Respiratory Deleted in Malignant Brain Tumours 1 (DMBT1) levels increase during lung maturation and infection

Infants and tracheal aspirates

The study group comprised 82 mechanically ventilated newborn infants (40 males, 42 females) treated at the Perinatal Center of the University of Heidelberg. The study was approved by the Ethics Committee and the parents agreed by informed consent. Gestational age is defined as time elapsed between the first day of the last menstrual period and the day of delivery. The mean values of the infants' gestational age derived from the maternal history and were confirmed clinically. The total study population had an averaged gestational age of 28·5 ± 4·4 weeks (range: 23–40 weeks; 23–25 weeks: 29; 26–29 weeks: 32; 30–36 weeks: 16; 37–40 weeks: 5). Twenty-two of the 82 infants were small-for-gestational-age (SGA) at birth, defined as birth weight below the 10th percentile (Bavarian growth charts).

A total of 205 tracheal aspirates were collected from 82 infants. The first sample was taken during the first 3 days from 40 infants and between days 4 and 120 from the other 42 infants. Subsequent samples were obtained every 2–4 days and stored immediately at −20°C. Sterile isotonic sodium chloride (0·9%; 0·5–1·0 ml) was instilled into the endotracheal tube and the trachea was suctioned once or twice after three breaths of manual or mechanical ventilation. The total protein concentration of the tracheal aspirates was measured by using the bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL, USA).

Production and purification of human recombinant DMBT1 (hrDMBT1)

Human rDMBT1 was produced and purified in the laboratory of the Institute of Molecular Biology and Cell Culture Technology, University of Applied Sciences Mannheim, as described elsewhere [18]. Briefly, the hrDMBT1 producer cell line T3 was induced with 10 μg/ml tetracycline for 48 h. After collecting the hrDMBT1-containing supernatant, the protein was isolated by mixing the supernatant with a suspension of Streptococcus gordonii. Phosphate-buffered saline (PBS) containing 10 mM ethylenediamine tetraacetic acid (EDTA) was used to elute bound hrDMBT1. The eluted protein was concentrated and applied to a size-exclusion chromatography column.

Monoclonal and polyclonal antibodies used for enzyme-linked immunosorbent assay (ELISA), Western blot and immunohistochemistry

The DMBT1-specific monoclonal antibody Hyb213-06 used for ELISA measurements and Western blotting was raised against DMBT1^gp340 isolated from broncheoalveolar lavage, as described by Holmskov [19]. The specificity of the antibody was confirmed by Western blotting [20]. The monoclonal antibody anti-DMBT1h12 was generated against a synthetic peptide (RTTDYASLIPSEVPLC) corresponding to amino acids 26–40 of DMBT1 [4]. The polyclonal anti-serum (anti-DMBT1p84) was made against the recombinant DMBT1 and its specificity was confirmed by a comparison of immunhistochemical sections with RNA in situ hybridization (data not shown).

Determination of the DMBT1 concentration in tracheal aspirates by ELISA

As concentration standard we used purified hrDMBT1. Microtiter plates (Microlon, F-Shape; Greiner Bio-One GmbH, Essen, Germany) were coated with tracheal aspirate fluids overnight at 4°C. After washing with Tris-buffered saline (TBS) containing 0·1% Tween 20 (TBS-T), the plates were incubated with the DMBT1-specific monoclonal mouse antibody Hyb213–06 (Antibodyshop, Gentofte, Denmark [19,20]) diluted 1 : 5000 for 1 h at room temperature. Subsequently, the wells were incubated with a horseradish peroxidase-conjugated anti-mouse antibody (1 : 5000) for 1 h at room temperature. The bound enzyme was detected by adding TMB-substrate solution [125 μg/ml 3,3‵,5,5‵-tetramethyl-benzidine; 125 μg/ml in 0·1 M citrate buffer pH 4·5 with 0·05% (v/v) H₂O₂]. After incubation for 20 min in the dark, the reaction was stopped by addition of 2 M HCl. The intensity of the dye reaction was measured at 450 nm in an ELISA reader (EL800; Bio-Tec Instruments Inc., Winooski, VT, USA). The DMBT1 concentration of each sample was determined at least three times, and the DMBT1 level was calculated as percentage of total protein concentration in order to adjust for differential dilution factors. Thus, DMBT1 was expressed as (μg DMBT1/μg total protein) × 100.

Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analyses

Proteins of the tracheal aspirate samples were separated under non-reducing conditions on 4–15% polyacrylamide gels (Phast Gels™, Amersham Biosciences, Freiburg, Germany). Afterwards, proteins were transferred onto polyvinylidine difluoride (PVDF) membranes (Millipore, Schwalbach, Germany), which were blocked subsequently at 4°C overnight with 10% (w/v) milk and 3% (v/v) bovine serum albumin in TBS containing 0·1% Tween 20 (TBS-T). All blots were made in duplicate to detect DMBT1 with the monoclonal antibody Hyb213–06 (Antibodyshop [19,20]), and to detect the secretory component of IgA (sIgA) with an anti-serum from rabbit against the secretory component (Dako, Hamburg, Germany) diluted 1 : 5000 and 1 : 1000 in blocking solution, respectively. As secondary antibodies, goat anti-mouse-IgG and goat anti-rabbit-IgG conjugated to alkaline-phosphatase (Santa-Cruz Biotechnology, Heidelberg, Germany) were used at a dilution of 1 : 10 000 in blocking solution. Incubation steps were performed for 2 h at room temperature, followed by several washing steps with TBS-T. The membranes were developed with 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium (BCIP/NBT) according to standard protocol.

Immunohistochemistry

Paraffin-embedded lung sections (thickness: 3–4 mm) of five stillborn infants and 12 neonates who died 1–42 days after birth was used for immunohistochemistry. The sections were examined at the Institute of Pathology (University of Heidelberg). The study was approved by the responsible Ethics Committee. Gestational ages at birth ranged from 20 to 41 weeks. None of these infants was included in the prospective analysis of DMBT1 in tracheal aspirates. For the detection of DMBT1 we used the monoclonal anti-DMBT1h12, as described earlier [21], and a polyclonal anti-serum (anti-DMBT1p84) generated against the human rDMBT1 in rabbit. Immunohistochemical studies with anti-DMBT1h12 were carried out manually, whereas those with anti-DMBT1p84 were carried out on an automated Ventana Discovery stainer (Ventana Medical Systems, Tucson, AZ, USA) according to a protocol described previously [21]. Peroxidase activity was detected with 3-amino-9-ethylcarbazole (AEC) as substrate (Sigma Chemicals, St. Louis, MO, USA). As a standard negative control, the antibodies were substituted by equal amounts of normal mouse IgG (Santa Cruz Biotechnology). All controls were negative. For image acquisition, digital analySIS 3·0 software, a 3CCD colour video 1705 camera (Sony Corp., Tokyo, Japan) and an Olympus BX-50 microscope (Olympus Optical, Hamburg, Germany) were used. We developed a score for determining the intensity and extent of DMBT1 signals. Sections displaying no signals were scored 0. DMBT1-positive tissues were scored semiquantitatively from 1 (weak staining) to 3 (highly intense staining). Each sample was scored independently by two investigators.

Statistics

The clinical data collected for each infant were correlated with pulmonary DMBT1 levels. Included in the analysis were gestational age, postnatal age, birth weight, SGA, gender, maternal infection, premature rupture of the membranes (>24 h), fetal lung maturation, asphyxia, respiratory distress syndrome (grade 1–4, according to Giedion et al. [22]), bronchopulmonary dysplasia (additional oxygen required at 36 weeks of corrected gestational age with abnormal radiological pulmonary findings), infection, survival, treatment with indomethacin and corticosteroids and operation.

Statistical analyses were performed with the sas software package, release 8·02. To provide a mathematical equation in order to estimate the DMBT1 levels in tracheal aspirates, we developed a general multiple model using the sas procedure proc reg. For this model, we included all potentially explanatory factors which were constant for each infant. We applied a backward selection method with a significance level of α = 0·05 to determine all significant variables in this multiple analysis.

Because the DMBT1 levels showed high interindividual variation between different infants, more analyses were conducted to examine associations between DMBT1 levels in tracheal aspirates and parameters that may vary within an individual infant. To analyse the effect of these variable parameters on individual DMBT1 levels, statistical models with only one explanatory factor were used, adjusted for each individual infant. Variable parameters were postnatal age, occurrence of infection, treatment with corticosteroids or indomethacin and operation. P-values smaller than 0·05 were regarded as statistically significant.

Pulmonary DMBT1 levels

Relative DMBT1 levels expressed as percentage of total protein concentration in tracheal aspirates studied at days 1–3 of postnatal life increased significantly (P = 0·042) from [mean ± standard error of the mean (SEM)] 0·68 ± 0·14 at 23–25 weeks to 2·94 ± 0·77 at 34–38 weeks of gestation (Fig. 1). Twenty-two infants were SGA and 60 were appropriate-for-gestational-age (AGA). SGA infants (gestational age 23–33 weeks) without infection studied at 1–10 days of life showed higher DMBT1 levels than AGA infants with comparable gestational and postnatal age (1·28 ± 0·25 versus 1·03 ± 0·12). Figure 2 shows that DMBT1 levels increased with increasing postmenstrual age (defined as gestational age plus chronological age).

The multiple models used for statistical analyses of factors associated significantly (P < 0·05) with the relative DMBT1 levels resulted in the following mathematical equation:

where y is the DMBT1 level expressed as percentage of the total protein concentration in tracheal aspirates. The variables gestational age in days (x₁; P < 0·0001), postnatal age in days (x₂; P < 0·0001) and SGA (x₃ = 1, if the infant was SGA and 0 if not; P = 0·0179) significantly influenced the DMBT1 level.

Relative DMBT1 protein levels increased upon infection

A total of 65 tracheal aspirates were collected from 36 preterm infants with infection during the observation time. These infections were accompanied by at least C-reactive protein (CRP) elevation and clinical signs of infections over a period of several days, and in most cases with positive pathological laboratory tests (e.g. neutropenia, leucocytosis, thrombopenia). However, in 42 of 65 tracheal aspirate samples (65%) infection was confirmed by evidence of positive bacterial cultures (blood, tracheal aspirate fluid, stomach aspirate fluid, catheter tip and smear of skin, ear and pharynx). The following bacteria were detected: Streptococcus species, Staphylococcus species, Staphylococcus haemolyticus, Staph. epidermidis, Klebsiella pneumoniae, Enterococcus faecalis, Escherichia coli, Enterobacter cloacae, Enterobacter aerogenes and Pseudomonas aeruginosa.

Infection increased the relative DMBT1 level, particularly in very immature infants (Fig. 2). After adjustment for each infant a significant association with postnatal age and infection was revealed, while no significant association with the remaining parameters, including indomethacin and corticosteroid treatment, was found (Table 1). Figure 3 shows time–courses for relative respiratory DMBT1 levels in four preterm infants with bacterial infections who had been studied before and during the infection. Three infants showed a marked DMBT1 response and one did not.

Western blotting of tracheal aspirate samples

To validate the results of ELISA measurements a subset of the tracheal aspirate samples was analysed additionally by Western blotting using the monoclonal antibody Hyb213-06. In this case the secretory component of IgA was used as reference protein to allow a comparison of the different samples. The results of Western blotting confirmed the specificity of the antibody used as well as reflecting the higher DMBT1 level with increased maturity and upon infection (Fig. 4).

Immunohistochemical analyses of DMBT1 levels in lung sections

Lung sections of nine infants without infection/chorioamnionitis and eight infants with infection/chorioamnionitis were used for immunohistochemical analyses. We used two antibodies for the assessment of DMBT1 levels in lung sections. Both antibodies showed an overlapping staining pattern with minor differences: the monoclonal antibody anti-DMBT1h12 stains preferentially DMBT1-variants produced by epithelial cells and the polyclonal antibody anti-DMBT1p84 stains preferentially variants produced by respiratory glands, and is therefore found in the mucus. Two of the neonates with infection had pneumonia, the other six infants had severe infections or septicaemia without clinical pneumonia. The levels of both variants increased with gestational age at death and were highly elevated in lung tissues of patients with infection/chorioamnionitis (Figs 5 and 6).