ negative monocytes, we observed (Table 1) that the percentage of original monocytes that were capable to adhere to a FN-coated surface did not match with the percentage of original monocytes capable of binding FITC-labelled FN or expressing the relevant FN-binding CD49d, e and CD29 integrins. Also the expression of such binding sites and integrins was still considerable in the fraction of FN-nonadhered monocytes (both numerically and regarding the intensity of expression). Moreover the number of FITC-FN, CD49d, e and CD29 positive cells in the FN-adhered fraction was certainly not enriched to 90–100%, though it was significantly increased for FITC-FN, CD49d and CD49e positive cells (see Table 1). The number of CD29 positive cells was even lower in the FN-adhered fraction, but values did not reach statistical significance. There were also no differences with regard to the MFI of integrin expression after FN adherence. Furthermore, repetitive steps of FN-adherence did not alter these integrin expression data in a noteworthy manner. These observations are in line with the previously described plasticity of integrin expression with regard to the number and avidity/affinity of these receptors after binding to ECM proteins [6,7]. This process of regulation of integrin expression is called ‘inside-out’ signalling.

Table 1.

Fibronectin-related integrin expression (percentage of monocytes positive in FACS analysis) of fibronectin-adhered and nonadherent monocytes of healthy donors (n = 10)

	Original monocytes	Fibronectin- adhered cells	Fibronectin- nonadherent cells
Fibronectin-adhering cells	47 ± 10%
Fibronectin-FITC+	36 ± 16%	55 ± 20%*	16 ± 10%†
CD49d+	84 ± 12%	93 ± 4%*	68 ± 12%†
CD49e+	67 ± 18%	86 ± 10%*	42 ± 14%†
CD29	70 ± 23%	51 ± 19%	n.t.
CD49b+	55 ± 17%	73 ± 7%	32 ± 12%†
CD49a+	14 ± 9%	17 ± 10%	8 ± 5%

^*P < 0·05 versus original monocytes.

^†P < 0·05 versus fibronectin-adhered monocytes.n.t. = not tested.

Table 1 also shows that CD49a and CD49b (α integrin chains important in laminin binding) were not significantly up-regulated after adherence to FN, though there were significant differences with the FN-nonadherent cells for CD49b.

Figure 3 gives the data for the percentages of patient monocytes with ‘fibronectin receptors’ (able to bind FITC-labelled FN) before and after FN adherence in relation to values of healthy control monocytes. Also the percentages of monocytes expressing the CD49e, CD49d, CD29, CD49b and CD49a integrins are shown. Before FN adherence there were no differences between healthy controls and thyroid autoimmune patients with regard to the percentage and expression levels of monocytes expressing FITC-FN binding sites or CD49e, d, and b integrins. Remarkably, more monocytes of thyroid autoimmune patients expressed CD49a. The number of CD29-expressing monocytes was lower, but this did not reach statistical significance. Figure 3 also shows that – although monocytes of thyroid autoimmune patients adhered equally well to FN-coated surfaces (see before) – after FN adherence there were fewer monocytes from these patients with ‘fibronectin receptors’ as compared to healthy control monocytes. The same phenomenon was seen for CD49e positive monocytes. CD49d expression was raised to the same extent in patient and control monocytes after FN adherence. CD29 and CD49b expression were neither significantly altered before nor after FN adherence in patients and controls. CD49a expression stayed high after FN adherence in monocytes of thyroid autoimmune patients as compared to those of healthy controls.

Fig. 3.

The percentages of monocytes positive in FACS analysis for ‘fibronectin receptors’ (FN-FITC, see text), CD49e, CD49d, CD29, CD49b and CD49a integrins. Mean ± s.d. are given of original monocytes (first two columns of a set; □ healthy controls, ▪ patients) and of FN-adhered monocytes (last two columns of a set; healthy controls, patients) of healthy controls and patients. a = P < 0·05 versus monocytes before FN adherence, b = P < 0·01 versus normal monocytes of healthy controls, c = P < 0·05 versus FN-adhered monocytes of healthy controls.

The development of functionally active veiled cells from fibronectin-adherent monocytes is lower in thyroid autoimmune patients

On a previous occasion we reported that the recovery of veiled cells (VCs) from normal, nonfibronectin-adhered monocytes is in the same range in Graves' patients as in healthy controls [1]. We experienced the same observation in a few patients tested (n = 4) in the present series of experiments, and values in the range of 22–37% of cells with long, actively moving veils could be generated from monocytes both in patients and healthy controls (Fig. 4). Figure 4 also shows that the percentages of VCs that could be generated from monocytes was higher in healthy controls after FN adherence of the cells, and values of 46% ± 8% (n = 8) were obtained. These generated VC were good stimulators of the T cell proliferation in an allogeneic MLR (Fig. 5).

Fig. 4.

The percentages of veiled cells (VCs) generated from blood monocytes prior to (□) and after () FN adherence. Means ± s.d. are shown of healthy controls before FN adherence (n = 4) and after FN adherence (n = 8), and of thyroid autoimmune patients under both conditions (n = 20). a = P < 0·01 versus monocytes before FN adherence, c = P < 0·001 versus FN-adhered monocytes of healthy controls.

Fig. 5.

The T cell stimulatory potential of the VC populations generated from FN-adhered monocytes. Data are expressed as the capacity of the VCs to induce T cell proliferation in an allogeneic MLR (3H-thyminidine incorporation, cpm, means ± standard errors). Data are given of 7 healthy controls (O) and of 18 thyroid autoimmune patients (•). The horizontal axis gives the VC to T cell ratios. At both ratios the differences between the controls and the patients are significant (P = 0·003 for the 1 to 10 ratio and P = 0·02 for the 1 to 5 ratio).

When FN-adhered monocytes of thyroid autoimmune patients were used, the recovery of VCs was much poorer (Fig. 4, 31% ± 10%, P < 0·01 n = 20), and the cell population was less capable to stimulate T cells in allo-MLR (Fig. 5).

It must also be noted that there existed a good correlation between the polarization capability of a given population of FN-adhered monocytes on the one side and its capability to generate VCs and the capability of that VC population to stimulate T cells in allo-MLR on the other side; correlation coefficients of 0·48 and 0·47 were found, respectively (n = 30, P < 0·01). This underlines the notion that the capability of fibronectin-adhered monocytes to rearrange their cortical cytoskeleton is closely linked to the capability of that population to generate veiled T cell-stimulating APCs.

Discussion

Integrins and chemoattractant stimuli play a pivotal role when monocytes enter the tissues from the bloodstream. Via integrin ligation and integrin signalling monocytes are able to firmly adhere to endothelial cells [18], to form uropods [19], to respond more vigorously to chemoattractants [8,9], to migrate through the vessel wall and through the connective tissue via ECM fibres [20] and to differentiate more efficiently into antigen presenting cells (APCs) and macrophages [21]. Integrin-induced stimulation of 2nd messengers and rearrangements of the cytoskeleton are essential in many of these processes (‘outside-in’ signalling, 6,7).

After binding to ECM-fibres, integrin expression itself and the avidity and affinity of integrins to ECM proteins is up-regulated [6,7]. The significance and biochemical pathways of this so-called ‘inside-out’ signalling are poorly understood [6,7]. It has been suggested that ‘inside-out’ signalling plays a role in the ability of migrating cells to align ECM-fibres [22].

Our experiments indeed show that greater numbers of monocytes from healthy controls expressed FN binding sites, CD49e and CD49d integrins after FN adherence. The FN-adhered cells also showed an enhanced chemotactic polarization and an enhanced capability to differentiate into veiled accessory cells.

With regard to monocyte dysfunction in thyroid autoimmune disease, we show here that monocytes of patients with recently diagnosed thyroid autoimmune disease are hampered in these integrin – and chemokine – mediated functions: the cells were less able to rearrange their cytoskeleton after FN adherence, to up-regulate the above-named integrins and to generate veiled, T cell-stimulating APCs. Monocytes of both hypo- and hyperthyroid patients showed these defects. It is known that the thyroid hormone status is able to influence the function of monocytes and monocyte-derived cells [23,24]. However since both hyper- and hypothyroid patients showed these abnormalities and since we reported on a previous occasion on monocyte cytoskeletal rearrangement disturbances in euthyroid TPO-Ab-positive women [2], we presume that our data indicate that the autoimmune status per se and not the thyroid hormonal status determines the above-described monocyte abnormalities.

Together with our earlier report on a poor homotypic clustering capability of veiled/dendritic cells in thyroid autoimmune disease [1], also a function dependent of integrins [25,26], the present data strongly point to the existence of defects in the integrin and chemokine signalling pathways in monocytes and monocyte-derived cells of thyroid autoimmune patients. There are previous reports on a defective and altered expression of integrins on leucocytes in thyroid autoimmune disease [27,28]. However, these reports mainly focus on a defective integrin expression on lymphocytes and the consequences of such defective expression for a hampered lymphocytic adherence to endothelial cells, which has indeed been established for high endothelial venules in the thyroid-draining lymph nodes of thyroid autoimmune patients [29].

It is perhaps difficult to conceive why a general disturbance in monocyte function – as described here – should occur in a disease in which the abnormality of the immune system is an immune response mainly focused toward thyroid gland antigens. In explaining this, the NOD mouse model can be helpful. In the majority of NOD strains, no thyroiditis but only autoimmune insulitis occurs. The NOD mouse is – despite a mainly focused immune reaction towards islet antigens – characterized by a general immune disturbance including an abnormal development of DC and other APC from precursors, leading to suppressor T cell defects [30–32]. Only when the thyroid is manipulated such that monocytes, macrophages and DC are attracted to the thyroid (via iodine intoxication, 33) is thyroid autoimmune disease precipitated. Thus, local factors together with a general immune disturbance lead to full-blown thyroiditis in this animal model. This is likely also the case in humans, as iodine precipitates autoimmune thyroid disease in human beings as well.

Integrin and chemoattractant signalling activates a plethora of molecules involved in intracellular signal transmission and cytoskeletal rearrangements. In Wiskott–Aldrich syndrome (WAS), a rare disease characterized by thrombocytopenia, immune dysregulation and also autoimmunity, there is an inherited deficiency of the so-called WAS protein (WASp, reviewed in [34]). WASp is a molecule important in the signalling cascade from the cell membrane to the myoactin cytoskeleton, probably at the level of the Rho GTP-ase CDC42 and the cytoskeletal organizing complex Arp2/3. This deficiency leads – amongst other things – to an inability of monocytes to respond to a chemoattractant signal with cell polarization [3]. Also the appropriate maturation of dendritic cells (DC) from monocytes is hampered, and WAS-DC are defective in their ability to migrate on ECM matrixes [35]. The here-presented data on monocytic defects in thyroid autoimmune patients are reminiscent of those reported in WAS, and therefore urge for a more detailed study on putative molecular abnormalities in the signalling cascade ‘integrins and chemokine receptors → cytoskeletal rearrangements’ in monocytes of patients with thyroid autoimmune disease. Interestingly in the NOD mouse an impaired cell membrane targeting of the Grb2-sos complex has been found [36] (the Grb2-sos complex mediates the signal from integrins to the MAP-kinase pathway). In BB-DP rats, another animal model of autoimmune thyroiditis, an abnormal expression of vav was found [37] (vav mediates the signal from integrin-activated tyrosine kinases to the Rho-family of GTP-ases).

The present report also expands our previous observations on abnormalities of monocyte-derived APCs in thyroid autoimmune patients. Here we show that the generation of VCs from fibronectin-adhered monocytes is lower in patients with a recently diagnosed thyroid autoimmune disease as compared to healthy controls. The here-used method of generating VC from monocytes yields a type of APC which is different from the so-called ‘immature dendritic cells (DC)’ generated according to a method first published by the group of Lanzavecchia [38]. In the latter method, monocytes are cultured under plastic-adherent conditions for one week in the presence of GM-CSF and IL-4. The here-generated VC show many characteristics known of dendritic/veiled cells in ex vivo/in vivo situations [13]: they are MHC class II positive, have various markers known of Langerhans cells (e.g. S100) and have a clearly increased capability in stimulating T cells as compared to monocytes. They are almost as good as ‘Lanzavecchia’ DC in this respect [11,12]. The VCs, however, hardly produce the immunostimulating cytokine IL-12. They do produce large quantities of the immunosuppressive and immunomodulating cytokine IL-10 (to be published). It is thus tempting to speculate that the here-described defects in VC development from integrin-activated monocytes represent defects in the generation of APC with immunosuppressive and tolerating capabilities. Interestingly, similar abnormalities in APC generation exist in BB-DP rats [39] and NOD mice [30]. In these animals, these defects indeed lead to a faulty activation of T cells with an immunosuppressive/regulatory capability [30,31,39]. Further research is, however, needed to explore the latter hypothesis.