Circulating matrix γ-carboxyglutamate protein (MGP) species are refractory to vitamin K treatment in a new case of Keutel syndrome


Karin Y. van Spaendonck-Zwarts, Department of Genetics, University Medical Center Groningen, University of Groningen, PO Box 30001, 9700 RB Groningen, the Netherlands.
Tel.: +31 503617229; fax: +31 503617231.


Summary. Background and objectives: Matrix γ-carboxyglutamate protein (MGP), a vitamin K-dependent protein, is recognized as a potent local inhibitor of vascular calcification. Studying patients with Keutel syndrome (KS), a rare autosomal recessive disorder resulting from MGP mutations, provides an opportunity to investigate the functions of MGP. The purpose of this study was (i) to investigate the phenotype and the underlying MGP mutation of a newly identified KS patient, and (ii) to investigate MGP species and the effect of vitamin K supplements in KS patients. Methods: The phenotype of a newly identified KS patient was characterized with specific attention to signs of vascular calcification. Genetic analysis of the MGP gene was performed. Circulating MGP species were quantified and the effect of vitamin K supplements on MGP carboxylation was studied. Finally, we performed immunohistochemical staining of tissues of the first KS patient originally described focusing on MGP species. Results: We describe a novel homozygous MGP mutation (c.61+1G>A) in a newly identified KS patient. No signs of arterial calcification were found, in contrast to findings in MGP knockout mice. This patient is the first in whom circulating MGP species have been characterized, showing a high level of phosphorylated MGP and a low level of carboxylated MGP. Contrary to expectations, vitamin K supplements did not improve the circulating carboxylated MGP levels. Phosphorylated MGP was also found to be present in the first KS patient originally described. Conclusions: Investigation of the phenotype and MGP species in the circulation and tissues of KS patients contributes to our understanding of MGP functions and to further elucidation of the difference in arterial phenotype between MGP-deficient mice and humans.


Matrix γ-carboxyglutamate (Gla) protein (MGP) is increasingly recognized as a potent inhibitor of arterial calcification with important clinical implications. MGP is present in the vascular wall and in cartilage, where it is synthesized by vascular smooth muscle cells and chondrocytes, respectively. It can undergo two post-translational modifications: γ-glutamate carboxylation, which is vitamin K-dependent, and serine phosphorylation. Uncarboxylated MGP is invariably found to accumulate in atherosclerotic lesions and areas of arterial calcification in tissue; high circulating levels are found in patients with a high risk of prevalent cardiovascular calcification [1–3]. The two post-translational modifications result in several MGP species (carboxylated, non-carboxylated, phosphorylated, and non-phosphorylated). Carboxylation, which is vitamin K-dependent, enhances the binding of MGP to calcium nuclei in hydroxyapatite [4,5], and is a prerequisite for its inhibition of bone morphogenetic protein-2 (BMP-2) [6–8], an osteogenic growth factor. Although the exact function of phosphorylation is unclear, it may regulate MGP secretion into the extracellular environment [9]. Several ELISA have been developed to measure different MGP species in the circulation.

Genetically determined loss of MGP function may provide important insights into the functions of the protein. MGP knockout mice were born without abnormalities, but developed severe calcification of all elastic and muscular arteries, eventually resulting in rupture of the calcified aorta and subsequent death within 2 months [10]. Histologic examination of MGP knockout mice revealed extensive calcification of the elastic lamellae at the age of 2 weeks, and the appearance of cartilage nodules in the calcified aorta at the age of 1 month [11]. They also displayed cartilage calcification at sites of proliferating chondrocytes – the lower end of the trachea, the main bronchi, and the growth plate – which resulted in short stature, osteopenia, and fractures [10].

Abnormal cartilage calcification is the hallmark of Keutel syndrome (KS), an extremely rare autosomal recessive disorder in humans, resulting from mutations in the MGP gene [12]. So far, only 26 patients with KS have been described in the literature [12–28]. In addition to abnormal cartilage calcification, KS is characterized by brachytelephalangism, peripheral pulmonary artery stenosis, hearing loss (sensorineural, mixed, and conductive), and facial abnormalities (i.e. mid-face hypoplasia and a depressed nasal bridge) [26]. Four MGP mutations have been reported thus far, all of which predict absent or non-functional MGP [12,26]. Remarkably, KS patients do not seem to suffer from extensive arterial calcification.

We describe a novel MGP mutation, found in a newly identified KS patient. His phenotype was carefully characterized, with specific attention being paid to signs of vascular calcification. In addition, this is the first patient in whom circulating MGP species have been characterized. These measurements were used to evaluate whether he would benefit from vitamin K supplements. In addition to these studies, we investigated the presence of MGP and other calcification inhibitors in tissues from the first KS patient originally described by Keutel et al. [13].

Materials and methods


A male patient was referred from the Department of Otorhinolaryngology to our clinical genetics outpatient clinic, and subsequently diagnosed with KS. This patient and his first-degree relatives (parents and two siblings) were included in our study. All participants gave their informed consent. To assess the effect of increased vitamin K intake on his circulating MGP levels, he was prescribed 10 mg of vitamin K1 (phylloquinone; Konakion; Roche Nederland, Woerden, The Netherlands) daily for 3 months at the age of 21 years. This is the standard form and dosage of vitamin K prescribed in The Netherlands.

No tissue from our patient was available for immunohistochemical staining. Fortunately, tissues from another male patient diagnosed with KS were available. This was one of the two siblings originally described by Keutel et al. [13]. Meier et al. [29] described the follow-up and post-mortem examination of this patient (their case 1) in 2001. He died at age 38 years, owing to right heart failure. He had previously suffered from increasing dyspnea and coughing, and seizures with ischemic cerebral lesions, and had been diagnosed with mediastinal seminoma, for which he had received chemotherapy.

Genetic analysis

Genomic DNA was isolated from blood samples (QIAamp DNA Blood mini kit, QIAGEN, Hilden, Germany) obtained from our newly identified patient with KS and his four first-degree relatives. The four coding exons, as well as exon–intron boundaries, of the MGP gene were amplified by PCR with the following oligonucleotides: exon1-F (CTCTCAACTGCTCTGGTTC), exon1-R (AAGTAAGCCAAAGTCAGAGGC), exon2-F (TTTCCTCTTCTTCCATCCCTG), exon2-R (CCCTCCCTGTTATATATCTTTC), exon3-F (TATTCACGGAAATATTTCCAGC), exon3-R (CAGATCTGTGATCTACACTG), exon4-F (GTATTTTTCCACTTTATCCTTC), and exon4-R (AAAATCAGGTGCCAGCCTC); they were then analyzed by bidirectional fluorescent direct sequencing. Results were confirmed in a second amplification product.

A skin biopsy of the patient was performed to obtain fibroblasts, from which mRNA was extracted. The effect of the identified MGP mutation on mRNA splicing was determined by performing RT-PCR on the obtained mRNA. The entire MGP coding sequence was amplified by RT-PCR and analyzed by bidirectional fluorescent direct sequencing.

Biochemical measurements

Blood samples were collected from our patient and, for comparison, from his first-degree relatives. For circulating MGP and osteocalcin (OC) measurements, blood was collected by venipuncture in serum (10 mL; BD Vacutainer Systems, Plymouth, UK) and in sodium citrate tubes (10 mL; BD Vacutainer Systems), and stored for 20 min at room temperature before centrifugation (15 min, 1580 × g). All data are means of duplicate measurements, unless stated otherwise.

Circulating MGP levels were quantified with four different MGP ELISA. Circulating non-phosphorylated MGP was measured with a kit from Biomedica (Vienna, Austria), in which the mAb against human non-phosphorylated MGP residues 3–15 is coated on the microtiter plate [30]. Circulating phosphorylated MGP was measured with a competitive ELISA, using a mAb raised against phosphorylated MGP residues 3–15 (VitaK BV, Maastricht, The Netherlands). The antibody was selected for its specificity towards phosphorylated MGP with standard techniques, as described for the other MGP antibodies [1]. For the phosphorylated MGP ELISA, the antibody was coupled to the microtiter plate. Serum sample or standard was supplemented with tracer (biotinylated phosphorylated MGP), transferred to the microtiter plate, and incubated overnight at 4 °C. The MGP concentration was calculated with the aid of a calibration curve of synthetic phosphorylated MGP. Biotinylated, synthetic phosphorylated MGP was made by Thinkpeptides (ProImmune, Oxford, UK).

Circulating uncarboxylated and carboxylated MGP were quantified with two sandwich ELISAs (VitaK BV). In these assays, the antibody against non-phosphorylated MGP residues 3–15 served as a detecting antibody. The biotinylated antibodies against uncarboxylated MGP residues 35–49 and carboxylated MGP residues 35–54, respectively, served as capture antibodies [31]. These assays seem to be particularly suited for assessing the effect of changes in vitamin K status on MGP levels.

The carboxylation status of circulating OC, a vitamin K-dependent protein present in bone, was assessed by measurements of uncarboxylated OC (ucOC) and carboxylated OC (cOC) with dual-antibody OC ELISAs (Takara, Shiga, Japan). The ucOC/cOC ratio was calculated from these measurements. A decrease in the ucOC/cOC ratio over time indicates an improved vitamin K status [32].


Immunohistochemistry for Keutel’s original first male patient was performed on sections of lung (from trachea to bronchi of the third generation), aorta, skin, and heart tissue embedded in paraffin. All sections were stained for hematoxylin and eosin (HE). Von Kossa and Elastica von Gieson staining were performed to visualize calcification and elastic fibers, respectively. Immunostaining for MGP was performed as described previously [1], with the same MGP mAbs as used in the MGP ELISAs directed against phosphorylated MGP, non-phosphorylated MGP, carboxylated MGP, and non-carboxylated MGP. Sections were also stained with rabbit anti-human fetuin-A (BioVendor, Heidelberg, Germany) and rabbit anti-human osteopontin (OPN) (Abcam, Cambridge, UK). For MGP, staining was performed with biotinylated sheep anti-mouse IgG (60 min at room temperature; Dako, Golstrup, Denmark) as a secondary antibody, followed by incubation with avidin-linked alkaline phosphatase complex (30 min at room temperature; Dako). Staining was then performed with the Vector Red Alkaline Phosphatase Substrate kit I (Vector Laboratories, Burlingame, CA, USA), yielding a red color. For fetuin-A and OPN, biotinylated goat anti-rabbit IgG was used as a secondary antibody, followed by incubation with peroxidase-labeled streptavidin (LSAB2 System-HRP; Dako), according to the manufacturer’s instructions. Staining was then performed with the Vector NovaRED substrate kit (Vector Laboratories), yielding a brown color. Sections were counterstained with hematoxylin. Controls for the immunoreactions were performed by omitting the primary antibody.


Phenotypic characterization

A 16-year-old male was referred to the Department of Otorhinolaryngology, University Medical Center Groningen, for evaluation of exertional dyspnea. His medical history included mild congenital bilateral peripheral pulmonary stenosis and respiratory problems, refractive to the use of bronchodilators. Psychomotor development was normal. He is the second child of consanguineous Turkish parents, living in The Netherlands. He has two sisters, and the younger one has a history of childhood epilepsy. All four first-degree relatives were apparently unaffected.

Physical examination of the patient revealed a slight inspiratory stridor on auscultation. There were no obvious dysmorphic facial features (Fig. 1). However, his auricles were strikingly stiff, his chest was asymmetric, and his hands showed shortening of all distal phalanges. His height was − 1 standard deviation (SD) for age (with target height plus 1 SD) according to Turkish growth charts.

Figure 1.

 Frontal and lateral view of our patient. Frontal (A) and lateral (B) views of the patient.

Lung function tests revealed obstructive upper airway disease. Laryngotracheoscopy showed a short and posteriorly tilted epiglottis, leading to difficult intubation. The vocal folds were remarkably short, with a large posterior chink. The cricoid cartilage was stenosed (Fig. 2A), providing the explanation for his exertional dyspnea. The trachea showed an irregular appearance (Fig. 2B). A computed tomography (CT) scan revealed extensive cartilage calcification of his larynx, trachea, main bronchi, costochondral junctions, and auricles. The severity of the stenosis of the cricoid cartilage was clearly shown by three-dimensional CT scan reconstructions (Fig. 3). Hearing thresholds were 10 dB (normal) for the right ear, with a 20-dB conductive hearing loss for the high tones in the left ear. An electrocardiogram showed no abnormalities. An echocardiogram showed normal dimensions and no functional abnormalities except for a small increase in the flow of the left pulmonary artery. Ophthalmologic evaluation did not reveal any abnormalities, and he did not have any skin abnormalities. No abnormalities were found with routine blood hematology and chemistry, including calcium, phosphate, liver function tests, and clotting tests. Overall, these features were most compatible with a diagnosis of KS) (Table 1), and mutation analysis of the MGP gene was performed to confirm this diagnosis.

Figure 2.

 Laryngotracheoscopy. Laryngotracheoscopy images illustrating the stenosed cricoid cartilage (A) and irregular appearance of the trachea of the patient (B), with images of the normal configuration for comparison (a, b). (A) Patient’s cricoid with irregular stenosed appearance. No visible tracheal rings caused by stenosis. (a) At the far left, the left true vocal cord can be seen, and in the center the almost circular appearance of cricoid. Several patches of mucus are present. The trachea is visible distally. (B) Patient’s irregular narrowed lumen of the trachea with tracheal rings shining through the mucosa. (b) Regular smooth tracheal rings. Note: Do not compare the picture size, but the configuration.

Figure 3.

 Three-dimensional computed tomography (CT) scan reconstructions. Three-dimensional CT scan reconstructions illustrating the severity of the patient’s cricoid cartilage stenosis (A), and calcifications of his rib cartilage (B), with images of the normal configuration for comparison (a, b). (A) CT neck, volume dataset, virtual rendering mode, anterior–posterior view: pronounced ossification of hyoid and extensive calcification of cricoid, thyroid and tracheal ring cartilage is visible in white. An air band of hypopharynx and proximal trachea is visible in red. (a) Patient of the same age with a small left lateral laryngocele. Normal ossification of hyoid. No cartilage calcifications are visible. (B) CT thorax, volume dataset, virtual rendering mode, anterior–posterior view: calcifications of bilateral rib arch cartilage (white) are visible medial to the regularly ossified ventral ribs. (b) Patient of the same age (status after sternotomy). No rib cartilage calcifications are visible. Note the extensive calcification of the cricoid, thyroid, hyoid and tracheal rings.

Table 1.   Clinical features of Keutel syndrome
 Our patientLiterature (n = 26)
  1. F, female; M, male; SD, standard deviation. Summary of the clinical features of our patient and the patients with Keutel syndrome described in the literature. *Reported heart defects are ventricular septal defect (n = 1), patent ductus arteriosus and foramen ovale (n = 1), and thick mitral valve leaflets (n = 1). †Reported cerebral defects are multiple calcifications (n = 2), leukodystrophy (with bilateral periventricular and subcortical white matter changes on brain magnetic resonance images after 3 years; n = 1), mild atrophy of the cerebral hemisphere (secondary to meningitis; n = 1), and encephalomalacia (n = 1).

Age at diagnosis (years [mean ± SD])168.3 ± 4.8
GenderM11 M, 15 F
Cartilage calcification
 Nose (septum, ala nasi)8/8
 Ear (outer and middle)+12/12
 Trachea and/or bronchi+20/20
 Epiphyseal stippling3/5
Characteristic facial features
 Midface hypoplasia14/14
 Broad, depressed nasal bridge21/21
Respiratory problems+14/18
Recurrent otitis media/sinusitis9/13
Hearing lossRight −, left +12/17
Peripheral pulmonary stenosis+14/17
Heart defects*3/15
Cerebral defects†4/12
Developmental delay5/15
Slightly short stature (length ≤10th percentile)10/21

Vascular evaluation was performed when the patient was 21 years old, to investigate whether any arterial calcification was present. He had no complaints consistent with cardiovascular disease. His blood pressure was 122/78 mmHg, and the ankle–brachial index was 108% on the right, and 112% on the left side; after exercise, these values were 82% and 79%, respectively (values above 140% may indicate arterial calcification [33]). Carotid intima–media thicknesses of the right and left common carotid arteries were both 0.45 mm, those of the right and left internal carotid arteries were 0.45 and 0.40 mm, respectively, and those of the right and left bulbus were 0.5 and 0.4 mm, respectively (ultrasound upper limit of normal range for adolescents [34]). No arterial calcifications were seen on a chest X-ray of the upper abdomen (hepatic, lienal and renal arteries). A cranial CT scan did not reveal any arterial calcifications of intracranial arteries of the circle of Willis. Multislice CT scanning of the coronary arteries revealed no calcification (calcium score of 0).

MGP mutation

The clinical diagnosis was confirmed by the identification of a novel, homozygous MGP mutation in this patient, c.61+1G>A, indicating nucleotide substitution of guanine by adenine in cDNA of the first nucleotide in the introns after base 61 (donor splice site) (Fig. 4). The effect of this mutation on mRNA splicing was determined by performing RT-PCR. This identified mutation eliminates the consensus donor splice site at the end of exon 1 (TG-gt is replaced by TG-at). Instead of this mutated splice site, a new donor splice site is used, 19 bases upstream of the usual one. Thus, the mutation results in the deletion of 19 bases in the cDNA, which gives rise to a frameshift. A novel stop codon in the 3′-untranslated sequence of the MGP gene is used, 49 bases downstream of the normal stop codon. The new predicted protein is composed of 113 amino acids (instead of 103), and only 14 amino acids of the original sequence are conserved in the sequence upstream of the deletion. Both parents and the older sister of the patient are heterozygous carriers of the mutation. His younger sister does not carry the mutation.

Figure 4.

 Matrix γ-carboxyglutamate protein (MGP) gene and identified mutations. Eexons are shown as boxes and introns as lines (not to scale). Oligonucleotides used for PCR and sequencing are indicated by arrows. The start codon is indicated by an arrow and the stop codon by a line, at the top of the exons. Nucleotide 1 has been counted as the transcription initiation site. Mutations previously reported are indicated, and the mutation reported here is in bold. Mutations affecting splice sites are indicated above the gene; other mutations are indicated below.

Circulating MGP

Four different assays were used to characterize circulating MGP. The circulating non-phosphorylated and phosphorylated MGP levels (Fig. 5A) of the patient and his first-degree relatives were within the normal range. He had a lower concentration of non-phosphorylated MGP than any of his relatives. In contrast, he had the highest concentration of phosphorylated MGP, which was 1.5–2-fold as high as those of his relatives. The circulating uncarboxylated and carboxylated MGP levels (Fig. 5B) were low to very low in all family members, including our patient.

Figure 5.

 Circulating matrix γ-carboxyglutamate protein (MGP) measurements. Circulating MGP species were measured in our patient’s father (parent 1, age 43 years), our patient’s mother (parent 2, age 42 years), our patient (age 19 years), his older sibling (sibling 1, age 20 years), and in younger sibling (sibling 2, age 15 years). Non-phosphorylated MGP and phosphorylated MGP (A), as well as uncarboxylated and carboxylated MGP (B), could be detected in blood samples from the patient and his family members. For reference, the normal range for each measured MGP species is given in brackets. The normal range, defined as the mean ±( 2 × standard deviation [SD]), was based on MGP measurements in 50 healthy subjects. The mean ± SD age of this reference population (28 ± 7 years) was comparable to that of the patient and family members (28 ± 14 years).

As MGP carboxylation depends on the presence of vitamin K, we aimed to investigate whether increased intake of vitamin K would influence the patient’s MGP levels. In theory, increased MGP carboxylation would enhance its calcification-inhibitory activity, which would have a beneficial effect in preventing the development of arterial calcification. The measurements of uncarboxylated and carboxylated MGP are suitable for detecting the effect of changes in vitamin K status on MGP levels. Uncarboxylated MGP, in particular, is sensitive to changes in vitamin K status, and its level decreases significantly within 12 weeks upon vitamin K supplementation in healthy subjects [31]. The patient received 10 mg of vitamin K1 daily for 3 months. The decrease in the ucOC/cOC ratio from 0.70 at baseline to 0.08 after 3 months of supplements confirmed that his systemic vitamin K status had improved. The results of MGP measurements are shown in Fig. 6. The uncarboxylated MGP levels remained unchanged; there was a decrease in carboxylated MGP levels after 3 months. This small decrease is within the range of intraindividual variation, and was not considered to be an effect of increased vitamin K intake. It was therefore concluded that the increased intake of vitamin K1 for 3 months had not affected the circulating MGP levels in our patient.

Figure 6.

 Effect of vitamin K supplementation on matrix γ-carboxyglutamate protein (MGP) levels. Circulating uncarboxylated and carboxylated MGP levels of the patient are shown at baseline and after 3 months of vitamin K1 (K1) supplementation (10 mg daily). The uncarboxylated MGP level remained unchanged. Although there was a very small decrease in the carboxylated MGP level, this was not considered to be clinically significant.

MGP in tissue

HE and Von Kossa staining of lung tissue from the first male patient described by Keutel et al. revealed concentric arterial calcification in the region of the internal elastic lamina of the small and medium-sized pulmonary arteries, which was often accompanied by intimal hypertrophy (Fig. 7A). This type of calcification was also seen in the coronary arteries and skin arteries. In the lung tissue, calcifications were also seen in the connective tissue surrounding the bronchi and vasculature. The presence of extravascular calcifications was also clearly demonstrated by Von Kossa staining of the skin, revealing ossifications and elastic fiber calcifications in the dermis (Fig. 7A). Staining for MGP was found to be positive in the arteries of the different tissues, with the predominant MGP species being phosphorylated MGP (Fig. 7B). Staining for fetuin-A was positive in the vasculature (not shown), as was OPN staining. Positive OPN staining was also seen in the aortic media, in which Von Kossa staining revealed minor calcifications along the elastic lamellae (Fig. 7B).

Figure 7.

 Immunohistochemical staining. (AI, AII) Hematoxylin and eosin (HE) and Von Kossa staining of lung tissue. (AIII, AIV) HE and Von Kossa staining of skin tissue. (BI, BII) Phosphorylated matrix γ-carboxyglutamate protein (MGP) staining of lung tissue. (BIII, BIV) Von Kossa and osteopontin staining of the aorta. Original magnifications × 100 (AI, AII, AIV, BI) and × 200 (AIII, BII–IV). HE staining of lung tissue obtained at post-mortem examination of the male patient originally described by Keutel et al. revealed extensive distortion of the normal anatomy, indicating arterial calcification (AI; arrows). Specks of calcification are also seen in the connective tissue surrounding the bronchi and arteries (AII; black, arrow). The presence of extravascular calcifications was also clearly demonstrated in the dermis of the right thigh, revealing ectopic calcifications (AIII and AIV; arrow) and elastic fiber calcifications (AIV; black, dotted arrow). The presence of MGP in the tissue was clearly demonstrated by staining of phosphorylated MGP in the media of a pulmonary artery, which appeared to colocalize with vascular smooth muscle cells (BI; red, arrow). A band of tissue with the morphological appearance of cartilage was present at the luminal side of an extensively calcified artery, with the chondrocyte-like cells positively stained for phosphorylated MGP (BII; red, arrow). Von Kossa staining of the aorta revealed calcium deposition along the elastic fibers in the aortic media (BIII; black, arrows). Osteopontin staining was positive in the aortic media (BIV; brown, arrow). C, cartilage; L, lumen; M, media.


This study was performed in two patients with KS, a rare autosomal recessive disorder resulting from mutations in the gene encoding MGP, which inhibits calcification and depends on vitamin K. First, we describe a novel homozygous MGP mutation in a newly identified KS patient. Second, this is the first report on circulating MGP species in the presence of a functional mutation, and it demonstrates high levels of phosphorylated MGP and low levels of carboxylated MGP. Third, we found that vitamin K supplements did not improve the patient’s MGP carboxylation. Finally, we were able to demonstrate the presence of phosphorylated MGP in the vasculature, obtained at post-mortem examination, of the first KS patient originally described by Keutel et al. [13].

Our patient presented with exertional dyspnea refractive to bronchodilatory drugs. Laryngotracheoscopy revealed cricoid stenosis and an abnormal trachea, and CT scanning showed abnormal cartilage calcification. Other features, including brachytelephalangism and peripheral pulmonary artery stenoses, were consistent with the diagnosis of KS. Mutation analysis revealed a novel MGP mutation (c.61+1G>A). This is the fifth MGP mutation identified so far. This particular mutation results in loss of the consensus donor splice site at the exon 1–intron 1 junction. The previously identified MGP mutations were two splice site mutations and two introducing a stop codon [12,26]. Although all of these mutations are expected to coincide with absent or non-functional MGP, circulating MGP species have not been previously characterized in KS patients.

In our patient, measurements of circulating MGP revealed a high level of phosphorylated MGP and a low level of carboxylated MGP as compared with both his heterozygous first-degree relatives and to sibling without the MGP mutation. The presence of different MGP species (non-phosphorylated, phosphorylated, non-carboxylated, and carboxylated) originates from two post-translational modifications that may occur: vitamin K-dependent carboxylation and serine phosphorylation. Phosphorylation of extracellular matrix proteins is thought to be important for control of biomineralization [35]. This has been extensively studied for OPN, a phosphorylated glycoprotein that inhibits soft tissue calcification. Non-phosphorylated OPN has a greatly reduced ability to inhibit hydroxyapatite formation as well as smooth muscle cell calcification [36–38]. It has been postulated that phosphorylation results in a conformational change that increases the protein’s affinity for calcium crystals, thereby limiting their growth and proliferation [35]. Moreover, it may facilitate the recruitment and activation of macrophages that remove pathologic calcium deposits [39]. For MGP, an additional function for phosphorylation might be the regulation of MGP secretion from the cell [9]. Moreover, it has been shown in experiments with vascular smooth muscle cells that phosphorylated MGP residues (without carboxylated MGP residues) may also partly inhibit calcification [40].

The phosphoserine domain, the Gla domain and the C-terminus seem to be important for normal MGP function. The high level of phosphorylated MGP in our patient may be explained by a relatively high level of synthesis of MGP, possibly induced by the high calcification load. The carboxylation of MGP is a prerequisite for its inhibition of BMP-2, and the low circulation levels of carboxylated MGP indicate that this mechanism of MGP action is likely to be impaired. MGP acts as a local inhibitor of calcification, as has been illustrated in MGP knockout mice, the phenotype of which can be rescued if MGP expression is selectively restored locally in the vascular smooth muscle cell [41]. However, increasing the circulating MGP level did not rescue the calcifying phenotype [41]. The finding of MGP in the circulation of our patient is therefore not clear proof that MGP still has the ability to inhibit calcification in the patient. However, it does confirm that MGP is synthesized in the tissues of our patient. It is of note that circulating MGP species may reflect various aspects of calcification inhibition by MGP in the tissue.

Vitamin K deficiency, resulting from either nutritional deficit or the use of vitamin K antagonists (e.g. coumarin derivatives), will lead to in undercarboxylation of MGP and thus impair its biological function [42]. In contrast, increased availability of vitamin K for MGP in tissue may enhance its calcification-inhibitory activity and possibly reduce the development or progression of calcification in cartilage, the arterial wall, or other soft tissues. We were able to measure circulating uncarboxylated MGP levels in our patient, which have previously been found to decrease when vitamin K supplements are given to healthy subjects. We therefore aimed to assess whether vitamin K would also affect MGP levels in our patient. If so, this would be a first indication that the patient might benefit from vitamin K supplements, and further exploration of the effect on clinical symptoms would be justified. However, no clinically relevant changes in either uncarboxylated of carboxylated MGP levels were seen after 3 months of vitamin K1 supplements. The question arises of why a beneficial effect from vitamin K supplements was lacking. We can speculate that carboxylation of MGP was not possible, owing to the mutation, as only the first 14 amino acids of the mutated protein are comparable with those of the mature MGP. In our sandwich assays we captured circulating MGP with an antibody recognizing N-terminal uncarboxylated residues 3–15, with the serines not phosphorylated. However, it is questionable whether MGP and carboxylated MGP can be quantified, as the region detected with the uncarboxylated MGP and carboxylated MGP antibodies is no longer present in the mutated protein. These data regarding mRNA analysis became available after the MGP measurements were completed. It should be noted that the detected uncarboxylated MGP and carboxylated MGP levels in our patient were around the lower detection limit of the assays, and that mainly phosphorylated MGP was measured. As the increased availability of vitamin K did not measurably affect MGP, it is unlikely that vitamin K supplements would improve the clinical symptoms or the long-term outcome in this particular KS patient.

Additional phenotypic characterization of our patient included a thorough vascular evaluation, which revealed no signs of arterial calcification. In contrast to the MGP knockout mice, which displayed severe arterial calcifications soon after birth, no clinical signs of vascular calcification have been reported in most KS patients. In 2001, Meier et al. reported on the follow-up of the siblings originally described by Keutel et al. [29]. The youngest sibling died at age 38 years from right heart failure, and a post-mortem revealed concentric calcification in several vascular beds, including the coronary, hepatic, renal, meningeal and cerebral arteries. This is the only KS patient in whom extensive arterial calcification has been described. Tissues were available from this patient for immunohistochemical staining, although no tissues were available from our own patient. We were able to confirm the finding of Meier et al. of concentric arterial calcification of several vascular beds. We also demonstrated the presence of MGP, predominantly in its phosphorylated form, in the vasculature of this patient. The finding of elastic fiber calcification in the dermis is in accordance with a report from Nanda et al. [43], who demonstrated fragmentation of mid-dermal elastic fibers in biopsy specimens of normal-looking skin from two patients with KS. These results suggest a possible role for MGP in preventing elastic fiber calcification in the skin.

The difference in the arterial phenotype between the MGP knockout mouse model and human KS patients is intriguing. It could be argued that the apparent absence of arterial calcifications in our patient is attributable to insufficient sensitivity of multislice CT scanning in detecting early calcifications. It has been shown in vascular tissue from young dialysis patients that apoptosis and damage of vascular smooth muscle cells, accompanied by increased tissue levels of uncarboxylated MGP, precede clinically overt calcification and even positive Von Kossa tissue staining for calcification [44]. Irrespective of this, the KS cases described so far may represent mutations accompanied by some residual calcification-inhibitory activity of MGP, which might be sufficient to prevent the development of arterial calcification. In both of the patients presented here, we found strong indications of the presence of phosphorylated MGP. Even if this is not the entire mature MGP, but fragments of MGP containing phosphoserine residues, binding of calcium crystals by MGP might still be possible.

In the MGP knockout mice, Speer et al. [45] demonstrated that OPN expression was upregulated and that it accumulated in the calcified arterial wall. In addition, mice deficient in both MGP and OPN showed twice as much calcification at the age of 2 weeks as MGP knockout mice, and they died approximately 2 weeks earlier, owing to vascular rupture [45]. It may therefore be hypothesized that other calcification inhibitors are upregulated in our patient that might prevent the development of arterial calcification. Our study is the first to confirm the hypothesis of Kaartinen et al. [46] hypothesis that KS patients lacking functional MGP have arteries containing mineral-inhibitory OPN, possibly allowing for the extended survival of KS patients as compared with MGP knockout mice. Patients with KS may not develop overt arterial calcification unless other risk factors are present: either the loss of another inhibitor, or the presence of a promoter of arterial calcification. These initiators might be absent in the relatively young KS patients described in the literature so far (mean age: 8.3 years), which is supported by the absence of laboratory abnormalities in 19 patients. In contrast, several potential triggers were present in the only KS patient with extensive arterial calcifications described so far (chronic bronchitis, and chemotherapy for recurrent malignancy) [29]. This patient was also older at the time of post-mortem examination than the patients described thus far (38 years).

In conclusion, we characterized the phenotype and circulating MGP species of a patient with KS, in whom we had identified a novel MGP mutation. In addition, we characterized MGP species in the tissues of the first KS patient originally described by Keutel et al. Phosphorylated MGP was found to be present in both patients, indicating that some residual MGP activity might be present. Obviously, further investigation of the function of MGP phosphorylation is needed. Vitamin K supplements did not affect MGP activity in our patient, but might have a beneficial effect for patients in whom carboxylation of MGP is possible. The investigation of KS patients’ phenotype and their circulating MGP species contributes to our understanding of MGP functions in various tissues, and helps to elucidate the difference in arterial phenotype between MGP-deficient mice and humans.


E. C. M. Cranenburg and K. Y. van Spaendonck-Zwarts: designed and performed the research, collected, analyzed and interpreted the data, and wrote the manuscript; L. Bonafe, L. Mittaz Crettol, and A. Superti-Furga: performed the genetic analysis; A. J van Essen, L. A. Rödiger, and F.G. Dikkers: performed diagnostic examinations; E. Alexandrakis: performed immunohistochemical analysis; C. Vermeer and L. J. Schurgers: performed MGP analysis; G. D. Laverman: performed diagnostic examinations and supervised data analysis and writing of the manuscript.


We thank the patient and his family members for participating in this study. We thank E. Magdeleyns and M. Herfs (research technicians of VitaK, Maastricht University) for excellent technical assistance with MGP measurements and immunohistochemical staining. We thank C. Chiesa Buzzi for excellent technical assistance with RT-PCR analysis. We thank J. Senior (science editor of the Department of Genetics, University Medical Center, University of Groningen) for providing editorial assistance to the authors during the preparation of this manuscript.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.