Elevated levels of CCR6+ T helper 22 cells correlate with skin and renal impairment in systemic lupus erythematosus

Systemic lupus erythematosus (SLE) is an autoimmune disease with a variety of pathological features. Our study investigated the potential role of CCR6+ T cells in organ impairment of SLE patients. We analyzed CCR6+/− T cell subset populations and compared the concentrations of IL-22, IFN-γ, TNF-α, and IL-17A cytokines in 67 patients with newly diagnosed SLE and 26 healthy controls. We found that SLE patients had elevated percentages of CCR6+ T, CCR6+ Th22, Th17, Th17.1, and CCR6− Th2 cell subsets, along with increased concentrations of IL-22, IFN-γ, TNF-α, and IL-17 cytokines. Higher levels of CCR6+ T and CCR6+ Th22 cells, along with plasma IL-22 were observed in SLE patients with sole skin and/or renal impairment. The percentage of Th22 cells also correlated with Revised Cutaneous Lupus Erythematosus Disease Area and Severity Index (RCLASI) and IgG levels, and inversely correlated with C3 levels in SLE patients with sole skin impairment. SLE patients with sole renal impairment showed a correlation between the percentage of Th22 cells and ESR levels. Our data indicated that CCR6+ Th22 cells may contribute to the pathogenesis of new onset SLE patients with skin or renal impairment, and CCR6 may, thus, be a possible therapeutic target for SLE treatment.

Interestingly, CCR6 + Th cells have recently been confirmed to play a pro-inflammatory role in autoimmune diseases 22,23 . Th17 cells expressing CCR6 appeared to be more pathogenic and accelerate organ impairment after renal injury 24 and arthritis 25 in various animal models. In addition, a genetic association has also been reported between CCR6 gene polymorphisms and susceptibility to lupus nephritis (LN) 26 . However, there have been few studies highlighting the relationship between CCR6 + Th cell sub-populations and SLE, especially in patients with organ impairment.
Thus, in our study, we aimed to determine the frequency of circulating CCR6 + /CCR6 − Th cells by flow cytometry in 67 new onset SLE patients and 26 age-and gender-matched healthy controls (HCs). In addition we also examined levels of IL-22, IL-17, TNF-α, and IFN-γ cytokines in parallel, and further assessed the expression correlation of these T cell subsets and cytokines with clinical parameters and severity index of SLE patients with varying organ impairment.

Result
Comparison of demographic and clinical characteristics of SLE patients. The comparison of 67 onset SLE patients and 26 matched HCs showed no significant difference in terms of age and gender. The levels of C-reactive protein(CRP) and white blood cell counts between SLE patients and HCs also demonstrated no difference, as shown in Table 1. However, SLE patients displayed significantly higher levels of IgG and erythrocyte sedimentation rate (ESR), while levels of complement factor (C)3, C4 were lower, as compared to HCs. In addition, we also observed varied SLE Disease Activity Index (SLEDAI) score among SLE patients, and some patients were positive for anti-double stranded(ds)DNA and anti-Smith(Sm) antibodies. SLE patients revealed increased frequency of CCR6 + , CCR4 + and CCR10 + Th cells. It has been previously demonstrated that different chemokine receptors, such as CCR4, CCR6, CCR10 and CXCR3, have varied expression on the surface of Th cell sub-populations 15,16,22 . Our receptor expression analysis indicated significantly increased frequencies of CCR6 + , CCR4 + and CCR10 + Th cells in SLE patients, in comparison to HCs. It is noteworthy that we did not observe any significant differences in Th cell sub-population expressing CXCR3 (Fig. 1). SLE patients demonstrated elevated proportions of CCR6 + Th22, Th17, Th17.1 and CCR6 − Th2 cell subsets. Next, different CD3 + CD4 + T cell subsets (Th22, Th17, Th17.1, Th9, Th2 or Th1) were assessed by using a gating strategy based on chemokine receptor expression ( Fig. 2A) 18 . The following chemokine receptor expression profile indicated specific subsets; CCR6 + CCR4 + CCR10 + phenotype represented Th22 cells, while CCR6 + CCR4 + CXCR3 − CCR10 − phenotype confirmed Th17 cells. Similarly, CCR6 + CCR4 − CCR10 -CXCR3 + phenotype represented Th17.1 cells, and CCR6 + CCR4 − phenotype indicated Th9 subset. Further comparison of these different subsets between SLE patient and HCs, indicated significantly higher frequencies of CCR6 + Th subsets (including Th22, Th17 and Th17.1) in SLE patients. However, we did not observe any significant difference in Th9 cells distribution between SLE patients and HCs (Fig. 2B). In addition, we also analyzed CCR6 − Th cell subpopulations representing either Th2 (CCR6 − CCR4 + CCR10 − CXCR3 − ) or Th1 (CCR6 − CCR4 − CCR10 − CXCR3 + ) profiles in SLE patients and HCs. Our results indicated that the proportion of Th2 cells was higher in SLE patients than HCs (Fig. 2B), while no significant difference was observed in Th1 cell subpopulations between these groups (Fig. 2B).

Positive correlation between CCR6 + Th22 cells and IL-22 plasma concentration in SLE patients.
In addition, we also determined the concentrations of plasma IL-22, IL-17, IFN-γ, and TNF-α in all the individual subjects, and observed that levels of plasma IL-22 (Fig. 3A), IL-17 ( Fig. 3B), IFN-γ (Fig. 3C), and TNF-α ( Fig. 3D) were notably elevated in SLE patients in comparison to HCs. We further noted that elevated levels of plasma IL-22 in onset SLE patients positively correlated with Th22 cell population (P = 0.0006 r = 0.4064; Fig. 3E). In contrast, we did not observe a similar correlation of plasma IL-22 concentration with frequencies of other Th cell subsets in SLE patients. Similarly, we also did not find any significant correlation between plasma IFN-γ levels and the

Correlation of Th cell subsets with clinic indicators in onset SLE patients.
We also examined the potential correlation between Th cell subsets and disease severity in new onset SLE patients. Our data revealed a positive correlation between the percentage of Th22 cells and SLEDAI score in SLE patients (P = 0.0002 r = 0.4391; Fig. 4A). Similarly, the percentage of Th17 cells also showed positive correlation with SLEDAI score (P = 0.0162 r = 0.2927; Fig. 4B) and levels of urinary protein (P < 0.0001 r = 0.6046; Fig. 4C). In addition, Th17.1 cell percentage correlated positively with ESR levels (P = 0.0369 r = 0.2595; Fig. 4D), but inversely correlated with C4 protein (P = 0.0012 r = −0.3871; Fig. 4E). However, no significant correlations were observed between the frequencies of other Th cell subsets with SLEDAI score, and levels of ESR, plasma CRP, Ig, C3, and C4 in SLE patients (data not shown).

SLE patients with skin and renal impairment showed enhanced CCR6 + T and Th22 cell populations.
In order to clarify the relationship of CCR6 + Th cell subsets with skin and renal impairment in new onset SLE patients, we categorized patients into four subgroups. The first group included SLE patients with sole renal impairment (SK, n = 13), while the second group consisted of SLE patients with skin involvement (SS, n = 20). The third group of SLE patients had both renal and skin impairment (KS, n = 18), and SLE patients with no kidney and skin involvement (NKS, n = 16) made up the fourth group. Our results indicated that patients from SK, SS and KS groups had significantly higher percentages of CCR6 + T and Th22 cell populations in comparison to NKS group or HCs (Fig. 5A,B). Similarly, Th17 cells were also elevated in SK and KS groups. However, the SS group did not show any significant changes in comparison to NKS and HC groups (Fig. 5C). In addition, other CCR6 + Th cells, like Th17.1 and Th9 did not show any significant changes (data not shown). Moreover, CCR6 − Th cells, Th1 (Fig. 5D) and Th2 (data not shown) also did not reveal any significant differences among these different group of patients. Th22 and Th17 cell populations correlated with clinical parameters in patients with sole LN or lupus skin disease.
As Th22 and Th17 cells were elevated in SLE patients with skin and renal impairment, we also analyzed their correlation with clinical parameters. The percentage of Th22 cells showed a positive correlation with revised cutaneous lupus disease area index (DAI) (P = 0.0046, r = 0.6068; Fig. 5E) and severity index (SI) (P = 0.0010, r = 0.5617 Fig. 5F). Consistent with this, we also observed a positive correlation of Th22 cells with IgG plasma levels (P = 0.0203, r = 0.5143; Fig. 5G), and an inverse correlation with C3 levels (P = 0.0245, r = 0.5008; Fig. 5H), specifically in patients with sole lupus skin disease. Interestingly, SLE patients with sole renal involvement showed a positive correlation of Th22 cells with ESR plasma levels (P = 0.0006, r = 0.8029; Fig. 5I), and Th17 cells with urinary protein levels (P <0.0001, r = 0.8956; Fig. 5J) and casts (P = 0.0214, r = 0.6287 Fig. 5K). Our results revealed that concentrations of plasma IL-22 were higher in patients from SK, SS and KS groups, in comparison to NKS and HC groups (Fig. 6A). Similarly, plasma IL-17 level was also significantly elevated in SK and KS group patients compared with NKS and HC group (Fig. 6B). In addition, IL-17 levels in the SS group patients were significantly elevated in comparison to HC group, but not the NKS group. Interestingly, plasma IFN-γ levels were significantly elevated in patients from all groups (SK, SS, KS & NKS) in comparison to HC group (Fig. 6C). This result indicated that changes in IFN-γ levels were not correlated with skin and renal impairment in SLE patients. Finally, patients from SS group showed maximum elevation of plasma TNF-α level, in comparison to NKS and NC group patients. Patients from SK and KS groups showed increased plasma TNF-α levels, but only compared to HC group patients (Fig. 6D).

Discussion
In this study, we identified that SLE patients have a higher frequency of circulating CCR6 + Th cell subsets, including Th22, Th17 and Th17.1. In addition, SLE patients also showed a higher percentage of CCR6 − Th2 cells, but revealed no difference in Th1 population, in comparison to healthy control subjects. These subsets showed a positive correlation with SLEDAI score, and, in addition, showed correlation with other clinical parameters such as urinary protein levels, ESR and C4 protein, in SLE patients. These observations suggest that CCR6 + Th cell subsets may play an important role in SLE pathogenesis. Consistent with our findings, previous studies have also established that SLE patients display elevated levels of CCR6 + Th22, Th17 cells and CCR6 − Th2 cells 27-29 . However, to the best of our knowledge, our study is the first to report elevated numbers of CCR6 + Th17.1 cells in SLE patients and this subset specifically correlated positively with ESR value, and negatively with C4 levels. The Th17.1 subpopulation is also termed as non-classic Th1 and their respective ontogeny is unclear 30 . This subpopulation may generate during Th17 lymphocyte transition under inflammatory conditions or derive from naive T cells directly in humans 31 , and seems to be particularly aggressive and more pathogenic 30 , especially when there is no response to glucocorticoids, due to expression of multi-drug resistance type 1 (MDR1) transporter gene 32 . This led us to hypothesize that Th17.1 cells may not only play a role in SLE pathogenesis, but also serve as a potential biomarker for the assessment of disease severity.
It has been shown previously that expression of CCR6 on T effector cells positively correlates with the severity of organ impairment in SLE patients 33 , and CCR6/CCL20 axis aids in recruitment of IL-17 producing cells into targeted tissue, and, thus, results in tissue injury 24,33 . To validate these findings and have further clarify the mechanism of CCR6 + Th cells in the SLE organ impairment, we stratified SLE patients in our study into different subgroups (SK, SS, KS and NSK) based on skin and renal impairment. The analysis of CCR6 expression on T cells in these different groups indicated that patients from SK, SS and KS group had higher frequencies of CCR6 + T cells compared with NKS group. This observation suggested that CCR6 + T cells might play pivotal role in skin and renal impairment during SLE.
Th22 cells express multiple skin homing receptors, including CCR6, CCR4 and CCR10 16,18 , which allow these cells to infiltrate into the skin, and thus highlighting the importance of this T cell subset in skin impairment. It has also been shown that Th22 cells and IL-22 overexpression are not only associated with psoriasis but also with various skin diseases, such as atopic dermatitis and scleroderma 34 . Consistent with these published reports, our data also indicated that SLE patients with skin only impairment (SS group) had elevated CCR6 + Th22 cells in comparison to NSK group. Also, the Th22 cells frequency in these patients positively correlated with disease area, severity of skin lesions, and IgG plasma concentrations, while negatively correlated with C3 protein levels. This co-relational analysis emphasized that CCR6 + Th22 cells might play a pertinent role in the pathogenesis of lupus involving skin issues. Furthermore, as IL-22 is the main effector cytokine of Th22 cells, we also observed elevated levels of plasma Il-22 in SS group patients. There existed a positive correlation between elevated level of plasma IL-22 and CCR6 + Th22 cells and SI value. Typically, IL-22 has been shown to be involved in inducing keratinocyte proliferation and causing epidermis hyperproliferation, resulting in premature keratinocytes 35 . The rate of keratinocyte apoptosis has also been noticed to be higher in lesional epidermis of cutaneous lupus erythematosus (CLE) patients, and the degree of apoptosis was same between SLE and CLE patients 36 . In this regard, TNF-α, also secreted by Th22 cells 15,16,37 , has been linked with improving keratinocyte responsiveness to IL-22 by upregulating the expression of IL-22 receptor along with additional signal transduction elements 38 . Interestingly, in our study, we also observed that plasma TNF-α level was increased in SS group patients, thereby suggesting that TNF-α may interact with IL-22 to play a critical role in the pathogenesis of lupus skin impairment. Based on these observations, we speculated that the possible mechanism of Th22 cell involvement in lupus skin disease is that an increased number of CCR6 + Th22 cells migrate into the inflammatory location in the skin lesion, and the chemokine receptors expressed by these cells interact with their ligands to produce specific cytokines, which, inturn, induce keratinocyte proliferation. However, to maintain a dynamic balance between apoptosis and proliferation, more keratinocytes undergo apoptosis, but due to impaired or delayed removal of apoptotic cells in SLE 39 , accumulated apoptotic keratinocytes undergoes secondary necrosis, and eventually cause local tissue inflammation. This may trigger a positive feedback loop, resulting in persistent inflammation.
In SLE, renal impairment is one of the other most severe and devastating manifestations. Thus, in our study, we compared the SLE patients with various renal impairment (SK, and NSK group patients). Interestingly, we observed a positive correlation between elevated Th22 levels and ESR in SK group patients. Since the patients in our study were in early stages of the disease, we proposed that Th22 cells might also play a role in the pathogenesis of LN in patients with early onset SLE. In SK group patients, we identified elevated plasma levels of IL-22 in comparison to patients in the NSK group. It has been noted that kidney cells express IL-22 receptor (IL-22R) and blockade of IL-22 results in alleviation of renal injury and improved the clearance of the pathogens in polymicrobial peritonitis. This indicated that IL-22 is detrimental for kidney regeneration 40 . Also, it has been shown that IL-22 could induce complement proteins 41 , and thus has the ability to mediate disease by indirectly affecting the glomerular endothelium in LN through complement. Moreover, TNF-α has also been observed to be elevated, and shows correlation with urinal protein level in SK group patients when compared to HC group. There have also been other studies that implicated TNF-α as an important pathogenic cytokine in LN, and it also contributed to the subsequent LN development 42,43 .
Another T cell subset, which showed higher frequency in patients from SK group, was Th17 cells. Elevated Th17 cell percentage positively correlated with both levels of urinary protein and cast in these patients. Furthermore, we also observed a significant increase in plasma IL-17 levels in SK group patients, which again showed a positive correlation with the percentage of Th17 cells, urinary protein level and casts. Overall, our results were consistent with previously published similar studies 44,45 and points to the notion that Th17/IL-17 axis might also be contribute to the pathogenesis of LN.
Surprisingly, we did not observe any significant difference in the levels of circulating Th1 cells between SLE patients and HCs, even though SLE patients had significantly higher levels of IFN-γ. This discrepancy may be accounted for due to; 1) Th1 cells may not be the major source of plasma IFN-γ, and this notion is supported by the fact that we did not observe any significant correlation between the plasma IFN-γ level and the percentage of Th1 cells. Thus, Other IFN-γ-producing cells, such as γδT cells, may be responsible for the elevated IFN-γ level in SLE. 2) We also observed an elevated percentage of Th17.1 in SLE patients, and it is highly possible that these cells might be contributing to the elevation of the IFN-γ level and the pathogenesis of SLE.
To summarize, our findings suggested that CCR6 + Th22 cells may play an important role in skin or renal impairment in SLE patients, and further proposed CCR6 as a probable therapeutic target for SLE treatment. It is important to mention that our study only detected CCR6 + Th cell subsets in the patients' blood, instead of identifying their infiltration to the target tissues. Furthermore, our study only focused on CCR6 + Th cell subsets and their relation with SLE pathogenesis, despite the evidence about the role of other Th cell subsets also. In addition, further studies about the mechanistic role of CCR6 + Th cells, especially the CCR6 + Th22 cell population on parenchymal cells, and their interaction in the tissues of target organs, would be warranted. In our future studies, we would also like to investigate the involvement of other T cells subsets, including γδT, Tfh and T reg cells in organs impairment including neurological system in SLE, by using larger sample size and multiple time points.

Methods
Patients and controls. Our study included a total of 67 new-onset SLE patients with clinical symptoms for less than three months and without any treatment. They were recruited between January and November 2016, from the in-patient department of the First Hospital of Jilin University, China. In addition, we also enrolled 26 healthy volunteers, as controls, matched for ethnicity, age, and gender, from the Physical Examination Center of the outpatient department in our hospital. SLE patients were diagnosed based on the revised criteria for the classification of SLE, by the American College of Rheumatology (ACR) 46 . SLEDAI was applied to assess the disease activity, and an SLEADI score of >6 represented active disease 47 . Moreover, based on ACR criteria, complicated LN was diagnosed if the patient displayed any one of the following signs: (i) persistent urinary protein (≥0.5 g/24 h); (ii) indication of active cellular casts; and (iii) positive biopsy of LN 48 . Also, the shape of a rash and/ or the histopathology of skin lesions indicated complicated lupus skin disease. Importantly, subjects with systemic sclerosis, myositis, other autoimmune disorders, cancer, recent infection or who had received glucocorticoid or immunosuppressive treatment within the past six months, were excluded from the study. All the individual subjects gave the written informed consent and experimental protocol was established according to the guidelines of the Declaration of Helsinki and approved by the Human Ethics Committee of Jilin University.
Clinical examination. The demographic and clinical characteristics of the subjects before treatment, including age and gender were obtained from hospital records. The routine laboratory tests, including determination of complete blood cell counts, concentrations of plasma C3, C4, ESR, CRP, IgG levels, anti-dsDNA and anti-Sm antibodies were measured as described previously 27 . Disease Area and Severity Index (RCLASI: DAI and SI) were calculated in the patients with lupus skin disease, and the urinary protein levels were determined in LN patients for 24 hour time periods.

Isolation of peripheral blood mononuclear cells (PBMCs).
Peripheral blood (6 mL) was collected from each participant at baseline after an overnight fasting, and PBMCs were isolated by density-gradient centrifugation at 800 g for 30 min using Ficoll-Paque Plus method (Amersham Biosciences, Buckinghamshire, UK). The isolated cells were then resuspended at a concentration of 1 × 10 6 per mL in RPMI-1640 culture medium (Invitrogen, Carlsbad, CA, USA) containing 10% fetal calf plasma.