Systemic and local evidence for complement involvement in chronic spontaneous urticaria

Abstract Background The pathogenesis of chronic spontaneous urticaria (CSU), including the mechanism of action of omalizumab, remain unclear. We hypothesized complement system involvement given the often fast clinical response induced by treatment, including omalizumab. Therefore, we assessed the role of various complement factors surrounding omalizumab treatment. Methods Thirty CSU patients (median age 42 [range 21–70]; 73 % female) with a median once daily Urticaria Activity Score over 7 days (UAS7) score at baseline of 31.5 points were enrolled. Treatment consisted of six administrations of 300 mg omalizumab every 4 weeks succeeded by a follow‐up period of 12 weeks. Four punch skin biopsies were taken per patient; at baseline from lesional skin, at baseline from nonlesional skin, and after 1 and 7 days from formerly lesional skin. Complement activity, including C1q, C3, C3bc/C3, C4, C4bc/C4, C5a, and Membrane Attack Complex in peripheral blood were analyzed and complement activation in the skin was determined by the analysis of C4d deposition. Results were related to the clinical response to omalizumab. Results Fifteen patients showed a UAS7 score of 6 or lower (median 0) at Week 24, 15 patients did not (median 16). Lesional skin biopsies at baseline revealed complement deposition (C4d) in blood vessels in the papillary dermis of 53% (16/30) of the patients, which suggests involvement of immune complexes in the pathogenesis of urticaria. Moreover, indication of increased complement activation in CSU was substantiated by increased C5a levels in peripheral blood compared to healthy controls (p = 0.010). The clinical effect of omalizumab could not be linked to the variation of complement components. Conclusions Both C4d deposition in lesional skin and elevated C5a levels in peripheral blood indicate the involvement of complement activation in the pathogenesis of CSU. No correlation was found between omalizumab and activation of complement indicative of independent processes in the immunopathogenesis of CSU.

complement indicative of independent processes in the immunopathogenesis of CSU.

K E Y W O R D S
blood, complement, omalizumab, skin, urticaria

| INTRODUCTION
Chronic spontaneous urticaria manifests as a burdensome skin disease with sudden onset, sometimes severe itching and wheals, that lasts for at least 6 weeks. Prevalence is estimated to be up to 1% at any time, with disease duration ranging from 1 to 5 years or even longer in more severe cases. 1 Omalizumab has a reported clinical response of over 50% within the first 2 days of treatment in CSU patients. 2 Depletion of free IgE by omalizumab leads to downregulation of the FcεRI on mast cells 3

and basophils 4 in patients with
allergic disease. 5 However, this downregulation alone cannot explain the fast clinical response to omalizumab. Hence, alternative mechanisms must be contributing to the rapid clinical efficacy of omalizumab.
Sera from patients with urticaria can induce degranulation of basophils, a process in which the presence of intact complement and patient IgG containing specific antibodies against IgE or the high affinity IgE receptor is essential. 6 By binding to IgE or FcεRI on mast cells, complement via the classical pathway can be activated and lead to the generation of C5a and C5b-9. 7 C5a can subsequently bind to the complement C5a receptor on mast cells and cause degranulation.
Cutaneous mast cells express the complement C5a receptor whereas mucosal mast cells do not. Additionally, the complement system is known for its rapid response upon activation. This may explain how IgG anti-FcεRI autoantibodies in combination with complement in patients with CSU can cause fast clinical symptoms which are limited to the skin and not mucosal tissue. 8 It has been reported that C1q, C2, C3, C4, and C5 levels in peripheral blood are within normal limits in chronic urticaria but no data have been published regarding complement degradation/ activation products in peripheral blood . [9][10][11] The effect of omalizumab treatment on peripheral blood complement levels in CSU patients has also never been studied. Furthermore, it is unknown whether complement activation occurs in the skin of patients with CSU. Complement activation in tissue can be evaluated by determination of C4d deposition: a well-studied marker and a characteristic feature of complement activation, which is, for instance, also included in the BANFF criteria for humoral rejection after kidney transplantation. 12 In this study, we investigated the role of the complement system and the effects of omalizumab treatment in CSU patients using the C4d marker in skin and peripheral blood samples. Additionally, we hypothesize that the efficacy of omalizumab in CSU may in part be accompanied by reduction of complement mediated inflammation.

| Design and population
This monocenter exploratory prospective cohort study was performed in the University Medical Center Utrecht, The Netherlands, from 2015 to 2017. Inclusion criteria were adult CSU patients with a significant disease activity defined as a once daily Urticaria Activity Score over 7 days (UAS7) ≥ 16 and a UAS7 ≥ 4 on the day of the first omalizumab administration despite treatment with antihistamines up to four times the daily dose. Exclusion criteria were based on one of the pivotal randomized controlled trials 13 and included a clearly defined underlying etiology for chronic urticaria (e.g., chronic inducible urticaria), a history of malignancy, known hypersensitivity to omalizumab, and pregnancy. Routine administration of immunosuppressants including prednisolone and ciclosporin 14 were discontinued with washout periods of 3 months prior to treatment with omalizumab. If prednisolone was used as rescue medication, a washout period of 2 weeks was maintained before the start of the study. After a screening period of up to 2 weeks, eligible patients started a 6-month treatment period, followed by a follow-up period of 3 months. The latter could be shortened upon patient-request if the UAS7 was projected to reach a score of 16 or higher. All patients provided written informed consent, and the study was approved by the local ethics committee (protocol number 15-167).

| Omalizumab and concomitant medication
All patients received six doses of 300 mg omalizumab every 4 weeks with follow-up starting at Week 25, 4 weeks after the last dose.
Leukotriene receptor antagonists or H2 blockers for indications other than CSU were permitted to be continued during the study.
Patients were allowed to use H1-antihistamines up to a maximum of four doses per day as rescue medication in addition to their concomitant medication, as well as prednisolone up to 30 mg. Due to worsening of the disease, 11 patients, of which 6 (55%) were pre-

| Assessments in blood samples
Blood samples were collected at the following time-points: at baseline, after 1, 2, 6, and 24 h, after 1 and 2 weeks, and 4 weeks after the first administration of omalizumab. Subsequently, blood was collected prior to each subsequent dose. Lastly, a venipuncture was performed at the last follow-up visit. For measurement of complement activation, ethylenediaminetetraacetic acid (EDTA) plasma, serum, and gel separated serum were used. EDTA blood and gel separated serum were put on ice immediately after venipunctures.
All serum samples were allowed to coagulate for 60 min. Serum and plasma were obtained by centrifugation and stored at −80°C.

| Measurements complement
Complement levels of C3 and C4 were determined in serum by an immunonephelometric method on a SPA+ turbidimeter, C5b-9 membrane attack complex formation via the classical complement activation route was measured in gel separated serum using a commercially available enzyme-linked immunosorbent assay (Euro-Diagnostica) according to the manufacturer's recommendations, and C5a was determined in EDTA plasma via Luminex xMAP technology (Luminex Corporation). Additionally, C1q in serum, and C3bc and C4bc in EDTA plasma were determined as previously described. 15  staining; ALPCO), and was graded from 0 (negative) to 3 (bright signal or fully surrounding blood vessel walls) ( Figure 1). As previously described, the original magnification was �400. 17

| Clinical assessments and patient-reported outcomes
Disease activity was measured throughout the study by using the UAS7. 18 Missing daily scores of the weekly disease activity scores after treatment was started were complemented by Last Observation Carried Forward method up to a maximum of 3 days. Missing follow-up scores were supplemented with data from each patient's clinical record following clinical visits if available. Weekly scores that could not be complemented with earlier mentioned methods were marked as missing and were not included in patient reported outcome results. Disease control was measured at baseline, 4 weeks after each administration, and at the last follow-up visit by using the urticaria control test. 19 Treatment response has been defined as a UAS ≤ 6 at Week 24 of treatment.

| Statistical analysis
Changes in inflammatory parameters in the skin were related to changes in levels of circulating complement components, by using Spearman Rank correlation. Inflammatory characteristic after treatment were compared to baseline, and/or to the previous measurement, using Wilcoxon matched pairs signed rank tests, or paired samples T test as appropriate. C3bc and C4bc activation ratios were determined by dividing the level of circulating C3bc or C4bc by the amount of C3 or C4, respectively, and multiplying the quotient by 100 to determine the percentage, as previously described. 15

| Inflammation and complement activation in lesional and nonlesional skin
Quantification of histological alteration found in skin biopsies of patients compared to healthy controls is presented in Table 1. Histological analysis demonstrated no significant differences between lesional and nonlesional biopsies at baseline or follow-up with regard to edema and cellular infiltration. Higher amounts of C4d deposition were significantly more frequently found in lesional skin compared to nonlesional skin (p = 0.033) ( Table 1). In the total 60 baseline skin biopsies, there was a significant correlation between C4d deposition and eosinophils scores (Spearman's ρ 0.358; p = 0.005). Table 2 shows that in a large portion of patients, peripheral blood levels of all complement components investigated were within normal ranges throughout the study. Most investigated complement component levels in peripheral blood were within normal ranges at both baseline and throughout the study. However, as shown in Additionally, no correlations were found between complement component levels in peripheral blood and C4d deposition in skin.

Total measurements Reduced Elevated values (%) values (%)
C1q  after the first administration were found (see Table A1). Additionally, no correlations were found between UAS7 after 1 week and both complement levels and difference in complement levels after 1 h.

| DISCUSSION
In this study, we investigated the role of the complement system in To the best of our knowledge, it is a novel finding that C4d is present in small blood vessel walls within the papillary dermis of a majority of CSU patients. This finding suggests that IgG or IgM autoantibodies in the skin are able to cause complement activation and supports the current hypothesis that IgG autoantibodies are involved in the pathogenesis of CSU. 3,22 Complement-fixing autoantibodies and complement deposition in the skin are also frequently found in systemic lupus erythematosus (SLE), which may point to common pathomechanisms in CSU and SLE. 23 In SLE, C4d was found not only in blood vessel walls (80% of patients) but also along the dermoepidermal junction (100% of patients). In the subacute cutaneous lupus erythematosus, deposits of C4d were detected within epidermal keratinocytes, and in pemphigus cases, intercellular C4d was found which roughly corresponded to the location of autoantibodies. 24 The location of C4d deposition-in the superficial dermisis where infiltration was seen most in patients with urticaria. It is not surprising that C4d deposits were, in lower amount, also present in  The relatively small patient numbers and the absence of a multiple testing correction for the different complement components is a limitation of this study. Therefore, additional research in larger study populations are needed. In conclusion, both C4d deposition in lesional skin and elevated C5a levels in peripheral blood indicate the involvement of complement activation in the pathogenesis of CSU.
No correlation was found between (response to) omalizumab and activation of complement, indicative of independent processes in the immunopathogenesis of CSU.

ETHICS STATEMENT
All patients provided written informed consent, including consent for publication. The study was approved by the local ethics committee (protocol number 15-167).