Association between mannan-binding lectin and impaired lung function in cystic fibrosis may be age-dependent

Abstract

An association between mannan-binding lectin (MBL) status and severity of lung function impairment in cystic fibrosis (CF) has been found in several studies, but not in others. To explore the possible basis for discrepancies in the literature, we related both MBL and l-ficolin concentrations to lung function and examined the results in relation to the age of the patients. For patients under 15 years of age, those with MBL < 200 ng/ml had better lung function than those with MBL > 200 ng/ml [median forced expiratory volume in 1 s (FEV₁), 99% versus 83%; P = 0·05]. For patients over 15 years of age, those with MBL < 200 ng/ml had poorer lung function than those with MBL > 200 ng/ml (median FEV₁, 44% versus 55%; P = 0·1). Also, for the over 15-year-olds, the proportion of patients with FEV₁ values below the median was greater in the MBL-insufficient subgroup (P < 0·04). In other words, relative deficiency of MBL appears to accelerate the age-related decline in lung function in CF patients. No corresponding relationships could be found between l-ficolin concentration and lung function. These findings and interpretation lend support to the potential value of MBL replacement therapy in a small minority of cystic fibrosis patients.

Introduction

Cystic fibrosis (CF) is an autosomal recessive genetic disorder defined by mutations in a transmembrane conductance regulator gene resulting in a defective epithelial chloride channel [1]. There is considerable variation in the phenotype and course of the disease, even among patients homogeneous for the most prevalent (F508) mutation; both environmental factors and unrelated modifier genes are thought to contribute to disease expression. Chronic Pseudomonas aeruginosa infection is considered a critical event in CF, accelerating lung disease and hastening the development of respiratory insufficiency. Chronic pulmonary infection and inflammation are the principal determinants of morbidity and mortality for CF patients, although life expectancy has improved in recent years with better medical management and the option of lung transplantation.

One putative modifier gene in CF is mbl-2. It encodes mannan-binding lectin (MBL), a plasma pattern recognition protein, allelic variants of which have been correlated with susceptibility to, and the course of, many disorders [2,3]. Genetic variations in mbl-2 are inherited as haplotypes, of which seven may be said to be common. These seven haplotypes give rise to 28 genotypes responsible for a huge (> 1000-fold) variation in functional (complement-fixing) MBL protein in the general population. In 1999, two independent groups found a remarkably similar relationship between variant alleles of MBL and reduced lung function in CF patients [4,5]. Garred et al. [4], with a large cohort of Danish patients (median age 16 years), reported reduced lung function in homozygotes and some heterozygotes for MBL variant alleles. Gabolde et al. [5], with a small number of French patients (mean age 19 years), obtained similar lung function data when MBL variant homozygotes were compared to MBL wild-type patients.

More recent studies of MBL in CF have been less consistent. Two independent groups confirmed the putative relationship in children 12–15 years of age [6] and young people (median age 18 years) [7], respectively, for both homozygotes and heterozygotes of MBL variant alleles. However, Davies et al. [8] found that severe MBL deficiency (homozygotes) was associated with poor lung function in adults, but MBL heterozygotes were not affected, nor was any relationship apparent at all in paediatric (mean age 8·5 years) patients. Carlsson et al. [9] failed to confirm the putative relationship in patients aged 4–45 years of age (median 20·5) at all, except for a subgroup colonized with Staphylococcus aureus. Most recently, by far the largest study (817 patients) found no relationship whatsoever between severity of CF pulmonary disease and MBL polymorphisms [10]. Therefore, collectively, most of the literature is consistent with the proposition that low circulating MBL is linked to poor lung function in white CF patients by the time they reach adolescence, yet that finding was not confirmed in the largest series studied, which must raise considerable doubt.

l-ficolin is another major plasma pattern recognition molecule which, like MBL, has the ability to initiate the lectin pathway of complement activation [11]. Unlike MBL, l-ficolin concentration varies only about fivefold in most healthy individuals, and this variation also appears to have a genetic basis [12]. Also in contrast to MBL, few clinical correlations studies have been performed, but low l-ficolin concentrations have been found in association with respiratory illness in Polish children, especially those children suffering from allergies [13].

In the present investigation, we sought to examine whether relative deficiencies of plasma pattern recognition molecules might be related to impaired lung function in CF patients and whether the age of the patients was a relevant factor in any such relationship.

Patients and methods

Patients

Patients with CF were recruited for donation of blood from the University of North Carolina paediatric and adult CF clinics after obtaining informed consent from patients and parents and assent from children old enough to read. The diagnosis of cystic fibrosis had been established by positive sweat chloride testing and presence of at least one CF mutation according to consensus guidelines [14]. The study was approved by the local Institutional Review Board. Serum was obtained by centrifugation of whole blood in standard serum collector tubes at 250 g for 10 min.

Lung function

Pulmonary function was determined using spirometry. Forced expiratory volume in 1 s (FEV₁) was chosen as the measure of lung function for analysis, expressed as a proportion (%) of the predicted values based on height, race, age and sex.

Assays

MBL and l-ficolin were assayed by enzyme-linked immunosorbent assay (ELISA) methods described previously [15]. Briefly, l-ficolin was captured using one monoclonal antibody (GN4) on the solid phase and detected using another biotinylated monoclonal antibody (GN5). MBL was captured by solid-phase mannan and detected with the monoclonal antibody 131–1; only functional (oligomeric) MBL is detected by that means.

Statistics

Statistical analysis was carried out using Prism for Windows software from GraphPad (San Diego, CA, USA). Differences between means were compared with an unpaired t-test and differences between medians with the Mann–Whitney U-test. Differences in proportions (contingency tables) were expressed as odds ratios and significance computed by Fisher’s exact test. Correlations were computed by both parametric (Pearson) and non-parametric (Spearman) methods.

Results

Patient demographics

A total of 149 patients with cystic fibrosis (79 male, 70 female) were included. Ages ranged from 4 weeks to 45 years (mean 12·8; median 10). Most (91%) were white, 6% were Africa American and the remainder were Hispanic or of mixed race. Of the 103 patients who were older than 5 years, spirometry results were available in 98. Mean FEV₁ was 69% of that predicted for height and weight, ranging between 20 and 135% predicted. FEV₁ was significantly higher in the 48 subjects younger than 15 years than in those older than 15 years (n = 50); 84·7 ± 3·0%versus 54·1 ± 3·3%, P < 0·0001.

P. aeruginosa infection was extremely prevalent, especially in the older subjects. Overall, 103 patients had either a current infection or had been infected previously with this organism, versus 47 in whom neither sputum or bronchoalveolar lavage cultures were ever positive for P. aeruginosa.

MBL insufficiency and lung function

MBL was measured in 148 sera from CF patients. Values ranged from undetectable to 5500 ng/ml, with a median value of 1700 (mean 1800). The distribution was typical of MBL, obviously non-Gaussian with a distinct subgroup of very low values. Values below 200 ng/ml were deemed to represent MBL deficiency for the purpose of analysis. (These sera constitute 18·9% of the samples and all but three of the values were < 100 ng/ml.) Lung function (FEV₁) data were available for 102 of the CF patients. When the MBL-deficient patients were compared with the others for lung function impairment, overall no difference was found (median values, 65% versus 71·5%; P = 0·28). However, it was noticed that there was a dramatic difference in the lung function data obtained from children and those patients who had reached adolescence and beyond. When MBL-deficient patients of 15 years of age and older (n = 12) are compared to MBL-sufficient patients (no age restriction), the median FEV₁ values (45% versus 71·5%) are significantly different (P < 0·009).

When lung function data for patients ≤ 600 ng/ml were compared to those from patients with > 600 ng/ml a significant difference was still evident, but this was clearly due to the contribution made by the lowest concentrations of MBL. While the same trend is present in patients with MBL levels between 100 and 600 ng/ml, the relationship did not reach statistical significance.

The data from patients under and over 15 years of age were analysed separately (Table 1). Whereas, in the younger group, lung function was closer to normal in the MBL-deficient patients, in the older group the MBL-deficient patients had poorer lung function, as expected. The most striking difference between the MBL groups is the change with age. The MBL-sufficient subjects showed a substantial drop in FEV₁ values (median values from 83% to 54%), but the MBL-deficient subjects showed a more dramatic swing from median values of 99% to 44% with age (Table 1).

Table 1.

Mannan-binding lectin (MBL) and lung function in cystic fibrosis (CF) patients.

	MBL (ng/ml)	Mean FEV1 (%)	Median FEV1 (%)
(a) Patients < 15 years old	< 200 (n = 8)	93·4	99
	> 200 (n = 43)	82·3 (P = 0·08)	83 (P = 0·05)
	> 600 (n = 38)	80·5 (P = 0·05)	82·5 (P = 0·03)
(b) Patients > 15 years old	< 200 (n = 11)	46·2	44
	> 200 (n = 39)	56·4 (P = 0·1)	54 (P = 0·13)
	> 600 (n = 32)	57·2 (P = 0·09)	54·5 (P = 0·11)

Comparisons (by Fisher’s exact test) are with the corresponding data for MBL < 200 ng/ml

An alternative statistical analysis is to compare the proportions of patients with below median lung function values in each group. For those < 15 years, three patients had values below and five above the median in the low (< 200 ng/ml MBL) group, and 20 patients had MBL concentrations on either side of the median in the sufficient (> 200 ng/ml) MBL group. That was expected and unremarkable. However, in the older patients, only two of the MBL values were above the median in the MBL-deficient group (eight were below the median). The corresponding figures for the MBL-sufficient group were 23 and 17. This difference in proportions is statistically significant (P < 0·04).

l -ficolin and lung function

l-ficolin was measured in 149 patients. Values ranged from 90 to 6100 ng/ml, with a median value of 3300 (mean 3290). The distribution was normal and similar to that found previously in blood donors. Overall, there was no significant correlation between l-ficolin concentrations and lung function (FEV₁) either for all patients (Pearson’s r = − 0·07; P = 0·48) or for patients over 15 years of age (r = − 0·16; P = 0·25).

Values of 2000 ng/ml OT less were deemed arbitrarily to be low for the purpose of statistical analyses; such values constitute the lowest 17% of the distribution. Overall, there was little difference in lung function between the 19 patients in the low l-ficolin category and the 84 others (median FEV₁, 74% versus 71·5%). Although there was a non-significant trend apparent when only low ficolin patients over 15 were compared to the remaining patients (FEV₁, 54% versus 71·5%; P = 0·13), no relationships were apparent when the data were analysed in an age-matched basis (Table 2).

Table 2.

l-ficolin and lung function in cystic fibrosis (CF) patients.

	l-ficolin (ng/ml)	Mean FEV1 (%)	Median FEV1 (%)
(a) Patients < 15 years old
	≤ 2000 (n = 7)	89·9	84
	> 2000 (n = 44)	83·1 (P = 0·21)	82·5 (P = 0·22)
(b) Patients >15 years old
	≤ 2000 (n = 11)	58·6	54
	> 2000 (n = 40)	53·75 (P = 0·27)	50 (P = 0·37)

Comparisons by Fisher’s exact test.

Only four patients with low l-ficolin also had MBL < 200 ng/ml; all were young children and only one had had a FEV₁ measurement (patients 2, 3 4 and 8: Table 3). Another four patients had low l-ficolin combined with MBL < 600 ng/ml. Three of those were over 15 years of age and had had lung function measured: all three had FEV₁ values below average (Table 3).

Table 3.

Patients with l-ficolin and mannan-binding lectin (MBL) insufficiency.

Patient (sex)	Age (years)	l-ficolin (ng/ml)	MBL (ng/ml)	FEV1(%)
1 (male)	< 1	500	500	n.d.
2 (female)	4	1200	0	n.d
3 (male)	3	1200	40	n.d
4 (female)	11	1300	0	101
5 (female)	33	1700	300	42
6 (male)	16	1900	500	54
7 (female)	30	1900	550	40
8 (male)	3	2000	0	n.d.

Associations with specific infecting organisms

No association between either MBL or l-ficolin insufficiency and any specific infection was apparent in our data. Neither current nor history of P. aeruginosa infection was associated with MBL or l-ficolin levels. MBL is known to bind S. aureus, but no relationship between MBL and current infection with that organism was apparent. However, S. aureus infection is not as persistent as P. aeruginosa and is therefore influenced strongly by recent antibiotic therapy. We also looked for relationships with particular respiratory disorders or features by identifying patients with each condition and comparing the proportion with low MBL or l-ficolin in that group with the patients as a whole. No statistically significant association was found, but the only positive finding concerned low l-ficolin and allergic rhinitis (Table 4). Five patients had both environmental allergy and allergic bronchopulmonary aspergillosis; two had low l-ficolin and two others had low MBL.

Table 4.

Mannan-binding lectin (MBL), l-ficolin and respiratory complications.

	Environmental allergy	ABPA	ARS
No of patients	25	15	11
Patients with MBL < 200 ng/ml	4 [0·8]	3 [1·1]	2 [0·9]
Patients with l-ficolin ≤ 2000 ng/ml	7 [2·1]	2 [0·7]	1 [0·5]

In patients who had symptoms suggestive of allergies, standard skin tests against 10 common allergens were performed. In practice, ‘environmental allergy’ refers to patients with allergic rhinitis with or without asthma. The group with acute respiratory symptoms (ARS) were so classified if their symptoms worsened during the previous month. ABPA: allergic bronchopulmonary aspergillosis. Odds ratios are given in square brackets.

Discussion

Disease association studies involving MBL are often conducted at the DNA level (allele, haplotype or genotype analysis) instead of measuring MBL protein concentration. Ideally, both genotype and protein data should be analysed. The principal justification for genetically based investigations is that they utilize a non-variable feature, whereas serum MBL is subject to internal environmental influences. However, genotype analysis is feasible only by grouping genotypes together (there are around 30 MBL genotypes), and any disease association might be the consequence of linkage disequilibrium with an unrelated susceptibility gene and therefore might have nothing whatever to do with MBL. Only serum MBL measurements permit a direct relationship between concentration (activity) and disease to be established. Moreover, although individuals who are homozygous (O/O) for structural alleles invariably have low circulating MBL, it is not generally appreciated that a substantial proportion of heterozygotes (A/O) have normal and, indeed, sometimes very high concentrations. It is also a myth that MBL is an acute-phase protein in the usual sense of that term [16]. It is true that populations of sick people often have slightly higher average MBL concentrations than blood donors, but in individuals MBL is generally very stable over time. All these comments refer to MBL detected by methods, such as our own, that detect functional (oligomeric) MBL; assays that detect the basic triplet subunit [17,18] are certainly unreliable for clinical studies. It is our view that serum MBL is a better single parameter for clinical correlation studies than haplotype/genotype, but we accept this is an area of controversy among experts.

The results described here are consistent with a weak relationship between MBL insufficiency and poor lung function in cystic fibrosis, but not for younger (< 15 years) paediatric patients. This is similar to the findings of Davies and coworkers [8], who compared young children (mean age 8·5 years) with mature adults (mean age 29·7 years). We suggest that the adverse relationship with MBL insufficiency may become apparent at an age threshold corresponding to adolescence or puberty, a threshold that does not translate precisely into years of age, rather than true adulthood. Although puberty may be delayed in patients with CF, that interpretation fits best with our data and could explain the apparent discrepancies in the literature, where different cohorts have included varying proportions of young and late teenagers.

Indeed, we made the surprising observation that lung function was actually better in young children with low MBL concentrations. Although we are the first to note this, in fact Davies et al. [8] also found slightly higher FEV₁ values in children who were homozygous for structural MBL mutant alleles. Our observation may not be as paradoxical as it first appears. Although MBL protects against pathogens and promotes clearance of apoptotic cell debris, high concentrations of MBL may cause tissue damage through excessive complement activation. The latter mechanism has been suggested in the context of diabetes [19], primary biliary cirrhosis [20] and rheumatic heart disease [21], for example. It is possible that, in children, the minority with very high levels of MBL experience more severe lung inflammation which manifests itself as a slight drop overall in FEV₁ values. The MBL-insufficient children are more susceptible to infections, however, and in time prolonged chronic infection will manifest itself increasingly as impaired lung function. This change in the net effect of MBL insufficiency seems to occur in the mid-teens, and may be accelerated by the onset of puberty. It may be pertinent that growth hormone [22], thyroxine [23] and glucocorticoids [24] can significantly affect circulating MBL levels in humans. It is also of interest that a very strong association between MBL deficiency and coronary artery lesions in Kawasaki disease was reported for children younger than 1 year, but the trend was weakly in the opposite direction for children aged 1 year or older [25]. Our finding that the MBL–disease relationship is reversed with a brief passage of life is therefore not without precedent.

The average MBL concentration in CF patients was rather higher than would be expected for healthy controls [26]. This was unsurprising, however, as it has been our experience that sick people generally have elevated MBL, and this presumably reflects an acute phase-type response to chronic illness and infection. Carlsson et al. [9] also reported that CF patients had higher levels of MBL than healthy controls, even when corrected for MBL genotypes. In contrast, the median and mean concentrations of l-ficolin were similar to those found previously in blood donors [26].

We found no indication that relative deficiency of l-ficolin accelerated the age-related decline in lung function in CF patients. Although it was not the purpose of this study to examine allergic complaints in relation to l-ficolin status and the numbers available were too low to provide conclusive data, yet it is interesting none the less that in patients who also had allergic respiratory disease, low concentrations of l-ficolin were over-represented. This is similar to our previous finding in children and young people with recurrent respiratory infections complicated by asthma or allergic rhinitis [13].

The most important contribution of this study to the literature is in highlighting the potential importance of age and physical development in understanding the relationship between MBL and lung function in the context of CF. Future studies should address this aspect carefully. If it can indeed be confirmed that MBL insufficiency during childhood contributes significantly to severe pulmonary impairment in adulthood, that would be a powerful and cost-effective justification for MBL replacement therapy [27].