SEARCH

SEARCH BY CITATION

Keywords:

  • cholesterol metabolism;
  • diabetes

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Objectives

Heterozygous mutations in hepatocyte nuclear factor 1α (HNF1α) cause maturity onset diabetes of the young 3 (MODY3), an autosomal dominant form of diabetes. Deficiency of HNF1α in mice results in diabetes, hypercholesterolaemia and increased bile acid (BA) and cholesterol synthesis. Little is known about alterations in lipid metabolism in patients with MODY3. The aim of this study was to investigate whether patients with MODY3 have altered cholesterol and BA synthesis and intestinal cholesterol absorption. A secondary aim was to investigate the effects of HNF1α mutations on the transcriptional regulation of BA metabolism.

Methods

Plasma biomarkers of BA and cholesterol synthesis and intestinal cholesterol absorption were measured in patients with MODY3 (= 19) and in matched healthy control subjects (= 15). Cotransfection experiments were performed with several promoters involved in BA metabolism along with expression vectors carrying the mutations found in these patients.

Results

Plasma analysis showed higher levels of BA synthesis in patients with MODY3. No differences were observed in cholesterol synthesis or intestinal cholesterol absorption. Cotransfection experiments showed that one of the mutations (P379A) increased the induction of the cholesterol 7α-hydroxylase promoter compared with HNF1α, without further differences in other studied promoters. By contrast, the other four mutations (L107I, T260M, P291fsinsC and R131Q) reduced the induction of the farnesoid X receptor (FXR) promoter, which was followed by reduced repression of the small heterodimer partner promoter. In addition, these mutations also reduced the induction of the apical sodium-dependent bile salt transporter promoter.

Conclusions

BA synthesis is increased in patients with MODY3 compared with control subjects. Mutations in HNF1α affect promoters involved in BA metabolism.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Maturity onset diabetes of the young (MODY) represents a spectrum of disorders characterized by autosomal dominant inheritance, early onset (usually before the age of 25 years) and β-cell dysfunction. The exact prevalence is unknown but MODY has been estimated to account for 1–2% of individuals with type 2 diabetes (T2D) [1]. Heterozygous mutations of the transcription factor 1 (TCF1) gene, encoding hepatocyte nuclear factor 1α (HNF1α), cause MODY3 [2]. More than 200 different mutations in the TCF1 gene have been reported [3, 4]. The clinical expression of MODY3 is highly variable which may, at least partially, be explained by the type and location of the mutation [4].

Bile acid (BA) synthesis serves as the major elimination route of excess cholesterol, to maintain cholesterol homoeostasis. In the liver, cholesterol is converted to 7α-hydroxycholesterol by cholesterol 7α-hydroxylase [also known as cytochrome P450 7A1 (CYP7A1)], the rate-limiting enzyme of the classic pathway of bile acid (BA) synthesis, which is then converted to 7α-hydroxy-4-cholesten-3-one (C4). Cholesterol is also oxidized by CYP27A1 and CYP7B1 in an alternative pathway that exists in all tissues. In humans, the classic pathway is mainly responsible for BA synthesis. It has been reported that serum levels of C4 reflect BA synthesis and correlate with the enzymatic activity of CYP7A1 [5-8]. Hence, determination of serum C4 levels allows investigation of BA synthesis in vivo without collection of liver biopsies.

Farnesoid X receptor (FXR) is regarded as the BA nuclear receptor as it regulates the expression of several key genes involved in BA synthesis, metabolism and transport [9, 10]. In the liver, FXR activation upregulates the expression of small heterodimeric partner (SHP), which then forms a heterodimer with liver-related homologue 1 (LRH-1) leading to repression of CYP7A1 [11, 12]. In the intestine, FXR activation induces the synthesis of fibroblast growth factor 19 (FGF19). Secreted FGF19 binds to its cognate receptor (FGFR4) which activates intracellular pathways that target the expression of CYP7A1 [13, 14].

BAs actively reabsorbed in the terminal ileum by the apical sodium-dependent bile salt transporter (ASBT/SLC10A2) contribute substantially to levels in the enterohepatic circulation. Mice lacking Asbt have increased faecal BA excretion and increased Cyp7a1 expression [15]. In humans, mutations in the ASBT gene lead to BA malabsorption [16], and patients with obstructive cholestasis have lower intestinal ASBT expression [17].

The role of HNF1α in the regulation of cholesterol homoeostasis has been investigated in Tcf1−/− mice [11]. These mice were reported to have defective BA transport, increased BA and cholesterol synthesis, and impaired HDL metabolism. By contrast, patients with MODY3 have been reported to have similar HDL cholesterol and apolipoprotein A1 levels as healthy control subjects [18], but whether these patients have altered BA and cholesterol synthesis remains unknown. The aim of this study was to investigate whether patients with MODY3 have altered BA synthesis, cholesterol synthesis and intestinal cholesterol absorption, and to investigate the effects of TCF1 mutations on some of the key promoters in BA metabolism.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

Study subjects

Blood samples were collected in the fasting state, and serum was obtained by centrifugation at 3000 g for 10 min and stored at −70 °C. The study participants included 19 patients with MODY3 and 15 healthy individuals matched on the basis of age and body mass index (Table 1). Patients with MODY3 included two carries of the P379A mutation (proline to alanine substitution at codon 379), three carriers of the L107I mutation (leucine to isoleucine substitution at codon 107), three carriers of the T260M mutation (threonine to methionine substitution at codon 260), eight carriers of the P291fsinsC mutation (insertion of a C at codon 291), two carriers of the R131Q mutation (arginine to glutamine substitution at codon 131) and one carrier of the InsTGGGGGT (insertion of TGGGGGT in the 5′-UTR). None of the participants was receiving lipid-lowering treatment, and all provided written informed consent to participate in the study. The study was approved by the Ethics Committee at the Department of Medicine, University of Lund, Sweden.

Table 1. Clinical characteristics of patients with MODY3 and control subjects
 ControlMODY3
  1. Values are presented as mean ± SEM, unless otherwise stated.

  2. BMI, body mass index.

  3. Student's t-test: **< 0.01 and, ***< 0.001, MODY3 versus control.

Subjects (male/female), n15 (7/8)19 (8/11)
Age (years)37.4 ± 0.438.4 ± 3.3
BMI (kg m−2)23.4 ± 0.923.4 ± 1.1
Cholesterol (mmol L−1)5.19 ± 0.25.49 ± 0.3
Triglycerides (mmol L−1)0.95 ± 0.11.12 ± 0.1
Glucose (mmol L−1)5.35 ± 0.18.86 ± 1.6 ***
Lathosterol/cholesterol (μmol mmol−1)0.58 ± 0.060.52 ± 0.04
Lanosterol/cholesterol (nmol mmol−1)42.9 ± 2.2749.9 ± 4.86
Sitosterol/cholesterol (μmol mmol−1)1.22 ± 0.161.25 ± 0.17
Campesterol/cholesterol (μmol mmol−1)2.44 ± 0.282.76 ± 0.43
C4/cholesterol (nmol mmol−1)7.30 ± 0.369.98 ± 0.54 **
FGF19 (pg mL−1)201 ± 47.4226 ± 37.9

Cells, plasmids and chemicals

HuH7 and Caco2 cells were purchased from American Type Culture Collection (Manassas, VA, USA). The human CYP7A1 promoter (−1886 to +24 bp cloned into the pGL2 basic vector) [19] was a generous gift from Professor Maurizio Crestani, University of Milan, Italy. The human FGF19 promoter (−1954 to +244 bp cloned into the pGL3 basic vector) [20] was a generous gift from Professor Thomas Langmann, University of Regensburg, Germany. The human ASBT promoter (−1688 to +599 bp cloned into the pGL3 basic vector) [21] was a generous gift from Professor Gerd A. Kullak-Ublick, University Hospital Zurich, Switzerland. The human FXR promoter (−886 to +129 bp cloned into the pGL3 basic vector) [22] was a generous gift from Dr Guiyu Lou, Department of Biochemistry and Molecular Biology, Third Military Medical University, Chongqing, China. The human SHP promoter (−2.2 kb to −243 bp cloned into the pGL2 basic vector) [23] was a generous gift from Professor Hueng-Sik Choi, Hormone Research Center, Chonnam National University, Kwangju, Republic of Korea. The HNF1α expression vector was a generous gift from Professor Pal R. Njølstad and Dr Lise Bjørkhaug Gundersen, Haukeland University Hospital, Norway. Standards for lanosterol and 7-α-hydroxy-4-cholesten-3-one were purchased from Research Plus Inc. (Barnegat, NJ, USA) and Steraloids Inc. (Newport, RI, USA), respectively. The internal standards D4-lathosterol and D6-campesterol:sitosterol were purchased from CDN Isotopes Inc. (Quebec, Canada). 7β-hydroxy-4-cholesten-3-one was purchased from Steraloids Inc. 2H3-lanosterol was synthesized in-house according to the methods of Lund et al. [24].

Serum analyses

Unesterified lathosterol was determined in 25 μL serum by isotope dilution mass spectrometry (MS), using deuterium-labelled lathosterol as the internal standard, as previously described [25]. Sitosterol and campesterol were extracted from 20 μL serum, using D6-campesterol:sitosterol as the internal standard, and analysed by gas chromatography (GC)/MS as previously described [26]. Lanosterol and C4 were extracted from 50 μL serum with isopropanol:heptan (4 : 1, w/w), in which the internal standards were dissolved and analysed using a liquid chromatography (LC)/MS/MS system as previously described [27]. Quantification of serum concentrations of FGF19 was performed using an enzyme-linked immunosorbent assay (ELISA) kit (MyBioSource, San Diego, CA, USA) according to the manufacturer's recommendations.

Mutagenesis

Each of the mutations found in the patients with MODY3 in this study was introduced into a human expression vector for HNF1α using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) according to the manufacturer's recommendations. Primers were designed with substituted or inserted bases indicated by the lower-case letters, using the web-based program Primer X (http://www.bioinformatics.org/primerx/).

For the P379A mutation: forward sequence, 5′-CCC TCC CCg CTG TCA GCA CCC TGA CA-3′; reverse sequence, 5′-TGT CAG GGT GCT GAC AGc GGG GAG GGG-3′. For the L107I mutation: forward sequence, 5′-GTG GTG GAG ACC aTT CTG CAG GAG G-3′; reverse sequence, 5′-CCT CCT GCA GAA tGG TCT CCA CCA C-3′. For the T260M mutation: forward sequence, 5′-GCT CCA ACC TCG TCA tGG AGG TGC GTG TCT AC-3′; reverse sequence, 5′-GTA GAC ACG CAC CTC CaT GAC GAG GTT GGA GC-3′. For the P291fsinsC mutation: forward sequence, 5′-GTA CAG CGG aCC aCC CcC CAG GGC CAG-3′; reverse sequence, 5′-CTG GCC CTG GgG GGt GGt CCG CTG TAC-3′. For the R131Q mutation: forward sequence, 5′-CAC AAC ATC CCA CAG CaG GAG GTG GTC GAT ACC-3′; reverse sequence, 5′-GGT ATC GAC CAC CTC CtG CTG TGG GAT GTT GTG-3′.

Cell experiments

HuH7 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 100 units mL−1 penicillin and 100 μg mL−1 streptomycin. Caco2 cells were grown in DMEM supplemented with 20% FBS, 100 units mL−1 penicillin and 100 μg mL−1 streptomycin. The medium was replaced every other day. Cells were maintained in 75 cm2 cell culture flasks and passaged at ~90% confluence. HuH7 cells were transfected using Lipofectin reagent (Invitrogen, Carlsbad, CA, USA) at a ratio of 3 : 1 (Lipofectin:DNA) as previously described [28]. pGL3 empty vector (Promega, Madison, WI, USA) was used to adjust for differences in amount of DNA added to the cells. Caco2 cells were transfected using Lipofectamine LTX reagent (Invitrogen) at a ratio of 1.5 : 1 (Lipofectamine:DNA) according to the manufacturer's instructions. Cell lysates were prepared in reporter lysis buffer (Promega) 24 h (Caco2 cells) or 48 h (HuH7 cells) post-transfection. β-galactosidase and luciferase activities were determined using kits, according to the manufacturer's recommendations (Promega). Transfection data are expressed as luciferase activity relative to β-galactosidase activity (103), and experiments were performed in triplicate, repeated at least twice.

Statistics

Values are presented as means ± SEM. Statistical analyses were performed using statistica version 6.0 (StatSoft, Tulsa, OK, USA). The Student's t-test was used to compare differences in serum parameters between patients with MODY3 and control subjects. anova followed by Tukey's honestly significant difference (HSD) test were used to compare differences in induction of promoters by HNF1α and MODY3 mutations. Correlations between serum C4 levels and CYP7A1, FXR and SHP promoter activities after cotransfection with the respective mutant expression vector were determined by least-squares regression analysis. A P-value of <0.05 was regarded as statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

BA synthesis is increased, but markers of cholesterol synthesis are not different in patients with MODY3 compared with control subjects

Serum biochemical markers were used to examine possible alterations in cholesterol homoeostasis in patients with MODY3 compared with healthy control subjects. Analyses of lathosterol and lanosterol, regarded as markers of hepatic and whole-body cholesterol synthesis [29], did not show any significant differences between control subjects and patients with MODY3 (Table 1). Serum levels of the plant sterols, sitosterol and campesterol, considered to indirectly reflect intestinal cholesterol absorption [30], were not significantly different between the two groups (Table 1). Serum levels of C4 reflect BA synthesis and are correlated with the enzymatic activity of CYP7A1 [5-8]. It is interesting that C4 levels were ~40% higher in patients with MODY3 compared with control subjects (< 0.01; Table 1). Serum concentrations of FGF19, an intestinal target of FXR that inhibits the expression of CYP7A1 [13, 14], were quantified using an ELISA. No significant differences in serum FGF19 levels were observed between the two groups (Table 1).

Serum C4 levels do not reflect CYP7A1 promoter activity

As mentioned, serum levels of C4 reflect BA synthesis and are correlated with the enzymatic activity of CYP7A1 [5-8]. Binding of HNF1α to its cognate HNF1 binding site in the human CYP7A1 promoter induces the promoter activity [31]. To investigate whether the different types of mutations in our patients with MODY3 may increase the ability to activate the human CYP7A1 promoter, each mutation was introduced into an expression vector for HNF1α and used for transient cotransfection of human hepatoma (HuH7) cells. However, it was not possible to investigate the InsTGGGGGT mutation, carried by one patient, by this method. Of interest, we found that only one of the mutations (P379A) increased the activation of the CYP7A1 promoter compared with HNF1α, whereas activation was reduced (L107I and R131Q) or abolished (T260M and P291fsinsC) by the other four mutations (Fig. 1). Thus, we plotted individual C4 levels, grouped according to the type of mutation (Fig. 2), and these levels were correlated with the CYP7A1 promoter activity after cotransfection with the respective mutant expression vector. However, no correlations were observed (data not shown) suggesting that the higher serum C4 levels in these subjects are not only caused by direct effects by HNF1α on the CYP7A1 promoter activity but also by indirect effects via other genes (e.g. FXR, SHP, FGF19 and ASBT) which in turn can affect the expression of CYP7A1. Hence, we investigated how these different types of mutations might affect other promoters involved in BA metabolism.

image

Figure 1. Transient cotransfection of HuH7 cells with the human CYP7A1 promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human hepatoma HuH7 cells along with the human CYP7A1 promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test: ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

image

Figure 2. Serum C4 levels in all participants. Serum levels of C4 reflect BA synthesis and are correlated with the enzymatic activity of CYP7A1. Levels in all individuals are shown, grouped by the type of mutation in the TCF1 gene. Median values are depicted by the solid lines.

Download figure to PowerPoint

Cotransfection experiments in HuH7 cells

In the liver, BAs activate FXR which in turn upregulates the expression of SHP leading to repression of CYP7A1 [11, 12]. Transient cotransfection of HuH7 cells with either the human FXR (Fig. 3) or SHP (Fig. 4) promoter along with mutant expression vectors showed no significant differences between the P379A mutation and HNF1α. By contrast, decreased activation of the FXR promoter (Fig. 3) and reduced repression of the SHP promoter (Fig. 4) were observed in the case of the other four mutations. In line with this, we did not find a significant correlation between serum C4 levels and FXR and SHP promoter activities when all constructs were used in the correlation analysis. However, there were strong correlations between serum C4 levels and FXR promoter activities (r = 0.82, < 0.05) and between serum C4 levels and SHP promoter activities (r = −0.79, < 0.05) when the four MODY3 mutations L107I, R131Q, T260M and P291fsinsC were included in the analysis.

image

Figure 3. Transient cotransfection of HuH7 cells with the human FXR promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human hepatoma HuH7 cells along with the human FXR promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test: ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

image

Figure 4. Transient cotransfection of HuH7 cells with the human SHP promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human hepatoma HuH7 cells along with the human SHP promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test: *< 0.05, **< 0.01 and ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

Cotransfection experiments in Caco2 cells

In the intestine, BAs also activate FXR which upregulates the expression of SHP leading to repression of CYP7A1 [11, 12]. However, intestinal FXR activation also upregulates FGF19, which also represses CYP7A1 [13, 14, 32]. Transient cotransfection of human intestinal Caco2 cells with either the human FXR (Fig. 5) or SHP (Fig. 6) promoter along with mutant expression vectors (0.5 or 1.0 μg) did not lead to any major differences at the lower concentration (0.5 μg) compared with HNF1α. However, at the higher concentration (1 μg), activation of FXR was not increased by the T260M and P291fsinsC mutations (Fig. 5). Transient cotransfection of Caco2 cells with the human FGF19 promoter along with mutant expression vectors revealed that HNF1α is a weak activator of the human FGF19 promoter (Fig. 7). Accordingly, only modest increases were seen with the L107I and R131Q mutations compared with HNF1α (Fig. 7). ASBT contributes substantially to the enterohepatic circulation, and increased Cyp7a1 expression has been reported in mice lacking Asbt [15]. Transient cotransfection of Caco2 cells with the human ASBT promoter along with mutant expression vectors showed no differences between the P379A mutation and HNF1α (Fig. 8). By contrast, activation of the ASBT promoter was reduced (L107I and R131Q) or abolished (T260M and P291fsinsC) by the other four mutations (Fig. 8).

image

Figure 5. Transient cotransfection of Caco2 cells with the human FXR promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human intestinal Caco2 cells along with the human FXR promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test: ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

image

Figure 6. Transient cotransfection of Caco2 cells with the human SHP promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human intestinal Caco2 cells along with the human SHP promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test did not show any significant differences, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

image

Figure 7. Transient cotransfection of Caco2 cells with the human FGF19 promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human intestinal Caco2 cells along with the human FGF19 promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test: *< 0.05, **< 0.01 and ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

image

Figure 8. Transient cotransfection of Caco2 cells with the human ASBT promoter. The mutations found in our patients with MODY3 were introduced into a human expression vector for HNF1α and used for cotransfection of human intestinal Caco2 cells along with the human ASBT promoter. Transfections were performed in triplicate and repeated at least twice. Values are presented as mean ± SEM. Tukey's HSD test ***< 0.001, HNF1α versus MODY3 mutations at either concentration of expression vector (0.5 or 1 μg).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

In this study, we used serum biomarkers to investigate possible alterations in cholesterol homoeostasis in patients with MODY3, compared with healthy control subjects. We found that BA synthesis was increased, whereas markers of cholesterol synthesis and intestinal cholesterol absorption were not different in patients with MODY3 compared with control subjects. Promoter studies in human hepatoma (HuH7) and intestinal (Caco2) cells showed that the mutations in our patients with MODY3 affect promoters involved in BA metabolism.

Cholesterol homoeostasis is maintained by balancing endogenous cholesterol synthesis and intestinal cholesterol absorption with excretion of biliary cholesterol and BAs. Dyslipidaemia, characterized by a high plasma triglyceride concentration, low HDL cholesterol and an increased level of small dense LDL particles, is very common in individuals with T2D [33]. Additionally, increased cholesterol synthesis has been reported in those with T2D [34]. Thus, monogenic forms of diabetes may also lead to unfavourable effects on lipid metabolism. However, MODY1 patients with a specific mutation (K99fsdelAA) in the gene encoding HNF4α have lower plasma triglyceride levels compared with control subjects [35], and we have previously shown that such patients have lower levels of very low-density lipoprotein and LDL esterified cholesterol [36]. In the same study, we also found that patients with MODY3 have similar lipoprotein profiles as healthy control subjects [36]. However, there was a trend towards lower LDL cholesterol levels, although – probably due to the small study sample size – this did not reach statistical significance. Unfortunately, in the present study, we were not able to analyse lipoprotein profiles as samples that have not undergone freeze-thawing are required (but were not available) for accurate measurements. Thus, whether the higher BA synthesis in patients with MODY3, without significant changes in cholesterol synthesis and cholesterol absorption, affects LDL cholesterol levels need to be further investigated. However, these results suggest that the dyslipidaemic phenotype reported in T2D is not present in individuals with MODY3.

The observed lipidaemic phenotype in patients with MODY3 is in contrast to mice lacking Tcf1 [11] in which increased cholesterol synthesis and impaired HDL metabolism were reported. However, there are major species-related differences between mice and humans with respect to cholesterol metabolism. For example, plasma cholesterol is mainly transported in LDL in humans and in HDL in mice, humans have much lower levels of cholesterol synthesis than mice [37], and the feed-forward regulation of Cyp7a1 by cholesterol (mediated by liver X receptor α) in mice is absent in humans [38]. Moreover, most patients with MODY3 have heterozygous mutations [39], whereas Tcf1+/− mice have been reported to be similar to wild-type mice [40]. Thus, the effects in Tcf1−/− mice are expected to be more severe and may not be clinically relevant.

For ethical reasons, we were not able to investigate the molecular mechanisms responsible for the increased BA synthesis in patients with MODY3 in vivo. In an attempt to gain some mechanistic insights, each mutation was introduced into an expression vector for HNF1α and used for transient cotransfection of human hepatoma (HuH7) and intestinal (Caco2) cells along with promoters for CYP7A1, FXR, SHP, FGF19 and ASBT. The human HNF1α protein consists of 631 amino acids (aa) subdivided into three functional regions: an amino-terminal dimerization domain (aa 1–32), a DNA-binding motif (aa 203–279) and a carboxyl-terminal transactivation domain (aa 281–631) [41]. Dimerization is necessary for interaction with specific cis-acting elements in promoters of target genes, as deletion of the dimerization domain prevents DNA binding. Thus, it has been suggested that the transactivation domain may be more tolerant of minor structural changes than the DNA-binding motif [42].

P379A is a missense mutation located in the transactivation domain, and it has previously been reported that the transactivation potential is decreased compared with HNF1α [43]. However, we found that this mutation increased the activation of the CYP7A1 promoter compared with HNF1α (Fig. 1), whereas no other major differences were observed for the FXR (Figs 3 and 5), SHP (Figs 4 and 6), FGF19 (Fig. 7) or ASBT (Fig. 8) promoters. The discrepancy between these findings and those of the previous study [43] may be due to the different cell lines used [i.e. mouse pancreatic (Min6) and African green monkey kidney (Cos7) cells, compared with our human hepatoma and intestinal cells]. Other possible reasons for the inconsistent results include species differences and the use of different promoters and expression vectors.

P291fsinsC, the most common mutation in HNF1α, is a frameshift mutation leading to synthesis of a 315-aa protein lacking most of the transactivation domain [44]. Consistent with this, we found that P291fsinsC reduced or abolished the activation of the CYP7A1 and FXR promoters in HuH7 cells as well as of the FXR and ASBT promoters in Caco2 cells. The reduced activation of FXR was followed by reduced repression, at the higher concentration (1.0 μg), of the SHP promoter in HuH7 cells (Fig. 4) but not in Caco2 cells.

R131Q is a missense mutation that is not defective in the DNA-binding motif but leads to decreased transactivation potential compared with HNF1α [45]. In line with this, we found that R131Q reduced the activation of CYP7A1 (Fig. 1) and FXR (Fig. 3) promoters in HuH7 cells and of the ASBT promoter in Caco2 cells (Fig. 8). However, similar activation of the FXR (Fig. 5) and SHP (Fig. 6) or increased activation of the FGF19 (Fig. 7) promoters were observed in Caco2 cells. Thus, it seems that the effects of this mutation are also dependent on the promoter and the cell line that are used.

L107I is a missense mutation that has been reported to result in loss of function of the protein and impaired transcriptional activity due to impaired DNA binding [46]. In the present study, the transfection experiments showed reduced activation of the CYP7A1 (Fig. 1) and FXR (Fig. 3) promoters in HuH7 cells and of the ASBT promoter in Caco2 cells (Fig. 8). By contrast, no major differences (FXR and SHP, Figs 5 and 6, respectively) or increased (FGF19, Fig. 7) activation were observed in Caco2 cells. Again, it appears that differences in promoters and/or cell lines may affect the results.

T260M is a missense mutation within the DNA-binding motif [3, 47]. We showed that this mutation abolished the activation of the CYP7A1 (Fig. 1), FXR (Fig. 3) and SHP (Fig. 4) promoters in HuH7 cells as well as of the ASBT promoter in Caco2 cells (Fig. 8). It is interesting that our results in Caco2 cells showed decreased activation of the FXR promoter (Fig. 5) at the higher concentration (1.0 μg), but similar effects on SHP (Fig. 6) and FGF19 (Fig. 7) promoters.

FXR is expressed at high levels in the human liver and intestine and may inhibit the expression of CYP7A1 by indirect mechanisms and not via direct binding to CYP7A1 [48]. These indirect mechanisms may be initiated by the FXR/SHP/LRH-1 pathway in the liver [11, 12] or the FXR/FGF19/FGFR4 pathway in the intestine [13, 14]. We used two human cell lines in this study, one hepatic and one intestinal, to investigate whether mutations in the gene encoding HNF1α may affect the expression of CYP7A1 via direct or indirect mechanisms. Only one of the MODY3 mutations (P379A) increased the CYP7A1 promoter activity compared with HNF1α, suggesting direct effects on CYP7A1 expression by this mutation. The other four mutations reduced the FXR promoter activity and then reduced repression of SHP, suggesting indirect effects of these mutations on the expression of CYP7A1. Although two of these mutations (T260M and P291fsinsC) significantly reduced FXR promoter activity in Caco2 cells, this was not followed by reduced SHP or FGF19 promoter activity in these cells, suggesting that the hepatic FXR/SHP/LRH-1 pathway may be responsible for the effects on CYP7A1 expression in subjects carrying these four mutations. This is also supported by the significant correlation between serum C4 levels and FXR and SHP promoter activities in HuH7 cells.

The mutations found in the patients with MODY3 in this study were within different domains of the HNF1α protein. However, whether the mutation was within the DNA-binding motif (i.e. T260M) or within the transactivation domain (i.e. P379A or P291fsinsC) did not affect the severity of the effect. Furthermore, the P379A and P291fsinsC mutations did not produce similar effects on the promoters used in this study. Instead, P379A seemed to have direct effects on the CYP7A1 promoter, which may, at least partially, contribute to the increased BA synthesis in individuals with this mutation. For the other four mutations (including P291fsinsC), reduced hepatic FXR activation followed by reduced repression of SHP may contribute to the higher degree of BA synthesis in individuals with MODY3. Thus, different MODY3 mutations may alter different targets; however, all these mutations lead to increased BA synthesis compared with controls.

Conflict of interest statement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

None of the authors has any conflicts of interest to declare.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References

This work was supported by grants from the Swedish Heart-Lung Foundation, and the Henning and Johan Throne-Holsts Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflict of interest statement
  8. Acknowledgements
  9. References