Somatic mutations in ATP1A1 and ATP2B3 lead to aldosterone-producing
adenomas and secondary hypertension
Felix Beuschlein1*, Sheerazed Boulkroun2,3, Andrea Osswald1, Thomas Wieland4, Hang N.
Nielsen5, Urs D. Lichtenauer1, David Penton6, Vivien R. Schack5, Laurence Amar2,3,7, Evelyn
Fischer1, Anett Walther4, Philipp Tauber6, Thomas Schwarzmayr4, Susanne Diener4, Elisabeth
Graf4, Bruno Allolio8, Benoit Samson-Couterie2,3, Arndt Benecke9, Marcus Quinkler10,
Francesco Fallo11, Pierre-Francois Plouin2,3,7, Franco Mantero12, Thomas Meitinger4,13,14,
Paolo Mulatero15, Xavier Jeunemaitre2,3,7, Richard Warth6, Bente Vilsen5, Maria-Christina
Zennaro2,3,7$, Tim M. Strom4,13$, and Martin Reincke1*
Affiliations:
1
Medizinische Klinik and Poliklinik IV, Ludwig-Maximilians-Universität München, Germany
INSERM, UMRS_970, Paris Cardiovascular Research Center, Paris, France
3
University Paris Descartes, Paris Cité Sorbonne, Faculty of Medicine, Paris, France
4
Institute of Human Genetics, Helmholtz Zentrum München, Neuherberg, Germany
5
Department of Biomedicine, Aarhus University, Aarhus C, Denmark
6
Medizinische Zellbiologie, Universität Regensburg, Germany
7
Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris, France
8
Department of Medicine I, Endocrine and Diabetes Unit, University Hospital Würzburg, Germany
9
Institut des Hautes Etudes Scientifiques, & Centre National de la Recherche Scientifique, Bures sur Yvette,
France
10
Clinical Endocrinology, Campus Mitte, University Hospital Charité, Berlin, Germany
11
Dept of Medicine, University of Padova, Italy
12
Endocrine Unit, Dept. of Medicine, University of Padova, Italy
13
Institute of Human Genetics, Technische Universität München, Munich, Germany
14
Munich Heart Allicance, München, Germany
15
Division of Internal Medicine and Hypertension, Department of Medical Sciences, University of Torino, Italy
2
$
equal contribution
*
Corresponding authors:
Martin Reincke, M.D.
Felix Beuschlein , M.D.
Medizinische Klinik und Poliklinik IV
Klinikum der Universität München
Ziemssenstr. 1
D-80336 Munich
Germany
p: xx49 (0)89 5160 2110
f: xx49 (0)89 5160 4467
e: martin.reincke@med.uni-muenchen.de
e: felix.beuschlein@med.uni-muenchen.de
Beuschlein et al.
1
INTRODUCTORY PARAGRAPH
2
Primary aldosteronism is the most prevalent form of secondary hypertension. To explore
3
molecular mechanisms of autonomous aldosterone secretion we performed exome sequencing
4
of aldosterone producing adenomas (APA). Somatic hot spot mutations in Na+,K+-ATPase
5
(ATP1A1) and Ca2+-ATPase (ATP2B3) genes were identified in 3/9 and 2/9 patients,
6
respectively. These ATPases are expressed in adrenal cells and control Na+, K+ and Ca2+
7
homeostasis. Loss of pump activity and a strongly reduced affinity for K+ was demonstrated
8
by functional in vitro experiments for ATP1A1 mutants. Electrophysiological ex vivo studies
9
on primary adrenal adenoma cells further provided evidence for inappropriate depolarization
10
of cells with ATPase mutations. In a large collection of 309 APA samples, 16 (5.2%) somatic
11
mutations were found in ATP1A1 and 5 (1.6%) in ATP2B3. Mutation carriers displayed male
12
dominance and increased plasma aldosterone and lower potassium compared with non-
13
carriers. In summary, dominant somatic mutations in two members of the ATPase gene
14
family converge in autonomous aldosterone secretion.
15
Beuschlein et al.
1
MAIN TEXT
2
Excessive autonomous aldosterone secretion by the adrenal gland, so called primary
3
aldosteronism causes drug-resistant and often life threatening arterial hypertension
4
accompanied by severe hypokalemia. Long-term consequences include increased risk for
5
stroke, myocardial infarction and atrial fibrillation. Primary aldosteronism is present in up to
6
7% of hypertensives in population based studies 1 and in up to 20% in patients with therapy-
7
resistance referred to specialized centers 2. Primary aldosteronism can be caused by bilateral
8
adrenal hyperplasia or a unilateral aldosterone producing adenoma (APA). Depending on the
9
patient population and applied diagnostic procedures, the proportion of primary aldosteronism
10
patients with APA can be as high as 60% 3. Recent reports identified mutations in the
11
potassium channel KCNJ5 as cause of familial and sporadic forms of primary aldosteronism
12
and estimated the proportion of APAs caused by those mutations at 30-40% 4,5.
13
To identify further genetic determinants of primary aldosteronism, we performed
14
exome sequencing in tumor and matched control tissue of 9 male patients affected by
15
hypokalemic primary aldosteronism without somatic KCNJ5 mutations (Supplementary Table
16
1). Sequencing revealed a low number of protein altering mutations (0-13 per adenoma;
17
mean: 4.1±1.4; Supplementary Table 2). Remarkably, within this small set of genetic hits, we
18
found multiple somatic variants in two members of the P-type ATPase gene family, ATP1A1
19
(Na+,K+-ATPase) and ATP2B3 (plasma membrane Ca2+-ATPase). ATP1A1 (NM_000701.7)
20
missense variants were present in 3 out of 9 adenomas (ATP1A1c.311T>G in two cases and
21
ATP1A1c.995T>G in one case) leading to a Leu104Arg and a Val332Gly substitution,
22
respectively. ATP2B3 (NM_021949.3) in frame deletions were present in 2 adenomas
23
(ATP2B3 c.1272_1277delGCTGGT and ATP2B3 c.1273_1278delCTGGTC) in both
24
instances resulting in deletion of 2 amino acids at position 425 and 426 (p.Leu425_Val426del;
Beuschlein et al.
1
Figure 1). Sequence comparison indicated the affected amino acids as highly conserved
2
among species as well as among different members of the P-type ATPase family
3
(Supplementary Figure 1).
4
We next sequenced the entire coding region of ATP1A1 and ATP2B3 in additional 100
5
adenomas (Supplementary Table 3) and identified 6 and 2 further somatic mutations in
6
ATP1A1 and ATP2B3, respectively. Four of the ATP1A1 mutations occurred at the same
7
position (c.311T>G) and two led to an in-frame deletion of 5 amino acids overlapping
8
position 311 (c.299_313delTCTCAATGTTACTGT, p.Phe100_Leu104del). Two adenomas
9
had an in-frame deletion of 2 amino acids in ATP2B3 overlapping with the two other
10
deletions (c.1277_1282delTCGTGG, p.Val426_Val427del). Targeted sequencing of the six
11
affected genomic positions in 199 additional adenomas revealed 7 and 1 further somatic
12
mutations in ATP1A1 and ATP2B3, respectively. The complete collection of APA samples
13
(n=308), thus, contained 21 (6.8%) ATPase mutations and 118 (38.3%) KCNJ5 mutations
14
(Supplementary Table 4). Concomitant KCNJ5 and ATPase mutations within the same tumor
15
were not observed.
16
None of the six different mutations were present in 1600 in-house exomes or the 1000
17
Genomes dataset. In addition, the two missense mutations were not present in the Exome
18
Variant Server dataset (http://evs.gs.washington.edu/EVS [v.0.0.14]). Although one of the
19
ATP1A1 mutations, c.311T>G, is listed in dbSNP, it is only described as a non-validated
20
candidate SNP derived from EST data and thus is rather unlikely to represent a germline
21
mutation. Exome and Sanger sequencing of the somatic mutations revealed that both the
22
reference and alternative alleles were present in tumor tissue. This observation is consistent
23
with a heterozygous state in case of the ATP1A1 mutations and the female carriers of an
24
ATP2B3 mutation. However, since ATP2B3 is located on the X-chromosome, these findings
25
suggest a polyclonal tumor composition in case of the 2 male carriers of an ATP2B3 mutation.
Beuschlein et al.
1
In fact, polyclonal composition has well been recognized as a feature of small adrenal
2
adenomas
3
normal adrenal within our group of APA patients. Of note, no germline ATPase mutation was
4
found in a cohort of 18 patients with familial aldosteronism type 2 8 or in 91 sporadic cases
5
with bilateral adrenal hyperplasia. In the normal adrenal, ATP1A1 was highly expressed in
6
the zona glomerulosa and to a lesser extent in zona fasciculata, while ATP2B3
7
immunoreactivity was similarly detectable in all three layers of the adrenal cortex. ATP1A1
8
and ATP2B3 expression at mRNA and protein levels were similar in APAs compared to the
9
normal adrenal. Similarly, within the APA group, ATP1A1 and also ATP2B3 mRNA levels
10
6,7
. No ATP1A1 and ATP2B3 mutations were found in the germline or adjacent
were found to be independent of the tumor’s mutational status (Supplementary Figure 2).
11
12
Function and structure of the Na+,K+-ATPase encoded by ATP1A1 has been unraveled
13
in exquisite detail over the last decades 9. It exchanges 3 cytoplasmic Na+ for 2 extracellular
14
K+ ions for each ATP being hydrolyzed
15
fluxes that generate the resting membrane potential and action potential. Through site-directed
16
mutagenesis and in vitro assays, individual domains within the Na+,K+-ATPase protein have
17
been associated with specific functional properties 11. Interestingly, there is good evidence for
18
a direct link between Na+,K+-ATPase and regulation of aldosterone secretion: Na+,K+-ATPase
19
blockage with the specific antagonist ouabain results in dose dependent stimulation of
20
aldosterone release from glomerulosa cells
21
Furthermore, angiotensin II decreases the Na+,K+-ATPase activity indicating its potential
22
contribution to angiotensin dependent aldosterone release
23
ATP1A1 knock-out animals are characterized by elevation of serum aldosterone in
24
comparison to wild type littermates
25
phenotype. However, up to now no germline or somatic ATP1A1 mutation has been
15
10
. The K+ and Na+ gradients created drive the ion
12
and glomerulosa cell growth in vivo
14
13
.
. Interestingly, heterozygous
although indirect effects might well apply to this
Beuschlein et al.
1
associated with a human disease. ATP2B3 also belongs to the ATPase family of genes and
2
encodes a plasma membrane Ca2+-ATPase (PMCA3) that is essential to clear Ca2+ from the
3
cytoplasm of eukaryotic cells and, thereby, plays a critical role in intracellular calcium
4
homeostasis 16. While several mutations of PMCA2 have been identified in mouse models
5
and patients with hearing loss
6
for human disease.
18
17
, no mutations in PMCA3 have been described as causative
7
8
Projection of the ATP1A1 mutations onto the resolved crystal structure of the
9
orthologous Squalus acanthias Na+,K+-ATPase (Supplementary Figure 1 B and Figure 2)
10
showed that all three mutations are either located in the trans-membranous α-helix M1 or the
11
juxtaposed α-helix M4, which have been suggested to interact and cooperate in K+ binding
12
and gating by interaction of Glu334 with Leu104 11. Since the atomic structure of the plasma
13
membrane Ca2+-ATPase ATP2B3 has not been determined, we used the known structure of
14
the homologous rabbit sarcoplasmic reticulum type Ca2+-ATPase (SERCA) for the
15
localization of the ATP2B3 mutations. Intriguingly, the APA associated deletion mutations in
16
ATP2B3 also involved the M4 transmembrane helix in the same region where the glutamate
17
homologous to Glu334 of ATP1A1 is positioned (Figure 2). In SERCA, this glutamate is a
18
crucial residue in Ca2+ binding (Figure 2 D). Thereby, the deletion in ATP2B3 is predicted to
19
cause a major distortion of the Ca2+ binding site. The multiple occurrence of the mutations in
20
these highly conserved regions involved in interaction with the transported cations in two
21
paralogs is suggestive of a gain-of-function effect.
22
23
To examine the functional consequences of mutations Leu104Arg and Val332Gly of
24
ATP1A1, COS cells were transfected with the corresponding mutated ouabain insensitive rat
25
cDNA and subjected to selection in the presence of ouabain inhibiting the endogenous COS
Beuschlein et al.
11
. No viable colonies were detected, indicating that the physiological Na+,K+-
1
cell enzyme
2
pump activity of the mutant rat enzymes was very low or insignificant. The mutants were
3
therefore expressed transiently in COS cells in the presence of siRNA to knock down the
4
endogenous Na+,K+-ATPase, and functional studies were carried out on isolated plasma
5
membranes. The ATPase activity determined under optimal conditions for wild type was
6
indeed insignificant for Leu104Arg and very low for Val332Gly (Figure 3 A). The mutant
7
enzymes were, however, well expressed and able to react with ATP and be phosphorylated in
8
a Na+ dependent reaction (Figure 3 B). Importantly, the apparent affinity for K+ was reduced
9
conspicuously in the mutants relative to wild type (∼500- and ∼50-fold, respectively, for
10
Leu104Arg and Val332Gly, Figure 3 C). The effect of Leu104Arg supports the previous
11
suggestion that Leu104 by van der Waals interaction positions Glu334, which is essential in
12
K+ binding and gating of the binding pocket 11. Substitution of the almost juxtaposed Val332
13
with a glycine without side chain may create a flexibility, likewise disturbing the function of
14
Glu334 (Figure 2A-C).
15
In ex vivo conditions, electrophysiological examination of primary cultured adenoma
16
cells with different underlying mutations revealed strikingly higher levels of depolarization in
17
ATPase mutated cells in comparison to cells from normal adjacent tissue. This indicates that
18
adenoma cells with ATPase mutations are profoundly altered. The finding was specific for
19
ATPase mutated adenoma samples, since such a strong depolarization was not observed in
20
KCNJ5 mutated adenoma samples or in adenoma cells without known mutation (Figure 3 D).
21
When extracellular Na+ was removed, primary cells hyperpolarized (Figure 3 E). As expected,
22
the hyperpolarization was most pronounced in cells from adenomas carrying mutated, Na+-
23
permeable KCNJ5 4. The effect of Na+ removal in primary adenoma cells with mutated
24
ATP1A1 was less pronounced. This suggests that the strong depolarization observed in cells
25
with mutated ATP1A1 is not primarily caused by an increased channel-like Na+ conductance
Beuschlein et al.
1
but possibly by disturbed intracellular ion composition (Figure 3 E). In HEK cells transfected
2
with Leu104Arg, the membrane voltage was depolarized in comparison to wild type ATP1A1
3
suggesting that the depolarization of primary adenoma cells is a specific consequence of the
4
mutation of ATP1A1 (Figure 3 F).
5
6
Whereas KCNJ5 mutations have been consistently reported to be more prevalent in
7
female patients 5,19,20 ATPase mutations were predominantly found in males (ATPases, 81.0%
8
(17/21) males vs. KCNJ5, 25.4% (30/118) males; p<0.0001). Consistent with a more severe
9
phenotype, carriers of ATPase mutated tumors had higher preoperative aldosterone levels
10
(median; interquartile range 397.8; 555.0 ng/l vs 286.6; 287.2 ng/l, in patients without
11
mutations, p=0.02) and significantly lower serum potassium concentrations (2.6; 0.8 mmol/l
12
vs 3.1; 0.8 mmol/l, p=0.013) than carriers of KCNJ5 mutations. Other clinical endpoints
13
including tumor size, blood pressure, serum sodium, and urinary albumin secretion were
14
similar between the groups (Supplementary Table 5).
15
16
We have been unable to detect germline mutations in ATP1A1 or ATP2B3 in patients
17
with familial forms or with the bilateral form of primary aldosteronism. Given the central role
18
of Na+,K+-ATPase and Ca2+-ATPase for generation of electrochemical gradients which are
19
required for electrical excitability, cellular uptake of ions, nutrients and neurotransmitters, and
20
for regulation of cell volume and intracellular pH, germline mutations in these genes are
21
predicted to underlie strong purifying selection. Interestingly, however, ATP2B1, one of the
22
four mammalian plasma membrane Ca2+-ATPase isoforms, was significantly associated with
23
systolic and diastolic blood pressure and hypertension in a recent genome wide association
24
study 21.
25
Beuschlein et al.
1
In summary, we show herein that somatic mutations in ATP1A1 and ATP2B3 are
2
present in roughly 7% of aldosterone producing adenomas. In both cases inactivation of the
3
pump function either indirectly - in the case of ATP1A1 - or directly - for ATP2B3 – are
4
predicted to increase intracellular Ca2+ levels, which in turn prime Ca2+ - dependent signalling
5
and aldosterone output (Figure 4). As the mutations identified in both ATPases are
6
constrained to specific and highly conserved functional domains further gain of function
7
mechanisms such as a pathological transport mode might be potential contributors to the
8
molecular phenotype. Thereby, these findings expand the spectrum of somatic mutations
9
leading to APAs to two members of the P-type ATPase pump family, extend our knowledge
10
on the molecular mechanism leading to aldosterone producing adenomas and point towards
11
novel potential therapeutic targets for the most frequent secondary form of arterial
12
hypertension.
13
Beuschlein et al.
1
ACKNOWLEDGMENT
Munich/Regensburg: This work has been made possible by a grant of the Else Kröner-
2
3
Fresenius-Stiftung
in
support
of
the
German
Conn’s
Registry-Else-Kröner
4
Hyperaldosteronism Registry (to MR). Further funding was received from the Deutsche
5
Forschungsgemeinschaft to FB and MR (Re 752/17-1) and RW (FOR1086). The work was
6
further supported by the German Ministry of Education and Research (01GR0802 and
7
01GM0867), the European Commission 7th Framework Program (261123, GEUVADIS), and
8
the German Center for Cardiovascular Research (DZHK).
9
Paris: The authors wish to thank Hervé Lefèbvre, Estelle Louiset (INSERM, U982,
10
Mont-Saint-Aignan and University Hospital of Rouen, Rouen, France) and Mathilde Sibony
11
(Assistance Publique-Hôpitaux de Paris, Hôpital Cochin, Paris, France) for providing control
12
adrenal samples. We thank the COMETE (COrtico and MEdullary adrenal Tumors) network
13
for providing tissue samples from APA. This work was funded through institutional support
14
from INSERM and by the Agence Nationale pour la Recherche (ANR Physio 2007, No.: 013-
15
01; Genopat 2008, No.: 08-GENO-021), the Fondation pour la Recherche sur l'Hypertension
16
Artérielle (AO 2007), the Fondation pour la recherche médicale (ING20101221177), the
17
Programme Hospitalier de Recherche Clinique (PHRC grant AOM 06179) and by grants from
18
INSERM and Ministère Délégué à la Recherche et des Nouvelles Technologies.
19
Aarhus: This work was supported in part by grants to BV from the Danish Medical
20
Research Council, the Novo Nordisk Foundation (Fabrikant Vilhelm Pedersen og Hustrus
21
Legat), and the Lundbeck Foundation.
22
Torino: This study was supported by grants from Fondi Ricerca Ex-60% MIUR
23
(Ministry of University, Scientific and Technological Research) 2012 and Compagnia di San
24
Paolo.
Beuschlein et al.
1
AUTHOR CONTRIBUTIONS
2
Sheerazed Boulkroun, Hang Nielsen, Urs Lichtenauer, David Penton, Vivien Schack,
3
Annet Walther, Philipp Tauber, Susanne Diener, and Benoit Samson-Couterie performed the
4
experiments; Andrea Oßwald, Thomas Wieland, Laurence Amar, Eveyln Fischer, Thomas
5
Schwarzmayr, Elisabeth Graf, and Arndt Benecke performed statistical analysis and analyzed
6
the data; Bruno Allolio, Marcus Quinkler, Francesco Fallo, Pierre-François Plouin, Franco
7
Mantero, and Paolo Mulatero contributed materials; Felix Beuschlein, Thomas Meitinger,
8
Xavier Jeunemaitre, Richard Warth, Bente Vilsen, Maria-Christina Zennaro, Tim M. Strom,
9
and Martin Reincke jointly supervised research, conceived and designed the experiments,
10
analyzed the data, contributed reagents/materials/analysis tools and wrote the paper
11
DATA ACCESS
12
Disease
causing
variants
will
be
submitted
to
ClinVar
13
(www.ncbi.nlm.nih.gov/clinvar/). Furthermore, exome data are available on request within a
14
scientific cooperation.
15
Beuschlein et al.
1
2
3
FIGURE CAPTIONS
Fig 1: Summary of somatic mutations in ATP1A1 and ATP2B3 identified in
aldosterone producing adenomas.
4
5
Fig 2: Structural position of mutated residues in Na+,K+-ATPase (A-C) and Ca2+-
6
ATPase (D). (A) Leu104 (“L104”) in transmembrane helix M1 and Val332 (“V332”) in M4
7
shown in relation to Glu334 (“E334”) in M4, which is crucial in binding of K+ (purple
8
spheres) and gating of the binding pocket. Leu104 positions Glu334 for occlusion of K+
9
(B) By replacing the hydrophobic leucine by a large, positively charged arginine, mutation
10
Leu104Arg (“L104R”) likely alters the position of Glu334, thus disturbing the gating
11
mechanism. (C) In mutant Val332Gly (“V332G”), the lack of side chain in glycine introduces
12
flexibility due to surrounding space, which may also influence the position of Glu334. Shown
13
as dashed lines is a likely hydrogen bond linking the two residues together (3.09 Å distance
14
between backbone oxygen and nitrogen). (D) Illustration of one of the two Ca2+ sites in
15
SERCA, which is equivalent to the Ca2+ site in PMCA. The purple sphere indicates the Ca2+
16
ion, which is shown together with the liganding residues (“V304”, “E309”, “N796” and
17
“D800”). Val304 in M4 is equivalent to the valine that together with a juxtaposed leucine is
18
deleted in the ATP2B3 mutant. Glu309 (“E309”) is equivalent to Glu334 of Na+,K+-ATPase.
9,11
.
19
20
Fig 3: Functional and electrophysiological examination of transfected cells and
21
adenoma primary cultures. (A) Plasma membranes were isolated from COS cells transfected
22
by cDNA encoding Leu104Arg, Val332Gly, and wild type rat Na+,K+-ATPase and siRNA
23
knocking down the endogenous Na+,K+-ATPase. The maximal Na+,K+-ATPase activity of the
24
exogenous rat enzyme determined in the presence of 130 mM Na+ and 20 mM K+ at 37 °C is
Beuschlein et al.
1
shown relative to that of the wild type, which was 0.58 ± 0.03 μmol/mg membrane protein/h
2
(SEM, n=5). (B) Quantification of relative phosphorylation from [γ-32P]MgATP in the
3
presence of 100 mM Na+. (C) K+ inhibition of phosphorylation from [γ-32P]MgATP in the
4
presence of 50 mM Na+ for wild type (open circles), Leu104Arg (filled circles), and
5
Val332Gly (filled diamonds). The phosphorylation level in the absence of K+ was taken as
6
100%. (D) Membrane voltages of primary cultured adrenal cells measured by whole cell
7
patch-clamp. Cells from normal adjacent tissue (65 cells of 14 patients) showed the most
8
hyperpolarized membrane voltages, cells from adenomas without mutations of KCNJ5 or the
9
two ATPases (KCNJ5-/ATPase-, 46 of 7) and adenoma cells with KCNJ5 mutations (33 of 5)
10
were slightly depolarized. Adenoma cells carrying mutations of ATP2B3 (7 of 2) or ATP1A1
11
(8 of 2) were significantly depolarized compared to adjacent tissue cells. (E) Effect of
12
removal of bath Na+ on membrane voltage (shown as delta Vm). As expected, adenoma cells
13
with the Na+-permeable mutant KCNJ5 channel showed the strongest effect of Na+ removal
14
on membrane voltage while less pronounced effects were present in adenoma cells carrying
15
mutations of ATP2B3 and ATP1A1. (F) Membrane voltages of HEK cells transfected with the
16
Leu104Arg mutant of ATP1A1 were depolarized compared to cells overexpressing the wild
17
type ATP1A1 (control). After removal of bath Na+, the membrane voltage was still more
18
depolarized in cells transfected with the Leu104Arg mutant suggesting a disturbed
19
intracellular ion composition and/or loss of net charge transport by the mutated pump.
20
Numbers of cells in parenthesis, asterisk indicates statistically different from adjacent cells
21
(p<0.05).
22
23
Fig 4: Proposed mechanism for autonomous aldosterone secretion in APAs with
24
somatic ATPase mutations. (A) Glomerulosa cell with hyperpolarized membrane potential
25
under baseline conditions. (B) Angiotensin II induced inhibition of K+ channels
22
and Na+,
Beuschlein et al.
1
K+ ATPase
2
increase of cytoplasmic Ca2+ levels. (C) Mutation of ATP1A1 resulting in alteration of K+
3
binding and loss of function followed by cell depolarization. (D) Mutation of ATP2B3
4
associated with impaired clearance of cytoplasmic Ca2+ levels. CYP11B2: aldosterone
5
synthase gene.
6
7
14
leading to cell depolarization, activation of voltage-gated Ca2+ channels and
Beuschlein et al.
1
ONLINE METHODS
2
Patient cohort
3
Patients with primary aldosteronism were recruited among seven different centers
4
from the APA working group of the European Network for the Studies of Adrenal Tumors
5
(ENS@T, www.ensat.org). Case detection and subtype identification were in accordance with
6
institutional guidelines. The diagnosis of adrenocortical adenoma was histologically
7
confirmed after surgical resection. All patients gave written informed consent for genetic
8
investigation within each individual institution. For initial exome sequencing nine patients
9
from the Munich cohort with absent germline or somatic KCNJ5 mutations were selected.
10
Baseline clinical and biochemical characteristics of these patients are summarized in
11
Supplementary Table 1.
12
Nucleic acid extraction
13
DNA or RNA was extracted from a total of 308 APAs with 17 paired peritumoral
14
adrenal cortices and 16 peripheral DNA samples, and 91 peripheral DNA samples from
15
patients with BAH. Tumor DNA was extracted using RNeasy DNA extraction kit (Qiagen,
16
Hilden, Germany); DNA from peripheral blood leukocytes was prepared using salt-extraction.
17
Total RNA was isolated from frozen tissue using Trizol (Invitrogen, Carlsbad, CA) and then
18
cleaned-up on silica columns using RNeasy Mini Kit (Qiagen). The integrity and quality of
19
the RNA were systematically checked using an Agilent 2100 bioanalyzer with the RNA6000
20
Nano Assay (Agilent Technologies, Santa Clara, CA). After DNaseI treatment (Invitrogen),
21
500 ng of total RNA were retro-transcribed using Superscript II RT (Invitrogen) and random
22
hexamers (Promega, Madison, WI).
Beuschlein et al.
1
Exome sequencing
2
Exomes were enriched in solution and indexed with SureSelect XT Human All Exon
3
50 Mb kits (Agilent). Sequencing was performed as 100 bp paired-end runs on HiSeq2000
4
systems (Illumina). Pools of 12 indexed libraries were sequenced on four lanes. Image
5
analysis and base calling was performed using Illumina Real Time Analysis. CASAVA 1.8
6
was used for demultiplexing.
7
8
Variant detection
9
BWA (v 0.5.9) with standard parameters was used for read alignment against the
10
human genome assembly hg19 (GRCh37). We performed single-nucleotide variant and small
11
insertion and deletion (indel) calling specifically for the regions targeted by the exome
12
enrichment kit, using SAMtools (v 0.1.18). Subsequently the variant quality was determined
13
using the SAMtools varFilter script. We used default parameters, with the exception of the
14
maximum read depth (-D) and the minimum P-value for base quality bias (-2), which we set
15
to 9999 and 1e-400, respectively. Additionally, we applied a custom script to mark all
16
variants with adjacent bases of low median base quality. All variants were then annotated
17
using custom Perl scripts. Annotation included information about known transcripts (UCSC
18
Known Genes and RefSeq genes), known variants (dbSNP v 135), type of mutation and - if
19
applicable – amino acid change in the corresponding protein. The annotated variants were
20
then inserted into our in-house database.
21
To discover putative somatic variants, we queried our database to show only those
22
variants of a tumor that were not found in the corresponding control tissue. To reduce false
23
positives we filtered out variants that were already present in our database, had variant quality
Beuschlein et al.
1
less than 40, or failed one of the filters from the filter scripts. We then manually investigated
2
the raw read data of the remaining variants using the Integrative Genomics Viewer (IGV).
3
4
ATP1A1 and ATP2B3 sequencing
5
DNA was amplified using intron-spanning primers. Bidirectional Sanger sequencing
6
was performed using the ABI BigDye Terminator v.3.1 Cycle Sequencing Kit. Primer
7
sequences (available on request) were determined using the Primer3 or ExonPrimer software
8
(http://frodo.wi.mit.edu/primer3/input.htm;
9
ExonPrimer.html).
http://ihg.helmholtz-muenchen.de/ihg/
10
11
12
Protein sequence analysis
Similarity
analysis
was
performed
using
the
ClustalW2
program
13
(http://www.ebi.ac.uk/Tools/msa/clustalw2/). Thereby, human ATP1A1 (GenBank accession
14
number NP_000692.2) and ATP2B3 sequences (NP_068768) were compared with sequences
15
from other species and with those of other human ATPases, respectively.
16
17
Modeling of protein structures
18
Structural figures were prepared using PyMOL software (www.pymol.org). The
19
Na+,K+-ATPase and sarcoplasmic reticulum Ca2+-ATPase structures shown have PDB codes
20
2ZXE and 1SU4, respectively.
21
Beuschlein et al.
1
Immunohistochemistry
2
Immunohistochemistry was performed on sections de-paraffinized in xylene and
3
rehydrated through graded ethanol. For antigen unmasking, slides were incubated in antigen
4
unmasking solution (Vector Laboratories, Burlingame, CA) for 30 min at 98 C. Endogenous
5
peroxidases were inhibited by incubation in 3% hydrogen peroxide (Sigma-Aldrich, St. Louis,
6
MO) in water for 10 min and nonspecific staining blocked with normal goat serum. Primary
7
antibodies [ATP1A1 (Abgent, AJ1524a; San Diego, CA); 1:1000; ATP2B3 (Sigma-Aldrich,
8
HPA001583 ); 1:100] were incubated overnight at 4 °C. Sections were washed, incubated 30
9
min with affinity-purified goat anti-rabbit (1:400; Vector Laboratories) antibodies, washed,
10
and incubated with an avidin-biotin-peroxidase complex (Vectastain ABC Elite; Vector
11
Laboratories) for 30 min. The slides were developed using diaminobenzidin (Vector
12
Laboratories) and counterstained with hematoxylin (Sigma). In the negative control reactions,
13
primary antibodies were omitted from the dilution buffer, which in all instances resulted in a
14
complete absence of staining. All microscopic examinations were done on a Leica
15
microscope. Quantification was performed on 3 different fields/tumor in n=3 ATP1A1
16
mutated APA and n=14 non-mutated APAs. Results represent means±SEM of 135-215 cells
17
expressed as percentage of each expression pattern.
18
Quantification of mRNA expression
19
mRNA expression data for 91 samples with genotype data were retrieved from a
20
pangenomic transcriptome analysis performed on 123 samples collected through the
21
COMETE network from patients operated on APA between 1994 and 2008 in the
22
Hypertension Unit at the Hôpital Européen Georges Pompidou in Paris. Procedures for data
23
acquisition and calculation have been described in details elsewhere 5.
Beuschlein et al.
1
Quantification of Na+,K+-ATPase activity in transfected cells
2
For in vitro studies, mutations were introduced into full-length cDNA encoding the
3
ouabain resistant rat α1-isoform of Na+,K+-ATPase, and the mutants and wild type were
4
expressed in COS-1 cells. The >100-fold difference between the ouabain affinities of the
5
exogenous rat Na+,K+-ATPase and the endogenous COS cell enzyme allows isolation of
6
stable cell lines under ouabain selection pressure, provided the exogenous enzyme is
7
functional in Na+ and K+ pumping. Because the mutants studied here were unable to support
8
cell growth in the presence of ouabain, thus indicating lack of pump function, we also used
9
siRNA co-transfection to knock down the endogenous enzyme, thereby allowing studies of
10
transiently expressed enzyme as an alternative. Leaky plasma membranes were assayed
11
functionally by previously described methods
12
37°C by following the liberation of Pi in the presence of 130 mM NaCl, 20 mM KCl, 3 mM
13
MgATP, 30 mM histidine buffer (pH 7.5), and 1 mM EGTA, together with 100 μM ouabain
14
to ensure complete knock out of the ouabain sensitive endogenous enzyme. Phosphorylation
15
was carried out for 10 s at 0°C with 2 µM AT32P in the presence of 100 mM NaCl, 3 mM
16
MgCl2, 20 mM Tris (pH 7.5), 100 μM ouabain, and 20 μg oligomycin/ml or in the presence of
17
50 mM NaCl, 3 mM MgCl2, 20 mM Tris (pH 7.5), 100 μM ouabain, and varying
18
concentrations of KCl with choline chloride added to maintain a constant ionic strength. The
19
32
20
radioactivity quantified by phosphor imaging. For quantification the number of repeats was
21
n=4-6. Background represented by the inactive phosphorylation site mutant Asp376Asn was
22
subtracted in all cases.
23
11
. Na+,K+-ATPase activity was determined at
P labeled Na+,K+-ATPase was separated by acid SDS gel electrophoresis and the
Beuschlein et al.
1
Preparation of primary cell cultures
2
For primary cultures from APAs and adjacent normal adrenal gland tissue, samples
3
were cleaned of surrounding fat, connective tissue and blood vessels. Thereafter, tissue
4
samples were minced into pieces smaller than 0.5 mm using a razor blade. Minced samples
5
were transferred into 15 ml Falcon tubes and spun down at 1200 rpm for 5 min. The pellet
6
was resuspended in 10 ml of digestion buffer containing 2 mg collagenase II (Biochrom,
7
Berlin, Germany) per ml of PBS and incubated at 37°C for no longer than 50 min in a shaking
8
water bath. Every 15 minutes, the tube content was pipetted up and down several times using
9
a 25 ml and subsequently a 10 ml pipette to support digestion. Collagenase was inactivated by
10
adding pure FCS to a minimum total concentration of 10%, where after the cells were
11
sequentially filtered through a 100 µm and a 70 µm nylon mesh, followed by a centrifugation
12
step as described above. The pellet was resuspended in erythrocyte lysis buffer and incubated
13
for 7 min at room temperature. After another centrifugation step, cells were resuspended in 1-
14
2 ml culture medium depending on the expected cell count (DMEM/F12 with 10% FCS, 3.1
15
g/l glucose, and 10 µl/ml Pen/Strep, all from Gibco), and sequentially filtered through a 70µm
16
nylon mesh. Cells were quantified using a Neubauer counting chamber and further processed
17
as described below.
18
Electrophysiological evaluation of primary cell cultures and transfected cells
19
Patch clamp recordings were performed using an EPC-10 amplifier without leak
20
subtraction (HEKA, Germany). The following solutions (pH 7.4; all concentrations in mM)
21
were used for primary adenoma cells: HEPES 5; NaCl 124.5; Na2HPO4 1.6; NaH2PO4 0.4;
22
Glucose 5; MgCl2 1; CaCl2 1.3; KCl 4.1. The extracellular solution for voltage measurements
23
in HEK cells contained: HEPES 5; NaCl 145; K2HPO4 1.6; KH2PO4 0.4; Glucose 5; MgCl2 1;
24
CaCl2 1.3. For the Na+-free solution, Na+ was replaced by N-methyl-D-glucamine. The
Beuschlein et al.
1
pipette solution was the following (pH7.2; all concentrations in mM): K-gluconate 95; KCl
2
30; Na2HPO4 4.8; NaH2PO4 1.2; glucose 5; MgCl2 2.3; CaCl2 0.762; EGTA 1; Na2-ATP 3.
3
For functional expression in HEK cells, full-length cDNA of wildtype rat ATP1A1 and
4
mutant ATP1A1Leu104Arg were subcloned into the bicistronic pIRES-CD8 expression vector.
5
One day before voltage measurements, HEK cells were transfected with wildtype rat ATP1A1
6
or its mutant ATP1A1Leu104Arg using Lipofectamine. Anti-CD8-labelled Dynabeads
7
(Invitrogen) were used to identify transfected cells.
8
9
Statistical analysis
10
If not stated otherwise, group results are expressed as median and interquartile range.
11
Data between groups were compared using Kruskal-Wallis test followed by 2-sided test for
12
pairwise comparison of two groups. The significance level of p<0.05 was considered to be
13
statistically significant. Statistical analysis was performed using standard statistical software
14
(SPSS 20).
15
16
Beuschlein et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
REFERENCES AND NOTES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Hannemann, A. et al. Screening for primary aldosteronism in hypertensive subjects: results from two
German epidemiological studies. European journal of endocrinology / European Federation of
Endocrine Societies 167, 7-15 (2012).
Eide, I.K., Torjesen, P.A., Drolsum, A., Babovic, A. & Lilledahl, N.P. Low-renin status in therapyresistant hypertension: a clue to efficient treatment. J Hypertens 22, 2217-26 (2004).
Rossi, G.P. et al. A prospective study of the prevalence of primary aldosteronism in 1,125 hypertensive
patients. J Am Coll Cardiol 48, 2293-300 (2006).
Choi, M. et al. K+ channel mutations in adrenal aldosterone-producing adenomas and hereditary
hypertension. Science 331, 768-72 (2011).
Boulkroun, S. et al. Prevalence, clinical, and molecular correlates of KCNJ5 mutations in primary
aldosteronism. Hypertension 59, 592-8 (2012).
Beuschlein, F. et al. Clonal composition of human adrenocortical neoplasms. Cancer Res 54, 4927-32
(1994).
Gicquel, C. et al. Clonal analysis of human adrenocortical carcinomas and secreting adenomas. Clin
Endocrinol (Oxf) 40, 465-77 (1994).
Mulatero, P. et al. KCNJ5 mutations in European families with nonglucocorticoid remediable familial
hyperaldosteronism. Hypertension 59, 235-40 (2012).
Morth, J.P. et al. Crystal structure of the sodium-potassium pump. Nature 450, 1043-9 (2007).
Kaplan, J.H. Biochemistry of Na,K-ATPase. Annual review of biochemistry 71, 511-35 (2002).
Einholm, A.P., Andersen, J.P. & Vilsen, B. Importance of Leu99 in transmembrane segment M1 of the
Na+, K+ -ATPase in the binding and occlusion of K+. The Journal of biological chemistry 282, 2385466 (2007).
Yingst, D.R., Davis, J., Krenz, S. & Schiebinger, R.J. Insights into the mechanism by which inhibition
of Na,K-ATPase stimulates aldosterone production. Metabolism: clinical and experimental 48, 1167-71
(1999).
Neri, G. et al. Ouabain chronic infusion enhances the growth and steroidogenic capacity of rat adrenal
zona glomerulosa: the possible involvement of the endothelin system. International journal of
molecular medicine 18, 315-9 (2006).
Hajnoczky, G. et al. Angiotensin-II inhibits Na+/K+ pump in rat adrenal glomerulosa cells: possible
contribution to stimulation of aldosterone production. Endocrinology 130, 1637-44 (1992).
Moseley, A.E. et al. Genetic profiling reveals global changes in multiple biological pathways in the
hearts of Na, K-ATPase alpha 1 isoform haploinsufficient mice. Cellular physiology and biochemistry :
international journal of experimental cellular physiology, biochemistry, and pharmacology 15, 145-58
(2005).
Di Leva, F., Domi, T., Fedrizzi, L., Lim, D. & Carafoli, E. The plasma membrane Ca2+ ATPase of
animal cells: structure, function and regulation. Archives of biochemistry and biophysics 476, 65-74
(2008).
Street, V.A., McKee-Johnson, J.W., Fonseca, R.C., Tempel, B.L. & Noben-Trauth, K. Mutations in a
plasma membrane Ca2+-ATPase gene cause deafness in deafwaddler mice. Nature genetics 19, 390-4
(1998).
Schultz, J.M. et al. Modification of human hearing loss by plasma-membrane calcium pump PMCA2.
The New England journal of medicine 352, 1557-64 (2005).
Akerstrom, T. et al. Comprehensive Re-Sequencing of Adrenal Aldosterone Producing Lesions Reveal
Three Somatic Mutations near the KCNJ5 Potassium Channel Selectivity Filter. PLoS One 7, e41926
(2012).
Azizan, E.A. et al. Microarray, qPCR, and KCNJ5 sequencing of aldosterone-producing adenomas
reveal differences in genotype and phenotype between zona glomerulosa- and zona fasciculata-like
tumors. The Journal of clinical endocrinology and metabolism 97, E819-29 (2012).
Ehret, G.B. et al. Genetic variants in novel pathways influence blood pressure and cardiovascular
disease risk. Nature 478, 103-9 (2011).
Spat, A. & Hunyady, L. Control of aldosterone secretion: a model for convergence in cellular signaling
pathways. Physiological reviews 84, 489-539 (2004).