Complement activation in Hidradenitis suppurativa: Covert low-grade inflammation or innocent bystander?

Abstract

Hidradenitis suppurativa (HS) is a chronic auto-inflammatory skin disease with a complex and multifactorial pathogenesis involving both the innate and adaptive immune system. Despite limited evidence for local complement activation, conflicting results have been published on the role of systemic complement activation in HS. It was hypothesized that complement was consumed in highly inflamed HS skin, trapping complement from the circulation. Therefore, the aim of this study was to evaluate this local complement deposition in HS skin lesions using routine and commonly used complement antibodies.Direct immunofluorescence for C1q, C3c, C4d, C5b-9, and properdin was performed on frozen tissue sections of 19 HS patients and 6 controls. C5a receptor 1 (C5aR1) was visualized using immunohistochemistry.

Overall, we found no significant local complement deposition in HS patients versus controls regarding C1q, C3c, C4d, C5b-9, or properdin on either vessels or immune cells. C5aR1 expression was exclusively found on immune cells, predominantly neutrophilic granulocytes, but not significantly different relatively to the total infiltrate in HS lesions compared with controls. In conclusion, despite not being able to confirm local complement depositions of C1q, C3c, C4d, or properdin using highly sensitive and widely accepted techniques, the increased presence of C5aR1 positive immune cells in HS suggests the importance of complement in the pathogenesis of HS and supports emerging therapies targeting this pathway.

Introduction

Hidradenitis suppurativa is a chronic auto-inflammatory skin disease with a complex and multifactorial pathogenesis (1). Both the innate and adaptive immune system are involved with upregulation of human beta-defensins, S100 proteins, complement components and a multitude of cytokines in lesional skin (2). Recent advances in HS treatment have seen the emergence of clinical trials with complement directed treatments such as vilobelimab (monoclonal antibody blocking C5a, InflaRx) and Avacopan (C5aR1 inhibitor, ChemoCentryx). However, the outcomes of these trials have been questioned due to the unclear primary outcome measure, HiSCR, that does not capture draining tunnels in a dynamic manner (3). Neither of these trials achieved its primary endpoint hampering the initial interpretation of the efficacy of these treatments in HS.

The rationale behind these clinical trials was based on the study by Kanni et al., who demonstrated systemic, late stage complement pathway activation in HS reflected by higher levels of C5a and C5b-9 in plasma of HS patients relative to controls (4). In contrast, our group found no increased circulating levels of C5a or sC5b-9 in patients with HS (5). Besides, concentrations of C3 and C3d as well as the C3d/C3 ratio were increased in patients with HS, but all, except for C3d, were within normal range. C3, C3b, iC3b and C4b were previously identified as downregulated differentially expressed proteins by Hoffman et al, who identified only C5a as an upregulated differentially expressed protein in the blood of HS patients (6). Remarkably, other inflammatory proteins with a known correlation to systemic complement activation such as IL-6 and CRP have not been found consistently elevated in plasma of HS patients (4, 5).

Besides systemic complement activation, several studies identified increased expression of complement genes using different RNA sequencing methods (7–9). Hotz et al. found increased RNA expression of C1q, C2, CR1, C3aR1, C5aR1, and Factor B and decreased expression of C7, Factor H, and Factor D in HS lesional skin (8). More recently, Gudjonsson et al. showed complement activation among the significantly enriched biological processes in both bulk RNA seq of HS lesional skin and plasma cells derived from single cell RNA sequencing of lesional skin (9). Moreover, this study demonstrated increased protein levels of C1q, C3b, C4d and complement receptors CR1 and CR2 using immunohistochemistry on formalin fixed paraffin embedded (FFPE) lesional slides of three HS patients (9).

To date, no pathogenic variants in complement genes have been found in HS. However, sufficiently large genome wide association studies have not yet been performed in HS. These studies might yield polymorphisms associated with complement genes as these have been found in HS-associated diseases such as psoriasis. Nonetheless, based on the pathogenesis of HS, activation of the complement pathway seems indisputable and is hypothesized to play a multifactorial role in the disease process (Figure 1). However, due to the conflicting results of the two published studies on systemic complement activation, it was hypothesized that complement might be consumed in highly inflamed HS skin, locally trapping complement from the circulation (4, 5). To support this theory, we aimed to provide high quality evidence of local complement activation in HS lesional skin by visualizing complement components using direct immunofluorescence (DIF) on frozen sections in accordance with routine diagnostic procedures for diseases such as auto-immune bullous diseases and IgA-vasculitis, using widely accepted specific complement antibodies.

Figure 1

Materials and methods

Patient selection

Lesional skin was obtained from HS patients who underwent routine excisional surgery in the Erasmus Medical Center Rotterdam, Rotterdam, the Netherlands. Skin considered surgical discard after mamma reduction surgery was used as control. Under the prevailing opt-out principle and ethical guidelines in the Erasmus MC, excised skin is considered surgical discard and exempt from informed consent or ethical approval. Patient baseline characteristics for HS patients were obtained from electronic patient files.

Sample preparation

Skin excision specimens were cut manually using the bread-loaf technique in which each specimen was cut transversally at 4 mm intervals. One central segment was frozen at -80 degrees and stored for DIF. The remaining segments were FFPE embedded, tissue sections from these segments were stained with hematoxylin and eosin (H&E) (or immunohistochemical staining) for light microscopic evaluation.

Direct immunofluorescence

Direct immunofluorescence was performed on the stored frozen material as previously described by Damman et al. (10). Additionally, frozen sections were stained with H&E for light microscopic evaluation. For detection of complement factors, DIF was performed by automated immunofluorescent staining using the Ventana Benchmark Discovery (Ventana Medical Systems Inc.). Wet slides were loaded and incubated with the primary antibody of interest for 32 min at 37 ⁰C followed by detection with either omnimap-anti rabbit or mouse, labeled with HRP and visualized with FAM (#760–243, Ventana). Slides were covered with anti-fading medium (DAKO, S3023). DIF intensity and area of vascular and immune infiltrate was scored by two independent pathologists (J.G. and J.D.). The DIF intensity was scored on a nominal scale of 0–3: none (0), weak (1), moderate (2), profound/bright (3). As the number of immune cells and vessels were found to differ between Hurley stages, the percentage positive vessels and immune infiltrate relative to the total number of vessels or immune cells was scored on a nominal scale: none (0), < 5% of positive (1), 5-50% positive (2), and > 50% positive (3). A cumulative score was calculated by multiplying the intensity x area. Markers that were included for analysis were C1q, C4d, properdin, C3c and C5b-9, covering classical (C1q, C4d), alternative (properdin), lectin and classical (C4d) and terminal pathway activation (C5b-9). See Supplemental Table 1 for antibodies used.

C5aR1 immunohistochemistry

Immunohistochemistry was performed for C5aR1 using a primary monoclonal antibody to human C5aR1, clone S5/1. Detection for C5aR1 was identical as described for DIF, except for the use of Diaminobenzidine (DAB) as detection substrate. The percentage positive immune cells relative to the total of immune cells was scored on a nominal scale: none (0), < 5% positive (1), 5-50% positive (2), and > 50% positive (3). This scoring was performed for the whole slide and per tunnel (when present) separately. For the infiltrate, a distinction was made between granulocyte or histiocyte positive C5aR1 cells, based on morphology.

Statistics

Differences in the nominally scored intensity of staining and the overall cumulative score for staining between HS patients and healthy controls were assessed using Chi-square tests. These were performed for both complement factors in blood vessels and on polymorphonuclear leukocytes (PMNs). Comparisons were performed for the overall HS group versus healthy controls as well as for each individual Hurley stage compared with controls and the other Hurley stages. Additionally, for C5aR1 analyses have been performed comparing the stained cell types as well as the effect of tunnels, stratified for different Hurley stages. Statistical analyses were performed using SPSS (IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 28.0.1.0 Armonk, NY.) and two-sided p-values ≤0.05 were considered statistically significant.

Results

Patient characteristics

In total, excisional samples from 19 HS patients and 6 controls were included. The majority of included HS samples were obtained from male patients (58%), all control samples were obtained from women. All Hurley stages were represented with 32% of samples obtained from patients with Hurley stage I, 42% from Hurley stage II patients, and 26% from Hurley stage III patients (Table 1). While all skin from healthy donors was derived from the mammae, HS samples were derived from the buttocks (47%), axillae (26%), groin (16%), or abdomen (11%).

Immune infiltrate in complement-stained tissue sections

In general, sections from Hurley stage I were characterized by moderate to severe deep dermal and subcutaneous active inflammation with influx of many neutrophils, sometimes with abscess formation and admixture of eosinophils. Hurley I lesions frequently showed granulomatous foreign body reaction on ruptured hair follicles. A few Hurley stage I lesions showed dermally located epithelial strands, potentially as remnants from a ruptured abscess, and dermal scarring. Hurley stage II sections exhibited moderate chronic lymphoplasmacellular and active inflammation, frequently associated with tunnel formation, moderate foreign body reaction and moderate scarring. Sections obtained from Hurley stage 3 lesions showed profound chronic lymphoplasmacellular inflammation, mild to moderate active inflammation, mild and focal granulomatous inflammation, and profound scarring.

Local complement deposition in HS lesions

C3c could only be detected in vessels of 3/19 HS patients (16%) with a moderate intensity in <5% of the total number of vessels, Table 2. Positive staining for C3c was only found in patients with Hurley stages 2 or 3, while C3c staining was absent in all Hurley stage I patients and all healthy controls (Table 3). Vascular C4d was found moderately to strongly positive in 12/19 HS (63%, Hurley I-III) in predominantly <5% of the total numbers of vessels. However, similar findings were also observed in 2/6 (33%) healthy controls. Vascular C5b-9 could be demonstrated in 5/19 of HS patients (26%), all of which were Hurley stage II or III, but also only in <5% of the total numbers of vessels present. C1q, a marker of early classical complement pathway activation, was detected in only 1/19 HS patients (Hurley stage III) in <5% of vessels, Tables 2, 3. Vascular properdin was not detected in any of the HS or control samples. Figure 2 shows representative DIF staining of the focal vascular depositions of C3c (Hurley stage III), C4d (Hurley stage III) and C5b-9 (Hurley stage II) in HS patients.

Figure 2

Immune cells throughout all HS sections were found to be negative for any of the complement components stained through DIF, regardless of disease stage.

Overall, statistical analyses revealed that there were no significant differences between either the intensity or cumulative scores for any C1q, C3c, C4d, or properdin in HS slides compared with healthy controls.

In contrast, C5aR1 expression was exclusively found on immune cells. Representative staining of C5aR1 in different Hurley stages can be found in Figures 3–5. C5aR1 expression was predominantly found on granulocytes and to a lesser extent on histiocytes and giant cells (Figure 6). A marked increase in C5aR1⁺ PMNs was seen in HS samples, particularly Hurley stage I samples, compared with controls. However, as staining was scored relative to the infiltrate size the difference was not statistically significant for any of the Hurley stages versus controls or between Hurley stages. There was no increased number of C5aR1⁺ immune cells found in sections with tunnels compared to sections without tunnels (Figure 7). Accompanying the localization of neutrophils, C5aR1 staining was most pronounced directly adjacent to ruptured cysts or dermal tunnels.

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Discussion

In recent years, conflicting results have been published on the role of systemic complement activation in HS. Although our group could not find activation of circulating complement components, others previously demonstrated significantly increased levels of C5a and C5b-9 in the circulation of HS patients (4). To explain this difference, it was hypothesized that complement might be consumed in highly inflamed HS skin, trapping complement from the circulation. The aim of this study was to evaluate this local complement deposition in the skin in frozen HS skin lesions using routine and widely used complement antibodies. Surprisingly, but in line with our previously published findings in plasma, we could not demonstrate significant local complement deposition in HS versus controls. Although we found focal endothelial C3c, C4d, and C5b-9 deposition, this was not significantly different from controls. Moreover, one could even argue the clinical relevance of these focal findings: it is well known that vascular complement deposition (and immunoglobulins) can be found in several types of dermatitis and even in normal skin, all in the absence of vasculitis. The finding of an almost lack of complement deposition in our study is in contrast to previous findings: using immunohistochemistry on three patients, Gudjonsson and colleagues demonstrated increased complement deposition on immune cells, particularly in the deeper layers of the dermis (9). The discrepancy could be explained by several factors: Firstly, there is a difference in tissue used, where we used frozen material, Gudjonsson et al. used FFPE. Complement staining tends to have the highest sensitivity and specificity on frozen material. Secondly, different antibodies were used, we worked with commonly accepted antibodies that are widely used in routine diagnostics and in the field of complement research. Thirdly, the life of lesions biopsied could influence results: comparable to immunoglobulins, complement factors have a certain tissue residency and complement depositions could be missed in older lesions as they are actively cleared. Lastly, low-grade complement activation in HS might be highly locally regulated by membrane bound complement regulator proteins such as soluble complement receptor 1 (SCR1/CD35), decay accelerating factor (DAF, CD55) and CD59.

It is worth mentioning that we found linear C4d, properdin and C5b-9 basement membrane (BM) deposits as well as focal arteriolar deposits of C3c, C4d and C5b-9. These findings parallel observations in the kidney in which arteriolar C5b-9 staining are considered an internal positive control and are of non-specific significance (11, 12). Based on several studies in the kidney it is hypothesized that small-to medium sized vascular and (tubular) BM staining for C3, C4 and C5b-9 in healthy subjects might be explained by localized cellular injury acquired during aging, as a consequence of tissue sampling, or on the epidermal BM as proteins involved in the first line of defense against pathogens (12, 13).

Although we did not find significant local cutaneous complement deposition, our findings however, do not exclude a role for complement in HS pathogenesis. Hypothetically, clear parallels could be drawn between the involvement of complement in HS and in granulomatosis with polyangiitis (GPA). In the latter, scarce immunoglobulin deposits in lesional biopsies (therefore labelled ‘pauci‐immune’) were originally considered as evidence for negligible complement activation (14). During the last decade however, up to one-third of patients were found to show complement deposition in their renal biopsies (15). The relevance of complement activation in GPA is illustrated by the fact that Avacopan has been registered for the treatment of GPA and shows clear beneficial results (16). Although there is a different pathogenesis underlying these diseases, a comparable adverse role of C5a-C5aR1-axis activation in GPA might also play a role in HS, despite almost negative DIF results. This is supported by preliminary results from a phase 2 randomized controlled trial in HS that demonstrated effectivity of Avacopan (30 mg twice-daily) in patients with Hurley stage III in secondary outcome measures; IHS4 and the abscess and inflammatory nodule (AN) count (17). Local activation of complement is further supported by our data showing expression of C5aR1 on multiple immune cells in HS lesional skin. Similarly, the vilobelimab (IFX‐1, a monoclonal antibody that selectively binds to C5a) demonstrated a significantly greater reduction in the number of draining tunnels in the highest dosed treatment group relative to the placebo group at week 16 (18).

Interestingly, both Avacopan and vilobelimab were shown to be more efficacious in patients with severe disease (Hurley stage III) and/or tunnels (17, 18). This is likely due to a difference in the contribution of complement in different disease stages. Acute lesions due to rupture of hair follicles resolve spontaneously, indicating that activation of the complement pathway in these lesions is likely adequately regulated by membrane bound regulator proteins and short lived. In contrast, severe, chronic lesions are characterized by non-resolving inflammation arising from partially epithelialized or ruptured tunnels. These lesions show active but predominantly chronic lymphoplasmacellar inflammation, prominent NETosis and decreased NET degradation, Figure 1 (9, 19). In these lesions, the large amount of locally secreted immunoglobulins (incl. autoantibodies) combined with the ongoing exposure to bacteria is likely a continuous driver of complement activation. It is therefore counterintuitive that we did not find an increased abundance of C5aR1 positive immune cells in slides with tunnels compared to those without. Additionally, proteins expressed by activated neutrophils during degranulation and excreted on NETs further attract and activate neutrophils in a potentially vicious cycle (20, 21). Inhibiting this vicious cycle of complement activation in severe disease by administration of Avacopan or vilobelimab is therefore likely to show greater clinical effect then inhibiting an already adequately regulated complement response in milder patients (17, 18). A large infiltrate in both mild and severe disease, however, could not only lead to complement activation it could also contribute to quick degradation of (otherwise long lived) complement components, potentially making them relatively difficult to detect on the protein level.

A novel finding was the association of C5aR1 staining on both histiocytes and giant cells. Loosely structured granulomatous inflammation and foreign body giant cells containing keratin fragments have previously been described in HS lesions (1, 22). However, this granulomatous inflammation likely also represents an attempt to limit the bacterial loads from ruptured or non-epithelialized tunnels. While the exact contribution of C5a to this process in HS remains unknown, data from other diseases indicate that C5a plays an important role in the mediation of chemotaxis and activation events required for the formation of a granulomatous response (23, 24).

In conclusion, despite not being able to confirm local complement depositions of C1q, C3c, C4d, properdin, or C5b-9 using highly sensitive and widely accepted techniques, the increased presence of C5aR1 on immune cells in HS indicates the importance of complement in the pathogenesis of HS and provide support for the emerging therapies targeting this pathway. Future studies assessing other complement receptors and membrane bound complement regulator proteins could shed more light on the activation of the complement pathway in HS lesional skin.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

• 1

  van StraalenKRPrensEPGudjonssonJE. Insights into hidradenitis suppurativa. J Allergy Clin Immunol (2022) 149(4):1150–61. doi: 10.1016/j.jaci.2022.02.003

  • CrossRef
  • Google Scholar

• 2

  VossenAvan der ZeeHHPrensEP. Hidradenitis suppurativa: A systematic review integrating inflammatory pathways into a cohesive pathogenic model. Front Immunol (2018) 9:2965. doi: 10.3389/fimmu.2018.02965

  • CrossRef
  • Google Scholar

• 3

  van StraalenKRIngramJRAugustinMZouboulisCC. New treatments and new assessment instruments for hidradenitis suppurativa. Exp Dermatol (2022). doi: 10.1111/exd.14609

  • CrossRef
  • Google Scholar

• 4

  KanniTZenkerOHabelMRiedemannNGiamarellos-BourboulisEJ. Complement activation in hidradenitis suppurativa: A new pathway of pathogenesis? Br J Dermatol (2018) 179(2):413–9. doi: 10.1111/bjd.16428

  • CrossRef
  • Google Scholar

• 5

  PrensLMArdonCBvan StraalenKRvan der ZeeHHSeelenMAJLamanJDet al. No evident systemic terminal complement pathway activation in hidradenitis suppurativa. J Invest Dermatol (2021) 141(12):2966–9 e1. doi: 10.1016/j.jid.2021.03.037

  • CrossRef
  • Google Scholar

• 6

  HoffmanLKTomalinLESchultzGHowellMDAnandasabapathyNAlaviAet al. Integrating the skin and blood transcriptomes and serum proteome in hidradenitis suppurativa reveals complement dysregulation and a plasma cell signature. PloS One (2018) 13(9):e0203672. doi: 10.1371/journal.pone.0203672

  • CrossRef
  • Google Scholar

• 7

  BlokJLLiKBrodmerkelCJonkmanMFHorváthB. Gene expression profiling of skin and blood in hidradenitis suppurativa. Br J Dermatol (2016) 174(6):1392–4. doi: 10.1111/bjd.14371

  • CrossRef
  • Google Scholar

• 8

  HotzCBoniottoMGuguinASurenaudMJean-LouisFTisserandPet al. Intrinsic defect in keratinocyte function leads to inflammation in hidradenitis suppurativa. J Invest Dermatol (2016) 136(9):1768–80. doi: 10.1016/j.jid.2016.04.036

  • CrossRef
  • Google Scholar

• 9

  GudjonssonJETsoiLCMaFBilliACvan StraalenKRVossenARJVet al. Contribution of plasma cells and B cells to hidradenitis suppurativa pathogenesis. JCI Insight (2020) 5(19):e139930. doi: 10.1172/jci.insight.139930

  • CrossRef
  • Google Scholar

• 10

  DammanJMooyaartALBoschTSeelenMADoornMBV. Lectin and alternative complement pathway activation in cutaneous manifestations of iga-vasculitis: A new target for therapy? Mol Immunol (2022) 143:114–21. doi: 10.1016/j.molimm.2022.01.011

  • CrossRef
  • Google Scholar

• 11

  DumontCMérouaniADucruetTBenoitGClermontMJLapeyraqueALet al. Clinical relevance of membrane attack complex deposition in children with iga nephropathy and henoch-schönlein purpura. Pediatr Nephrol (2020) 35(5):843–50. doi: 10.1007/s00467-019-04445-x

  • CrossRef
  • Google Scholar

• 12

  KoopmanJJEvan EssenMFRennkeHGde VriesAPJvan KootenC. Deposition of the membrane attack complex in healthy and diseased human kidneys. Front Immunol (2020) 11:599974. doi: 10.3389/fimmu.2020.599974

  • CrossRef
  • Google Scholar

• 13

  Basset-SeguinNDersookianMCehrsKYanceyKB. C3d,G is present in normal human epidermal basement membrane. J Immunol (1988) 141(4):1273–80.

  • Google Scholar

• 14

  JennetteJCWilkmanASFalkRJ. Anti-neutrophil cytoplasmic autoantibody-associated glomerulonephritis and vasculitis. Am J Pathol (1989) 135(5):921–30.

  • Google Scholar

• 15

  ChenMDahaMRKallenbergCG. The complement system in systemic autoimmune disease. J Autoimmun (2010) 34(3):J276–86. doi: 10.1016/j.jaut.2009.11.014

  • CrossRef
  • Google Scholar

• 16

  JayneDRWMerkelPASchallTJBekkerPGroupAS. Avacopan for the treatment of anca-associated vasculitis. N Engl J Med (2021) 384(7):599–609. doi: 10.1056/NEJMoa2023386

  • CrossRef
  • Google Scholar

• 17

  ChemoCentryx. Chemocentryx announces positive topline results of phase ii aurora clinical trial of avacopan in the treatment of hidradenitis suppurativa (HS). (2020).

  • Google Scholar

• 18

  inflaRx. Inflarx reports additional analysis of the shine phase lib results for ifx-1 in hidradenitis suppurativa. (2019).

  • Google Scholar

• 19

  ByrdASCarmona-RiveraCO'NeilLJCarlucciPMCisarCRosenbergAZet al. Neutrophil extracellular traps, B cells, and type I interferons contribute to immune dysregulation in hidradenitis suppurativa. Sci Transl Med (2019) 11(508):eaav5908. doi: 10.1126/scitranslmed.aav5908

  • CrossRef
  • Google Scholar

• 20

  Carmona-RiveraCO'NeilLJPatino-MartinezEShipmanWDZhuCLiQZet al. Autoantibodies present in hidradenitis suppurativa correlate with disease severity and promote the release of proinflammatory cytokines in macrophages. J Invest Dermatol (2022) 142(3 Pt B):924–35. doi: 10.1016/j.jid.2021.07.187

  • CrossRef
  • Google Scholar

• 21

  de BontCMBoelensWCPruijnGJM. Netosis, complement, and coagulation: A triangular relationship. Cell Mol Immunol (2019) 16(1):19–27. doi: 10.1038/s41423-018-0024-0

  • CrossRef
  • Google Scholar

• 22

  van der ZeeHHde RuiterLBoerJvan den BroeckeDGden HollanderJCLamanJDet al. Alterations in leucocyte subsets and histomorphology in normal-appearing perilesional skin and early and chronic hidradenitis suppurativa lesions. Br J Dermatol (2012) 166(1):98–106. doi: 10.1111/j.1365-2133.2011.10643.x

  • CrossRef
  • Google Scholar

• 23

  BordersCWCourtneyARonenKPilar Laborde-LahozMGuidryTVHwangSAet al. Requisite role for complement C5 and the C5a receptor in granulomatous response to mycobacterial glycolipid trehalose 6,6'-dimycolate. Scand J Immunol (2005) 62(2):123–30. doi: 10.1111/j.1365-3083.2005.01643.x

  • CrossRef
  • Google Scholar

• 24

  SwaisgoodCMRupleLMCulverDA. Complement anaphylatoxins C3a and C5a increase in the bal of mice at the peak of lung granuloma formation using the mycobacterial antigen soda,. Am J Respir Crit Care Med (2011) 183:A4520.

  • Google Scholar