Chemokines and common variable immunodeficiency; possible contribution of CCL19, CCL21 and CCR7 to immune dysregulation

Authors


B. Fevang, Research Institute for Internal Medicine, Rikshospitalet, University of Oslo, N-0027 Oslo, Norway.
E-mail: Borre.Fevang@medisin.uio.no

Summary

Common variable immunodeficiency (CVID) is a heterogeneous syndrome characterized by defective immunoglobulin production and high frequency of bacterial infections, autoimmunity and manifestations of chronic inflammation. The homeostatic chemokines CCL19 and CCL21 and their receptor CCR7 are associated with modulation of inflammatory responses. CVID patients have decreased proportions of CCR7+ T cells in peripheral blood and we hypothesized a further dysregulation of CCL19/CCL21/CCR7 in CVID. Serum levels of CCL19 and CCL21 were compared in CVID patients and controls. T cell expression of CCR7 was related to clinical characteristics in CVID patients. Spleens extirpated from CVID patients were analysed for expression of CCL19, CCL21 and CCR7. Peripheral blood mononuclear cells (PBMC) from CVID patients and controls were analysed for cytokine response on stimulation with CCL19 and CCL21. The main findings were: (i) CVID patients have raised serum levels of CCL19 and CCL21 independently of features of chronic inflammation; (ii) CCL19 and CCR7 have similar expression in spleens from CVID patients and controls, while CCL21 is variably down-regulated in spleens from patients; (iii) T cell expression of CCR7 is particularly low in patients characterized by chronic inflammation in vivo; and (iv) PBMC from CVID patients had attenuated cytokine response to stimulation with CCL19 and CCL21. CVID patients have raised circulatory levels of CCL19 and CCL21, and an attenuated cytokine response to stimulation with these chemokines. Because CCR7, CCL19 and CCL21 are key mediators balancing immunity and tolerance in the immune system, the abnormalities of these mediators might contribute to the profound immune dysregulation seen in CVID.

Introduction

Common variable immunodeficiency (CVID) is a heterogeneous syndrome characterized by failure of B cell differentiation and defective immunoglobulin (Ig) production leading to recurrent bacterial infections, particularly in the respiratory tract. Although reduced Ig secretion from B cells is the hallmark of CVID, other immunological abnormalities such as T cell dysfunction and monocyte/macrophage hyperactivity are seen in a considerable proportion of patients. These abnormalities may be of importance for both B cell deficiency and for some of the clinical manifestations in these patients [1].

Chemokines, a group of cytokines that attracts and activates leucocytes into inflamed tissue, have been associated with the pathogenesis of a number of diseases, ranging from atherosclerosis to human immunodeficiency virus (HIV) infection [2]. While most chemokines have been linked to inflammatory processes in peripheral tissue, the homeostatic chemokines CCL19 and CCL21 and their corresponding receptor CCR7 have been associated with development and maintenance of secondary lymphoid organs as well as the entry of lymphocytes and dendritic cells to secondary lymphoid tissue [3,4]. Recently, however, reports have pointed to a broader role for these homeostatic chemokines, including modulation of inflammatory and anti-inflammatory responses in lymphoid and non-lymphoid tissue. An imbalanced regulation of CCL19 and CCL21 has been suggested to be involved in the pathogenesis of various inflammatory disorders including rheumatoid arthritis, inflammatory bowel diseases and atherosclerosis [5–9]. While we and others have shown a decreased proportion of CCR7+ T cells in peripheral blood from CVID patients, characterizing a shift towards an inflammatory T cell phenotype [10–12], there are no reports regarding CCL19 and CCL21 in CVID.

To elucidate further the possible role of homeostatic chemokines in CVID, we analysed the levels of CCL19 and CCL21 in CVID patients and healthy controls, as well as the effect of these chemokines on cellular cytokine responses from these two groups of individuals.

Materials and methods

Patients

Sixty-two patients with CVID, diagnosed according to the criteria of the World Health Organization expert group for primary immunodeficiencies [1], attending the Section of Clinical Immunology and Infectious Diseases, Rikshospitalet University Hospital, Oslo, Norway, were included in the study (Table 1). The patients did not have any clinically apparent infection when recruited and did not receive corticosteroids or prophylactic treatment with antibiotics. In patients who received intravenous immunoglobulin (IVIG) substitution, typically with an interval of 3–4 weeks, the blood samples were taken immediately before IVIG administration. We have described previously a subgroup of patients (CVIDHyper) characterized by chronic inflammation in vivo as reflected by splenomegaly and increased serum levels of inflammatory markers [13–15]. Patients were classified as CVIDHyper if they had splenomegaly (>13 cm on ultrasonographic or CT scan examination) and serum neopterin levels greater than the mean ± 4 standard deviations (s.d.) of healthy controls (11·5 nmol/l), while other patients were classified as CVIDNorm. The diagnosis of bronchiectasis was based on typical findings on high-resolution computerized tomography (CT) scan of the thorax [16]. Twenty-seven sex- and age-matched healthy individuals were included as controls (Table 1).

Table 1.  Characteristics of the study group.
 Controls (n = 27)CVIDNorm (n = 41)CVID Hyper (n = 21)P-value*
  • *

    Categorical data compared using the χ2 test. Continuous data compared using the Kruskal–Wallis test.

  • Patients classified according to the EUROCLASS trial [19]. Data available from 21 patients. SmB: switched memory B cells; smB: smB ≤ 2% of B cells; smB+: smB > 2% of B cells.

  • Cases include patients which are splenectomized

  • §

    § Cases include any history of idiopathic thrombocytopenic purpura, autoimmune thyroiditis, haemolytic anaemia and autoimmune colitis.

  • Some patients on intravenous immunoglobulin (IVIG) received supplementary subcutaneous immunoglobulin therapy (SCIG). CVID, common variable immunodeficiency.

Demographics
 Gender (male/female, %)48/5251/4957/430·823
 Age (years, median and 25th to 75th percentiles)38 (33–50)46 (33–54)41 (29–52)0·440
Immunological characteristics (median and 25th to 75th percentiles if not stated otherwise)
 CD3+ cell count (×106/ml) 1386 (1009–2131)947 (775–1720)0·047
 CD4+ cell count (×106/ml) 735 (499–1007)632 (340–950)0·271
 CD8+ cell count (×106/ml) 638 (421–1051)460 (277–866)0·086
 CD19+ cell count (×106/ml) 246 (186–495)110 (37–178)<0·001
 Proportion of smB/smB+ patients (%) 29/7150/500·378
 IgG (g/l) 2·10 (1·25–3·90)1·90 (1·50–3·05)0·982
 IgA (g/l) 0·18 (0·10–0·30)0·13 (0·06–0·29)0·470
 IgM (g/l) 0·20 (0·10–0·34)0·20 (0·19–0·30)0·566
Clinical symptoms
 Bronchiectasis (%) 49570·533
 Splenomegaly (%) 22100<0·001
 Autoimmune disease (%)§ 44620·180
Therapy
 SCIG/IVIG (%) 83/1752/480·011

Informed consent to blood sampling was obtained from all subjects. The study was conducted according to the ethical guidelines at our hospital, which comply with the Helsinki Declaration, and was approved by the regional ethical committee.

Blood sampling protocol

Peripheral venous blood was drawn into pyrogen-free tubes without additives. The tubes were immersed immediately in melting ice and allowed to clot before centrifugation at 1500 g for 10 min. All serum samples were stored at –80°C and thawed <3 times.

Isolation of cells

Peripheral blood mononuclear cells (PBMC) were obtained from heparinized blood by Isopaque-Ficoll (Lymphoprep; Nycomed Pharma, Oslo, Norway) gradient centrifugation. Further isolation of CD3+ and CD14+ cells for polymerase chain reaction (PCR) analyses was performed as described previously [17]. For flow cytometry (see below), PBMC were cryopreserved in liquid nitrogen [18].

Real-time quantitative reverse transcription–polymerase chain reaction (RT–PCR)

Total RNA was extracted from monocytes and T cells that had been stored in liquid nitrogen using RNeasy columns (Qiagen, Hilden, Germany), subjected to DNase I treatment (RQI DNase; Promega, Madison, WI, USA) and stored in RNA storage solution (Ambion, Austin, TX, USA) at –80°C. Primers were designed using the Primer Express software, version 2·0 (Applied Biosystems, Foster City, CA, USA) for CCL19 [forward primer (FP): 5′-CCTCAGCCTGCTGGTTCTCT, reverse primer (RP): 5′-CAGCAGTCTTCAGCATCATTGG-3′] and CCL21 (FP: 5′-CCAAGCTTAGGCTGCTCCAT, RP 5′-TGCACATAGCTCTGCCTGAGA-3′). Quantification of mRNA was performed using the ABI Prism 7500 (Applied Biosystems). The relative standard curve method was used to calculate the relative gene expression. SyBr green assays were used for quantification and the specificity of the SyBr green assays was assessed by melting-point analysis and gel electrophoresis. Gene expression of the housekeeping gene β-actin (Applied Biosystems) was used for normalization (FP: 5′-AGGCACCAGGGCGTGAT, RP: 5′-TCGTCCCAGTTGGTGACGAT).

Flow cytometry

Cryopreserved PBMC were thawed as described elsewhere [18]. Staining of PBMC for evaluation of CCR7 expression on CD3+ cells and proportion of switched memory B cells was performed as described previously [10,17]. Flow cytometry was performed using a FACSCalibur instrument with CellQuest software (Becton Dickinson, San Diego, CA, USA).

Cell cultures for evaluation of cytokine release

Freshly isolated PBMC were resuspended in RPMI-1640 (Gibco, Paisley, Scotland, UK) with 2 mM L-glutamine and 25 mM HEPES buffer and 10% fetal calf serum (TCS BioSciences, Buckingham, UK) and seeded in 96-well plates (2 × 106 cells/ml; 0·2 ml/well; Costar, Cambridge, MA, USA) with or without the CCL19 and CCL21 (final concentration 100 ng/ml for both; R&D Systems, Minneapolis, MN, USA), and at the same time a suboptimal dose of phytohaemagglutinin (PHA; Murex, Dartford, UK; final concentration 22·5 ng/ml) as determined by previous assays or a combination thereof. Cell-free supernatants were harvested after 20 h and stored at –80°C until analysis.

Immunohistochemistry

Extirpated spleens from CVID patients with intractable thrombocytopenia and splenomegaly [n = 9, median weight 1157 g (range 425–2300 g)] were compared to spleens extirpated from patients during aortic surgery due to peri-operative spleen rupture [n = 3, median weight 93 g (range 77–107 g)]. Histologically, red and white pulp contributed differently to spleen enlargement in individual CVID patients: in four patients increased spleen size was caused by hyperplasia of the red pulp, in two cases by an enlarged white pulp due to follicular hyperplasia and in three cases, both red and white pulp equally contributed to the increased spleen size. Slides were stained with monoclonal mouse anti-human antibodies to CCR7 and CCL19 and polyclonal goat anti-human CCL21 (all from R&D Systems). Non-specific IgG isotypes were used as negative controls.

Enzyme immunoassays (EIAs)

Concentrations of CCL19, CCL21, tumour necrosis factor (TNF)-α, interleukin (IL)-8, IL-10, interferon (IFN)-γ and monocyte chemoattractant protein 1 (MCP-1) were measured by EIAs obtained from R&D Systems. Serum levels of von Willebrand Factor (vWF) were analysed by EIA as described previously [17]. The inter- and intra-assay coefficients of variation were <10% for all EIAs.

Statistical methods

For comparison of two groups, the Mann–Whitney U-test was used. When more than two groups of individuals were compared, the Kruskal–Wallis test was used. If a significant difference was found, the Mann–Whitney U-test was used to calculate the difference between each pair of groups. Paired data were analysed using the Wilcoxon signed-rank test. Coefficients of correlation were calculated by Spearman's rank test. Data are given as median and 25th to 75th percentiles unless stated otherwise. Probability values are two-sided and considered significant when <0·05.

Results

Serum levels of CCL19 and CCL21 in CVID patients and healthy controls

As shown in Fig. 1a and b, serum levels of CCL19 and CCL21 were raised significantly in CVID compared to healthy controls. Both patients with and without bronchiectasis had higher levels of CCL19 than controls, with particularly high levels in those with bronchiectasis (P = 0·007 versus other CVID patients) (Fig. 1c). In contrast, no such association with bronchiectases was seen for CCL21 (Fig. 1d). The CVIDNorm and CVIDHyper subgroups showed similar levels of both CCL19 and CCL21 (Fig. 1e and f). There was a strong correlation between serum levels of CCL19 and CCL21 within the patient group (r = 0·37, P = 0·006) (Fig. 2a), but no correlation between these chemokines and the proportion of CCR7+ T cells within the whole CVID group (r = 0·20, P = 0·29 and r = 0·02, P = 0·91; CCL19 and CCL21, respectively) (Fig. 2b and c). No differences were seen in the levels of CCL19 and CCL21 in subgroups of CVID based on levels of switched memory B cells (data not shown). There were similar levels of the endothelial cell activation marker vWF in patients and controls, and there was no correlation between levels of CCL19/CCL21 and vWF within the patient group (data not shown). There were no differences in neopterin levels between patients with and without bronchiectasis [S-neopterin (median and interquartile range) 11·16 (7·714–19·41) nmol/l versus 10·11 (6·474–21·37) nmol/l, P = 0·706].

Figure 1.

Serum levels of CCL19 (a) and CCL21 (b) in common variable immunodeficiency (CVID) patients (n = 62) and healthy controls (n = 27). Serum levels of CCL19 (c) and CCL21 (d) in CVID patients with (n = 32) and without (n = 30) bronchiectases. Serum levels of CCL19 (e) and CCL21 (f) in CVID patients classified as CVIDHyper (n = 21) and CVIDNorm (n = 41) (see Materials and methods for details). Horizontal line represents median value. *P < 0·01 versus patients without bronchiectases. **P < 0·001 versus controls.

Figure 2.

Correlation between serum levels of CCL19 and CCL21 in common variable immunodeficiency (CVID) patients (a). Correlation between expression of CCR7 on CD3+ cells and serum levels of CCL19 (b) and CCL21 (c) in CVID patients.

mRNA expression of CCL19 and CCL21 in monocytes from CVID subgroups

CD14+ monocytes from randomly selected CVIDNorm (n = 15) and CVIDHyper (n = 11) patients showed a trend towards enhanced expression of CCL19 mRNA compared to controls, most notable in the CVIDHyper group (Fig. 3a). In contrast, T cells from neither CVID patients nor controls had detectable levels of CCL19 mRNA (data not shown). Similarly, no transcript of CCL21 was found in T cells or monocytes from controls or patients. There was no significant correlation between mRNA expression of CCL19 and serum levels of CCL19 in CVID patients (r = 0·345, P = 0·091). Furthermore, there was no significant relationship between mRNA levels of CCL19 and expression of CCR7 (r = –0·195, P = 0·409).

Figure 3.

The mRNA levels of CCL19 in monocytes from healthy controls and common variable immunodeficiency (CVID)Norm (n = 15) and CVIDHyper (n = 11) patients as assessed by real-time reverse transcription–polymerase chain reaction (RT–PCR) in relation to the housekeeping gene β-actin (a). Expression of CCR7 on CD3+ T cells in healthy controls and two subgroups of CVID patients [CVIDNorm (n = 17) and CVIDHyper (n = 14); see Materials and methods for details] as assessed by flow cytometry (b). Horizontal line represents median values.

Expression of CCR7+ in T cells from CVID subgroups

We and others have previously shown low expression of CCR7 in T cells from CVID patients compared to healthy controls [10–12]. In the present study we extend these findings, showing that the decrease in the proportion of CCR7+ T cells was particularly pronounced in the CVIDHyper subgroup (P = 0·01, CVIDHyperversus CVIDNorm) (Fig. 3b). In contrast, no association was seen with levels of switched memory B cells (data not shown).

Stimulatory effect of CCL19 and CCL21 in PBMC cultures

Cell cultures of PBMC from healthy controls (n = 8) and randomly selected CVID patients (n = 10) were stimulated with CCL19 and CCL21 with or without a suboptimal dose of PHA (22·5 ng/ml) for 20 h. In healthy controls, stimulation with CCL19 and CCL21 alone attenuated the secretion of TNF-α (Fig. 4a). A similar effect was seen of CCL19, but not of CCL21, on the release of IL-8 (Fig. 4b). Neither CCL19 nor CCL21 had any effects on the release of IL-10, IFN-γ and MCP-1 (data not shown). In contrast, when PBMC from healthy controls were co-activated with a suboptimal concentration of PHA, CCL21 enhanced significantly the release of IL-8, IFN-γ, TNF-α and IL-10 (Fig. 4c–f). A similar pattern was seen when the PHA-activated cells were stimulated with CCL19, but the enhancing effects were more modest, reaching statistical significance for TNF-α (Fig. 4c–f). In contrast to these responses in healthy controls, stimulation of PBMC from CVID patients with CCL19 or CCL21 either alone or in combination with PHA gave no significant responses (data not shown).

Figure 4.

Supernatant levels of tumour necrosis factor (TNF)-α (a) and interleukin (IL)-8 (b) in peripheral blood mononuclear cell (PBMC) cultures of healthy controls (n = 8) stimulated with or without CCL19 (100 ng/ml) or CCL21 (100 ng/ml) for 20 h. Supernatant levels of IL-8 (c), interferon (IFN)-γ (d), TNF-α (e) and IL-10 (F) in PBMC cultures of controls stimulated with phytohaemagglutinin (PHA) (22·5 ng/ml) with or without CCL19 or CCL21. Data are mean ± standard error of the mean. Open bars = unstimulated or PHA stimulation only. Grey bars = stimulation with CCL19 alone or CCL19 with PHA. Black bars = stimulation with CCL21 alone or CCL21 with PHA. *P < 0·05 versus unstimulated or PHA stimulated cells only. **P < 0·01 versus unstimulated or PHA stimulated cells only.

Expression of CCL19, CCL21 and CCR7 in spleens from CVID patients

Extirpated spleens from CVID patients (n = 9) and controls (n = 3) were examined for expression of CCL19, CCL21 and CCR7 by means of immunohistochemistry. Spleens from eight CVID patients and all controls showed marked and similar expression of CCL19 and CCR7 in lymphoid cells of the white pulp, while endothelial cells and macrophages in the red pulp showed a more moderate expression of these markers. One CVID patient had similar expression of CCL19 and CCR7 in both white and red pulp. CCL21 was expressed primarily in dendritic cells in the normal spleen with moderate to strong intensity and there were high proportions of CCL21+ cells among macrophages and endothelial cells (Fig. 5a). In contrast, while five CVID patients showed a similar expression of CCL21 as the controls, four patients had a strikingly lower expression of CCL21 in dendritic cells (Fig. 5b). Importantly, there was no up-regulation of either CCL19 or CCL21 in spleens from CVID patients, and no down-regulation of CCR7. Isotype controls for CCL19, CCL21 and CCR7 were negative in all specimens.

Figure 5.

Immunohistochemical detection of CCL21 in extirpated spleens from one control (a) and one common variable immunodeficiency (CVID) patient (b). Staining of paraffin section from the control spleen shows strong expression of CCL21 by many elongated dendritic cells (dark brown) in the T cell area (T) of the white pulp, while the B cell area (B) is devoid of CCL21-positive cells apart from a very weak staining of follicular dendritic cells in the germinal centre (upper left corner). Staining of paraffin section from the spleen of the CVID patient reveals only very few scattered CCL21-positive dendritic cells in the T cell area (T) of the white pulp. For details of the immunohistochemical procedure, see Material and methods. Both images were taken with a ×20 objective and show identical magnification.

Discussion

Previously, we have shown involvement of the chemokine system in CVID [17], and there are extensive studies implying the involvement of CCL19 and CCL21 in a range of immune responses [5–9]. Recent studies also suggest that a balanced interaction between CCL19, CCL21 and CCR7 is necessary for the induction of peripheral tolerance and the regulation of the immune response by CD4+CD25+ T cells (regulatory T cells) [3,19,20].

While CCL19 and CCL21 are expressed mainly in lymphoid tissue, circulating levels of CCL19 and CCL21 possibly also reflect expression in peripheral inflamed tissue. CVID patients are prone to both recurrent and chronic bacterial infections, providing inflammatory foci that might contribute to levels of CCL19 and CCL21. Bronchiectases are associated frequently with bacterial infection in CVID and while both patients with and without bronchiectases had raised levels of CCL19 compared to controls; this was most notable in patients with this complication. It is therefore possible that the recurrent pulmonary bacterial infections seen in CVID patients with bronchiectases contribute to the high levels of CCL19 seen in this subgroup of patients. Longitudinal data on serum levels of CCL19 and CCL21 in CVID, including data during acute infections, which would be of interest, were not available in this study.

Subgroups of CVID patients are characterized by manifestations associated with chronic inflammation, including splenomegaly. In our study, however, the presence of splenomegaly, contributing to increased lymphoid mass in the CVIDHyper group, was not correlated with serum levels of CCL19 and CCL21. Furthermore, there was no up-regulation of these chemokines in spleen tissue from the CVID patients as assessed by immunohistochemistry. In fact, some CVID patients showed lower expression of CCL21 in their spleens than controls, even if the lack of objective measurements of chemokine expression in the spleen samples is a limitation of these analyses. In contrast, monocytes from the CVIDHyper group tended towards enhanced expression of CCL19, and a possible contribution from monocytes to serum levels of CCL19 in this subgroup cannot be excluded. The relatively low number of samples calls for some caution regarding the interpretation of these results. CCL21 can be produced by the endothelium and activation of endothelial cells could influence serum levels of CCL21. However, we found no evidence of endothelial activation in CVID as reflected by enhanced levels of vWF, and no correlation between levels of CCL21 and vWF. Thus, even if the raised serum levels of CCL19 might be related to inflammatory cells associated with bronchiectases as well as monocyte activation in the CVIDHyper subgroup, the serum levels of CCL19 and CCL 21 seem elevated, at least in part, independently of features of chronic inflammation in CVID. Our findings also suggest that CCL19 and CCL21 may be regulated differently, as has also been suggested by others [5]. The similar levels of neopterin in patients with and without bronchiectases and the similar prevalence of bronchiectases in the CVIDNorm and CVIDHyper subgroups suggest that these partitions reflect different clinical and immunological features of the CVID population.

Three studies have shown a down-regulation of CCR7 on T cells in groups of CVID patients [10–12]. In the present study we extend these findings by showing particularly low proportions of CD3+CCR7+ in the CVIDHyper patients, suggesting that this characteristic is another inflammatory feature of this subgroup. Interestingly, while expression of CCR7 is affected markedly in the CVIDHyper group, the levels of CCL19 and CCL21 are not and, moreover, there is no correlation between serum levels of CCL19 and CCL21 and T cell expression of CCR7. Thus, even if CCL19 has been shown to modify cellular CCR7 expression [21], our study suggests that expression of CCR7 on circulating T cells is regulated at least partly independently of its ligands in CVID. In contrast to the diminished CCR7 expression on circulating T cells in CVID, spleens from CVID patients show a similar expression of CCR7 compared to controls, suggesting that the down-regulation of CCR7 in CVID does not affect all lymphoid compartments.

Stimulation of CCR7+ cells with CCL19 and CCL21 in various settings has shown important modulation of cell responses by these chemokines [9,22,23]. In our study, stimulation of PBMC from healthy controls with CCL19 and CCL21 alone elicited none or an attenuated response while co-stimulation of CCL19 or CCL21 with PHA induced a moderately enhanced release of both inflammatory and anti-inflammatory cytokines. In contrast to these modulatory effects in PBMC from healthy controls, cells from CVID patients gave no significant response to CCL19 and CCL21. This suggests that the dysregulation of CCR7 in circulating mononuclear cells also has functional consequences, even if the in vivo relevance of our in vitro data must be interpreted cautiously.

In conclusion, we extend previous reports on dysregulation of CCR7 expression in CVID and report raised circulatory levels of its ligands CCL19 and CCL21, as well as an attenuated modulatory response to stimulation with these chemokines. As CCR7, CCL19 and CCL21 are key mediators balancing immunity and tolerance in the immune system, the abnormalities of these mediators might contribute to the profound immune dysregulation seen in CVID.

Disclosure

The authors have no conflict of interest to declare.

Acknowledgements

We thank Azita Rashidi, Hogne Røed Nilsen and Marit F Klingvall for excellent technical assistance.

Ancillary