Article Text

Download PDFPDF

Original research
Telangiectasia-ectodermal dysplasia-brachydactyly-cardiac anomaly syndrome is caused by de novo mutations in protein kinase D1
  1. Svenja Alter1,
  2. Andreas David Zimmer1,
  3. Misun Park2,
  4. Jianli Gong2,
  5. Almuth Caliebe3,
  6. Regina Fölster-Holst4,
  7. Antonio Torrelo5,
  8. Isabel Colmenero6,
  9. Susan F Steinberg2,
  10. Judith Fischer1
  1. 1 Institute of Human Genetics, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany
  2. 2 Department of Pharmacology, Columbia University, New York, New York, USA
  3. 3 Institute of Human Genetics, Christian-Albrechts University Kiel & University Hospital Schleswig-Holstein, Kiel, Germany
  4. 4 Department of Dermatology, Christian-Albrechts University Kiel & University Hospital Schleswig-Holstein, Kiel, Germany
  5. 5 Department of Dermatology, Hospital Infantil Universitario Niño Jesús, Madrid, Spain
  6. 6 Department of Pathology, Hospital Infantil Universitario Niño Jesús, Madrid, Spain
  1. Correspondence to Professor Dr Judith Fischer, Institute for Human Genetics, Medical Center – University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany; judith.fischer{at}; Professor Susan F Steinberg, Department of Pharmacology, Columbia University, New York, New York, USA; sfs1{at}


Background We describe two unrelated patients who display similar clinical features including telangiectasia, ectodermal dysplasia, brachydactyly and congenital heart disease.

Methods We performed trio whole exome sequencing and functional analysis using in vitro kinase assays with recombinant proteins.

Results We identified two different de novo mutations in protein kinase D1 (PRKD1, NM_002742.2): c.1774G>C, p.(Gly592Arg) and c.1808G>A, p.(Arg603His), one in each patient. PRKD1 (PKD1, HGNC:9407) encodes a kinase that is a member of the protein kinase D (PKD) family of serine/threonine protein kinases involved in diverse cellular processes such as cell differentiation and proliferation and cell migration as well as vesicle transport and angiogenesis. Functional analysis using in vitro kinase assays with recombinant proteins showed that the mutation c.1808G>A, p.(Arg603His) represents a gain-of-function mutation encoding an enzyme with a constitutive, lipid-independent catalytic activity. The mutation c.1774G>C, p.(Gly592Arg) in contrast shows a defect in substrate phosphorylation representing a loss-of-function mutation.

Conclusion The present cases represent a syndrome, which associates symptoms from several different organ systems: skin, teeth, bones and heart, caused by heterozygous de novo mutations in PRKD1 and expands the clinical spectrum of PRKD1 mutations, which have hitherto been linked to syndromic congenital heart disease and limb abnormalities.

  • protein kinase D1 (PRKD1)
  • telangiectasia
  • ectodermal dysplasia
  • brachydactyly
  • cardiac anomaly

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Additional information is available from the corresponding authors on reasonable request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The PRKD1 (HGNC:9407) gene encodes the protein kinase D1 protein, the founding member of a family of three structurally related stress-activated enzymes that regulate a large number of fundamental cellular processes involved in cell differentiation and proliferation,1 cellular apoptosis,2 immune regulation,3 epithelial-to-mesenchymal cell transition,4–6 cardiac contractility and cardiac hypertrophy,7 angiogenesis8 and the pathogenesis of certain cancers.9 10 Moreover, PKD1 is involved in actin remodelling in migrating tumour cells by regulating cofilin activity through phosphorylation of slingshot (SSH) phosphatase SSH1L.11 (Note: with regard to nomenclature, PRKD1 should not be confused with the gene PKD1 (HGNC:9008), that is associated with polycystic kidney disease. This manuscript uses PRKD1 when referring to the gene and (to conform with conventions in the biochemical literature) PKD1 when referring to the protein enzyme.)

PKD isoforms share a common modular domain structure consisting of a C-terminal kinase domain and an N-terminal regulatory domain consisting of tandem C1A/C1B motifs that anchor full-length PKD to diacylglycerol-containing/phorbol ester-containing membranes and a pleckstrin homology (PH) motif that participates in intramolecular autoinhibitory interactions that regulate enzyme activity12 (figure 1). PKD1 activation is generally attributed to growth factor-dependent mechanisms that promote diacylglycerol accumulation and thereby co-localise PKD1 with allosterically activated novel protein kinase C (nPKC) isoforms at lipid membranes. nPKCs then trans-phosphorylate PKD1 at conserved serine residues (Ser744 and Ser748) in the activation loop.13 Activated PKD1 autophosphorylates at Ser910 in a PKD consensus phosphorylation motif in a PDZ domain-binding motif at the extreme C-terminus. While PKD1-Ser910 autophosphorylation was initially viewed as an obligatory autocatalytic reaction that faithfully reflects PKD1 activity in cells, it has subsequently become evident that this autocatalytic reaction can occur without an associated increase in PKD1 activity towards heterologous substrates—that Ser910 autophosphorylation does not necessarily provide a valid surrogate readout of PKD1 activity in cells.14

Figure 1

Schematic representation of protein kinase D1 (PKD1) (HGNC:9407). Schematic representation of PKD1 protein domains and de novo mutations identified so far (asterisks). The mutations p.(Gly592Arg) and p.(Arg603His) were identified in the patients described in this paper (bold).

While PKD1 is a ubiquitous kinase15 and early studies reported that homozygous PRKD1 knockout in mice is embryonic lethal, there has been considerable interest in PKD1’s specific role in cardiac development. Shaheen et al studied consanguineous families as a strategy to expose novel recessive congenital heart disease genes16 and showed that a homozygous truncating mutation in PRKD1 leads to the generation of a catalytically inactive protein (that contains the entire N-terminal regulatory domain but only the first 35 residues at the N-terminus of the kinase domain) in patients with truncus arteriosus.16 A separate study from Sifrim et al identified heterozygous de novo missense mutations in PRKD1 in three patients with syndromic congenital heart disease associated with developmental delay as well as mild skin and limb abnormalities.17 Two of the patients had identical kinase domain p.(Gly592Arg) missense mutations and presented with pulmonic stenosis (P1) or atrioventricular septal defects (P3), whereas the second patient (P2) had a missense mutation at p.(Leu299Trp) (a residue in the regulatory C1B domain) and presented with atrioventricular septal defects (figure 1). These genetic studies in humans provide compelling evidence that PRKD1 plays also a role in the pathogenesis of certain congenital cardiac disorders.

Materials and methods

Informed consent

Informed consent was obtained from all individual participants included in the study.

Exome sequencing

Whole-exome sequencing (2×150 bp) was performed using the Nextera Rapid Capture Exome Kit (Illumina) and a MiSeq System (Illumina). Reads were aligned against the human genome sequence (hg19) using bwa18 (0.7.5a) and genotype calling was performed with GATK19 (3.8 nightly-2017-12-06-1). Variants were annotated using ENSEMBL Variant Effect Predictor20 (V.88) and Gemini21 (V.0.20.1) was used to analyse and filter the variants. All six exomes have a mean sequencing depth >47.9x in coding exons (Ensembl Release 87, GRCh37) and at least 96.0% of the coding regions were covered (>1 read).


PKD-Ser(P)916 (catalogue number #2051), PKD1 (catalogue number #90039), CREB-Ser(P)133 (catalogue number #9198) and CREB (catalogue number #9197) antibodies from Cell Signalling Technology were used. Ser916 in rodent PKD1 corresponds to Ser910 in the human PKD1 sequence. HA-agarose from Roche Applied Science was used. Phorbol 12-myristate 13-acetate (PMA) from Sigma and a recombinant human CREB-maltose binding protein fusion construct from Biosource were used. All other chemicals were reagent grade.

Plasmids and HEK293 cell culture

PKD1 constructs harbouring single residue substitutions were generated using the QuikChange mutagenesis system (Agilent Technologies) and then validated by sequencing. Expression vectors were introduced into HEK293 cells (maintained in Dulbecco’s Modified Eagle's Medium) with 10% fetal bovine serum) using the Effectene transfection reagent (Qiagen) according to the manufacturer’s instructions. After 24 hours, cells were lysed in RIPA buffer containing 1 mM sodium orthovanadate, 10 µg/mL aprotinin, 10 µg/mL leupeptin, 10 µg/mL benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 5 µM pepstatin A and 0.1 µM calyculin. Cell lysates were used for immunoblotting or in vitro kinase assays followed by immunoblot analysis according to standard methods as described previously.14

In vitro kinase assays

In vitro kinase assays (IVKAs) were performed according to methods described previously.14 Assays were performed with PKD1 immunoprecipitated from 150 µg of starting cell extract in 110 µL of a reaction buffer containing 30 mM Tris-Cl, pH 7.5, 5.45 mM MgCl2, 0.65 mM EDTA, 0.65 mM EGTA, 0.1 mM dithiothreitol, 1.09 mM sodium orthovanadate, 0.1 µM calyculin, 0.55 µM protein kinase inhibitor, 217 mM NaCl, 3.6% glycerol, 66 µM ATP and 1 µg recombinant human CREB-maltose binding protein fusion construct. Incubations were for 30 min at 30°C in the absence or presence of 89 µg/mL phosphatidylserine (PS) and 175 nM PMA as indicated.


We identified de novo mutations in PRKD1 by whole-exome sequencing in two unrelated patients who showed similar clinical features including generalised telangiectasia, ectodermal dysplasia, brachydactyly and congenital heart defects (online supplementary table S1). Patient A was a woman aged 19 years, with a history of widespread, persistent skin erythema that had already been present at birth. She had an atrial septal defect with a spontaneous closure at 7 years, a pulmonary stenosis and a mild tricuspid insufficiency. At the age of 6, medullary nephrocalcinosis had been detected (Cacchi-Ricci syndrome/medullary sponge kidney), with hypocalcaemia, hypomagnesaemia, hypercalciuria and hypocitraturia. On examination, the height and weight were normal for her age, and her mental development was normal. Her nasal bridge was broad, and showed mild maxillary and mandibular hypoplasia and high palate. The skin was covered with diffuse erythema due to millions of telangiectasias that were present over the entire skin surface and lips (figure 2). She had a mild hypotrichosis, with retraction of the frontal line and sparse, curly hair. Eyebrows and eyelashes were sparse. She had a reduced number of permanent teeth that were carious, which might had been caused by enamel defects (figure 2). The patient also had brachydactyly with shortened thumbs and metacarpals, especially of the third finger (figure 2). Other abnormal features on examination were sternal deformity, scoliosis and joint hypermobility, with recurrent patellar luxation. Laboratory studies disclosed autoimmune thyroiditis and slightly elevated parathyroid hormone levels. Imaging revealed an arteriovenous fistula in the pulmonary lingual lobe. A skin biopsy showed numerous telangiectatic capillary vessels in the papillary dermis (figure 2).

Supplemental material

Figure 2

Clinical presentation of patient A. The whole body surface is covered by millions of tiny telangiectasias (A, B, C, G). Loss of teeth, reduced number of permanent teeth with abnormal enamel and premature caries (D). Brachydactyly with different degrees of brachymesophalangy, brachytelephalangy and brachymetacarpia. Notably, the index finger is least affected due to preserved metacarpal length (E, F). Dilatation of capillaries in the upper dermis, without changes in mid and deep dermis (H&E, original magnification 40x) (H).

Patient B was born at 37 weeks with normal measurements. Early development was normal. He walked at the age of 14/15 months and spoke his first words at the age of 12 months. At the age of 2 years, he started to develop recurrent middle ear infections. He was affected by bilateral cholesteatoma, which resulted in hearing loss. At the age of 3 5/12 years haemodynamically irrelevant pulmonary valve stenosis was diagnosed. At the age of 3 years, deciduous teeth were lost after minor trauma. At the age of 5 years, he had already lost five teeth. At that time, he had sparse hair, which had never been cut and sparse eye brows. On physical examination, mild frontal bossing and pectus excavatum were noticed. At the age of 8 years, shortening of hands and feet was noted. X-ray imaging showed multiple abnormalities including shortening of the third-fifth metacarpal and cone epiphysis. Moreover, he showed a partial syndactyly of toes two and three. Bone age was advanced. Clinically, the fourth and fifth metatarsal was shortened as well. Metabolic investigations of pseudopseudohypoparathyroidism were normal. Due to dry skin, missing permanent teeth and poor hair growth, genetic testing for ectodermal dysplasia and trichorhinophalangeal syndrome were performed, which yielded normal results. At the age of 14 years, he developed widespread telangiectasia (figure 3).

Figure 3

Clinical presentation of patient B. Clinical presentation of patient B at the age of 8 years (A, C, E, F) and at the age of 18 years (B, H–J). On physical examination (at age 8 years), pectus excavatum and genua vara were noticed (A). At the age of 8 years, brachydactyly type E and short thumbs with short first toe (B–F; H–J). X-ray imaging (C) showed multiple abnormalities including brachymetacarpalia 1 and 3–5, dysplastic metaphyses and epiphyses of metacarpalia 3–5, brachytelephalangy, cone epiphyses (MC 1, MP 2 and 5), brachymesophalangy 2 and 5. Notably, the index finger is least affected due to preserved metacarpal length. He presents partial syndactyly of toes two and three (F, I). Reduced number of permanent teeth (G) loss of teeth after minor trauma due to short roots. At the age of 15 years he developed widespread multiple pinhead sized telangiectasia, partly confluent to angioma like small papules, the whole integument is involved (J). MC, metacarpal; MP, metaphyses.

Trio whole-exome sequencing analyses of the patients and their parents showed that both patients carry a heterozygous de novo mutation in PRKD1 (NM_002742.2). Patient A carries the mutation c.1808G>A, p.(Arg603His) in exon 13 and patient B carries the mutation c.1774G>C, p.(Gly592Arg) in exon 12 of the PRKD1 gene. The genotypes of both patients and their parents were validated by Sanger sequencing (online supplementary figure S1). While c.1808G>A is recorded in the database dbSNP (rs776034417) (build 152)22 and in the ExAC database with one heterozygous count (allele frequency: 0.0008 %) (0.3.1),23 c.1774G>C could not be found in ExAC database (0.3.1)23 19, gnomAD (V.2.1.1),24 dbSNP (build 152)22 or HGMD professional (2019.1).25 InterVar26 (queried 23 May 2019) classified c.1808G>A as ‘uncertain significance’ and c.1774G>C as ‘likely pathogenic’ (online supplementary table S2). Very interestingly, a single nucleotide exchange at the same position (c.1774G>A) resulting in the same amino acid exchange (p.(Gly592Arg)) has been described in two patients with syndromic congenital heart defects.27 Four out of six mutations in PRKD1 described thus far are located within the catalytic domain of PRKD1 and two of these result in an identical amino acid exchange (p.(Gly592Arg)). Strikingly, glycine 592 is the second glycine residue in the highly conserved glycine-rich loop (or G-loop, GxGxxG) in the kinase core. The G-loop plays a critical role to bind and orient ATP in a position that is optimal for catalysis.28 ,29 While all three G-loop glycine residues are highly conserved across protein kinase family members, substitutions at the first and third glycine residue are tolerated in some eukaryotic protein kinases. In contrast, the second glycine residue in the G-loop is conserved in more than 99% of eukaryotic protein kinases28 (online supplementary figure S2). Modelling studies based on structural data suggest that other amino acids at this position (as found in the p.(Gly592Arg) mutant) would result in steric clash and disrupt protein kinase activity.28 The functional consequences of the kinase domain p.(Arg603His) mutation identified in patient A (at a position that is not conserved across protein kinase family members and is not directly involved in ATP-binding or catalysis) is less predictable.

Supplemental material

Supplemental material

Supplemental material

We used a cell-based approach with HEK293 cells that heterologously overexpress HA-tagged WT-PKD1 and PKD1-Arg603His as an initial strategy to examine whether an Arg603His substitution influences PKD1 activity. WT-PKD1 is a lipid-dependent enzyme that autophosphorylates at Ser910 and phosphorylates CREB at Ser133 only in the presence of PS/PMA (figure 4A). WT-PKD1 is recovered with little-to-no Ser910 autophosphorylation or phosphorylation of CREB (a known downstream PKD1 substrate) under resting conditions; PKD1-Ser910 autophosphorylation and CREB-Ser133 phosphorylation increase following treatment with PMA (a pharmacological activator of PKD). In contrast, PKD1-Arg603His is a lipid-independent enzyme that displays similarly high levels of Ser910 autophosphorylation and CREB-Ser133 phosphorylation in resting and PMA-treated cells. We show that WT-PKD1 and PKD1-Arg603His enzymes display similar activity profiles when used in IVKAs, whereas PKD1-Arg603His displays high levels of constitutive/lipid-independent autocatalytic and CREB kinase activity (figure 4B).

Figure 4

Protein kinase D1 (PKD1) mutants display altered catalytic activity. Panel A: Left: control uninfected HEK293 cells or cells that heterologously overexpress WT-PKD1 or PKD1-R603H were treated with vehicle or phorbol 12-myristate 13-acetate (PMA) (200 nM, 30 min) and then subjected to immunoblot analysis for PKD1 protein expression, PKD1-Ser910 autophosphorylation and CREB-Ser133 phosphorylation. Quantification of signals of PKD1-Ser910 autophosphorylation and CREB-Ser133 phosphorylation of uninfected cells, WT-PKD1 or PKD1-R603H expressing cells are shown in the middle. Right: WT-PKD1 and PKD1-R603H were subjected to in vitro immunocomplex kinase assays (IVKAs) in the absence or presence of phosphatidylserine (PS)/PMA and immunoblot analysis was used to track PKD1-Ser916 autophosphorylation as well as PKD1 phosphorylation of recombinant CREB (added as a heterologous substrate). All results were replicated in two separate experiments. Panel B: WT-PKD1, PKD1-G592R and PKD1-K612W (kinase-dead) enzymes heterologously overexpressed in HEK293 cells were subjected to IVKAs in the absence or presence of PS/PMA and immunoblot analysis was used to track PKD1-Ser910 autophosphorylation as well as PKD1 phosphorylation of recombinant CREB (added as a heterologous substrate). A representative experiment is depicted on the left and autophosphorylation at PKD-S910 (n=6) and PKD1-dependent phosphorylation of CREB at S133 (n=3) is quantified on the right; in each case values are normalised to the activity of WT-PKD1.

We used a similar IVKA approach to examine the functional consequences of the Gly592Arg substitution. We show that WT-PKD1, PKD1-Gly592Arg and PKD1-Lys612Trp (a kinase-dead mutant included as a control in these experiments) are recovered from HEK293 cells with similar low levels of anti-PKD1-pSer910 immunoreactivity, which reflects either a low level of autophosphorylation or more likely the fact that the anti-PKD1-pSer910 phospho-site specific antibody is not entirely specific—it recognises the non-phosphorylated Ser910 epitope weakly and produces some signal at high levels of transgene expression (figure 4B). WT-PKD1 displays lipid-dependent Ser910 autocatalytic activity and CREB-Ser133 kinase activity, whereas PKD1-Lys612Trp is catalytically inactive. Surprisingly, the PKD1-Gly592Arg mutant (which also had been predicted to be catalytically inactive) retains a considerable amount of lipid-dependent autocatalytic activity, but only a low level of CREB-Ser133 kinase activity (ie, it effectively autophosphorylates, but shows little activity towards a heterologous substrate).


We present two patients, each carrying one different de novo mutation in PRKD1, with a similar clinical phenotype. One mutation has been described in two patients with syndromic congenital heart defects.27 The phenotype of these patients (patient 1 and 3 in Sifrim et al 17) comprised a congenital pulmonary valve abnormality and an atrioventricular septal defect, respectively. Moreover, they both had mild skin changes (including dry and thin skin in patients 1 and 3, fragile nails in patient 3), sparse scalp hair (patients 1 and 3) and dental abnormalities (premature loss of primary teeth in patient 1 and small, widely spaced teeth in patient 3). In addition, they both have been described to have some abnormalities regarding fingers and toes (broad thumb and short digit for patient 1 and syndactyly for patient 3).27 Our two patients have novel additional features, such as the diffuse skin telangiectasia, and bone anomalies (mainly brachydactyly due to metacarpal and metatarsal defects). Remarkably, the brachydactyly of our patients’ hands presents very similar, whereas their feet are quite different. Teleangiectasia differ from features reported in patients 1 and 3 by Sifrim et al 17. As the skeletal features in the two patients described herein became obvious around the age of 8 years and the patients described by Sifrim et al 17 were younger, the skeletal phenotype may evolve in them as well.

The implicit assumption in previous studies (that did not perform biochemical characterisations) was that a disease-causing de novo PRKD1 mutation would be inactivating, an assumption based at least in part on literature showing that homozygous PRKD1 knockout in mice is embryonic lethal.30 However, these studies in knockout mouse models have to be interpreted with caution. First, based on recent evidence that PKD1 forms homodimers or heterodimers with other PKD family members (PKD2 or PKD3),31 a targeted PRKD1 deletion might be predicted to impact the expression, activation and or localisation of PKD2 or PKD3. This in turn is predicted to disrupt the normal coordinate actions of these PKD isoforms in the control of normal cardiovascular development and as drivers of pathophysiological responses to cardiac stress. Second, there is growing body of evidence that PKDs localise to multiprotein complexes32–35 and that they can play additional more complex kinase-independent roles in cellular signalling responses.36 37 Hence, the pathophysiological consequences of a gene knockout (loss of protein and activity) versus the loss of activity with preserved either truncated or full-length protein might be predicted to differ.

The observation that the two patients in this study with mutations that result in very distinct enzymologies (either activating Arg603His or principally functionally inactivating Gly592Arg de novo PRKD1 mutations) displayed very similar clinical symptoms was surprising. However, it is worth noting that a similar dichotomy has been identified in studies of the mechanisms driving the pathogenesis of ‘RASopathies’, where patients with Noonan syndrome (NS) and LEOPARD syndrome (LS) present with similar cardiac abnormalities but harbour either gain-of-function mutations (NS) or loss-of-function mutations (LS) in the PTPN11 gene (reviewed in Lauriol and Kontaridis38). A mechanism to explain the similar congenital disorders identified in patients with either activating or inactivation PKD1 mutations is uncertain. In theory, the constitutively active PKD1-Arg603His enzyme could become trapped in non-productive complexes with other PKD isoforms and thereby prevent transduction of receptor-dependent PKD1 signalling responses, but other mechanisms are possible and deserve further study.

Finally, the precise mechanisms that contribute to the syndromic phenotypes described in the few known patients with de novo PRKD1 mutations are not obvious. While some insights can be gleaned from studies in reductionist cell-based systems or more physiologically relevant genetic models in mice where PKD1 has been implicated in diverse fundamental biological processes (including cell migration, proliferation and differentiation) and PKD1 has emerged as a key regulator of the complex/coordinate events required for normal angiogenesis, immune regulation, cardiovascular development39 bone remodelling and skeletal development,40 41 the precise role of PKD1 in the human developmental phenotypes identified in our patients remains more elusive. The considerable complexities in PKD1-driven biology and the strong interest in PKD1 as a target for the therapy of certain clinical disorders provide a strong rationale for future studies that elucidate the role of PKD1 in normal human development.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information. Additional information is available from the corresponding authors on reasonable request.



  • SFS and JF are joint senior authors.

  • Contributors SA, JF and SFS conceived the project. SA, ADZ, SFS and JF wrote the manuscript. SA, AC, RF-H, AT, IC, SFS and JF contributed to the acquisition and interpretation of data. SA, ADZ, MP, JG and SFS contributed to the analysis of data. All authors contributed to the critical revision of the manuscript and approved the manuscript.

  • Funding This work is supported by NIH, NHLBI grant HL112388, and by the faculty of Medicine of the University of Freiburg.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.