SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Circulating mannan-binding lectin (MBL) levels are elevated in type 1 diabetes. Further, high MBL levels are associated with the development of diabetic nephropathy. In animals, a direct effect of MBL on diabetic kidney changes is observed. We hypothesized that MBL levels and detrimental complement activation increase as a consequence of diabetes. We measured plasma MBL before and 7 weeks after inducing diabetes by streptozotocin. Mice have two MBLs, MBL-A and MBL-C. Diabetes induction led to an increase in MBL-C concentration, whereas no change during the study was found in the control group. The increase in MBL-C was associated with the increasing plasma glucose levels. In accordance with the observed changes in circulating MBL levels, liver expression of Mbl2mRNA (encoding MBL-C) was increased in diabetes. Mbl1expression (encoding MBL-A) did not differ between diabetic and control animals. The estimated half-life of recombinant human MBL was significantly prolonged in mice with diabetes compared with control mice. Complement activation in plasma and glomeruli did not differ between groups. We demonstrate for the first time that MBL levels increase after induction of diabetes and in parallel with increasing plasma glucose. Our findings support the previous clinical observations of increased MBL in type 1 diabetes. This change may be explained by alternations in both MBL production and turnover.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Diabetes is the leading cause of end-stage renal disease in Western countries [1-3]. Increasing evidence indicates a significant role of the complement system in diabetic nephropathy [4]. In particular, the lectin pathway initiator mannan-binding lectin (MBL), also known as mannose-binding lectin, has been shown to be of major importance [4, 5]. A clear association between high levels of MBL and increased risk of microalbuminuria and diabetic nephropathy is found in patients with diabetes [6-8]. As the within-subject variations in MBL levels are very small compared with the large variations between-subjects, MBL may serve as a significant risk indicator of diabetic nephropathy [9]. We have previously found a direct cause–effect relationship between MBL deficiency and attenuation of diabetic kidney changes in a mouse model of type 1 diabetes [10, 11]. The exact mechanism underlying the effect of MBL in the pathogenesis of diabetic nephropathy remains unknown, but studies indicate that increased complement activation may happen in diabetes [12, 13]. Furthermore, circulating levels of MBL are elevated in type 1 diabetes compared with healthy individuals [14-16]. In vitro studies found no effect of insulin on liver cell MBL production [17]. The MBL2 gene encoding for MBL expression in humans is highly polymorphic with three common point mutations at three alleles in exon 1 and promoter-region polymorphisms at two alleles leading to variations in MBL level [18]. Mutations in exon 1 decrease ligand avidity of MBL possibly because of alternations in MBL oligomerization [19]. Together, the large between-subject variation of functioning MBL (1,000-fold) compared with little within-subject variation (2–3-fold) makes MBL a potential risk marker. However, in diabetes, the significance of MBL genotype is conflicting in different studies [7, 8, 20]. To the best of our knowledge, no study has measured MBL level before and after the debut of diabetes.

Thus, we hypothesized that the MBL level increases as a consequence of diabetes within a specific genotype. Also, changes in MBL production or half-life may account for alternations in circulating MBL level. We tested this with the use of a mouse model of diabetes. Furthermore, we estimated the decree of complement activation in both circulation and kidney glomeruli that potentially could link the complement system to diabetic nephropathy [13, 21].

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Animals

Eight weeks old female C57BL/6JBomTac mice were purchased from Taconic (Ry, Denmark) and divided into two groups; control (n = 24) and diabetes (n = 16). Animals were housed in 6–8 per cage with free access to water and standard chow (Altromin 1324, Lage, Germany). Cages were placed in a room with an artificial light cycle (dark 7.00 p.m. to 7.00 a.m.), constant temperature of 21 ± 1 °C, and a humidity of 55 ± 5%. A substudy was conducted to describe the time of MBL-C protein increase. To this substudy, we used C57BL/6JBomTac at same age and sex and from the same supplier as describe previously. We divided 14 animals into control (n = 6) and diabetes (n = 8). Both the method of diabetes induction and the duration of diabetes were identical in the main study and substudy. Body weight and blood glucose were measured every week. Animals with signs of illness, persistent ketonuria or more than 15% weight loss were excluded. Furthermore, to validate the results in a spontaneous model of type 1 diabetes, MBL levels were measured in eight 26 weeks old male Akita insulin knock-out mice (C57BL/6-Ins2Akita/J, The Jackson Laboratory, Bar Harbor, ME, USA) and compared with eight age-matched wild-type control. The studies complied with the Danish regulations for care and use of laboratory animals.

Induction of diabetes

Animals of the diabetes group were injected i.p. on five consecutive days by 55 mg/kg body weight streptozotocin (STZ, Sigma-Aldrich, St. Louis, MO, USA) freshly dissolved in cold sodium citrate acid buffer (10 mm, pH 4.5). Animals were fasted for 4 h prior to the injections. Re-injections with STZ 55 mg/kg were performed until diabetes was established (blood glucose ≥ 15 mm). All animals in the diabetic groups had a total of seven STZ injections. Four animals from the diabetic group of the main study and two animals from the diabetic group of the substudy were excluded from due the inadequate response to STZ. Controls were injected with buffer.

Blood glucose and urine analysis

Blood glucose was measured from tail-capillary blood by Contour™ (Bayer Diabetes Care, Kgs. Lyngby, Denmark). Urine was tested for glucose and ketone bodies by Combur5 Test D (Roche Diagnostics GmbH, Mannheim, Germany).

Collection of samples

Animals were sampled both at the start and at the end of the study. As some assays required large volumes of plasma, not all analyses could be made on every animal. Before initiation of the study, we therefore made the following subdivision within each cage of both control and diabetic animals. All animals with odd identification numbers were selected for the measurements of MBL half-life estimates and C5a/C5a desArg measurements, whereas animals with even identification numbers were selected for the measurements of MBL-A, MBL-C and immunohistochemistry. Animals were numbered successively in the cages. Blood samples from odd numbered animals were added Nafamostatmesylate (protease inhibitor also known as Futhan or FUT-175, Framingdale, NY, USA) to inhibit ex vivo activation of complement system.

All blood samples were drawn by heparinized capillary tubes from the retro-orbital plexus to potassium EDTA tubes (Sarstedt, Nümbrecht, Germany). The samples were kept at 0–4 °C during preparation of plasma. Plasma was snap-frozen in liquid nitrogen and stored at −80 °C until analysed. At termination of the study, animals were anaesthetized by 0.5 mg/g body weight ketamine i.p. and 0.2 mg/g body weight xylazine i.p. (Ketaminol® Vet and Narcoxyl® Vet, respectively, Intervet, Skovlunde, Denmark) and plasma was collected and treated as described previously. Kidneys, liver and heart were dissected and weighed. Left kidney was divided into two pieces horizontally through the hilus and both pieces fixed in 10% formalin for 24 h and embedded in paraffin. The poles of right kidney and three pieces of liver were snap-frozen in liquid nitrogen and stored at −80 °C for later analyses of gene expression.

Quantitative RT-PCR

The quantifications by RT-PCR were performed as previously described [22]. In brief, total cellular RNA was extracted from liver tissue by a 6100 Nucleic Acid PrepStation (Applied Biosystems, Foster City, CA, USA). The quality of ribosomal RNA was estimated by agarose gel. Reverse transcription from RNA to DNA was performed with a Multiscribe Reverse Transcriptase kit (Applied Biosystems). The polymerase chain reaction was performed in triplicates of each sample in wells containing RNA, TaqMan Universal PCR MasterMix and amplifying primers (assay id Mm00495413_m1 and Mm00487623_m1) purchased from Applied Biosystems. Ribosomal 18S were used as housekeeping gene. Liver RNA was used as negative control. Data were analysed with the ABI Prism 7000 Sequence Detector Software from Applied Biosystems. The relative quantification of target gene was calculated as described in the Users Bulletin 2, 1997 from Perkin-Elmer (Perkin-Elmer Cetus, Norwalk, CT, USA) [23].

MBL-A and MBL-C assays

Plasma concentrations of MBL-A and MBL-C were measured in ligand-based time-resolved immunofluorescence assay (mannan-coated plates) as previously described [24]. In brief, plates were coated with mannan in carbonate-coating buffer. Ligand-bound MBL-A or MBL-C was detected with biotinylated rat anti-mouse MBL-A antibodies (MAb 13H6) or biotinylated rat anti-mouse MBL-C antibodies (MAb 14D12).

Half-life

To estimate half-life of MBL, we intravenously injected a bolus of 25 μg recombinant human MBL (rhMBL) [25] dissolved in 200 μl Tris-buffered saline. After 15 min, 2 h, 8 h and 24 h blood samples were drawn by heparinized capillary tubes from the retro-orbital venous plexus to tubes coated with EDTA. Samples were spun at 1,500 g for 15 min for plasma separation. Plasma was stored at −20 °C until analysed. The concentration of rhMBL was determined as previously described in an assay not giving a signal with mouse MBLs [19, 26]. Briefly, a sandwich-type time-resolved immunofluorescence assay with the same murine monoclonal anti-human MBL antibody (clone 131-1) used as capture antibody and as detection antibody (europium-conjugated). The slope of concentration-to-time curve was estimated in a random coefficient model including the information of repeated measurements per animal.

When plotting the loge to the plasma concentration of rhMBL against time, the curve approached a linear relationship with time. The concentration to a given time can thus be expressed as loge(Ctime) = −k × time + loge(C0), where Ctime denotes the concentration at a given time, k the elimination constant and slope and C0 the intercept with the y-axis. From this equation, t½ can be described as t½ = loge(2)/k. The elimination constant, k, is estimated by mixed-effect linear regression of loge-transformed concentration.

C5a/C5a desArg assay

Plasma was prepared as described previously in EDTA and protease inhibitor to avoid ex vivo activation of complement. The C5a concentration was determined as described previously modified for time-resolved immunoflourences assay [27]. In brief, plates (FlouroNunc Maxisorp, Nunc, Roskilde, Denmark) were coated overnight at 4 °C with 2 μg/ml rat anti-mouse C5a (558027 BD Biosciences, Franklin Lakes, NJ, USA) in carbonate-coating buffer, pH 9.6. According to the manufacturer, this capture antibody recognizes C5a and C5a desArg, but not C5. Plates were blocked with 1 mg/ml BSA in PBS. Purified recombinant mouse C5a desArg (HC1102, Hycult Biotech, Uden, the Netherlands) and samples were diluted in PBS/Tw with 1 mg/ml BSA. Biotin-labelled rat anti-mouse C5a (558028, BD Pharming) and Eu3+-streptavidin in addition with enhancement buffer (Perkin Elmer, Wallac, Waltham, MA, USA) were used for detection on time-resolved fluorometry on a 1232 Delfia fluorometer (Perkin Elmer).

Deposition of membrane-attack complex (MAC)

Two micrometre thick slices of paraffin-embedded kidney were stained with rabbit polyclonal anti-C5b-9 (MAC) antibody (ab55881, Abcam, Cambrigde, United Kingdom) after heat-induced epitope retrieval. Envision FLEX+ Rabbit Linker followed by Envision Flex+/Horse-radish peroxidase with Envision Flex+ DAB Substrate Working Solution (Dako, Glostrup, Denmark) was used for visualization. Observer-blinded semi-quantitative measurement of MAC-stained fraction of glomeruli was made by point counting by light microscopy. At 400 times magnification, the relative distribution MAC-stained tissue to glomerulus tissue was estimated.

Statistical analysis

Student's t-test was used to compare start and end values (changes during study) and differences in gene expression after control for equal variance and normal distribution of data. Estimations of half-life were made based on the kinetic assumptions described previously. To use the statistical power of information from paired data for individual animals, multilevel mixed-effects linear regression analyses were applied for the estimation of MBL half-life and in the substudy for MBL changes over time in relation to glucose changes. In the results sections, differences between groups are given as mean or median (95% confidence interval). stata 11 (StataCorp LP, College Station, Tx, USA) was used for the statistical analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Animal characteristics

At start of study, the body weight did not differ between the control group and the diabetic group. However, at termination of study, diabetic animals had a significantly lower body weight than controls. This difference was already observed at the time of MBL half-life determination (please see Table 1).

Table 1. Mean body weight
 Control (n = 24)Diabetes (n = 12)Student's t-test
  1. Mean body weight (g) and 95% confidence intervals in control and diabetic group at start and end of study. In addition, body weight at the time of mannan-binding lectin (MBL) half-life estimation is presented. T-statistics for pair wise comparison between groups are shown for each time-point.

Start20.5 (20.1; 21.0)20.9 (20.4; 21.4)NS
End22.0 (21.6; 22.4)19.3 (18.5; 20.1)P < 0.001
At estimation of MBL half-life22.4 (21.9; 22.8)19.7 (19.0; 20.3)P < 0.001

Diabetes-induced increase in circulating levels of MBL-A and MBL-C

Diabetes-induced changes in MBL-A and MBL-C plasma concentration were calculated from samples before and 7 weeks after induction of diabetes.

For MBL-A, no change during the study (end concentration minus start concentration) was found either within the control group (n = 11) or within the diabetic group (n = 7) (Fig. 1A). However, when comparing changes in MBL-C concentration during the study, MBL-C concentration increased in median by 3.6-fold (2.9; 4.5-fold) during the study in the diabetic group (n = 10, P < 0.0001), whereas no changes during the study was found in the control group (n = 7) (Fig. 1B). When comparing change in MBL-A concentration during the study (delta value in diabetic group versus delta value in control group), the change in MBL-A during the study differed between the diabetes group and the control group (P = 0.02, Fig. 1C). Likewise, MBL-C change during the study differed significantly when comparing the diabetic group with the control group (P < 0.0001, Fig. 1D).

image

Figure 1. Plasma levels of mannan-binding lectin (MBL)-A and MBL-Cat start and end of the study. (A) MBL-A (control: n = 11, diabetes: n = 7). (B) MBL-C (control: n = 10, diabetes: n = 7). Test of change from start to end of study was performed by Student's t-test. (C) MBL-A change from start to end of study (control: n = 11, diabetes: n = 7). Test of different change over time was performed by Student's t-test. (D) MBL-C change from start to end of study (control: n = 10, diabetes: n = 7). Test of different change over time was performed by Student's t-test.

Download figure to PowerPoint

Relationship between increase in MBL and plasma glucose

To describe the kinetics of the MBL-C increase further, we conducted an additional substudy. In control mice (n = 6) and diabetic mice (n = 6), we drew small plasma samples every week through a 7-week diabetes period. Methods and materials are identical to what is described earlier for the main study.

As in the main study, change over time in plasma MBL-A (end concentration minus start concentration) differed between diabetes and control mice (P = 0.01, Fig. 2A). When analysing diabetes and controls separately, a weak increase in MBL-A concentration was observed in diabetes (P < 0.0001), whereas no change in MBL-A concentration was found in control animals, although P-value was borderline significant (P = 0.05).

image

Figure 2. Substudy including weekly measurements of plasma mannan-binding lectin (MBL) and plasma glucose (control: n = 6, diabetes: n = 6). (A) Lines indicate MBL-A concentration in individual animal. Test of parallel curves was performed by multilevel mixed-effects linear regression (control versus diabetes). (B) Lines indicate MBL-C concentration in individual animal. Test of parallel curves was performed by multilevel mixed-effects linear regression (control versus diabetes). (C) Lines indicate mean of MBL-A values from Fig. 2A and therefore no indication of statistical error is given. (D) Lines indicate mean of MBL-C values from Fig. 2B and therefore no indication of statistical error is given. Superimposed are mean curves of plasma glucose. Test of interaction between glucose and MBL-C was performed by multilevel mixed-effects linear regression.

Download figure to PowerPoint

The changes in MBL-C concentration during the study differed clearly between diabetic and control animals (P < 0.0001, Fig. 2B). MBL-C concentration increased by a median of 2.8-fold (2.4; 3.4-fold) in diabetic animals (P < 0.0001) compared with a median of 1.2-fold (1.1; 1.3-fold) increase during the study (end concentration minus start concentration) in the control group (P = 0.007).

As we hypothesized that induction of diabetes would cause a change in MBL level, we proceeded by testing if plasma glucose had impact on the observed MBL levels. For this test, it was assumed that the intervention (induction of diabetes) could lead to a change in MBL level and that MBL concentration would not impact blood glucose. No association between change in blood glucose and change in MBL-A concentration was found (Fig. 2C). However, we did find that increasing blood glucose was significantly associated with an increase in plasma MBL-C concentration when analysing diabetic and control animals together (P = 0.03, Fig. 2D). This association between blood glucose and MBL-C level was significantly stronger in diabetic animals compared with control animals (interaction between group term and glucose, P = 0.004), although statistical significant in both groups (P = 0.04 in control group and P = 0.002 in diabetes group).

Genetic model of type 1 diabetes

To validate the observations of an increased MBL-C concentration in diabetes, we measured MBL-A and MBL-C levels in the Akita insulin gene knock-out model of the type 1 diabetes. Like in STZ-diabetes, Akita diabetic mice (n = 8) had an elevated mean MBL-C level of 67.5 μg/ml (56.2; 80.9 μg/ml) compared with wild-type controls with a mean of 30.9 μg/ml (27.1;35.2 μg/ml) (n = 8), P < 0.0001 (data not shown). MBL-A level did not differ between the control group and the diabetes group.

Hepatic transcription of MBL

Mannan-binding lectin is predominately produced by hepatocytes. Thus, to identify the mechanism underlining the increase in circulating MBL-C in diabetes, we quantified transcription of the two murine MBL genes in liver tissue. Transcription of Mbl2 mRNA (encoding MBL-C) was increased by a median of 1.3-fold (1.1; 1.5-fold) in diabetic animals (n = 11) compared with controls (n = 15), P = 0.006 (Fig. 3). No difference was found in level of Mbl1 mRNA transcription (encoding MBL-A) between the control group and the diabetes group.

image

Figure 3. Liver expression of mannan-binding lectin (MBL) mRNA with + indicating the median. (A) Expression of Mbl1mRNA (encoding MBL-A), control: n = 15, diabetes: n = 12. (B) Expression of Mbl2mRNA (encoding MBL-C), control: n = 15, diabetes: n = 11. Test of difference between groups was performed by Student's t-test.

Download figure to PowerPoint

Estimation of MBL half-life

By analysing the decline in rhMBL concentration after intravenous injection, we estimated the half-life of MBL in the two groups (Fig. 4). Based on these data, the median half-life of MBL was significantly prolonged in diabetes animals (n = 5), that is, 14.3 h (11.3; 19.4 h) compared with 11.8 h (11.0; 12.9 h) in control animals (n = 12), P = 0.03. Body weight was lower in diabetic animals when the half-life study was performed (Table 1). As the injected mass of rhMBL relative to body weight was therefore larger in diabetic animals, we performed additional analyses of MBL half-life controlling for body weight (data not shown). After this adjustment, half-life was still prolonged in diabetes compared with control animals (P = 0.03).

image

Figure 4. Elimination of recombinant human mannan-binding lectin (MBL) injected intravenously to the time zero (control: n = 12, diabetes: n = 5). Markers indicate mean ± SD of log-transformed values. Test of differing slopes was performed by multilevel mixed-effects linear regression.

Download figure to PowerPoint

Complement deposition in glomeruli

By semi-quantitative measurement of deposition of membrane-attack complex (MAC) in glomeruli, we estimated complement activation. We did not find any evidence of increased deposition of MAC in the diabetes group compared with the control group (data not shown).

Plasma level of complement activation products

Upon complement activation, C5 is cleaved and the C5a fragment is released to the circulation. We could not detect any difference in the C5a concentration in plasma from diabetic mice compared with the control mice, although certain reservations needs to be taken due to technical issues with large variation within animal and a large number of animals with C5a concentrations below level of detection (data not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Our study shows for the first time a marked increase in MBL-C concentration after the induction of experimental type 1 diabetes. This finding was repeated in a substudy and furthermore validated in a genetic model of type 1 diabetes. By weekly measurements of MBL concentration, we describe the temporal association between increasing plasma glucose and increasing MBL-C concentration. Exposing the data to statistical analysis, plasma glucose could explain the change in MBL-C level when assuming an uni-directional relationship between the intervention of the study (induction of diabetes) and observed change in MBL concentration. To explore the origin of the increasing MBL-C level, we evaluated liver expression of MBL mRNA. In good agreement with the observations on MBL-C in circulation, liver MBL-C mRNA expression from theMbl2 gene was upregulated in diabetic animal at end of the study period. MBL-C is transcribed from the Mbl2 gene in analogue with human MBL from MBL2. In addition, as we asked whether STZ possible had diabetes-independent toxic effects, we measured MBL in the Akita insulin knock-out mouse which represents a genetic model of type 1 diabetes. In this model, we confirmed the finding of elevated circulating MBL-C levels in diabetes. In preparation of the study, we intended to treat with insulin to normalize blood glucose in an extra group of diabetic animals. However, in pilot studies, this proved very difficult to obtain and was therefore not included in the design.

For the second murine isoform of MBL, MBL-A, a small difference was found when comparing the change over time in the diabetes group with the control group. However, this relationship was much less pronounced as for MBL-C. Furthermore, no association was observed between change in MBL-A concentration and change plasma glucose concentration. In addition, MBL-A mRNA transcribed from the Mbl1 gene did not differ between groups. Together, these observations indicate that the elevated MBL level in patients with type 1 diabetes is a secondary effect to diabetes. This is in good agreement with a large human study that does not find association between MBL genotype and prevalence of type 1 diabetes [20]. In additions, a lack of insulin seems not to explain changes in MBL concentration based on previous in vitro studies [17]. Furthermore, some human studies find HbA1C and MBL concentration to be positively correlated although only weakly [6, 8]. The increase in circulating MBL-C concentration found in our study may well also be explained by alternations in MBL elimination. As the route for MBL elimination is not known, we made an approximation of MBL half-life. As the kidneys may participate in elimination of MBL, the study was terminated after very short-term diabetes duration, where renal impairment is not yet developed [28]. By measurements of declining concentration of recombinant human MBL, we estimated the elimination constant in the two groups and calculated the half-life by assuming first order kinetics. From these rough estimates, we found that prolonged elimination of MBL might in part add to the increased circulating MBL level. We speculate that altered protein glycosylation in diabetes may allow transient or permanent MBL-binding and thereby delay MBL elimination [29].

Most of the publications of MBL in diabetes have focused on the association with diabetic nephropathy. There are an increasing amount of studies that indicate a strong association between increased MBL level and increased prevalence of micro- or macroalbuminuria [6, 8, 15, 20, 30]. Hovind et al. [7] described a strong association between MBL levels and the future risk of microalbuminuria. It is obvious to speculate that an increased MBL level may lead to an increased activation of the complement system through the lectin pathway. It has been hypothesized that the increased protein glycosylation in poorly controlled diabetes can provide ligands in a sufficient density and composition to enable MBL-binding to human cells [29, 31, 32]. Also, glycosylation may lead to dysfunctional complement regulatory proteins that normally protect human cells from complement self-attack [13, 21]. It may therefore be hypothesized that diabetes would lead to increased complement activation. However, both at systemic level and at glomerulus level, we did not find any evidence of increased complement activation, although reservations needs to be taking especially regarding plasma measurements of C5a. The low number of valid measurements of C5a concentration was the result of large intra-assay CV in some animals and C5a concentrations below the limit of quantification. Also, as ex vivo complement activation is very dependent of temperature of samples, we could not repeat these measurements on the existing samples. Furthermore, a large sample volume is required, which limits the possibilities for repetition. Regarding immunohistochemistry, it should also be noted that the duration of diabetes in the animals were relatively short.

In general, it is a limitation to the study that samples from all animals could not be included in every assay, which would have increased statistical power. As stated in the methods section, the selection of animals for the different assays was made at random and before the induction of diabetes. To further control for selection bias, this randomization was made within each animal cage. Sample size was estimated in advance based on a pilot study to provide adequate statistical power in the MBL analysis under consideration of the randomization as described previously.

In conclusion, we demonstrate for the first time that MBL-C is upregulated in response to induction of diabetes both at the level of liver mRNA transcription and at protein level. In our substudy, we found evidence of a tight correlation between the increase in plasma glucose and increase in MBL-C concentration over time. Furthermore, from approximation a prolonged MBL half-life might be a consequence of diabetes that contributes to an increased circulating level. The strongest evidence from this study indicates that the elevated MBL level in diabetes is caused by increased production. Future studies should also explore the cellular mechanism leading to increased production.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

Mette Hagensen, Aarhus University is thanked for the donation of Akita plasmas. The great support from Mogens Erlandsen, Aarhus University with the statistical analyses is appreciated. Henrik Hager and Jens Randel Nyengaard Erlandsen, Aarhus University are thanked for assistance with immunohistochemistry. The financial support from the A.P. Møller Foundation for the Advancement of Medical Science, the EFSD-JDRF-NN Research Grant and the Lundbeck Foundation (Lundbeck Foundation Nanomedicine Centre for Individualized Management of Tissue Damage and Regeneration (LUNA)) is much appreciated.

Author contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References

JAØ contributed essential in the conception and design of the study, acquisition of data, analysis, interpretation of data, drafting and revising the article critically for important intellectual content and the final approval of the version to be published. MB, TKH, ST and AF contributed essentially in the conception and design of the study, acquisition of data, analysis, interpretation of data, revising the article critically for important intellectual content and the final approval of the version to be published. FDH: contributed to the conception and design of the study, acquisition of data and by revising the article critically for important intellectual content and the final approval of the version to be published.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Declaration of interests
  9. Author contributions
  10. References