α4β1⁺ and α4β7⁺ CD4⁺ T cell numbers increase and CLA⁺ CD4⁺ T cell numbers decrease in systemic sclerosis

Abstract

We studied the expression of adhesion molecules affecting recirculation and homing on peripheral blood CD4⁺ T cells of patients with systemic sclerosis (SSc), in order to evaluate whether the distribution of tissue targeted subsets could reflect the participation of internal organs or the extent of cutaneous involvement [i.e. limited cutaneous (lc) and diffuse cutaneous (dc)]. Peripheral blood mononuclear cells (PBMC) from 51 patients with SSc and 19 sex- and age-matched controls were investigated by cytofluorimetric analysis for lymphocyte subpopulations carrying the following surface molecules: CD3, CD4, CLA, α4β7 and α4β1. Standard routine biochemistry and clinical examinations were also performed in all patients. We found that both α4β1⁺ and α4β7⁺ cells within the CD4⁺ T cell population were significantly increased, while CLA⁺ CD4⁺ T cells were significantly reduced in SSc, compared to healthy donors. Significantly lower absolute numbers of α4β7⁺ cells were found in lc- compared to dc-SSc. Patients with oesophageal involvement had high numbers of α4β7⁺ cells, while those with nephritis also showed low levels of CLA⁺ cells. Lung involvement was related directly to α4β1⁺ cell numbers and inversely to α4β7⁺ CD4 cell numbers. Taken together, our findings demonstrate that distinct CD4⁺ T cell populations with selective homing properties show changes from normal distribution in SSc, and such changes are related to clinical expression and organ involvement in the course of the disease.

Introduction

Systemic sclerosis (SSc) is a connective tissue disorder characterized by vascular changes and immunological abnormalities leading to progressive widespread fibrosis of the skin and other visceral organs [1,2]. Clinically, SSc can be divided into two major subtypes, the first characterized by a diffuse (dc-) skin fibrosis with a severe involvement of internal organs, such as lung, gastrointestinal tract, heart and kidney, and the second characterized by a limited (lc-) cutaneous involvement [3]. The progressive fibrosis is a consequence of multiple and partially unknown events that lead to increased biosynthesis of matrix proteins by connective tissue cells. Inflammatory cells are recruited into the perivascular dermal matrix of lesional skin in scleroderma and T cells predominate in these infiltrates, which are often found early in the course of disease [4]. Circulating T cells display an activated phenotype in SSc, as demonstrated by the increased expression of CD25, HLA-DR and CD29 [5], and may promote fibrosis through the production of several soluble mediators [6–9].

It is well established that activated T lymphocytes have different capacities to traffic in mucosal or non-mucosal (e.g. cutaneous) sites. Much of this migration occurs through high endothelial venules (HEVs), and is regulated by multi-step cascades involving sequential lymphocyte/endothelial adhesion and activation events [10–12]. The tissue-selective homing is regulated at the level of lymphocyte recognition of post-capillary venular endothelial cells via the interaction of differentially expressed lymphocyte homing receptors and their endothelial cell ligand.

The T lymphocyte molecule thought to participate in ‘skin selective’ lymphocyte homing is called the cutaneous lymphocyte-associated antigen (CLA), a sialyl Lewis^x-related carbohydrate dependent epitope, which interacts with E-selectin in endothelial cells [13,14]. On the other hand, lymphocytes expressing the integrin α4β7 bind the mucosal vascular addressin MAdCAM-1 [15,16] selectively expressed on HEV in mesentheric lymphnodes, intestinal lamina propria and Peyer's patches [17], allowing the selective homing to both intestinal secondary lymphoid tissue and intestinal effector sites. α4β1, expressed by almost all lymphocytes at moderate to high levels, and on monocytes and eosinophils, is a receptor for the immunoglobulin family adhesion ligand vascular cell adhesion molecule-1 (VCAM-1) for fibronectin and for the α4 integrin chain itself [18]. VCAM-1 is induced on the endothelium during inflammation. Recently it has been demonstrated that VCAM-1 is expressed on HEVs in bronchus-associated lymphoid tissue (BALT) at a much higher level than on HEVs in other secondary lymphoid tissues. Therefore the α4β1/VCAM-1 pathway may support the recruitment of activated T cells from the bloodstream into BALT [19].

Taken together, the expression of CLA, α4β7 and α4β1 might play a critical role in defining lymphocytes subsets that participate in mucosal versus non-mucosal immune responses [20]. It has been suggested that aberrant expression of cell adhesion molecules on lymphocytes, monocytes and endothelial cells may drive the accumulation of activated leucocytes in the tissues of SSc, which is a crucial step involved in fibrosis induction [21].

We examined the distribution of tissue targeted CD4⁺ T cell subsets in peripheral blood mononuclear cells (PBMC) to address the possible use of homing molecules expression as tools to discriminate between limited and diffuse form of SSc and as correlates of internal organ involvement.

Patients and methods

Patients

Fifty-one consecutive patients (49 females and two males) and 19 healthy subjects (15 females and four males) were included in the study. The mean age ± standard deviation (s.d.) in the total patient group was 53 ± 1·1 years; the mean age of the controls was 48 ± 11 years. All scleroderma patients were first evaluated by experienced physicians in our division and met the American College of Rheumatology (formerly the American Rheumatism Association) criteria for the classification of systemic sclerosis (scleroderma) [3,22]. Patients were classified as either dc-SSc, showing cutaneous thickening proximal to the elbow, or lc-SSc, characterized by skin sclerosis restricted to the distal extremities.

Clinical assessment

A complete history, physical examination and laboratory evaluation were obtained in each patient. The disease duration was measured from the onset of the first symptom other than Raynaud's phenomenon consistent with SSc (puffy hands, digital pitting scars, sclerodactily with or without proximal scleroderma, dyspnoea, dysphagia).

Pulmonary involvement was defined as a forced vital capacity <70% of predicted, as measured by a dry spirometer, or by other confirmatory evidence of fibrosis such as bibasilar fibrosis on chest radiograms and/or amorphous or reticulonodular infiltrates in high resolution computed tomography scans (HRCT). The HRCT was scored as follows: 0 = no modifications, 1 = mild alterations (ground glass and/or linear opacities of 1–2 mm thickness), 2 = moderate alterations (with linear opacities of 2–3 mm thickness) and 3 = severe alterations (linear opacities of >3 mm and the presence of honeycombing). Pulmonary hypertension (PH) was evaluated indirectly by Doppler echocardiography, according to Denton et al. [23]. Renal involvement was considered as ‘present’ when at least one of the following conditions was documented: hypertensive renal crisis attributable to SSc, or persistent renal impairment (serum creatinine value >2 mg/dl plus diastolic blood pressure >110 mmHg) or otherwise inexplicable presence of blood and/or protein in the urine. Oesophageal dismotility was diagnosed by manometry. Three degrees of oesophageal motor disorders were identified: normal, moderate (uncoordinated peristalsis) and severe (both aperistalsis and decreased low oesophageal sphincter pressure). Joint involvement was considered as ‘present’ in the presence of symmetric synovitis, flexion contractures and tendon friction rubs.

The study was approved by the Istituto Dermopatico dell'Immacolata (IDI-IRCCS) Institutional review board, and informed consent was obtained from all patients.

PBMC samples

PBMC from both patients and healthy donors were isolated from venous blood by density gradient sedimentation on Lymphoprep (Pharmacia, Uppsala, Sweden), resuspended in phosphate buffered saline (PBS) (Gibco, Grand Island, NY, USA) supplemented with 2% v/v fetal calf serum (FCS, Gibco, Paisley, Scotland, UK) at 10⁶ cells/ml.

The following monoclonal antibodies (MoAb) were used: fluorescein isothiocyanate (FITC)-conjugated anti-CD3, CLA, CD49d and CD29, R-phycoerythrin (PE) conjugated CD49d and β7, peridin chlorophyll protein (PCP) conjugated anti-CD4 (Becton Dickinson Immunocytometry Systems, Mountain View, CA, USA).

Flow cytometric analysis

Three-colour flow cytometry analysis was performed as described previously [24]. Briefly, 2 × 10⁵ lymphocytes from PBMC were delivered into the test tube containing 20 µl of each MoAb (FITC, PE, PCP or PE-Cy5 conjugated) and placed in an ice water bath for 20 min. After two washes with PBS containing 2% FCS and 0·02% NaN₃, cells were resuspended in 300 µl PBS containing FCS and NaN₃ for immediate flow cytometry analysis. The following MoAbs were used: CD4 (RPA-T4 mouse IgG_1k), CLA (HECA-452 rat IgM_k), CD29 (MAR4 mouse IgG_1k), CD49d (9F10 mouse IgG_1k) and integrin β7 (FIB504 rat IgG_2ak), all purchased from Becton Dickinson Immunocytometry Systems). Irrelevant mouse IgG_1k, rat IgM_k and rat IgG_2ak, MoAb (Becton Dickinson) FITC, PE, PCP or PE-Cy5 conjugated were used as negative controls. Quantitative analysis for three-colour flow cytometry was carried out using a FACSCalibur instrument (Becton Dickinson). At least 5000 events were acquired in list mode, and all the data were analysed by cellquest software (Becton Dickinson). Live lymphomonocytes were electronically gated by forward- and right-angle scatters, CD4⁺ cells were then gated and the percentage and intensity of expression of β1 (CD29) and β7 cells were evaluated as histograms within the CD49d (α4)⁺ subpopulation (Fig. 1). A single fluorescence analysis was performed to quantify CD4⁺ cells and absolute number calculated from total white blood cell (WBC) count and percentage lymphocytes (gating by scatter and CD45 control).

Fig. 1.

The three colour flow cytometry of CD4/α4β1 and α4β combination on a healthy donor and SSc patient PBL. Live lymphocytes were gated on forward and side light scatters. CD4⁺ cells were then gated, and the α4 (CD49d) cells within the CD4⁺ subset were selected. The histograms represent the CD29 (β1) and integrin β7 expression within the CD4+CD49d+ T cells. (a) Healthy donor; (b) SSc subject.

Statistical analysis

Data were entered into an electronic database and analysed using the SPSS/PC+ statistical package [25]. Socio-demographic and clinical differences between patients with lc-SSc and dc-SSc were investigated using the χ² test. The levels of the variables of interest in the two classes of SSc patients as well as in controls were summarized showing the mean ± s.d., in order to provide the most easily interpretable summary measure. However, differences between the levels of these variables among lc-SSc, dc-SSc and controls were tested using the non-parametric Mann–Whitney U-test.

The association between T cell subpopulations and the different internal organs involved was studied in univariate and multivariate analysis. In univariate analysis, we first used the above-mentioned non-parametric test to compare variables in subjects with or without organ involvement; then, each variable of interest was dichotomized (as ‘high’ and ‘low’ levels) using the median value observed among the controls as cut-off point to study the proportion of subjects with organ involvement in the two groups thus obtained. Differences in the two groups were analysed using the χ² test. Finally, to see whether the association between organs involved and different biological variables was present after simultaneously adjusting for the other variables of interest, we performed multiple linear regression analysis for the clinical variables with linear scores (i.e. lungs and oesophagus) and multiple logistic regression for the clinical variables with dichotomous scores (i.e. joints and kidney).

Although multiple pairwise comparisons were performed, no adjustments were made to control for overall type I error, according to the recommendations of Rothman [26,27] and Savitz [28], and therefore we set the significance cut-off level at P < 0·05.

Results

Patients' characteristics

Thirty-two patients (all female) had lc-SSc and 19 (17 females and two males) had dc-SSc (Table 1). The disease duration from the onset time to study entry was 10·8 ± 8·9 years (mean ± s.d.; range 1–35 years). None of the patients had a recent history of infection or other inflammatory diseases, including connective tissue disease. No patients had been treated with disease-modifying therapy for 2 months. All patients had Raynaud's phenomenon after exposure to low temperature. Antinuclear antibodies, demonstrated by immunofluorescence, were present in the sera of all patients. At the time of the investigation, anticentromere antibodies (ACA) were detected in 13 patients with lc-SSc, while anti-topoisomerase I antibodies (scl70) were present in 11 with lc-SSc and 17 with dc-SSc. The patients included in the study had been evaluated for organ involvement at least 1 month before blood collection. Forty patients had respiratory involvement. PH was found in 17 patients. Kidney involvement was found in 17 patients. Thirty-one patients had oesophageal dismotility. Joint involvement was observed in 32 individuals.

Table 1.

T cell subsets within the CD3⁺ CD4⁺ population

	HD (19)	SSc (51)
CD4+ %	53 ± 6·5	53 ± 8·6
(Abs. no.)	(1223 ± 341)	(1001 ± 381)*
CD4+ CLA+	7 ± 3·6	4 ± 2**
(Abs. no.)	(83 ± 52)	(37 ± 24)**
CD4+ α4β1+	59 ± 8·8	70 ± 10·5**
(Abs. no.)	(725 ± 237)	(696 ± 272)
CD4+ α4β7+	20 ± 6·5	29 ± 10·5**
(Abs. no.)	(246 ± 112)	(271 ± 144)

^*P = 0·029;

^**P < 0·0001.

Adhesion molecules evaluation by flow cytometry

We investigated by FACS analysis the distribution of well-characterized and antigenically defined tissue targeted T cell subpopulations within CD4⁺ lymphocytes, i.e. gut committed α4β7⁺ cells, preferentially BALT committed α4β1⁺ cells and skin-associated cells expressing the CLA. CD4⁺ lymphocytes absolute numbers in SSc patients were significantly reduced in comparison to healthy donors, while the percentage levels were identical. We found a significant reduction of CD4⁺ cells bearing CLA in SSc (Table 1) in comparison to healthy donors, when considering percentages (P < 0·001) and absolute numbers (P < 0·001). On the other hand, the percentages of both α4β1⁺ and α4β7⁺ cells within the CD4⁺ T cell population were increased significantly (Table 1 and Fig. 1) when compared to controls. Interestingly, a direct correlation between the absolute number of CD4⁺ CLA⁺ lymphocytes and CD4⁺ α4β7⁺ cells (r 0·303, P = 0·0455) or CD4⁺ α4β1⁺ cells (r 0·479, P < 0·0001) was found. Finally, a significantly (P < 0·05) higher absolute number of α4β7⁺ cells was found in dc-SSc, compared to lc-SSc (Table 2). Using receiver operating characteristic (ROC) curve analysis we identified 247 cells as the optimal cut-off to identify dc-SSc with a sensitivity of 74% and specificity of 76% (with a likelihood ratio of 3·1).

Table 2.

T cell subsets within CD3⁺ CD4⁺ population in present (yes) or absent (no) organ involvement

			Lung		Oesophagus		Kidney
	lc-SSc	dc-SSc	No	Yes	No	Yes	No	Yes
CLA+	3·9 ± 2·2	3·3 ± 2·1	4·3 ± 2·1	3·5 ± 2·1	3·9 ± 2·1	3·5 ± 2·1	4·1 ± 1·9	2·8 ± 2·1*
(Abs. no.)	36 ± 24	38 ± 26	43 ± 26	36 ± 24	39 ± 23	36 ± 26	41 ± 21	31 ± 29
a4b1+	71·1 ± 9·3	67·8 ± 11·7	66 ± 13	70·9 ± 8·5	66·9 ± 10·4	71·9 ± 9	69·6 ± 10	67·2 ± 10·2
(Abs. no.)	670 ± 284	731 ± 263	657 ± 272	727 ± 277	695 ± 251	722 ± 296	709 ± 241	693 ± 351
a4b7+	26·9 ± 31·6	30·9 ± 13·9	27·7 ± 9	29·2 ± 11·5	24·8 ± 7·8	31·6 ± 11·8*	25·9 ± 8·7	31·5 ± 11·4
(Abs. no.)	235 ± 93	324 ± 179*	252 ± 101	285 ± 159	250 ± 96	296 ± 173	248 ± 112	309 ± 181

^*P < 0·05.

Association of immunological parameters with internal organ involvement

We then assessed the association of CD4⁺ T cell subsets with the prevalence of internal organ involvement in patients with SSc.

We found a significantly lower percentage of CLA⁺ CD4⁺ T cells in patients with kidney involvement, while high percentages of α4β7⁺ CD4⁺ T cells were observed in patients with oesophageal dismotility (Table 2). The cut-off level of 30%α4β7⁺ CD4⁺ T lymphocytes was found to be optimal to discriminate the presence of oesophagus involvement, with a sensitivity of 60% and specificity of 82·4% (with a likelihood ratio of 3·4).

When we performed a multivariate linear regression analysis, we found that the presence of pulmonary fibrosis was inversely related to the percentage of α4β7 cells (standardized linear regression coefficient β =–0·455; P = 0·013), but directly correlated to that of α4β1 CD4⁺ T cells (standardized linear regression coefficient β = 0·506; P = 0·006). Furthermore, we observed that the absolute number of α4β7 CD4⁺ lymphocytes was directly correlated to oesophageal involvement (standardized linear regression coefficient β = 0·563; P = 0·045), while both the absolute number (standardized logistic regression coefficient β = 0·0097; P = 0·046) and percentage of this subset (standardized logistic regression coefficient β = 0·01038; P = 0·0481) were also directly related to kidney involvement.

Discussion

Perivascular infiltrates of inflammatory cells are a hallmark of lesional tissues in SSc. These infiltrates can be observed often before fibrosis occurs and are constituted mainly of T cells [4]. Both CD4⁺ T cells and CD8⁺ T cells are found in the skin of patients with SSc, but CD4⁺ T cells represent the principal subset [4]. In this work we investigated the distribution of the skin and mucosal committed CD4⁺ T lymphocytes subsets in the peripheral blood from patients with SSc.

We found a significant increase of both α4β1 and α4β7-expressing CD4⁺ cells in the peripheral blood of SSc patients (Table 1 and Fig. 1). Recently, T cells overexpressing α4 (CD49d) have been demonstrated to be autoreactive and able to proliferate in response to APCs in a major histocompatibility complex (MHC) class II-dependent manner in the absence of nominal Ag [29]. In addition to their adhesion and T cell migration functions, the integrins are increasingly recognized for their role in T cell activation [30]. Taken together, the increased levels of α4-expressing cells observed might represent an indirect proof of autoreactive T cell proliferation in SSc patients.

We observed high levels of α4β7 CD4⁺ cells in patients with oesophagus dismotility (Table 2), while in the case of a predominant lung involvement, an inverse correlation with this population was found. Furthermore, α4β7 CD4⁺ cells absolute number allows the discrimination between limited and diffuse forms of the disease (Table 2). Peripheral blood lymphocytes display heterogeneous patterns of in vivo trafficking, determined in part by differential expression of cell surface homing receptors or adhesion molecules that control their interactions with tissue-selective vascular addressins at sites of lymphocyte recruitment from the blood [11]. The integrin molecule α4β7 is expressed selectively on a subset of human CD4⁺ memory T cells with gut tropism [31]. The importance for homing to this mucosal site may be explained by its ability, in contrast to α4β1, to bind the vascular addressin MAdCAM-1. MAdCAM-1 is expressed selectively on venules in mucosal tissues, defining sites of lymphocytes extravasation into Peyer's patches, appendix, mesenteric lymphnodes, intestinal lamina propria and lactating mammary gland. Conversely VCAM-1, the ligand of α4β1, is induced in diverse sites of inflammation but it is rarely observed on venules at mucosal sites. Vascular MAdCAM-1 and VCAM-1 expression therefore helps to define the physical segregation of mucosal versus non-mucosal immune responses in vivo by controlling the homing of specialized memory and effector cell subsets [20]. The regulated expression of α4β7 and of α4β1 plays a critical role in defining lymphocytes subsets that participate in mucosal or non-mucosal immune responses. In this light, the high percentages of α4β7-expressing cells in patients with oesophagus involvement might allow to suggest a direct role of this cell subset, with intestinal tropism, in the pathogenesis of dismotility. However, it should be taken into consideration that we measured the distribution of α4β7 and α4β1 CD4⁺ T cell subset in the peripheral blood and it would also be necessary to study such lymphocyte subsets in situ before any definitive conclusions are drawn.

We found that the increase of α4β1 CD4⁺ T cells in peripheral blood was related to the presence of pulmonary fibrosis. It has been demonstrated recently that high endothelial venules (HEVs) in BALT express VCAM-1 at a much higher level than on other secondary lymphoid tissues [19]. The preferential expression of VCAM-1 by BALT HEVs suggests that lymphocytes may migrate to BALT mainly using the α4β1 integrin/VCAM-1 pathway. In fact, antibodies directed against α4 integrin and VCAM-1 block the migration of total T cells, but not naive T and B cells, to BALT [32]. In contrast, α4β7 integrin and MAdCAM-1 are not involved in lymphocyte homing to BALT. Interestingly, Ihn et al. [33] demonstrated that elevated serum levels of soluble VCAM-1 in SSc correlated with the presence and severity of pulmonary fibrosis, thus implying ongoing activation of endothelial cells in SSc. The association between α4β1 CD4 lymphocytes and the presence of pulmonary fibrosis should be further investigated in situ to define whether it has a relevant pathogenetical role in this condition.

Finally, we have found a significant decrease of skin-associated CD4⁺ T cells, defined by the cutaneous lymphocyte antigen CLA, in SSc patients, irrespective of the extent of skin involvement (Table 1). Interestingly, the lower levels of circulating CLA⁺ CD4⁺ cells were found in patients with nephritis (Table 2). In several T cell-mediated skin diseases CLA is an important skin homing epitope for T cells. Patients with severe atopic dermatitis have a significantly expanded proportion of CLA⁺ T cells [34,35] and expression of CLA among CD8⁺ T cells is significantly elevated in patients with psoriasis [36]. Our finding suggests that the migration of T cells to skin in systemic sclerosis does not reflect an increase of circulating skin-homing CD4⁺ cells. CLA epitopes are not necessarily responsible for T cell binding to E-selectin and subsequent T cell migration to the skin, and therefore other surface molecules may be involved in SSc patients [37–39].

Our study revealed a profound imbalance in the expression of adhesion molecules by circulating CD4⁺ T cells from patients with SSc. These findings show that different subset changes are associated with different organ involvement, suggesting their possible role in the clinical picture. Considering other recent data pointing to abnormal lymphocyte migration patterns in the pathogenesis of SSc [40], our results may prove useful for further research at the molecular level.