Investigating the humoral immune response in chronic venous leg ulcer patients colonised with Pseudomonas aeruginosa

Abstract

The ability to manage the bioburden in chronic wounds is most likely coupled to the humoral immune response of the patient. We analysed markers of systemic immune response in patients with chronic venous leg ulcers (CVLUs) colonised (no‐systemic infection) with the opportunistic pathogen Pseudomonas aeruginosa. Sera from 44 clinically non infected patients with CVLUs were analysed for total IgM and IgG isotype 1–4, complement C3, mannose‐binding lectin (MBL), interleukin (IL)‐6, C‐reactive protein (CRP) and specific anti‐P. aeruginosa antibodies against exotoxin A, elastase and alkaline phosphatase. Concentrations of IL‐6 versus CRP intercorrelated (β = 2·43 95% CI (1·34–4·34)), but were independent of P. aeruginosa colonisation. MBL deficiency (MBL < 500 ng/ml) correlated to high serum levels of IgG₁ (P = 0·038) consistent with a compensatory mechanism, but not related to presence of P. aeruginosa in the ulcers. Twenty‐four patients (54·5%) were culture positive for P. aeruginosa, also conferring significantly high serum levels of complement C3 (P = 0·014), but only two of these had positive titres for antibodies against exotoxin A. All patient sera were negative for antibodies against elastase and alkaline phosphatase. Fluorescent in situ hybridization analysis on randomly selected culture‐positive patients could not establish unambiguous presence of P. aeruginosa biofilms in the ulcers. A multiple regression model showed P. aeruginosa and systemic CRP as significant factors in deterioration of ulcer healing rate.

INTRODUCTION

A wound on the skin represents a breach in the barrier between the blood stream and the outer environment, thereby compromising the integrity of the immune defence and its responses to potential invading pathogens. Chronic wounds harbour a multitude of microorganisms (1), and the presence of Pseudomonas aeruginosa as a resident microorganism in chronic venous leg ulcers (CVLUs) has been significantly associated with ulcer enlargement (2). In an open ‘battlefield’ where the innate and adaptive humoral immune systems are constantly challenged with microbial intruders (3), immune responses would be expected to reflect elevated serum titres of specific antibodies directed against common wound pathogens such as P. aeruginosa.

The complement system is an important part of the innate humoral immunity and plays a crucial role in the rapid recognition and clearance of pathogenic intruders (4). There are three mechanisms of complement activation known as the classical‐, the lectin‐, and the alternative pathway. The classical pathway is linked to the adaptive immune response and triggered when serum immunoglobulins (IgM and IgG) have bound specifically to antigens on bacterial surfaces (5). In chronic wound inflammation and infection, IgM and IgG in particular are dominating the immune profile and serve as principal opsonins against P. aeruginosa (6).

Mannose‐binding lectin (MBL) initiates the lectin pathway by binding to carbohydrate patterns on a broad range of pathogenic microorganisms (7), thus enhancing phagocytosis through direct opsonisation and activation of complement with subsequent fixation 8, 9. Deficiency of functional MBL is the most common congenital human immune deficiency state with approximately one third of the Caucasian population carrying genotypes conferring low MBL levels; however, no absolute serum cut‐off value has been defined 10, 11. Although most MBL‐deficient individuals appear healthy, an increased risk of infection and infectious diseases is associated both in humans 11, 12 and animal models 13, 14. Furthermore, deficiency of MBL has been associated with impairment of the spontaneous and necessary separation of the eschar in a mouse MBL‐null burn wound model (15). In a recent study of chronic leg and foot ulcer patients, serum MBL levels were shown to differ between wounds of different aetiologies, with CVLU patients having a significantly higher frequency of MBL deficiency (16).

The third ‘alternative’ pathway of the complement system interacts with the lectin pathway and is initiated by binding of spontaneously hydrolysed complement component C3 to foreign surface structures (4). C3 serves a key position in the complement cascades of all three pathways. Cleavage of C3 constitutes the first common step leading to opsonisation of the target organism or ultimately the assembly of a multiprotein pore structure called the membrane attack complex causing bacterial elimination by complement‐mediated lysis 4, 17. Furthermore, spatial and species distribution of bacteria and a biofilm mode of growth have previously been addressed as contributing factors in the persistence of chronic wounds 18, 19, 20, 21. The complexity of the bacterial communities within CVLU has recently been associated with impaired healing making interspecies synergies a possible virulence factor (22). The presence of bacterial biofilms has been shown in acute wounds in animal models 23, 24 and more recently also in a range of chronic wounds in humans 19, 25.

Figure 1 outlines the principal aims and design of this pilot study. (a) First, we measured selected components of the adaptive humoral immune responses in sera from 44 clinically non infected CVLU patients to investigate for potential interplay between the classical and lectin pathway of the complement system in case of MBL deficiency. (b) Next, it was assessed whether the presence of P. aeruginosa interfered with complement activation expressed as altered serum levels of C3. We also investigated whether virulence factors from P. aeruginosa gave rise to a specific antibody response. (c) We further attempted to visualize whether the discrepancy between P. aeruginosa cultivation from CVLUs and recovery of specific virulence factor antibodies in serum could be attributed to a biofilm mode of growth, thus modulating the host immune response (26). (d) Finally, we measured C‐reactive protein (CRP) based on its characteristic as a systemic biomarker of inflammation (27) mainly stimulated by the pro‐inflammatory interleukin (IL)‐6 (28). Both serum CRP and MBL were used together with P. aeruginosa culture results and C3 serum levels as covariates against ulcer healing rate in a multiple regression model. Our goal was ultimately to assess whether the presence of wound pathogen P. aeruginosa together with systemic markers of complement activation and inflammation showed influence on CVLU healing.

Figure 1.

Outline of study design. Potential association between the IgG/M humoral immune response in 44 clinically non infected chronic venous leg ulcer patients and mannose‐binding lectin deficiency (a). Presence of Pseudomonas aeruginosa and the impact on general complement activation expressed as C3 serum levels (b). Detection of specific P. aeruginosa virulence factors (exotoxin A, elastase and alkaline phosphatase) in serum and assessment of biofilm mode of growth (c). Pro‐inflammatory cytokine IL‐6 and acute‐phase protein C‐reactive protein as systemic markers of inflammation and their correlation to presence of P. aeruginosa and ulcer healing rate (d). Dashed lines show tested hypotheses in this study.

METHODS

Patient material

At the Copenhagen Wound Healing Centre, Bispebjerg University Hospital, sera were obtained following scientific ethical committee approval and informed patient consent from 44 non clinically infected patients with CVLUs. Patients with ulcers persisting for more than 3 months were included. The inclusion criteria for the study were as follows: ankle brachial pressure index above 0·6, no antibiotic treatment 14 days prior to inclusion or during the study. Patients with diabetes, pregnant or breastfeeding women were excluded. All patients were treated with standard compression therapy during the study. Ulcer healing rate (U _HR, % per week) was calculated as follows:

where U _incl and U _end denote ulcer area in cm² at times, t (weeks), of inclusion and end of study, respectively. Wound microflora was extensively characterised and correlated to ulcer size at inclusion (2).

Serological analysis

Patient sera (n = 44) were analysed for the following immune factors: Total immunoglobulins IgG (subclasses 1–4), IgM (IUPAC: NPU 19806), complement component C3 (IUPAC: NPU 19742) by rate‐nephelometry (Beckman Coulter IMMAGE, GMI, Inc., Ramsey, MN) and ultra sensitive CRP by TRACE (Time‐Resolved Amplified Cryptate Emission) technology (Kryptor, Brahms Hennigsdorf, Germany). MBL (IUPAC: NPU 19843) analysis was performed essentially as described previously (29). Briefly, MBL was quantified by a semi‐automated time‐resolved immunofluorescence assay using the monoclonal antibody Hyb 131‐01 (Statens Serum Institut, Denmark) as catcher, biotinylated Hyb 131‐01 as detector and streptavidin‐Eu‐chelate for quantification. The assay was run on an AutoDelfia platform (Perkin Elmer, Shelton, CT); 500 ng/ml was used as cut‐off value for MBL deficiency (10). Cytokine levels of IL‐6 were measured using a BioPlex™ protein analyzer according to the manufacturer's guidelines (Bio‐Rad Laboratories; Hercules, CA; catalog number: 171‐A11080). Detection limit was approximately 10 pg/ml. Specific anti‐P. aeruginosa IgG antibodies against three antigens (i.e. exotoxin A, elastase and alkaline phosphatase) were tested by enzyme‐linked immunosorbent assay (ELISA) as described by the manufacturer (Mediagnost^® E15, Reutlingen, Germany). Serum samples from n = 8 healthy volunteers were included as controls.

Histology

Biopsies (4 mm) were taken at inclusion for the study. Wound biopsies from 10 blindly selected patients were further investigated. Biopsies were fixed in formalin and stored at −80°C and subsequently embedded in paraffin, cryo‐sectioned (3 µm) and stored at −80°C until use. The tissue sections were de‐paraffinized with xylene for 3 × 10 minutes and in 96% ethanol for 10 minutes, and the fixed bacterial cells were permeabilized with a lysozyme solution of 10 mg/ml, 100 mM Tris–HCl (pH 8·0) and 50 mM EDTA for 1 hour at 37°C. The lysozyme solution was removed by washing with 0·9% NaCl, brief immersion in 96% alcohol and then air‐dried.

Fluorescent in situ hybridization with DNA oligonucleotide probes

The de‐paraffinized tissue sections were analysed by fluorescent in situ hybridization (FISH) using DNA probes targeting the 23S and 16S ribosomal rRNA regions. DNA probe solutions at concentrations of 2·5 ng/µl were prepared in dilution buffer [0·9 M NaCl, 100 mM Tris (pH 7·2), 0·1 % SDS]. A closed chamber moisturised with dilution buffer was prepared and preheated prior to hybridization. The samples were incubated at 48°C for 3 hours with the following bacteria‐specific probes: P. aeruginosa was detected by the 23S Psae‐probe (5′‐TCT CGG CCT TGA AAC CCC‐3′) conjugated with a fluorescein isothiocyanate (FITC) (5′) fluorophore (30). Glass slides with tissue sections were incubated in preheated washing buffer I [0·9 M NaCl, 100 mM Tris (pH 7·2), 0·1 % SDS] for 15 minutes at 48°C and preheated washing buffer II [0·9 M NaCl, 100 mM Tris (pH 7·2)] for 15 minutes at 48°C. The glass slides were washed in cold, demineralized H₂O, dried and mounted in Vectashield^® 4′,6′‐diamidino‐2‐phenylindole (DAPI) Mounting Medium Vector (Vector Laboratories, Inc., Burlingame, CA) under a cover slip glass. The cover slip glass was sealed with nail polish, and the samples were examined under an epifluorescence microscope (Olympus^® BX‐71, Tokyo, Japan) equipped with FITC and DAPI filters. The photos of fluorescence staining were obtained using the imaging software Cell^P (Olympus^®).

Statistics

All analyses were carried out using IBM PASW Statistics 18 (SPSS Inc., Chicago, IL). A Mann–Whitney's U‐test was applied to correlate MBL deficiency and the presence of P. aeruginosa to serum levels of IgG₁ and complement C3, respectively. The significance of P. aeruginosa, MBL deficiency, CRP and C3 on ulcer healing rate was analysed by multiple regression. P < 0·05 was considered significant.

RESULTS

Ulcers healing rate, P. aeruginosa colonisation and antibodies against virulence factors

The main dataset of this study is outlined in Table 1. All 44 patients previously enrolled in the Gjodsbol et al. (2) study had an ulcer healing rate of 4·04% (95% CI 1·49–6·58) per week. Of the 24 patients harbouring P. aeruginosa, only two had positive serum titres for antibodies against exotoxin A. The remaining 20 patients who were negative for cultivation all had negative titres for P. aeruginosa‐specific virulence factors, except a single patient showing borderline detection for exotoxin A. All patient sera had negative titres for both elastase and alkaline phosphatase antibodies. No antibodies against P. aeruginosa were detected in sera from healthy controls.

Table 1.

General characteristic of the 44 ulcer patients

	Ulcer patients (n = 44)	Normal range * , †
Age, ‡ median (range)	78 (29–91)	–
Study period, ‡ weeks	4–8	–
Ulcer healing rate, † % per week	4·04 (1·49–6·58)	–
Number of bacterial species cultivated from ulcers, ‡ median (range)	6 (2–14)	–
Number of Pseudomonas aeruginosa culture‐positive patients ‡ (also positive for P. aeruginosa virulence factors § )	24 (2)	–
Number of patients negative or borderline for P. aeruginosa virulence factors	19 (1)	–
Number of MBL‐deficient patients ¶	11	–
Immunoglobulins, † mg/ml
IgM	1·09 (0·888–1·28)	1·56 (0·56–3·52)
IgG1	8·18 (7·25–9·11)	5·0 (2·8–8·0)
IgG2	3·60 (3·07–4·13)	3·0 (1·2–5·7)
IgG3	0·994 (0·805–1·18)	0·64 (0·24–1·25)
IgG4	0·718 (0·426–1·01)	0·35 (0·05–1·25)
Complement proteins †
MBL, µg/ml	2·18 (1·56–2·80)	n/a **
C3, mg/ml	1·33 (1·22–1·44)	1·25 (0·83–1·77)
CRP, µg/ml	35·5 (21·5–49·5)	0·64 (0·08–3·11)
Cytokine †
IL‐6, ng/ml	0·0352 (0·022–0·0484)	n/a **

^*Normal ranges based on reference values from healthy, adult blood donors 31, 32, 33.

^†Data given as mean (95% CI).

^‡Original data from Gjodsbol et al. (2).

^§ P. aeruginosa‐specific titres for antibodies against exotoxin A, elastase and alkaline phosphatase were measured. All patients were serum negative for elastase and alkaline phosphatase.

^¶Cut‐off value: 500 ng/ml.

^**Data not available.

MBL, immunoglobulins, and complement C3 in serum

The concentrations of MBL, immunoglobulins, and complement C3 in patient serum samples were measured. No significant correlation was found between MBL deficiency (<500ng/ml) and the presence of cultured P. aeruginosa in the ulcers (data not shown). Serum levels of the above mentioned immunological factors from ulcer patients to referenced ranges of healthy, adult blood donors 31, 32, 33 are presented in Table 1. In our dataset, IgG₁ stands out as being the only serum parameter with a mean recorded slightly higher than the upper normal 95% CI interval boundary. We found that an elevated IgG₁ serum level significantly correlated to MBL deficiency (Figure 2A, P = 0·038), thus implying a compensatory role of the classical complement pathway for the lectin pathway in the patients. Moreover, we observed significantly higher concentrations of C3 in serum from patients with ulcers that were culture positive for P. aeruginosa (P = 0·014; Figure 2B) compared with ulcers that were culture negative for this pathogen.

Figure 2.

Differences in serum levels of (A) immunoglobulin IgG₁ versus mannose‐binding lectin (MBL) status (MBL < 500 ng/ml, deficiency, n = 11; MBL ≥ 500 ng/ml, sufficiency, n = 33) and (B) complement component C3 versus Pseudomonas aeruginosa wound swab cultivation (negative/positive). Boxplot shows median (centreline), upper and lower quartiles (box width), 1·5× interquartile range (whiskers), and outliers not included in the analysis (open circles). Mann–Whitney's U‐test, significance level P < 0·05.

Spatial distribution of bacteria in venous leg ulcers

Tissue sections from the patients' ulcers were investigated for the presence of bacterial aggregates and spatial distribution of P. aeruginosa using fluorescently labelled DNA probes and for extracellular polysaccharide (EPS) material using DAPI contained in the mounting medium. This was performed on randomly selected histological material from a previous study (2), and a representative image is shown in Figure 3. The findings indicate that P. aeruginosa formed microcolony‐like aggregates (Figure 3C). The DAPI stain which is a general DNA stain did not yield the desired visualisation of the bacterial EPS mainly constituted by DNA (34), so the presence of EPS could not be confirmed (results not shown). We recorded 80% concordance between P. aeruginosa swab culture and FISH on histological wound sections.

Figure 3.

Detection of Pseudomonas aeruginosa and total bacterial flora by FISH in a representative chronic venous leg ulcer biopsy. (A) Haematoxylin and eosin stain of wound biopsy. (B) Detection of rod‐shaped P. aeruginosa with the 23S Psae‐probe conjugated with a FITC fluorophore (5′). (C) Magnified inset of (B) showing rod‐shaped green fluorescent P. aeruginosa aggregating in cluster‐like biofilm structures at the wound surface. Dashed line represents the segregation of background fluorescent tissue and the colonising P. aeruginosa (arrows). Size bars: 1 mm (A), 50 µm (B) and 10 µm (C).

IL‐6 and CRP as systemic markers of inflammation

We measured serum levels of the pro‐inflammatory cytokine IL‐6 and its effector protein CRP as systemic indicators for inflammation. We confirmed a significant correlation between log‐transformed values of IL‐6 and CRP (P = 0·005), which showed a dose–response effect (Figure 4). However, the presence of P. aeruginosa did not influence the degree of inflammation characterized by the IL‐6/CRP relationship.

Figure 4.

Linear regression showing dose dependency between log‐transformed values of the pro‐inflammatory cytokine IL‐6 and the effector protein, C‐reactive protein – a general serological marker of inflammation. Parameter estimate (95% CI): β = 2·43 (1·34–4·34), P = 0·005, r = 0·511. Dashed lines show 95% confidence bands. The degree of inflammation was independent of whether Pseudomonas aeruginosa could be cultured from the ulcers or not (P = 0·570).

Multiple regression models

We performed a multiple regression (Table 2) using the presence of P. aeruginosa, serum levels of C3, log(CRP) and MBL as covariates and ulcer healing rate as dependent variable. Using backwards elimination of parameters, we established three models of differing simplicity. Model 1 contains all four covariates with P. aeruginosa colonisation as a significant factor for deterioration of ulcer healing (P = 0·022). Elimination of C3 generated model 2, where both P. aeruginosa (P = 0·047) and log(CRP) (P = 0·048) had a negative influence on ulcer healing. MBL deficiency could not be used as an explanatory variable, and excluding this provided the simplest model 3 where neither P. aeruginosa colonisation nor the systemic inflammation marker CRP was useful in explaining ulcer healing rate in the patients.

Table 2.

Multiple regression models assessing the impact of Pseudomonas aeruginosa, MBL deficiency, ^* C3 complement and CRP levels on ulcer healing rate

Model †		β‐estimate	95% CI		P‐value
			Lower	Upper
1	Intercept	5.06	−10.3	20.4	0.504
	MBL deficiency	3.79	−1.86	9.44	0.181
	P. aeruginosa	−6.22	−11.5	−0.959	0·022
	C3	5.37	−3.37	14.1	0.218
	log(CRP)	−5.58	−11.4	0.208	0.058
2	Intercept	12.4	2.75	22.1	0.014
	MBL deficiency	3.12	−2.47	8.72	0.263
	P. aeruginosa	−4.98	−9.88	−0.078	0·047
	log(CRP)	−5.87	−11.7	−0.052	0·048
3	Intercept	14.4	5.37	23.4	0.003
	P. aeruginosa	−4.68	−9.56	0.212	0.060
	log(CRP)	−5.72	−11.5	0.110	0.054

MBL, mannose‐binding lectin; CRP, C‐reactive protein. Values in bold: significant variable, P < 0·05.

^*MBL deficiency defined as <500 ng/ml in serum.

^†Model reduction with backwards elimination of parameters.

DISCUSSION

To our knowledge this is the first study to address the immunological significance of P. aeruginosa colonisation in patients with CVLUs. We observed an increased mean level of IgG₁ in MBL‐deficient patients compared with the MBL‐sufficient group (Figure 2A). IgG₁ is a powerful opsonin and initiator of complement activation. Therefore, a compensatory role of the classical pathway for the lectin pathway could be proposed. However, the elevated mean IgG₁ level does not account for the lack of correlation between MBL sufficiency and the presence of P. aeruginosa, because we did not detect any specific humoral antibody response against P. aeruginosa antigens in the majority of the patients. Intriguingly, a recent study on burn injuries in MBL‐null mice showed that MBL modulates not only the release of inflammatory factors, such as cytokines and chemokines, but also cell adhesion molecules, growth factor‐binding proteins and matrix metallo‐proteinases (15), which are intricately linked with wound healing and chronic wounds 35, 36. Thus, MBL may play other roles in wound healing and chronic wounds besides bioburden control.

A central component of the entire complement system, C3, was detected in significantly higher concentrations in P. aeruginosa‐colonised wounds compared with ulcers that did not harbour this pathogen (Figure 2B). This observation is interpreted as an enhancement of complement activation, primarily independent of the classical and lectin pathway, because we could not report significant alterations in serum levels of immunoglobulins or MBL in response to P. aeruginosa colonisation.

The elastase produced by P. aeruginosa is known to have proteolytic activity against C3 in vitro 37, 38, but in our study, no antibodies were found against this virulence factor. In fact, only 2 of 24 P. aeruginosa‐positive serum samples were positive for antibodies against exotoxin A, which is in accordance with previous findings by Danielsen et al. (39). In that study, 2 of 10 CVLU patients were serum positive for antibodies against exotoxin A, but none against elastase and alkaline phosphatase.

We are only able to account for the 2 months inclusion period of the patients where no clinical infections had been detected, but no information whether a previous invasive infection or other immunological insults had occurred was available. However, it is not regarded as a problem because we are reporting negative antibody titres, eliminating the incidences of false positives. The commercially available ELISA kit used in this study was designed for the analysis of the P. aeruginosa infection status of cystic fibrosis patients and differentiates between infected and uninfected patients (40). A possible source of false‐negative results might be found in the current controversies and difficulties revolving around P. aeruginosa serology (41). This phenomenon was observed in one patient who showed borderline exotoxin A antibody level, yet exhibiting a culture‐negative result. Alternative antibody and antigen options for serological detection of P. aeruginosa should therefore be explored in future studies.

The discrepancy between P. aeruginosa culture and serology can largely be ascribed to two factors. One is the proposed immune modulatory abilities of P. aeruginosa by quorum‐sensing molecules as reviewed by Pritchard. (42), which is beyond the scope of this text. The other is the persistence of P. aeruginosa in the wound environment as microcolonies or biofilms. Bacterial biofilms have been observed in animal model studies of acute wounds 43, 44 and were shown to be significantly more tolerant to antibiotic therapy and host immune defences than individual bacterial cells in vitro (45). Moreover, bacteria embedded in a self‐secreted matrix are less likely to expose their surface antigens to macrophages and other host cells exhibiting phagocytotic activity (46). In our study, we reported 80% concordance between culture and molecular data, that is swab culture and FISH on histological wound sections for P. aeruginosa. The microbial niches that exist within chronic wounds (47) pose a significant challenge when investigating the causative effects of bacteria in relation to non healing wounds. A biopsy only represents a fraction of the total wound surface and therefore may be misguiding when assessing the presence of pathogens and the extent of microcolony/biofilm formation in the wound.

Systemic biomarkers of inflammation, the pro‐inflammatory cytokine IL‐6 and the associated acute‐phase protein CRP were internally positively correlated, but not to the presence of P. aeruginosa. However, it is likely that the ongoing inflammation is unspecific as a reaction to a multitude of the bacterial pathogens in the wounds. A median of six species (range 2–14) were originally cultivated from the ulcers (2), but no attempt was made to correlate categorised wound bioburdens to the degree of inflammation, because this would require a substantially larger dataset.

We performed multiple regression models outlined in Table 2 to assess whether P. aeruginosa colonisation in CVLUs in combination with selected serological markers for complement activation and inflammation were influential in the deterioration of ulcer healing rate. The output of model 1 only yielded P. aeruginosa as a significantly contributing factor with an area reduction of 6·2% per week compared with ulcers that were culture negative for this pathogen. When C3 was eliminated in model 2, log(CRP) then became significant because C3 alone was dependent on P. aeruginosa (Figure 2B). We could not establish evidence that MBL deficiency was related to the healing rate of CVLUs in this pilot study. However, a large cohort study by Bitsch et al. (16) showed a higher frequency of MBL deficiency among CVLU patients compared with healthy controls, thus suggesting MBL‐supplemented serum therapy as a subject for a controlled clinical trial. This notion is further encouraged by a case report where an MBL‐deficient patient suffering from both chronic leg and surgical wounds experienced improved healing after receiving substitution therapy with plasma‐derived MBL (48). We do not dispute these observations; however, a direct causal effect between MBL deficiency and deteriorated CVLU healing still remains to be established, eventually by a similar cohort study including MBL genotyping of each patient. Removing MBL as an explanatory variable in model 3 rendered neither P. aeruginosa nor log(CRP) influential in CVLU healing rate. It is important to emphasize that the above models should not be regarded as an attempt to develop new prognostic markers for CVLU healing. For this purpose, local inflammatory factors in wound fluid are much more informative and useful (49). Rather, we attempted to investigate selected markers of complement activation and inflammation in CVLUs colonised with P. aeruginosa.

In conclusion, our data consisted with a compensatory role for IgG₁ in response to MBL deficiency and an elevated serum concentration of complement C3 in P. aeruginosa culture‐positive ulcers. The latter may be different in the wound fluids of patients colonised by elastase‐producing pathogens, but we did not detect serum antibodies against this specific virulence factor. FISH analysis showed P. aeruginosa organized as aggregate‐like structures at the wound surface, but we could not establish unambiguous evidence of biofilm morphology. The degree of inflammation as shown by the positive correlation between IL‐6 and CRP did reflect a similar correlation with the presence of P. aeruginosa, which may be the result of a broader immune reaction to a multitude of microorganisms harboured in the ulcers. This pilot study was the first attempt to elucidate the interaction of the humoral immune system with P. aeruginosa colonisation in clinically non infected CVLU patients. Larger clinical studies are desirable to further investigate these aspects, preferably including patients with clinically infected wounds.

ACKNOWLEDGEMENTS

We would like to thank Kristine Gjødsbøl, Statens Serum Institut for providing culture data, serum and biopsy samples of CVLU patients. Project nurses Hanne Vogensen and Lone Haase, Copenhagen Wound Healing Center, Bispebjerg University Hospital for sample collection and patient recruitment. Furthermore, we thank Lone Rabøl and Mai‐Britt Bernitt, Department of Clinical Biochemistry, Statens Serum Institut for TRIFMA and rate‐nephelometry analysis. ASA was supported by a clinical PhD scholarship from the University of Copenhagen, Faculty of Health Sciences. JNJ was supported by a grant from The Danish Research Council for Technology and Production 274‐05‐0435 to KAK, Statens Serum Institut.