Mobilization of lymphatic endothelial precursor cells and lymphatic neovascularization in primary Sjögren's syndrome

Abstract Although lymphatic neovascularization may be a key feature of chronic inflammation, it is almost unexplored in primary Sjögren's syndrome (pSS). A recent study revealed a pro‐lymphangiogenic function of interleukin (IL)‐17, a leading player in pSS pathogenesis. The aims of the study were to investigate lymphangiogenic mediators and lymphatic vasculature in pSS, as well as their possible association with IL‐17. Circulating lymphatic endothelial precursor cells (LEPCs) and Th17 cells were enumerated in pSS patients and healthy donors. VEGF‐C and IL‐17 levels were assessed in paired serum samples. Lymphatic vasculature, VEGF‐C/VEGF receptor (VEGFR)‐3 and IL‐17 were evaluated in pSS minor salivary glands (MSGs) and compared with normal and non‐specific chronic sialadenitis (NSCS) MSGs. Circulating LEPCs were expanded in pSS and correlated with circulating Th17 cells, IL‐17 and VEGF‐C. In pSS MSGs, a newly formed lymphatic capillary network was found within periductal inflammatory infiltrates and the number of interlobular lymphatic vessels was significantly increased compared with normal and NSCS MSGs. Strong VEGF‐C expression was detected in pSS ductal epithelial cells and periductal inflammatory cells. Numerous VEGFR‐3+ infiltrating mononuclear cells were exclusively observed in pSS MSGs. VEGFR‐3 expression was strongly increased in lymphatic capillaries of pSS MSGs. IL‐17+ inflammatory cells were preferentially observed around lymphatic vessels in pSS MSGs. This study supports the notion that lymphvasculogenesis and lymphangiogenesis are active in pSS, thereby unmasking a novel aspect of disease pathogenesis. In addition, our results suggest another possible pathogenic role of IL‐17 in pSS, further supporting its therapeutic targeting in this disease.


Introduction
Primary Sj € ogren's syndrome (pSS) is a systemic autoimmune disorder characterized by chronic inflammation of exocrine glands leading to impaired secretory function [1]. Minor salivary gland (MSG) biopsies of patients with pSS, routinely evaluated for diagnostic purposes, have to display, according to commonly used classification criteria [2], cell aggregates with >50 periductal/ perivascular mononuclear cells (i.e. foci), a pattern defined as focal lymphocytic sialadenitis (FLS). In a subgroup of patients, such inflammatory infiltrates organize in highly specialized ectopic lymphoid structures that resemble germinal centres characteristic of secondary lymphoid organs.
Since neovascularization is a crucial event during inflammation, it has been put forward the hypothesis that it may be also a pathogenic hallmark of pSS [3]. Expansion of the blood vessel microvascular bed within inflamed tissues aims at fuelling inflammatory response via recruitment of pro-inflammatory cells and recirculation of proinflammatory soluble mediators. Similarly, lymphatic neovascularization is a key feature of acute and chronic inflammation. However, whether its role is either protective or deleterious is still a matter of debate. Although an increased blood vessel density that parallels the extent of glandular inflammation and an activation of VEGF-A/VEGF receptor (VEGFR)-2 and neuropilin-1 co-receptor pro-angiogenic system have been recently described in pSS [4][5][6], the lymphatic vessel counterpart is almost unexplored. Only two studies attempted to quantify the lymphatic vascular bed extension in pSS MSGs through the detection of mature lymphatic endothelial cells (ECs), but they yielded conflicting results reporting either no difference or an increase in lymphatic capillaries in patients compared with controls (the latter defined as MSGs without foci) [7,8]. During adulthood, lymphatic vessels originate not only from lymphangiogenesis, namely from pre-existing lymphatic vasculature, as previously thought, but also from lymphvasculogenesis, the generation of novel vessels via precursor cells. The latter event has been recently confirmed and points out the importance of bone marrow-derived lymphatic endothelial precursor cells (LEPCs) in postnatal lymphvasculogenesis [9]. Similarly to their blood vascular counterpart (i.e. EPCs), LEPCs express CD34 and CD133 on their surface, but while EPCs display VEGFR-2, and LEPCs express VEGFR-3 (also known as Flt4) [10]. LEPCs isolated from healthy adult peripheral blood are able to proliferate in vitro and give origin to mature ECs expressing lymphatic markers such as LYVE-1 and podoplanin [10]. Lymphatic neovascularization via both lymphangiogenesis and lymphvasculogenesis is mainly orchestrated by the VEGF-C/VEGFR-3 pathway. Interestingly, VEGFR-3 silencing in LEPCs seems to reduce their proliferation, thereby preventing their differentiation into mature lymphatic ECs in vitro [11]. During inflammatory processes, VEGF-C is abundantly released by immune cells, for example, macrophages, leading to mobilization of LEPCs from bone marrow. An expansion of circulating LEPCs has been previously described in spondyloarthritis and spondyloarthritis associated with Crohn's disease [12], but no data are currently available in pSS.
Growing evidence suggests that interleukin (IL)-17 and, therefore, IL-17-producing T cells are leading players in pSS pathogenesis [13,14]. Interleukin-17 participates in both induction and perpetuation of glandular inflammation and, likely, also in ectopic lymphoid neogenesis [15,16]. A recent study, employing a mouse cornea micropocket model, demonstrated that IL-17 is able to induce lymphatic neovascularization via the VEGFR-3 pathway fostering LEPC proliferation [17]. It is, therefore, conceivable that IL-17 may display such pro-lymphangiogenic function also in pSS.
On these premises, the aims of the present study were (i) to investigate circulating LEPCs and their possible association with IL-17 in pSS and (ii) to characterize the lymphatic vasculature and the expression of lymphangiogenic mediators in MSGs from patients with pSS compared with non-specific chronic sialadenitis (NSCS) and normal MSGs.

Patients and healthy donors
Fifteen female patients with pSS classified according to the American-European criteria [2] and 15 age-matched healthy female donors (HD) were enrolled for analyses on peripheral blood samples. Clinical and serological records were collected at the time of enrolment. Disease activity was measured using the EULAR Sj€ ogren's syndrome disease activity index (ESSDAI) [18]. All patients were receiving topical medications for sicca symptoms and hydroxychloroquine 200 mg/day. None of the patients was taking corticosteroids or immunosuppressive therapies. The study was approved by the local ethics committee, and written informed consent was obtained from each participant in accordance with the declaration of Helsinki.

Results
Circulating LEPCs are expanded in pSS and correlate with serum VEGF-C, serum IL-17 and circulating Th17 cell percentage Demographic, clinical and serological features of the cohort of patients with pSS enrolled for analyses on peripheral blood samples are summarized in Table 1. We first sought to investigate the proportion of circulating LEPCs in pSS patients with respect to HD. Lymphatic endothelial precursor cells were identified within the lymphocyte gate (Fig. 1A) as CD34 + CD133 + VEGFR-3 + cells (Fig. 1B-E). Patients with pSS displayed significantly higher proportions of circulating LEPCs compared to HD (P < 0.0001; Fig. 1F). A representative plot of one HD and one patient with pSS is displayed in Figure 1D and E, respectively. In addition, the percentage of circulating LEPCs was directly correlated to that of Th17 cells, identified as CD4 + IL-17 + lymphocytes (Spearman's rho = 0.60, P = 0.0173; Fig. 2). Subsequently, we assessed the concentration of VEGF-C in paired serum samples and observed comparable levels of this molecule in pSS and HD (Fig. 3A). Of interest, however, VEGF-C concentration was directly correlated to LEPC percentage in patients with pSS (Spearman's rho = 0.62, P = 0.01; Fig. 3B). Finally, the concentration of IL-17 resulted directly correlated to the percentage of circulating LEPCs in paired serum samples from patients with pSS (Spearman's rho = 0.53, P = 0.04).

Lymphatic vascularization and expression of lymphangiogenic mediators are increased in MSGs from patients with pSS
Immunohistological analyses were carried out on labial MSG biopsies from 12 patients with pSS (i.e. displaying FLS) and 16 sicca syndrome non-pSS controls. Of these, eight participants displayed normal MSGs and eight displayed a certain degree of MSG inflammation (i.e. NSCS), but no evidence of FLS. Representative images of  For specific detection of lymphatic vessels, we performed both immunoperoxidase-based immunohistochemistry (Fig. 4D-L) and immunofluorescence (Fig. 4M-O) using the mouse monoclonal antibody D2-40 that reacts with a fixation-resistant epitope in podoplanin, a mucin-type transmembrane protein expressed at high levels in lymphatic ECs but not in blood vascular ECs [22]. As displayed in Figure 4M-O, MSG lymphatic vessels were consistently and strongly immunostained by D2-40, while CD31 + blood vessels were consistently negative for D2-40. In line with previous observations in human parotid gland and MSGs [7], lymphatic vessels were absent from acinar regions in control MSGs (Fig. 4D, E, G and H). Indeed, lymphatic vessels were only detected in the interlobular connective tissue, especially around excretory ducts, in both normal and NSCS MSGs ( Fig. 4F and I). In pSS MSGs, a newly formed lymphatic capillary network was found within periductal inflammatory infiltrates replacing the secretory units ( Fig. 4J and K). Moreover, the lymphatic vessel network was expanded in the interlobular connective tissue of pSS MSGs (Fig. 4L). In fact, the number of interlobular lymphatic vessels was significantly increased in pSS MSGs compared with normal and NSCS MSGs (both P < 0.01; Fig. 4M-P).
As far as lymphangiogenic mediators are concerned, VEGF-C was barely detectable in ductal epithelial cells and microvessels of both normal and NSCS MSGs (Fig. 5D and E). Conversely, in pSS MSGs, a strong expression of VEGF-C was detected in ductal epithelial cells, microvessels and periductal inflammatory cells (Fig. 5F). As displayed in Figure 6A, densitometric analysis showed that VEGF-C immunofluorescent staining intensity was significantly increased in ductal epithelial cells of pSS MSGs compared with normal and NSCS MSGs (both P < 0.01). Numerous VEGFR-3 + infiltrating mononuclear cells were exclusively observed in pSS MSGs (Fig. 5G-I). In addition, the expression of VEGFR-3 was strongly increased in lymphatic capillaries of pSS MSGs with respect to normal and NSCS MSGs (Fig. 5J-L). Indeed, VEGFR-3 immunofluorescent staining intensity was signif-icantly greater in pSS MSG lymphatic ECs than in normal and NSCS MSG lymphatic ECs (both P < 0.01; Fig. 6B). Finally, IL-17 + inflammatory cells were preferentially observed around lymphatic vessels in pSS MSGs, while no IL-17 expression could be detected either in normal or NSCS MSGs (Fig. 5M-O).

Discussion
Since lymphatic neovascularization is a key feature of chronic inflammation, a role of this process in inflammatory autoimmune diseases has been speculated [3]. In this study, we investigated for the first time lymphvasculogenesis and lymphangiogenic mediators, namely LEPCs and VEGF-C/VEGFR-3 axis [9][10][11]23], in pSS and demonstrated that they are increased in this disease. In addition, we reported an increased and anatomically aberrant lymphatic neovascularization in MSGs from patients with pSS. The only two available studies that focused on lymphatic neovascularization in MSGs were limited to the assessment of glandular mature lymphatic ECs and yielded conflicting results [7,8]. In fact, while Yazisiz et al. reported an increase of lymphatic capillaries within the mononuclear cell infiltrate of MSGs from eight patients with pSS [8], McCall and Baker failed to identify any differences between patients and controls [7]. As far as MSGs are concerned, our present study not only assessed lymphatic vessel distribution but also analysed for the first time the expression of the pro-lymphangiogenic VEGF-C/VEGFR-3 axis. Interestingly, the increased expression of VEGFR-3 consistently observed in lymphatic endothelium further indicates an activation of the pro-lymphangiogenic programme within the inflammatory milieu of pSS MSGs.
Besides being involved in tissue fluid homeostasis and lipid uptake in the gastrointestinal tract, lymphatic vessels also play a role in immune surveillance driving the recirculation of immune cells toward secondary lymphoid organs. In normal conditions, salivary gland lymphatic vessels are selectively localized within the interlobular connective tissue, being absent in lobules and acinar regions [7,24]. Our  study revealed that pSS salivary gland tissues display a consistent increase of interlobular lymphatic vessels. In addition, it was of great interest the finding that lymphatic vessels were also evident where the mononuclear cell infiltrates replaced acinar structures. Such de novo lymphatic network may account for a recirculation of immune cells from MSGs and might even help autoreactive lymphocytes in reaching extraglandular sites. Moreover, the strong glandular expression of VEGF-C by ductal epithelial cells may be an additional clue for the active role of glandular epithelium in the scenario of disease pathogenesis, possibly by recruiting LEPCs from bone marrow to generate new glandular lymphatic vessels via lymphvasculogenesis. In addition, we observed strong VEGF-C expression in pSS periductal inflammatory cells, presumably monocytes/macrophages. Interestingly, monocytes/macrophages have been proposed to participate in lymphangiogenesis in at least two ways, namely as a source of VEGF-C after appropriate stimulation or by transdifferentiation into lymphatic ECs that integrate into the growing capillaries [25]. Indeed, as potential lymphangioblastic precursors, a large pool of CD14 + monocytes was identified in the circulation. These monocytes were found to constitutively express VEGFR-3 and turn on expression of podoplanin and several other lymphatic markers when cultured for a prolonged period [25]. In this context, it is noteworthy the presence of numerous VEGFR-3 + infiltrating mononuclear cells found in pSS MSGs.
Of note, our study also sheds light on a possible association between lymphvasculogenic/lymphangiogenic mediators and IL-17 axis in pSS. Interleukin-17 is a pro-inflammatory cytokine representing a leading player in the pathogenesis of systemic autoimmune diseases, including pSS [13,26]. Interleukin-17 signalling in immune and non-immune target cells induces a cascade of events that culminates in the release of pro-inflammatory mediators. In the matter of lymphatic neovascularization, a previously published study provided evidence that IL-17 also displays a pro-lymphangiogenic function by stimulating LEPCs and fostering their differentiation into mature lymphatic ECs [17]. In pSS, IL-17 is able to trigger FLS [13] and is associated with the extent of glandular inflammation [16,27] and, possibly, with ectopic lymphoid neogenesis [16]. Therefore, such a cytokine appears to represent an intriguing therapeutic target in this disease. Herein, we demonstrated that circulating IL-17 and Th17 cells are directly correlated to LEPCs, thereby suggesting a possible pro-lymphangiogenic/lymphvasculogenic activity of this cytokine in pSS. Taken this association and the role of IL-17 in pSS, it is reasonable to postulate that aberrant glandular lymphatic neovascularization in pSS is another facet of disease pathogenesis.
In conclusion, our study demonstrates for the first time that lymphvasculogenesis and lymphangiogenesis are active in pSS and suggests another pathogenic role of IL-17 in this disease. We are aware that functional in vitro studies will be required to confirm that IL-17 may directly affect LEPC activation and differentiation in pSS. However, we believe that the present data unmasked a novel aspect of disease pathogenesis, providing the basis for further investigation on this issue in larger patient cohorts and identifying another possible rationale for therapeutic targeting of IL-17 in pSS [28].