Hormonal regulation of mannan-binding lectin synthesis in hepatocytes

Abstract

Activation of the complement system via the plasma protein mannan-binding lectin (MBL) provides a first line of defence against infections. The plasma level of MBL is, in part, determined genetically, but may also be influenced by different hormones in vivo. Here we study the hormonal regulation of MBL synthesis from the human hepatocyte cell line HuH-7. Cells were exposed to medium with growth hormone (GH), hydrocortisone, insulin-like growth factor (IGF)-1, insulin, interleukin (IL)-6 or thyroid hormones (T3 or T4). After 3 days the concentration of MBL in the culture supernatants was determined and the amount of mRNA for MBL was measured, relative to mRNA for β2 microglobulin. GH, IL-6, T3 and T4 significantly increased MBL synthesis in a dose-dependent manner, while hydrocortisone, insulin and IGF-1 had no effect. T3 caused a fourfold increase at 1 nM of T3 (P < 0·001) and at 100 nM of T3 the production was increased more than eightfold. The effect of T4 was less potent, reaching an eightfold increase at 1 µM of T4 (P < 0·001). GH augmented the production of MBL threefold at a concentration of 100 ng/ml (P = 0·018) with no further effect up to 10 µg/ml, whereas IL-6 caused only a very weak increase in MBL production. MBL mRNA levels were stable during the first 24 h of T3 stimulation but increased significantly between 24 and 48 h. The results suggest that MBL synthesis in humans may be increased by thyroid hormone and GH, whereas it does not exhibit a classical IL-6-dependent response.

Introduction

The ability to distinguish between self, altered self or non-self is of paramount importance for the immune system. The immune system may be divided into innate and adaptive immune systems, where the complement system is an important part of the innate immune system [1,2]. Activation of the complement system via the plasma protein mannan-binding lectin (MBL, also referred to as mannose-binding lectin) occurs when MBL recognizes patterns of carbohydrates, such as that presented on many microorganisms. This is a beneficial activity, but many of the different products produced during such activation are potentially harmful for the body itself. Thus, the system is under tight control, on the level of a balanced activity of activators and inhibitors of the system and on the level of synthesis of the proteins involved in the system [1,2]. We expect that synthesis of components of the complement system may be under hormonal influence, and in the present investigation we examined the influence of a series of hormones on the synthesis of MBL.

MBL is a soluble protein produced mainly by hepatocytes. MBL is an oligomeric protein composed of subunits, where each subunit is assembled from three identical polypeptide chains [3]. In plasma MBL exists in sizes from a single subunit up to an oligomer of six or more subunits [4,5]. When MBL binds to a suitable carbohydrate surface, e.g. on microorganisms, complement activation may be initiated through the activation of MBL-associated serine proteases (MASP-1, MASP-2 and MASP-3) [6].

MBL deficiency is the most common congenital immunodeficiency, with a prevalence of 10% in the normal population [3,7]. Low levels of MBL result in deficiency in opsonic activity, i.e. less complement is deposited and killing is thus less efficient [8]. From an evolutionary viewpoint, the high prevalence of gene mutations among otherwise healthy individuals points to a balanced role of MBL in the sense that MBL deficiency in certain situations may confer some selective advantages. Experimental data have demonstrated that MBL is involved in complement-mediated injuries induced by altered self tissues, i.e. after ischaemia and reperfusion of kidney, heart or intestinal tissue [9–11]. Clinical findings also indicate that the proinflammatory effects of high concentrations of MBL may be a pathogenic factor in the development of complications in patients with type 1 diabetes [12,13].

The plasma concentration of MBL varies between less than 10 ng/ml up to 10 000 ng/ml [3,7]. This level is largely determined genetically by the presence of three single nucleotide polymorphisms, named alleles B, C and D, in exon 1 of the gene encoding MBL (MBL2). The level of MBL is also influenced by polymorphisms in the promoter region. However, despite identical MBL genotypes one can still find up to a 10-fold interindividual difference in the plasma concentration of MBL [3,7]. It has been noted previously that the promoter region is characterized by sequences common to promoters of acute phase proteins and elements common to genes of other proteins produced by hepatocytes [14,15]. However, clinical investigations have reported only a minor, sluggish increase with regard to acute phase responses, i.e. a twofold increase after 1–2 weeks [16]. In clinical trials it has been observed that growth hormone and thyroid hormones influence the concentration of MBL in plasma [17,18]. Otherwise, little is known about factors that influence MBL levels and that may explain the differences in MBL concentrations between individuals with identical genotypes.

We hypothesize that different hormones might influence the activity of the innate immune system and in the present study we investigated whether different hormones could influence the synthesis of MBL from hepatocyte cell lines.

Materials and methods

Cell lines

The cell lines HepG2 and Chang liver cells were purchased from LGC Promochem, Boras, Sweden [American Type Culture Collection (ATCC) cat. no. HB-8065 and CCL-13, respectively], while HuH-7, HuH-1 and HLF were obtained from Health Science Research Resources Bank, Sennan-shi, Japan (cat. no. JCRB0403, JCRB0199 and JCRB0405, respectively).

MBL2 genotyping

Genomic DNA was extracted from frozen cells using the QIAamp DNA blood mini kit (Qiagen, Hilden, Germany). As described previously, LightCycler-based real-time polymerase chain reaction (PCR) assay with fluorescent hybridization probes was used for genotyping the cell lines at codons 52 (D), 54 (B) and 57 (C) in exon 1 of the MBL2 gene and promoter variants at positions −550 (H/L), −221 (X/Y) and +4 (P/Q) [19]. Because the three polymorphic sites in the promoter are in linkage disequilibrium with the structural gene mutations, the genotype is expressed by the haplotypes LYPB, HYPD and LYQC, together with four wild-type (A variant) haplotypes LYPA, HYPA, LYQA and LXPA.

Hormones and interleukin (IL)-6

A series of different hormones were purchased: growth hormone (GH) (Humatrope, Lilly, Indianapolis, IN, USA); hydrocortisone (Solu-Cortef, Pfizer, Cambridge, MA, USA); insulin-like growth factor (IGF)-1 (Peprotech, London, UK); insulin (Actrapid, Novo Nordisk, Bagsværd, Denmark); 3,3′,5′-triiodo-l-thyronine sodium salt (T3) (T5516, Sigma-Aldrich, St Louis, MO, USA); l-thyroxine sodium salt pentahydrate (T4) (T 0397, Sigma-Aldrich); and IL-6 (R&D Systems, Oxon, UK; cat. no. 38220000). Two different lots of each hormone preparation were tested to reduce the theoretical risks of biased results originating from possible endotoxin contamination.

Growth and stimulation conditions

The cell lines were cultured in 5% CO₂ and 95% atmospheric air at 37°C in Dulbecco’s modified Eagle’s medium (DMEM) supplemented to complete culture medium with NaHCO₃, HEPES (5·938 g/l), 10% (v/v) heat-inactivated fetal calf serum (cat. no. 10270–106, Invitrogen, San Diego, CA, USA), 1% (v/v) glutaMAX (cat no. 35050–038, Invitrogen), penicillin (50 international units/ml) and streptomycin (0·5 mg/ml). When cells were divided for subculture they were washed with 140 mM NaCl, 2·7 mM KCl, 1·5 mM KH₂PO₄, 8·1 mM Na₂HPO₄, pH 7·4 [phosphate-buffered saline (PBS)] and detached with PBS containing 10 mM ethylenediamine tetraacetic acid (EDTA) and 0·05% (w/v) trypsin (cat no. 15090–046, Invitrogen).

Measurements of protein synthesis

Cells in logarithmic growth phase were seeded in six-well plates (9·4 cm² bottom surface) (TPP, Trasadinge, Switzerland) at 20 000 per cm², with 2 ml of medium per well. After 5 days 100% confluence was reached and the medium was changed to medium, with hormone added at concentrations covering a wide range from very low to supraphysiological levels. After an additional 3 days of growth the cells were detached with EDTA and trypsin, as above, and the number of cells was counted using a haemocytometer. The medium was used for measurements of protein production. The synthesis is given as production per 10⁶ cells to compensate for possible increased or decreased numbers of cells in the wells after the 3-day period.

Measurement of human serum albumin (HSA)

The concentration of HSA in the culture supernatants was measured using an inhibition assay in the format of a time-resolved immunofluorometric assay (TRIFMA). Microtitre wells were coated with 500 ng HSA (Statens Serum Institut, Copenhagen, Denmark) per well (in 100 µl) and residual binding sites blocked with 200 µg of bovine serum albumin (BSA) (cat. no. A-4503, Sigma-Aldrich) diluted in 100 µl 10 mM Tris, 145 mM NaCl, 15 mM NaN₃ and pH 7·4 [Tris-buffered saline (TBS)]. The wells were washed subsequently in TBS containing 0·05% Tween 20 (TBS/Tw); all the following washes and incubation buffers were TBS/Tw. The samples were diluted in buffer with 0·1% (w/v) BSA and rabbit anti-human albumin antibodies (cat. no. A001, DakoCytomation, Glostrup, Denmark) added to a final concentration of 0·156 µg/ml. After incubation in the wells and wash, the amount of anti-albumin antibodies bound to the HSA coat was detected with biotinylated donkey anti-rabbit Ig (cat. no. RPN 1004, Amersham Bioscience, Uppsala, Sweden) followed by europium-labelled streptavidin (PerkinElmer, Boston, MA, USA). The amount of europium bound was measured by time-resolved fluorometry on a 1232 Delfia fluorometer (Perkin Elmer). A standard curve was made from signals obtained when adding a 3·5-fold dilution series of HSA starting at 350 µg per ml and ending at 0·005 µg per ml.

Measurement of MBL

The MBL concentration was determined by a TRIFMA, as described previously [7]. In brief, microtitre wells coated with mannan were used to catch MBL in the sample. Subsequently, bound MBL was detected using europium-labelled monoclonal anti-MBL antibody. The detection limit was 30 pg/ml.

Measurement of cytokines

The concentrations of IL-2, -4, -6, -8 and -10, tumour necrosis factor (TNF)-α, interferon (IFN)-γ and granulocyte–macrophage colony-stimulating factor (GM-CSF) were quantified on a Luminex® 2000 by means of assays from Bio-Rad Laboratories, Hercules, CA, USA (cat. no. 171-A11080). The detection limit for the cytokines was 2 pg/ml.

Measurement of C-reactive protein (CRP)

CRP measurements was performed with the use of a kit from Bender MedSystems, Vienna, Austria (cat. no. BMS288INSTCE). The detection limit was 78 pg/ml.

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting

To analyse the oligomeric state of MBL in the culture supernatants, MBL from 500 µl of supernatant was purified on five microtitre wells coated with anti-MBL antibody (131-1, Immunolex, Copenhagen, Denmark). After incubation of supernatant in the wells and wash, the MBL was eluted with 15 mM Tris, 70 mM NaCl, 1·5% (w/v) SDS, 4 M urea, 0·5% (v/v) glycerol and 0·0005% (w/v) bromphenol blue. The samples were boiled for 3 min, and analysed non-reduced on a 4–20% polyacrylamide gel. The All-blue precision plus protein standard (Bio-Rad) was used as a marker. After electrophoretic transfer of proteins onto a 0·45 µm polyvinyl difluoride (PVDF) membrane (Hybond-P; Amersham Biosciences), the membrane was developed with biotinylated monoclonal anti-MBL antibody (131-1) followed by horseradish peroxidase-labelled streptavidin (P0397, DakoCytomation). Subsequently, luminescence reaction was produced using the Supersignal® West Dura kit (cat. no. 37071, Pierce Biotechnology Inc., Rockford, IL, USA) and the signal was recorded with a Kodak Image Station 1000. MBL was also purified from 2·8 µl serum (diluted with 500 µl of complete culture medium) from a person with a genotype matching the HuH-7 cells.

Quantitative real-time reverse transcriptase–polymerase chain reaction (RT–PCR)

The MBL2 mRNA expression levels were quantified using a two-step RT–PCR based on detection of a fluorescent signal produced proportionally during the amplification of a PCR product, employing TaqMan chemistry and the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). For the measurement of mRNA contents, total RNA was extracted from cells with the RNeasy mini kit (Qiagen) followed by DNAse treatment with deoxyribonuclease I, amplification grade (Invitrogen). Subsequently cDNA synthesis was performed using a first-strand cDNA synthesis kit from Roche Molecular Biochemicals (Mannheim, Germany) using oligo-d(T) 18 primers. Assay-on-demand gene expression products from Applied Biosystems were used for quantitative real-time PCR gene expression for MBL2 mRNA (cat. no. Hs01551501_m1) using beta 2 microglobulin (β2m) mRNA (cat. no. Hs99999907_m1) as a housekeeping gene for normalization of the system. The amount of cDNA is expressed as the threshold cycle (Ct), which is the number of cycles needed to gain 50% of maximal fluorescence.

To express the production of MBL2 mRNA as the number of copies, we initially calibrated against a plasmid containing inserted MBL2 cDNA. We found this approach suboptimal due to the highly purified nature of the plasmid compared to cDNA produced from the cells.

Mononuclear cells were isolated from blood from a HYPA/LYQA individual on Ficoll-Paque (cat. no. 17-1440-02, Amersham Biosciences) using the manufacturer’s procedure. Monocytes were isolated from the mononuclear cells by allowing the monocytes to adhere for 3 h to the walls of a cell culture flask. RNA was purified from the monocytes as well as from whole blood using the procedure used above for the cultured cells. The presence of MBL mRNA was tested, again with β2m mRNA as positive control, using procedures similar to those used for cultured cells.

Measurement of thyroid hormones and MBL in plasma samples

MBL concentrations were measured as described previously [7] in samples from 17 healthy controls with the MBL2 genotype HYPA/LYQA. Thyroid hormones (total T3 and total T4) and thyroid-stimulating hormone (TSH) were measured by immunofluorescent methods (Immulite, DPC, Los Angeles, CA, USA). Free thyroid hormones thyroxin (fT4) and triiodothyronine (fT3) were measured by ultrafiltration and RIAs as described previously [13].

Statistical analysis

Data are expressed as means ± s.e.m. P-values refer to unpaired t-tests, and P < 0·05 was considered significant.

Results

MBL genotypes and MBL production of hepatocyte cell lines

Five different hepatocytic cell lines were analysed initially for MBL synthesis and MBL genotype. The genotypes of the cell lines were determined to be HYPA/HYPA for HuH-7, LYPB/LYQA for HuH-1, HYPA/LYQA for HLF and LYPB/LYQC for Chang liver. The genotype of HepG2 cells has been determined previously to be A/B for the structural polymorphisms [20]. For unknown reasons, it was not possible to determine the genotype of the HepG2 cells using the method described in the present report.

When testing the growth media for the presence of MBL we observed measurable production of MBL from only the cell lines HuH-7 and HuH-1. From HuH-7 the production resulted in a concentration of 0·75 ± 0·06 ng/ml after 3 days’ growth and in HuH-1 the concentrations were only just above background. We consequently chose to use HuH-7 cells in the following experiments.

Hormonal influence on MBL production in HuH-7 cells

To analyse for the hormonal influence on MBL synthesis, HuH-7 cells were stimulated for 3 days with GH, hydrocortisone, IGF-1, insulin, T3 or T4 or with the cytokine IL-6, and the culture supernatants analysed subsequently for MBL and HSA. The synthesis is given as production per million cells to compensate for possible increased or decreased number of cells in the wells after the 3-day period. GH, IL-6, T3 and T4 significantly increased MBL synthesis in a dose-dependent manner, while hydrocortisone, insulin and IGF-1 had no effect on the production of MBL (Fig. 1). The thyroid hormones T3 and T4 had the strongest influence on MBL production. At concentrations from 1 pM to 100 pM there was no effect, but at 1 nM of T3 there was a fourfold (P < 0·001) increase in production and at 100 nM of T3 the production was increased more than eightfold (P < 0·001) (Fig. 1f). Equivalent to this was an eightfold increase with 1 µM of T4 (P < 0·001) (Fig. 1g). Increasing concentrations of GH from 100 pg/ml to 100 ng/ml augmented the production of MBL threefold (P = 0·018) and at GH concentrations from 100 ng/ml to 10 µg/ml the production of MBL was three times higher than that of non-stimulated production (Fig. 1a).

Fig. 1.

Influence of hormones on mannan-binding lectin (MBL) production from the hepatocyte-like cell line HuH-7. The production of MBL (ng per million cells) after 3 days of synthesis is shown on the ordinate, while the concentration of hormone added is depicted on the abscissa. The bars represent the mean plus the standard error of the mean of two separate experiments and each concentration of hormone was tested in duplicate. (a) growth hormone (GH); (b) hydrocortisone; (c) insulin-like growth factor (IGF)-1; (d) interleukin (IL)-6; (e) insulin; (f) thyroid hormone T3; (g) T4; and (h) combinations of T3 (100 nM), GH (100 ng/ml), hydrocortisone (10 µg/ml), insulin (10 mU/ml) and IL-6 (10 ng/ml). Note the changed scale on the ordinate in (h).

MBL has been classified as an acute phase protein based on the finding of a twofold increase in plasma levels 2 weeks after major surgery [16]. To test whether MBL synthesis is responding to IL-6, as would be expected of a classical acute phase reactant, this cytokine was added to the cultures. There was a weak but significant dose-dependent increase in MBL production, and at 1 µg/ml of IL-6 the production increased twofold (P = 0·01) (Fig. 1d). However, this concentration of IL-6 is vastly higher than that usually employed (approximately 10 ng/ml) for studying acute phase reactions in vitro. The physiological relevance of the observed response on IL-6 is thus doubtful.

To evaluate a general effect on protein synthesis, the production of HSA was also measured. The average HSA synthesis was 13·7 ± 1·0 µg per million cells, and none of the hormones had any significant effect on HSA production (data not shown). This indicates the absence of cytotoxic effects of the added hormones.

The concentrations of CRP in the culture supernatants from HuH-7 cells were below detection limit (78 pg per ml) at all times, with and without hormone stimulation (data not shown).

To study possible additive effects of hormones we subsequently mixed T3 with the other hormones (excluding IGF-1 and T4). The individual effects of the hormones were as we observed previously; a stimulatory effect of T3 and GH, while hydrocortisone, IL-6 and insulin did not alter production (data not shown). When GH was combined with T3 a significant increased synthesis was seen, from 4·18 ± 0·18 for T3 alone (GH alone was 1·6 ± 0·02) to 7·33 ± 0·49 ng per 10⁶ cells (P = 0·01) (Fig. 1h). Addition of hydrocortisone, insulin or IL-6 to T3 did not alter the production significantly.

Hormonal effects on the oligomerization of MBL

MBL is found in plasma as oligomers of a structural subunit that is itself made of three identical polypeptide chains. To characterize the appearance of the MBL produced by HuH-7 cells and to study if any of the hormones altered the oligomerization of MBL, Western blot analyses using non-reduced samples from the cell cultures were performed (Fig. 2). In each lane MBL from 500 µl of cell culture medium, purified on an anti-MBL-coated surface, was applied. MBL was also purified under identical conditions from serum from a subject with a MBL2 genotype (HYPA/HYPA) matching the HuH-7 cells. The MBL from the culture supernatants was seen in all cases as a band at larger than 250 kDa, corresponding to MBL-II [21] and the addition of hormones did not change this pattern. MBL purified from serum is seen as three bands representing MBL-I, MBL-II and MBL-III, as described previously by us [21]. MBL-I and MBL-II represent trimer and tetramer of the structural subunit, respectively [22].

Fig. 2.

The effect of hormones on the oligomerization of MBL polypeptide chains. Mannan-binding lectin (MBL) purified by antibody-based affinity purification from 500 µl culture supernatant, or from 2·8 µl or 0·28 µl serum was applied to the well of a sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) gel. The hormones used for stimulating synthesis are given at the top of the figure, as also are the amounts of MBL present in the different supernatants. After electrophoresis and blotting the membrane was developed with anti-MBL antibody. Molecular weight markers (kDa) are given on the left and the positions of the serum forms of MBL are indicated.

Kinetics of MBL and MBL mRNA production

To investigate whether T3 influences MBL mRNA levels and to study the kinetics of MBL production from HuH-7 cells the supernatant and cells were harvested 6, 9, 12, 24, 48 and 72 h after addition of medium alone or medium with 100 nm of T3. RNA was extracted from the harvested cells and used in a real-time quantitative RT–PCR analysis. The number of cells was counted and an equal amount of cells were used for the RNA purification. Following cDNA synthesis the amount of products were measured by TaqMan chemistry during the amplification. β2m was used as housekeeping gene to adjust for amount of added RNA. In order to examine the suitability of this approach RNA was purified from dilutions of HuH-7 cells (grown without hormones added). The dilution curves (Ct versus cell number) of the two mRNAs were found to be parallel, indicating that it is relevant to use the ratio between the two mRNAs (Fig. 3a). The production of β2m mRNA increased twofold after the first 24 h, and from 24 to 72 h there was a modest decline (Fig. 3b). This development in β2m mRNA amount was not influenced by the addition of T3 (Fig. 3b). It was concluded that β2m, in accordance with other published findings [23] was appropriate as the housekeeping gene. The ratio between MBL mRNA and β2m mRNA was stable during the first 24 h, whether or not the cells were under the influence of T3. Between 24 and 48 h there was an increase in the MBL/β2m mRNA ratio (Fig. 3c). This increase was larger under the influence of T3, and thus created a significant difference in the MBL2/β2m mRNA ratio (Fig. 3c). From 48 h and onwards to 72 h after the start of synthesis this ratio remained constant. With regard to the concentration of MBL in the culture medium, it was below the sensitivity of the MBL assay (30 pg/ml) during the first 12 h. From 24 h to 72 h there was a gradual increase in MBL up to 16 pg per hour per 10⁶ cells from cells grown in medium only, while the addition of 100 nM T3 increased the production to 115 pg per hour per 10⁶ cells (Fig. 3d). The concentration of HSA was also followed, and no difference in HSA production was seen between medium with or without T3 added.

Fig. 3.

The kinetic of mannan-binding lectin (MBL) and MBL mRNA production. (a) Standard curves of mRNA for β2m and MBL. RNA was purified from the number of cells given on the abscissa and the threshold cycle is given on the ordinate. The upper line is the curve for MBL mRNA and the lower line for β2m mRNA. (b) The amount of β2m mRNA given as percentage of the value at 0 h. At each time-point RNA was purified from an equal number of cells. The curves represent the mean plus the standard deviation from two experiments. (c) The amount of MBL mRNA during synthesis is given here in relation to β2m mRNA, to adjust for total amount of RNA. (d) MBL production. The concentration of MBL (given on the ordinate) in the supernatant of HuH-7 cells after growth for different time periods (given on the abscissa). The production of MBL was studied with (○) or without (•) 100 nM of thyroid hormone T3 added to the culture medium. The curves indicate the mean plus the standard deviation of two experiments.

It has been suggested previously that MBL might be synthesized from monocytes [24], but we did not find any MBL mRNA to be present in monocytes or in RNA isolated from full blood (the presence of β2m mRNA was used as positive control) from a person with wild-type genes.

Cytokine production

The influence of hormones on the synthesis of the cytokines IL-2, -4, -6, -8, -10, GM-CSF, IFN-γ and TNF-α from the HuH-7 cell line was investigated. A stable level of IL-8 was observed, with concentrations between 300 and 750 pg/ml in the supernatants (not shown). None of the hormones had any effect on the IL-8 levels.

There was a barely detectable production of IL-6 with a concentration of just above 2 pg/ml in the supernatants. This production was unaffected by addition of any of the hormones. This was a several-fold lower concentration than the concentration reached when IL-6 was added directly to the medium, e.g. when 100 000 pg/ml of IL-6 was added to the wells (4800 pg/ml was detected after 3 days of culture). This indicates that the increased MBL production from GH, T3 and T4 was not secondary to stimulation of IL-6 production from the cells.

In general, the production of IL-2, -4 and -10, GM-CSF, IFN-γ and TNF-α was below the detection limit. The exception was when supraphysiological concentrations (see above) of IL-6 was added to the culture medium, in which case there was a dose-dependent increase in production of IL-4 to 7·2 pg/ml, TNF-α to 7·6 pg/ml, GM-CSF to 6·5 pg/ml and IFN-γ to 35 pg/ml, respectively.

MBL concentrations in vivo and thyroid status

To study whether the in vitro data described above are reflected in a human situation, MBL levels and concentrations of TSH, total T3 and total T4 were examined in plasmas from 17 healthy controls with the MBL2 genotype HYPA/LYQA. No association was seen between the levels of MBL and TSH (r² = 0·02, P = 0·62), whereas MBL levels were correlated positively with both total T3 (r² = 0·23, P = 0·05) and total T4 (r² = 0·34, P = 0·01) (Fig. 4).

Fig. 4.

The association between mannan-binding lectin (MBL) levels and total thyroid hormone T4 (TT4) in 17 healthy individuals with identical MBL2 genotype. TT4 levels are given on the abscissa and the corresponding MBL levels on the ordinate. The line drawn on the figure was obtained by linear regression (r² = 0·34, P = 0·01).

Discussion

Despite a well-described genetic influence on MBL levels there are unexplained differences in MBL concentrations between individuals with the same genotype [7]. We tested the impact of a range of different hormones and IL-6 on MBL production from the HuH-7 cell line and found that thyroid hormones in particular, but also GH and to a lesser extent IL-6, influence production. These findings are in line with recent clinical data showing significant variations in MBL levels with variations in GH and thyroid hormone status [17,18].

We studied hepatocyte cell lines as MBL production in healthy humans has been confirmed mainly in liver cells, although the presence of MBL mRNA and of MBL was indicated in biopsies from intestinal tissue from patients with coeliac disease [25], and low levels of MBL2 mRNA was also reported in RNA from the small intestine and in testis [26]. The expression levels in the latter tissues were less than 1% of the MBL mRNA levels from liver cells. Confirming the hepatic origin of MBL, by studying liver transplantations it was found that plasma MBL levels were strictly dependent upon the genotype of the donor liver [27]. It has been suggested previously that MBL might be synthesized from monocytes [24], but we did not find any MBL mRNA to be present in monocytes or in RNA isolated from full blood.

We identified MBL production from the hepatoma cell lines HuH-7 and HuH-1. The production was higher in HuH-7 in accordance with the fact that the genotype of this cell line was HYPA/HYPA, known to be associated with high MBL levels, compared to the genotype LYPB/LYQA (which is associated with medium levels of MBL) for HuH-1. There was no measurable MBL production in any of the other screened cell types. This may be expected for Chang liver cells, as they have a genotype expected to give low levels of MBL, but not of HLF, where a genotype was found associated normally with higher MBL levels.

It has been reported previously that addition of 20 ng/ml IL-6 or 0·4 µg/ml dexamethasone to HuH-7 cells results in up-regulation of MBL mRNA, whereas the addition of IL-1 results in a down-regulation of MBL mRNA synthesis [28]. The changes in mRNA were measured after 12–24 h of stimulation via RT–PCR followed by Southern blot hybridization, a technique that lacks the precision of today’s real-time quantitative PCR. In the present report we do see a relative increase in MBL mRNA levels, but not until after 24 h of stimulation, and we do not find MBL protein production to be influenced by 100 ng/ml IL-6 or from 10 µg/ml of hydrocortisone (equivalent to 0·38 µg/ml dexamethasone).

We have reported previously that GH, but not IGF-1, administration to humans increased the concentration of MBL [17]. Patients with an increased production of GH (acromegaly) were found to have a higher MBL concentration than normal individuals, and this concentration was lowered when they were treated with a GH-receptor antagonist. The stimulatory effect of GH is also seen when treating patients with Turner syndrome [29]. In the present report we found a significant increase in MBL production when stimulating the HuH-7 cells with GH (100 ng/ml), while IGF-1 had no effect. GH may thus act directly on the hepatocytes in vivo. In vivo the concentration of GH fluctuates between 0·05 and 50 ng/ml.

In the present study the glucocorticoid, hydrocortisone and insulin showed no effect on MBL production. In one previous study another glucocorticoid, dexamethasone, was reported to down-regulate the activity of the MBL promoter [15]. In another study where MBL mRNA was measured, dexamethasone resulted in an up-regulation of MBL mRNA [28]. With regard to insulin, no previous in vitro studies exist, but we [30] and others [31] have reported that patients with insulin-dependent (type 1) diabetes have approximately 1·3-fold higher MBL levels than healthy people. We have speculated that this could be due to lower hepatic insulin levels in diabetics, i.e. that insulin depresses the production of MBL. In another study, the MBL levels of patients admitted to intensive care, where half of them were selected for intensive insulin treatment, was followed [32]. During the stay at the intensive care unit, the MBL levels increased in general, but the insulin-treated group had a slightly (but significant) smaller increase in MBL levels. We found no effect of adding insulin to the HuH-7 cells either alone or in combination with T3, suggesting that MBL production is not influenced directly by insulin.

When IL-6 was added to the cell culture, we observed only a very modest increase in synthesis of MBL and only when using IL-6 at 1 µg/ml. This is considerably more than the 50 ng/ml systemic IL-6 concentrations associated with an acute phase response. In the literature MBL is often referred to as an acute phase protein [33], and we have published previously that MBL levels in patients undergoing hip operations increased approximately twofold, but not until 2 weeks after the operation [16], and in a study of 451 critically ill patients MBL levels were increased two- to threefold 7 days after the patients were admitted to intensive care [32]. In comparison, a classical acute phase protein, CRP, peaked with 10–100-fold increased concentrations several days earlier. Another study found that in patients with sepsis and septic shock, the majority of patients do not display an MBL acute phase response [34]. These results, together with the findings in the present report, underscore that viewing MBL as a classical acute phase protein greatly stimulated by IL-6 is misleading rather than revealing. However, it may be that the modest increase in MBL during an acute phase response is due to hormonal changes. During acute illness there is an increased production of hormones from the pituitary gland, e.g. the GH level is increased [35].

The strongest hormonal influence we observed was induced by the thyroid hormones T3 and T4, indicating that they have an important role in regulating the MBL production. It has been reported recently that hyperthyroid patients had significantly higher MBL levels before than after treatment for the disease [18]. Hypothyroid patients, on the other hand, had low MBL levels during disease, which increased after T4 substitution. The in vitro data in the present study suggest a direct effect on the liver cell by thyroid hormones as the explanation to the in vivo findings. We found that a comparable effect of T4 was seen at a concentration 100 times higher than that needed of T3, which is in concordance with T3 being the most potent of the two in vivo [36]. The concentration of T3 (1 nM) and T4 (100 nM) significantly increasing the MBL levels, are similar to the levels of total T3 and T4 found in plasma, i.e. approximately 2 nM and 100 nm, respectively. The stimulatory effect of thyroid hormones on MBL level is not unique to MBL; for example, cDNA microarray analysis and in vitro protein measurements have revealed that a broad range of other plasma proteins are also influenced in a stimulatory or inhibitory manner by thyroid hormones [36,37].

While thyroid hormones may explain part of the intragenotype variation, they may also provide an explanation to other changes of MBL levels; for example, during pregnancy thyroid hormone levels are increased [38], and it has been described that the MBL levels increase twofold during pregnancy [39].

We analysed whether there was a synergistic effect of other hormones on the T3-mediated effect, and we found that the stimulatory effect on MBL production by T3 was augmented by addition of GH to the medium. This indicates that the stimulatory mechanisms of T3 and GH were either unsaturated or independent of each other.

The promoter region of MBL2 was studied by Madsen and colleagues [40] and recently in more detail by Naito and coworkers [15]. It contains a TATA box, a CAAT box, type 1 and type 2 IL-6 response elements, a glucocorticoid response element and hepatocyte nuclear factor 3 (HNF-3) recognition sites. To analyse the activity of the MBL2 promoter, Naito et al. made a reporter gene construct which they transfected into the human hepatoma cell line HepG2. Observing the effect of deletion of parts of the promoter they found that in particular deletion of some of the HNF-3 sites had a negative effect on the promotor activity. In cows GH has a strong stimulatory effect on HNF-3 production [41], and it may be that the same is true in humans. MBL production may thus be promoted via an up-regulation of HNF-3, but the mechanism by which GH and thyroid hormones affect the MBL promoter remains to be elucidated.

The stimulatory effect of T3 on MBL levels was reflected in an increase in MBL mRNA, indicating that the effect was due to an effect on transcription. The MBL mRNA/β2m mRNA ratios rose with and without T3 in the time-span from 24 to 48 h after start of the cell culture, while an increase in β2m mRNA was seen from 12 to 24 h. This increase in β2m mRNA synthesis is induced probably by stress caused by the change in medium. The time lapse before increase in MBL mRNA synthesis is seen indicates that the effect of the hormones is not a direct effect on assembly of transcription factors. Rather, the hormones are probably stimulating up- or down-regulation of other factors.

With regard to cytokine production by HuH-7 cells only IL-8 could be detected at levels above the detection limit of the assays, and the level was not influenced by the presence of hormones or IL-6. The exception was when IL-6 was added to the culture supernatant in very high concentrations, which resulted in a marginal but dose-dependent and statistically significant increase in IL-2, -4, TNF-α, GM-CSF and IFN-γ. IL-6 production was barely increased to a detectable level when stimulating with GH, T3 and T4, indicating that the increased MBL production was not secondary to stimulation of hepatic IL-6 production.

When looking at one isolated genotype, e.g. HYPA/LYQA, an interindividual fivefold variation in MBL levels was seen previously [7]. Environmental or hormonal influences on this variation have been suggested logically, but have not yet been explored. We investigated whether thyroid hormone levels are associated with MBL levels in healthy people with an identical genotype. We found MBL levels ranging from 1050 ng/ml to 5168 ng/ml and a significant positive correlation between MBL levels and both total T3 and T4. Larger study groups may be relevant, but the data indicate that variations in total T4 may explain 34% of the observed variation in MBL levels in healthy subjects with identical genotypes.

Our in vitro findings confirm clinical observations of hormonal influence on MBL synthesis. Compared to the strong impact of MBL2 genotypes these hormonal inputs may, in healthy individuals, merely represent a fine-tuning of circulating levels, although the fluctuations in MBL levels seen in patients with hyper- and hypothyroidism are substantial. Rather than being attributable to a classical IL-6-dependent acute phase response, the slow and modest increase in MBL levels seen in critically ill patients and after major surgery may well be secondary to changes in hormone levels.