OUP user menu

Low mannose-binding lectin function is associated with sepsis in adult patients

Damon P. Eisen, Melinda M. Dean, Peter Thomas, Penny Marshall, Natalie Gerns, Sue Heatley, Josephine Quinn, Robyn M. Minchinton, Jeff Lipman
DOI: http://dx.doi.org/10.1111/j.1574-695X.2006.00144.x 274-282 First published online: 1 November 2006


Mannose-binding lectin (MBL) is an innate immune system pattern recognition molecule that kills a wide range of pathogens via the lectin complement pathway. MBL deficiency is associated with severe infection but the best measure of this deficiency is undecided. We investigated the influence of MBL functional deficiency on the development of sepsis in 195 adult patients, 166 of whom had bloodstream infection and 35 had pneumonia. Results were compared with 236 blood donor controls. MBL function (C4b deposition) and levels were measured by enzyme-linked immunosorbent assay. Using receiver operator characteristics of MBL function in healthy controls, we identified a level of <0.2 U µL−1 as a highly discriminative marker of low MBL2 genotypes. Median MBL function was lower in sepsis patients (0.18 U µL−1) than in controls (0.48 U µL−1, P<0.001). MBL functional deficiency was more common in sepsis patients than controls (P<0.001). MBL functional deficient patients had significantly higher sequential organ failure assessment (SOFA) scores and higher MBL function and levels were found in patients with SOFA scores predictive of good outcome. Deficiency of MBL function appears to be associated with bloodstream infection and the development of septic shock. High MBL levels may be protective against severe sepsis.

  • innate immunity
  • deficiency
  • mannose-binding lectin
  • sepsis


Severe sepsis and septic shock are common clinical syndromes that have a high mortality. In a recent Australian and New Zealand study the calculated annual incidence of severe sepsis was 0.77 per 1000 of the population (Finfer et al. 2004). About 12% of intensive care unit admissions were for such patients with about 30% of these patients dying in hospital. These figures are not dissimilar to those of the United States and the United Kingdom (Angus et al. 2001; Padkin et al. 2003).

Our most primitive barriers against infection include the physical and mechanical barriers, soluble molecules and cellular defences of the innate immune system. The innate immune response to infection is typified as rapid, co-operative and efficient and is mediated through recognition of pathogen-associated molecular patterns. Severe sepsis arises when the innate immune system fails to control initial infection and a dysregulated inflammatory response results (Oberholzer et al. 2001). Genetic variability in the innate immune system may contribute to altered susceptibility to sepsis.

Mannose-binding lectin (MBL) is a central player in the innate immune response. This C-type serum lectin, produced by the liver, binds microbial surface carbohydrates and mediates opsonophagocytosis directly and by activation of the lectin complement pathway. Reduced MBL levels and function result from MBL2 structural gene and promoter gene polymorphisms. Three polymorphisms in the structural gene at codons 52, 54 and 57 encode for variant alleles referred to as D, B and C (and collectively as O) (Sumiya et al. 1991; Lipscombe et al. 1992; Madsen et al. 1994) with the wild-type gene being A. Promoter gene polymorphisms at positions −550 and 221 of the 5′ flanking region of the MBL2 gene encode for alleles H/L and X/Y that also influence MBL levels (Minchinton et al. 2002). MBL deficiency is common, as indicated by a recent Australian study of normal blood donors, in which 30% were found to be heterozygous for structural gene mutations and a further 8% were homozygous or had double mutations of the structural genes (Minchinton et al. 2002). Furthermore, 24% of this group were MBL deficient using a composite definition of low MBL levels and function (Eisen et al. 2004). MBL deficiency has been associated with susceptibility to severe bacterial infection in numerous clinical studies (Eisen & Minchinton, 2003). For example, carriage of MBL variant alleles was associated with meningococcal disease susceptibility in paediatric patients (Hibberd et al. 1999) and bacteraemic pneumococcal pneumonia has been associated with MBL deficiency in adults (Roy et al. 2002).

We have explored the possible association between MBL deficiency and increased susceptibility to severe sepsis through an unmatched case-controlled clinical study. To substantiate the basis for choosing a single measure of MBL function as the definition of deficiency, levels in healthy controls were analysed to determine the most discriminative level between those with high and low MBL2 haplotypes.



The Royal Brisbane and Women's Hospital is a 750-bed tertiary referral centre serving the northern region of Brisbane, Queensland, Australia. Over the period 19 December 2001 to 27 June 2003, patients with proven sepsis due to bloodstream infection were approached for consent for involvement in the study. A smaller number of patients with proven pneumonia were also included at the study outset while enrolments of bloodstream-infected patients were assessed. Only patients with definitive evidence of infection were considered for study entry. Patient diagnoses that were included were: bloodstream infection (isolation of at least one pathogenic bacteria or yeast from blood cultures), community-acquired pneumonia [lobar or bronchopneumonic infiltrate on chest X-ray (CXR) in patients with neutrophil leukocytosis] or bacterial meningitis (polymorph pleocytosis in cerebrospinal fluid with positive culture). Patients with blood cultures growing only skin contaminants were excluded. The clinical investigators verified the presence of CXR changes in patients with admission diagnoses of pneumonia. MBL parameters had been previously measured in 236 Queensland blood donors who were used as unmatched healthy controls to define a normal population (Minchinton et al. 2002). The Royal Brisbane and Women's Hospital Human Research Ethics Committee and the Queensland Guardianship Administration Tribunal gave approval for the hospital study.

Information on bloodstream isolates was collected; in some instances multiple pathogenic isolates were present in the same specimen. In patients with severe community-acquired pneumonia, respiratory pathogens grown from blood cultures were identified as the cause of the pneumonic illness. Clinical data were acquired by investigators blinded to MBL results (P.T., P.M. and J.Q.). The principal investigator (D.P.E.), who reviewed microbiological results for all possible patients, was blinded to all MBL genotyping and phenotyping results. Demographic information was collected on all patients. Apache II scores (Knaus et al. 1985) were calculated on all case patients in the first 24 h of their presentation with sepsis. Sequential organ failure assessment (SOFA) scores (Moreno et al. 1999) were calculated over the first week of illness. Survival was recorded at 28 day postsepsis onset.

MBL assays

All MBL assays were performed on EDTA samples of blood taken at the time of diagnosis of bloodstream infection or pneumonia.

C4b deposition assay

This assay demonstrates deposition of C4b following activation of MBL by binding with solid-phase purified mannan. This method was originally described for detection of C3b and C3bi (Super et al. 1989) and modified for C4b deposition (Valdimarsson et al. 1998). Using serum dilutions of 1 : 20 or higher, the C4b deposition assay provides a highly significant correlation between deposition of C4b and MBL levels determined in a mannan-binding enzyme-linked immunosorbent assay. There was no activation of either the classical or the alternative complement pathways using solid-phase coupled mannan and no correlation between MBL levels and antimannan antibodies in blood donor samples (Super et al. 1990). Following binding of MBL to solid-phase mannan, plates were washed and MBL-deficient serum was added to enable complement activation. Details of this method were previously described (Minchinton et al. 2002). One microlitre of Antibody Shop MBL standard (AntibodyShop) was arbitrarily assigned one unit of C4b deposition activity. The between-run coefficient of variation (CV) for this assay was 9.4% using an internal plasma control with calculated C4b deposition of 0.54 U µL−1.

Mannan-binding assay

MBL level was quantified (µg mL−1) for each sample using a mannan-binding assay based on the method of Holmskov (1993) except that biotinylated HYB 131-01 monoclonal anti-MBL (AntibodyShop) was used to tag bound MBL. Briefly, standards, controls and patient plasma were applied to mannan-coated microtitre plates for 90 min. Plates were then washed and bound MBL detected using human anti-MBL monoclonal antibody (HYB 131-01). This method was also previously described in detail (Minchinton et al. 2002). The CV over eight assays using an internal plasma control containing 2 µg mL−1 MBL was 6.1% at 1 : 25 sample dilution and 8.8% at 1 : 100 sample dilution.

MBL genotyping

DNA was extracted from buffy coats using the QIAamp DNA kit (QIAGEN). Five polymorphisms in the MBL2 promoter (−550 H/L, −221 X/Y) and the first exon (codons 52, 54 and 57) were determined using the PCR and sequence-specific primers as previously described (Mullighan et al. 2000; Minchinton et al. 2002).

Statistical analysis

We defined MBL deficiency on the basis of a functional assessment of MBL activity using the C4b deposition assay. Receiver operator characteristics (ROCs) of C4b deposition as a predictor of high (A/A, HYA/O and LYA/O) MBL2 genotypes (Mullighan et al. 2002) in healthy blood donor controls were plotted using Stata 9 (StataCorp). Comparisons of discrete outcomes including frequency of MBL-deficient individuals among cases and healthy controls and the relationship between MBL deficiency and shock and survival were performed using the chi-squared test. Comparisons of continuous values such as C4b deposition and mannan binding were performed using Mann Whitney and Kruskal Wallis tests in two or more groups. Statistical analyses were performed using Minitab® Release 14 (Minitab Inc.). Hardy Weinberg equilibrium of MBL2 genotypes was assessed using Arlequin software package v2.000 (Schneider et al. 2000).



A total of 195 patients with documented bloodstream infection or community-acquired pneumonia were enrolled as cases. These consisted of 166 patients with documented bloodstream infection, 35 with pneumonia and 12 with bacterial meningitis. Seven of the pneumonia patients and 11 of the 12 patients with meningitis were bacteraemic. The median age was 61 years (range 16 95); 110 patients were male and 85 female. Patients with mean arterial pressures of <70 mmHg were defined as having septic shock as information on blood pressure response to intravenous fluid administration was not available in all cases. Thus, 114 patients had septic shock and 81 had sepsis alone. The organisms grown from the blood cultures of the patients are detailed in Table 1. MBL status had previously been documented (Minchinton et al. 2002) in 236 blood donor controls and these acted as healthy controls. This group was representative of Queensland blood donors, who have a median age of 45 years (range 18 70) and male to female ratio of 1.11 : 1.

View this table:
Table 1

Sites of infection and causative agents

Bloodstream infection isolatePolymicrobial infectionPneumoniaMeningitis
Staphylococcus aureus361Lobar21N. meningitidis6
Streptococcus pneumoniae12Multilobar7S. pneumoniae1
Other Streptococcus species81Bronchopneumonia7H. influenzae1
Enterococcus faecalis44S. aureus1
Streptococcus pyogenes3E. faecalis1
Other Staphylococcus species11Nocardia nova1
Listeria monocytogenes1Aseptic1
Escherichia coli332
ESCAPPM16Aetiological agent
Klebsiella pneumoniae12S. pneumoniae6
Neisseria meningitidis93S. aureus1
Pseudomonas aeruginosa72
Pseudomonas sp.4
Salmonella sp.3
Haemophilus influenzae2
Klebsiella sp.2
Campylobacter jejuni1
Nocardia nova1
Proteus mirabilis11
Candida albicans21
Other Candida species3
  • Cerebrospinal fluid white cell count 6500 × 106 L−1, glucose 4.7 mM L−1, protein 3.9 g L−1.

  • Enterobacter sp., Serratia marcescens, Citrobacter freundii, Acinetobacter sp., Providencia stuartii, Proteus vulgaris, Morganella morganni.

MBL deficiency definition

The ROC curve of high MBL2 genotypes as predicted by C4b deposition results in healthy individuals is shown in Fig. 1. A C4b deposition of >0.2 U µL−1 was 85% sensitive and 83% specific for high MBL2 genotypes. The area under the ROC curve was 0.92. The positive likelihood ratio for a result >0.2 U µL−1 was 4.87. The area under the ROC for MBL level as a predictor of MBL2 genotype was 0.93 and an MBL level >0.25 µg mL−1 was 96% sensitive and 73% specific for high MBL2 genotypes. However, its positive likelihood ratio of 3.69 was less than that for the chosen C4b deposition cut-off. Therefore, the definition of MBL deficiency used in this study is C4b deposition <0.2 U µL−1.

Figure 1

Receiver operator characteristic curve of high mannose-binding lectin (MBL) genotypes as predicted by MBL functional assay.

MBL deficiency in case and control groups (Table 2)

View this table:
Table 2

MBL measurements and MBL2 genotypes in septic patients and blood donor controls

ShockSepsisAll cases of sepsisControls
Median C4 deposition (U µL−1)
C4 deposition < 0.2 U µL−157/11443/81100/19567/236
Odds ratio C4 < 0.2 U µL−1 (95% CI)2.66 (1.75–4.04)
Comparison all cases and controlsP<0.0001
Median MBL (µg µL−1)1.431.051.211.55
MBL level < 0.25 µg mL−125/11417/8142/19541/236
Odds ratio MBL < 0.25 µg µL−1 (95% CI)1.31 (0.79–2.17)
Comparison all cases and controlsP=0.27
Low MBL2 genotypes16/10518/6534/17046/236
  • Differs from healthy controls, P<0.001.

  • Not significantly different from healthy controls.

  • LXA/LXA, LXA/O, O/O.

C4b deposition was significantly lower in case subjects with bloodstream infection and pneumonia than in healthy controls. Similarly, the frequency of deficient C4b deposition levels was significantly higher in cases than in healthy controls. Overall, case patients with septic shock could not be distinguished from those with sepsis without shock on the basis of C4b deposition. C4b deposition was closely correlated with MBL levels in the cases studied herein (Pearson's coefficient R=0.673, P<0.001). There was a nonsignificant trend to lower MBL levels in cases compared with controls. There was no difference in the frequency of MBL levels >0.25 µg mL−1, indicative of high MBL2 genotypes, in cases and controls (P=0.27).

MBL allelic frequencies in septic patients and control groups (Table 3)

View this table:
Table 3

MBL2 structural gene and promoter genotype in septic patients and blood donor controls

MBL2 genotypeCasesLow genotypesControlsLow genotypes
Total A/A103136
Total A/O6182
Total O/O661818
Total17034 (20%)23646 (19.5%)
  • A, wild-type MBL2 allele; O, variant MBL2 allele.

MBL genotype was available for 170 of the 195 case subjects. In the remainder, amplification of the MBL genes from the available samples was unsuccessful despite repeated attempts. Genotype data had previously been derived in all of the healthy controls. The allelic frequencies in the cases (A 0.79, B 0.13, C 0.02, D 0.06) and healthy controls (A 0.75, B 0.14, C 0.03, D 0.08) were identical and MBL2 genotypes were in Hardy Weinberg equilibrium, indicating the genetic homogeneity of the study groups. There was no difference in the frequency of carriage of MBL2 variant alleles (A/O or O/O) between the septic cases and control groups. Similarly, there were no differences in the frequencies of low MBL2 genotypes (LXA/LXA, LXA/O and O/O) between the study groups.

Comparison between MBL-deficient and MBL-sufficient case patients (Table 4)

View this table:
Table 4

Case patient characteristics and clinical outcomes in relation to MBL functional deficiency, MBL level and frequency of low MBL2 genotypes

MBL functionally deficient (C4<0.2 U µL−1) n=100PMBL functionally sufficient (C4>0.2 U µL−1) n=95POdds ratio95% CILow MBL2 genotypesPOdds ratio95% CI
M : F49 : 5161 : 340.0322/95 : 12/750.24
WCC (106 L−1) day
Septic shock vs. sepsis
Median C40.055 : 0.091<.010.48 : 0.530.96
Median MBL levels0.35 : 0.390.913.28 : 3.460.71
Disease outcomes16/105 : 18/65<0.050.470.20–1.07
    Frequency of shock57/10043/950.11.60.88–2.9416/105 : 18/65<0.050.470.20–1.07
    Apache II16.6115.910.5NS
    SOFA day 04.364.270.87NS
    SOFA day 34.313.150.05NS
    SOFA day 53.513.120.5NS
    Death13/10014/950.720.830.36–2.091/34 : 22/136<0.050.160–1.05
  • LXA/LXA, LXA/O, O/O.

  • Not significant.

  • Sequential organ failure assessment.

  • Fisher's exact test, two-tailed.

MBL functionally deficient patients with septic shock had significantly lower C4b deposition than those who were not shocked. MBL levels also trended to be lower in these MBL functionally deficient, septic shock cases. Among MBL functionally sufficient patients, no similar differences were seen in MBL function or levels between the shocked and normotensive septic groups. There was a trend to higher frequency of low MBL2 genotypes in normotensive than shocked patients. Clinical indices and outcomes were compared in MBL functionally deficient and sufficient sepsis patients. MBL functional deficiency was significantly associated with higher SOFA scores at day 3 with no significant differences seen in Apache II and day 0 or day 5 SOFA scores. Low MBL2 genotypes were not significantly associated with these disease indicators (data not shown). MBL functional deficiency was not associated with death due to sepsis but there was a nonsignificant trend to improved survival with carriage of low MBL2 genotypes [one death in 34 patients with low MBL2 genotypes compared with 22 deaths in 136 patients with high MBL2 genotypes, P<0.05, odds ratio 0.16 (95% CI 0.0 1.05)].

Influence of MBL on SOFA scores and associations with survival

The median day 3 SOFA score was significantly higher in nonsurvivors than in survivors (5 vs. 2, P=0.03). MBL function was significantly different in patients with SOFA scores above and below 5 at day 3 (0.13 vs. 0.21 U µL−1, P<0.02). The original SOFA score validation study (Vincent et al. 1998) showed that the mean SOFA score was 8 in those who died and 4 in survivors. Using these prognostic breakpoints to analyse MBL activity in our patients showed significant differences in both MBL level and function. For example, at day 3, MBL levels (1.37 vs. 0.51 µg mL−1, P<0.01) and function (0.22 vs. 0.08 U µL−1, P<0.001) were higher in patients with SOFA scores of <8 than in those with SOFA scores of 8+. Similarly, patients with day 3 SOFA scores of 4 or below had higher MBL function (0.21 vs. 0.13 U µL−1, P<0.02) and there was a trend to higher MBL levels (1.40 vs. 0.85 µg mL−1, P=0.07) than in those with SOFA scores of greater than 4. Multivariable analysis showed that MBL function remained a significant predictor of day 3 SOFA scores <8 (P<0.02) when age and white cell counts were taken into account.

Associations between bloodstream isolates and MBL function (Table 5)

View this table:
Table 5

Associations between MBL activity, bloodstream isolates and site of disease

Median C4b (U µL−1)PMedian MBL (µg mL−1)P
Bloodstream isolates
    Streptococcus pneumoniae : other bacteraemia0.07 : 0.22<0.0010.37 : 1.33<0.01
    Streptococcus pneumoniae : other gram-positive bacteria0.07 : 0.23<0.0010.37 : 1.62<0.01
    E. coli : other bacteraemia0.15 : 0.22<0.050.73 : 1.400.09
    E. coli : other gram-negative bacteria0.15 : 0.300.010.73 : 1.830.01
    Staphylococcus aureus : other bacteraemia0.27 : : 1.210.45
    Staphylococcus aureus : other gram-positive bacteria0.27 : : 0.660.06
    N. meningitidis : other bacteraemia0.46 : : 1.210.25
    Gram-positive : gram-negative bacteria0.18 : : 1.280.3
    Streptococcus pneumoniae : nonbacteraemic pneumonia0.07 : : 1.200.04
    Pneumonia : meningitis : bacteraemia0.16 : 0.20 : : 1.27 : 1.210.9*
    Bacteraemia : nonbacteraemia0.19 : 0.1711.22 : 1.140.9
  • Kruskal–Wallis test.

Median C4b deposition levels were significantly lower in patients with Streptococcus pneumoniae and Escherichia coli sepsis than in patients with other gram-positive and gram-negative bacteraemias, respectively. Patients with pneumococcal bloodstream infection had significantly lower MBL function and levels than patients with nonbacteraemic pneumonia. Interestingly, Staphylococcus aureus, the commonest cause of bacteraemia, and Neisseria meningitidis were found in patients with median C4b deposition above the deficient range. Low MBL2 genotypes were not significantly associated with specific organisms or sites of disease (data not shown).


This is the first study of the influence of MBL deficiency on the development of sepsis to include patients purely on the basis of severe, documented infection and to assess MBL activity on a functional as well as quantitative basis. We have shown a strong association between functional MBL deficiency and sepsis. Our use of ROC curves to identify a discriminative level of MBL function predictive of MBL2 haplotypes represents an advance in the assessment of MBL deficiency. Applying our ROC-derived C4b deposition cutoff to sepsis cases gave a positive likelihood ratio of 20, significantly higher than the MBL level cutoff of <0.25, which had a positive likelihood ratio of only 3.12. Patients with bloodstream infection and pneumonia showed markedly reduced ability to activate complement through the MBL pathway compared with healthy controls. Among MBL functionally deficient patients, lower levels of C4b function were present in patients with septic shock. Although there was no significant association between MBL functional deficiency and death in our study, important associations with SOFA scores were found. Higher MBL C4b deposition function and mannan binding levels were seen in patients with SOFA scores predictive of good outcome (Vincent et al. 1998). Noncomplement-binding, monomeric, dimeric and trimeric MBL with higher disassociation constants can still be measured in the mannan binding assay, albeit at a low efficiency (McGreal et al. 2004; Dean et al. 2006). This might contribute to the absence of association between MBL level and sepsis. We did not demonstrate higher than expected carriage of MBL variant structural alleles or promoter haplotypes in the sepsis patients studied. This reflects the wide range of MBL activity within A/O patients, particularly those with the A/B genotype (Minchinton et al. 2002; Dean et al. 2006), and the small number of O/O patients in our study. The potential effect of low MBL2 genotypes in protecting against death in our study is questionable given that only one event occurred in this group and that it is the converse of the results seen in other studies of MBL deficiency in systemic inflammatory response syndrome (SIRS) patients (Garred et al. 2003).

Our major finding of significantly reduced MBL function in patients with sepsis does not appear to be due to the impact of the inflammatory cascade in this group. We have previously reported the temporal change in MBL activity in the sepsis patients described in this study (Dean et al. 2005). There was no consistent acute phase response, either positive or negative, in MBL activity. Patients with low MBL levels or function mostly maintained those levels. Patients with MBL2 variant alleles who were MBL deficient at the onset of their sepsis were unable to increase either their MBL level or function into the normal range. This lack of normalization of MBL levels in A/O patients under infective stress has previously been reported in paediatric oncology patients (Neth et al. 2001). A small but significant early decrease in MBL levels after colectomy and subsequent modest increase at day 8 postoperatively has been demonstrated. This contrasted with marked increases in acute phase reactants C-reactive protein and interleukin-6 (IL-6) (Ytting et al. 2006). While a recent report from a mouse model shows that levels of both MBL and C4b cleavage were significantly decreased after development of sepsis (Windbichler et al. 2004), complement components are added to the C4b deposition assay used here, ruling out sepsis-mediated complement consumption. Lower C4b deposition was seen in MBL functionally deficient patients with septic shock compared with those with sepsis and no shock. This same relationship was not seen among MBL-sufficient septic patients, further suggesting the C4b deposition deficiency is predisposing to and not merely reflecting shock.

Given our inclusion criteria of proven, mostly bloodstream, infection, we have been able to observe directly the association between MBL deficiency and sepsis due to different bacteria. Streptococcus pneumoniae sepsis was strongly associated with MBL deficiency with lower MBL function and levels than other patients with bacteraemias and nonbacteraemic pneumonia (the majority of which will be pneumococcal). Escherichia coli sepsis was also associated with MBL functional deficiency. Interestingly, we found no association between N. meningitidis or Staphylococcus aureus sepsis and MBL deficiency. This is despite previous clinical studies linking meningococcal sepsis with MBL deficiency in children and recent data that show MBL-null mice are highly susceptible to lethal Staphylococcus aureus sepsis. As is frequently the case, it is difficult to generalize from animal models and disease patterns in children to illness characteristics in adults.

Other clinical studies support the association we have shown between MBL deficiency and SIRS. Among Danish patients with SIRS, MBL variant alleles were found more frequently in those with documented sepsis. It appeared that patients with the lowest MBL levels were most likely to develop septic shock with a fatal outcome. (Garred et al. 2003). Patients homozygous for MBL variant alleles were predisposed to invasive pneumococcal infection in a case-controlled study of Caucasian patients (Roy et al. 2002). Neither of these studies directly correlated MBL level or functional assays with the risk of sepsis. In paediatric patients with organ failure, MBL deficiency, assessed by genotype or low MBL levels, was significantly associated with the development of SIRS. Increasing proportions of patients with localized infection, sepsis and septic shock had MBL deficiency (Fidler et al. 2004). MBL levels were also studied in a randomized intensive care unit study of treatment with intensive or conventional insulin regimens. Although admission MBL levels were normal overall, low MBL levels were associated with death in those patients conventionally treated with insulin. Multivariate analysis showed that the insulin regimen, but not baseline MBL levels, predicted survival (Hansen et al. 2003). Another study of adult patients revealed an association between MBL2 genotype and MBL levels (measured in fewer than half the subjects) and the presence of severe shock although no information on infective organisms was provided (Gordon et al. 2006).

MBL appears to be an important modulator of the inflammatory cascade. Proinflammatory cytokines are needed for an effective immune response to infectious disease. MBL has been shown to enhance production of tumour necrosis factor (TNF) (Huffnagle & McNeil, 1999) by mononuclear cells bound to microorganisms, preventing disseminated infection. Septic shock, however, results from dysregulated production of TNF. In an animal model of MBL deficiency, MBL-null mice had reduced early levels of proinflammatory cytokines in Staphylococcus aureus sepsis but higher levels than control wild-type mice after 24 h (Shi et al. 2004). Production of TNF and IL-6 by monocytes incubated with N. meningitidis was decreased by the presence of MBL in a dose-dependent fashion (Jack et al. 2001). Gram-negative lipopolysaccharide-containing mannose heteropolymers have the potential for causing the most profound shock in experimental animals. MBL binds most avidly to mannose heteropolymer containing lipopolysaccharide (Zhao et al. 2002). Proof of the role of MBL in protection against severe infection comes from the study of MBL-null mice showing reduced survival from Staphylococcus aureus bloodstream infection compared with MBL wild-types (Shi et al. 2004).

MBL replacement therapy will potentially be available in the near future. Both plasma-derived (Laursen, 2003) and recombinant MBL (Vorup-Jensen et al. 2001) are being prepared for clinical trials. Possible uses include elective MBL replacement therapy in bone marrow transplantation or acute treatment of septic patients with MBL functional deficiency. Hypersupplementation of MBL levels may also be beneficial as MBL does not demonstrate predictable positive acute phase activity and we have shown that high MBL levels predict low SOFA scores.

This study is the first to select patients for documented, mostly bloodstream, infection and show an association between MBL functional deficiency and sepsis. Of all the innate immune effector molecules, MBL is potentially best suited to the therapy of infectious diseases. Infusions of MBL have a good safety profile and have a reasonably long half-life (Valdimarsson et al. 2004). Physicians managing infectious diseases are faced with major challenges from increasingly resistant pathogens such as penicillin-resistant pneumococci and multiresistant gram-negative bacteria such as E. coli. A new immune-based therapy would not be expected to select for antimicrobial resistant organisms and would be a welcome addition in this context. Newly accumulating information about the innate immune system's role in prevention of sepsis may help reduce the high mortality associated with this condition.


D.P.E. receives support from The Clinical Centre for Research Excellence in Infectious Diseases, Victorian Infectious Diseases Service, Royal Melbourne Hospital. This study was supported by a grant from the Cooperative Research Centre for Vaccine Technology. We acknowledge helpful statistical advice from Drs Emma McBryde and Jim Black.


  • Editor: Artur Ulmer


View Abstract