Angiotensin II Type 1 Receptor Autoantibodies in Primary Aldosteronism

Abstract

Primary aldosteronism (PA) is the most common form of endocrine hypertension. Agonistic autoantibodies against the angiotensin II type 1 receptor (AT ₁ R-Abs) have been described in transplantation medicine and women with pre-eclampsia and more recently in patients with PA. Any functional role of AT ₁ R-Abs in either of the two main subtypes of PA (aldosterone-producing adenoma or bilateral adrenal hyperplasia) requires clarification. In this review, we discuss the studies performed to date on AT ₁ R-Abs in PA.

Introduction

Primary aldosteronism (PA) displays a prevalence of around 5–10% in patients with hypertension increasing up to 20% in patients with treatment-resistant hypertension ^{1 2 3} . Patients with PA have an elevated risk of cerebrovascular and cardiovascular events relative to patients with hypertension with matched cardiovascular risk profiles ^{4 5 6} . Different subtypes of PA have been defined which affect one or both adrenal glands. The main subtypes are a unilateral aldosterone-producing adenoma (APA) or bilateral adrenal hyperplasia (BAH, also called idiopathic hyperaldosteronism). Other sporadic forms include unilateral adrenal hyperplasia and the very rare occurrences of aldosterone-producing carcinoma. Hereditary forms of familial hyperaldosteronism are rare and the genetic basis of the 4 described forms of familial hyperaldosteronism (FH types I–IV) has been identified ^{7 8} . A number of somatic mutations, mainly in ion channels and ATPases, have been reported which are likely to drive the aldosterone excess in the majority of APAs. Increasing interest in autoantibodies led to studies reporting a potential role for autoantibodies against the G-protein-coupled receptor (GPCR) angiotensin II type 1 receptor in PA.

G-Protein-Coupled Receptors

The largest superfamily of receptors in the human genome are GPCRs that are located in the plasma membrane of nearly all cell types ⁹ . The research of Kobilka and Lefkowitz about the molecular structure and function of GPCRs, especially β2-adrenergic G-protein-coupled receptors, was rewarded in 2012 by the Nobel prize in chemistry ¹⁰ . GPCRs are responsible for signal transduction to regulate numerous essential functions by mediating extracellular signals from hormones, neurotransmitters or environmental stimulants to the intracellular metabolic pathways ⁹ . Their structure consists of seven transmembrane-spanning helices bound by intra- and extracellular loops ¹¹ . On the extracellular side, GPCRs are targeted by their ligands but also by autoantibodies, which may induce agonistic receptor stimulation or inhibition dependent on the autoantibody binding site to the first and second or third extracellular loops, respectively ⁹ .

Binding of extracellular agonists activate the receptor by initiating a conformational change that induce further signal transduction pathways ¹² . Conformational changes result in the hetero-trimeric G-protein to exchange GDP for GTP at its Gα subunit ¹³ . The Gα subunit with GTP and the Gβγ subunit both dissociate from the receptor, resulting in the activation of specific signaling pathways such as, adenylyl cyclase (via cAMP generation) and phospholipase C [via diacylglycerol (DAG) and inositol trisphosphate (IP ₃ ) production] ¹³ . The Gα subunit bound to GTP hydrolyses GTP back to GDP to reassociate with the Gβγ subunit ¹³ . Receptor signaling determination is mediated by G-protein-coupled receptor kinases (GRK) which phosphorylate the activated receptor to bind to a distinct scaffold protein for subsequent internalization into the cells ¹³ . These scaffold proteins are arrestins and are known to desensitize GPCRs and induce G-protein independent signaling ¹⁴ . Following internalization the receptor is either degraded or recycled back to the plasma membrane ¹³ . Besides its involvement in GPCR internalization, arrestins are able to interact directly with activated GPCRs resulting in a clear conformational change that could initiate further downstream signaling pathways ¹³ . There is some evidence for biased agonism of GPCRs towards β-arrestin-mediated signaling ^{15 16} .

The angiotensin II type 1 and 2 receptors (AT ₁ R and AT ₂ R) are GPCRs with opposing functions in blood pressure regulation and sodium excretion ^{17 18 19} . The two subtypes share 34% sequence homology and stimulate different signaling pathways to elicit distinct and counter-regulatory biological functions ¹⁷ . AT ₂ R is highly expressed in the fetal state although lower levels are present in the adult brain, heart, kidney and the adrenal ¹⁷ . In some diseases, AT ₂ R is upregulated acting as an anti-inflammatory and repairing factor for wound healing or after cardiac or vascular events ^{20 21} . In contrast, the AT ₁ R is widely distributed, for example, in the adrenal gland, liver, kidney, fat, brain, placenta, spleen, or thyroid, and its physiological role as a component of the renin-angiotensin-aldosterone system (RAAS) is well characterized via binding of its cognate ligand angiotensin II for blood pressure regulation, vasoconstriction, inflammatory response and vascular and cardiac hypertrophy ²¹ .

The Discovery of Autoantibodies Against Angiotensin II Type 1 Receptor

The prevalence of autoimmune diseases in the population is around 2.5% although autoantibodies are also often present in healthy individuals ^{22 23} . Impaired B cell tolerance can allow autoantibody-producing B cells with medium or low binding affinity to self-antigens to escape from elimination or further anergy during B cell maturation thus becoming autoantibody-secreting plasma cells ²² . However, the pathogenic role of autoantibodies is mostly unknown.

The role of autoantibodies against the AT ₁ R (AT ₁ R-Abs) in hypertension has been investigated over the last decades. In 1999, Wallukat et al. described the presence of AT ₁ R-Abs in patients with preeclampsia, which is discussed further below ²⁴ . This discovery was extended one year later by Fu and coworkers who observed elevated AT ₁ R-Ab levels in 33% patients with malignant secondary hypertension, but also found increased levels of AT ₁ R-Abs in 14% of the control group ²⁵ . Soon afterwards, an association of AT ₁ R-Abs and renal graft failure was observed by many research groups, especially in patients without donor-specific human leukocyte antigen (HLA)-antibodies ^{26 27 28 29} . Dragun et al. reported a comparable prevalence of kidney rejections associated with either AT ₁ R-Abs (3.6%) or donor-specific HLA-antibodies (3.2%) in a cohort of 278 kidney transplantations including 119 rejections ²⁹ . High levels of AT ₁ R-Abs (>17 units) without additional presence of donor-specific antibodies have been ascribed to an increased risk for antibody-mediated kidney rejection ²⁷ . Other studies reported on AT ₁ R-Ab levels of >9 U/ml or >10 U that have been shown to elevate the risk of graft loss ^{28 30} . Notably, in some patients (17–47%) AT ₁ R-Abs were already present before transplantation ^{26 27 28 30} . A case report of Jobert et al. described a 28-year old male recipient of a well-matched renal allograft lacking anti-HLA-antibodies but with a high AT ₁ R-Ab level of 14.1 U/ml prior transplantation who developed acute vascular rejection four days post-transplantation ³¹ . The authors hypothesized that the AT ₁ R-Abs are the only factor that could have led to the allograft rejection due to the otherwise well-matched HLA-type and virus status ³¹ . Treatment with anti-thymocyte globulin, methylprednisolone, candesartan and plasma exchange was successful suggesting probably plasma exchange as a considerable option in patients with AT ₁ R-Abs ³¹ . However, AT ₁ R-Abs also appeared after kidney transplantation in prior AT ₁ R-Ab-negative patients without an associated worse clinical outcome ³² . Of note is that Taniguchi et al. stressed that such studies report an association of AT ₁ R-Ab levels and allograft failure and do not explain any causal relationship ²⁶ . The occurrence of autoantibodies is widely distributed in transplantation medicine but also in autoimmune diseases such as Huntington, multiple sclerosis or systemic sclerosis ^{33 34} . A summary of diseased states widely reported as associated with increased AT ₁ R-Ab levels is shown in the Fig. 1 .

Fig. 1.

Potential pathological effects of agonistic AT ₁ R autoantibodies: Autoantibodies to the AT ₁ R have been described in a number of different pathological states as indicated. Continuous lines represent a direct effect, dashed lines indicate an indirect effect, dotted lines denote putative effects. Figure compiled by the authors using elements of https://smart.servier.com/ (licensed under Creative Commons Attribution 3.0 Unported License).

Comparison of Different Assays for AT1R-Ab Characterization

AT ₁ R-Abs in patient serum have been widely measured by enzyme-linked immunosorbent assays (ELISA). For such ELISAs, the target antigens are immobilized on a solid phase, mainly on a microplate to detect binding autoantibodies from serum samples ²² . As antigens serve the human full-length AT ₁ R or peptides comprising known epitopes (AFHYESQ or ENTNIT) for AT ₁ R-Abs in the AT ₁ R second extracellular loop ²⁹ . The bound autoantibodies are then either directly labelled with reporter enzymes (direct ELISA) or by a secondary labelled antibody coupled to a reporter enzyme (indirect ELISA) for quantification. Internal laboratory assays have been developed as well as commercially available kits ( Table 1 ). Piazza et al. compared in a study with patients diagnosed with APA two of the commercially available Sandwich-ELISA kits used for AT ₁ R-Abs determination ^{35 36} by CellTrend and Cusabio ³⁷ . The results of both assays were consistent and both demonstrated higher AT ₁ R-Ab levels in patients diagnosed with an APA compared with healthy controls ³⁷ . In contrast, the same CellTrend kit and a second kit from ELISA-Creative Diagnostics revealed the detection of contrasting AT ₁ R-Ab levels depending on the assay used for patients with pre-eclampsia relative to controls ³⁸ . This highlights the inherent drawbacks of assays based on the detection of AT ₁ R-Ab concentrations and not AT ₁ R bioactivity. To investigate if AT ₁ R-Abs activate the AT ₁ R, assays have been developed, which measure cultured spontaneously beating neonatal rat cardiomyocytes in response to immunoglobulins G (IgGs) and the perfused rat cremaster arteriole assay to measure losartan-sensitive antibody-mediated vasoconstriction ^{24 39} . In vitro experiments using genetically engineered cells (for example, Chinese hamster ovary cells (CHO), or human bone osteosarcoma cells (U2OS)) stably transfected with human AT ₁ R) to measure the functional activation of the AT ₁ R in response to whole serum or purified IgGs have also been used ^{37 38 39 40} . Upon receptor activation, the transfected U2OS or CHO cells mediate chemiluminescent or fluorescent signals that can be quantitatively measured. Of note is that many cell-based assays used to quantify AT1R-Ab agonistic activity measure responses mediated by the β-arrestin signaling pathway. In addition, the production of aldosterone or increased expression of the aldosterone synthase gene ( CYP11B2 ) in response to treatment with IgG fractions or whole serum has been measured using human adrenocortical carcinoma cell line (HAC15) ³⁷ .

Table 1 Angiotensin II type 1 receptor autoantibody measurements in primary aldosteronism.

ELISA
Indirect ELISA using immobilized AT 1 R peptides of extracellular loop 2
Rossitto et al. 2013 51	AT 1 R-Ab level in patients with PA (n=46) higher than in NT (n=45) AT 1 R-Ab level in patients with APA (n=26) 2-fold higher than IHA (n=20)
Kem et al. 2014 39	elevated AT 1 R-Ab levels in 4 of 13 patients with PA (31%)
Li et al. 2015 40	elevated AT 1 R-Ab levels in 5/12 (42%) of BAH, 3/13 (23%) of APA and 1/15 (7%) of NT
Sandwich-ELISA with full-length AT 1 R
Sabbadin et al. 2018 35	Human angiotensin II receptor 1 antibody, ATIIR1 Ab ELISA Kit (Cusabio, Wuhan, China) elevated AT 1 R-Ab titers in PA (n=44) than in NT (n=18) AT 1 R-Ab level comparable between APA and BAH
Piazza et al. 2019 37	Human angiotensin II receptor 1 antibody, ATIIR1 Ab ELISA Kit (Cusabio, Wuhan, China) AT 1 R-Ab levels in patients with APA (n=12) compared to NT (n=7)
Piazza et al. 2019 37	CellTrend, Luckenwalde, Germany higher AT 1 R-Ab levels in patients with APA (n=27) compared to NT (n=7)
Williams et al. 2019 38	CellTrend, Luckenwalde, Germany equal levels of AT 1 R-Abs in APA (n=40), BAH (n=40), PH (n=40), NT (n=25)
Williams et al. 2019 38	Creative Diagnostics equal levels of AT 1 R-Abs in APA (n=40), BAH (n=40), PH (n=40), NT (n=25)
Functional AT 1 R agonist measurements
Cell-based functional assays
Williams et al. 2019 38	SERUM: without losartan: no group differences between APA (n=40), BAH (n=40), PH (n=40), NT (n=25) with losartan: AT 1 R-bioactivity of APA, PE and NT reduced in comparison to BAH Purified IgG: without losartan: AT 1 R-bioactivity of BAH higher than of APA; AT 1 R-bioactivity of BAH, PH, PE all higher than NT with losartan: AT 1 R-bioactivity not abolished by losartan
Kem et al. 2014 39	SERUM: elevated AT 1 R-bioactivity through serum from PA patients compared to NT, mainly suppressed by losartan/candesartan
Li et al. 2015 40	SERUM: 15/25 (60%) of patients with PA showed autoantibody-mediated AT 1 R-bioactivity: 9/12 (75%) patients with BAH; 6/13 patients with APA (46%) NT were negative for AT 1 R-Ab induced AT 1 R-bioactivity losartan inhibited AT 1 R-bioactivity
Piazza et al. 2019 37	SERUM: values below threshold; AT 1 R-bioactivity from APA patient serum does not differ from NT
Perfused rat cremaster arteriole assay
Kem et al. 2014 39	SERUM: higher cremaster arteriole contractility in patients with PA (n=3) than NT which could be inhibited by losartan
CYP11B2-mRNA expression in HAC15 cells
Piazza et al. 2019 37	Purified IgG: purified IgG from patients with an APA (n=10) induced a 40% increase in CYP11B2-mRNA expression

Ab: Antibody; APA: Aldosterone producing adenoma; AT ₁ R: Angiotensin II type 1 receptor; BAH: Bilateral adrenal hyperplasia; ELISA: Enzyme-linked immunosorbent assay; HAC15: Human adrenocortical cell line; NT: Normotensive controls; PH: Primary hypertension.

AT ₁ R-Abs and Pre-Eclampsia

Pre-eclampsia is the sudden appearance of hypertension in pregnant women after 20 weeks’ gestation week that is associated with increased risks of long-term hypertension, stroke, cardiovascular morbidity and proteinuria for the mother and uteroplacental dysfunction, preterm birth, fetal distress and fetal death for the unborn child ^{24 41} . In 1999, Wallukat et al. were the first who reported the presence of AT ₁ R-Abs in pregnant women with pre-eclampsia compared to controls ²⁴ . The purified IgG fractions from patients with preeclampsia demonstrated losartan-suppressible AT ₁ R stimulation and identified the amino acid sequence AFHYESQ in the secondary extracellular loop of the AT ₁ R as the possible binding site for these autoantibodies ²⁴ . Peptides against the AFHYESQ epitope successfully abolished AT ₁ R-Ab mediated activation of the AT ₁ R suggesting a potential target for the treatment of patients with pre-eclampsia ^{42 43} . Despite postpartum persisting AT ₁ R-Ab levels in 17% of women with previous pre-eclampsia ⁴⁴ , the drop in AT ₁ R-Ab levels after giving birth and the correlation of AT ₁ R-Ab levels with the severity of the disease indicated a role for AT ₁ R-Abs in pre-eclampsia ^{24 45} . Numerous studies subsequently supported the findings of Wallukat et al. with reports of AT ₁ R-Abs in pregnant women with pre-eclampsia ^{43 46 47} . Diverse studies have reported a role for AT ₁ R-Abs in pre-eclampsia in mediating intracellular Ca ²⁺ release ⁴³ , induction of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and reactive oxygen species (ROS) production leading to the activation of the transcription factor nuclear factor Kappa B (NFκB) activation ⁴⁶ and initiation of vasoconstriction ⁴⁷ , all of which were inhibited by losartan. Further in vivo experiments ascribed a causative role for AT ₁ R-Abs in the development of pre-eclampsia. Zhou et al. infused pregnant mice with human total IgG or affinity purified AT ₁ R-Abs which resulted in the characteristic symptoms of pre-eclampsia including proteinuria, hypertension and abnormal appearance of the placenta ⁴² . In addition, Wenzel et al. injected angiotensin II, purified rabbit AT ₁ R-Abs or both in pregnant rats which induced a phenotype resembling pre-eclampsia when a combination of both, angiotensin II and purified AT ₁ R-Abs were applied ⁴⁸ . In contrast to the aforementioned findings, in a population of Mexican-Mestizo women diagnosed with pre-eclampsia no AT ₁ R-Abs were found ⁴⁹ . Despite an apparently clear role for AT ₁ R-Abs in pre-eclampsia, a recent retrospective analysis of 485 women with pre-eclampsia concluded that AT ₁ R-Ab levels alone are not sufficient to predict hypertension in such patients, and additional risk factors for hypertension should be considered ⁵⁰ .

AT ₁ R-Abs in Primary Aldosteronism

Recently, AT ₁ R-Abs were also detected in patients diagnosed with PA suggesting a possible role of AT ₁ R activation in aldosterone production in some patients. Rossitto et al. described for the first time AT ₁ R-Abs in patients with PA which was subsequently reported by other research groups ^{39 40 51} . In this study, the AT ₁ R-Ab levels of serum from 46 patients with PA (26 APA, 20 BAH) as well as 62 patients with primary hypertension (PH) was measured by an indirect ELISA assay using an immobilized peptide of the second extracellular loop of AT ₁ R. Thirteen pregnant women with preeclampsia and 45 normotensive control patients (NT) were used as positive and negative controls, respectively. Patients with PA and PH showed significantly higher AT ₁ R-Ab levels than NT, with higher levels in patients with PA than in PH ⁵¹ . In 92.3% of patients with APA, AT ₁ R-Abs were detected comprising a concentration 2-fold higher than of patients diagnosed with BAH or PH, despite comparable blood pressure levels ( Table 1 ) ⁵¹ . Interestingly, the AT ₁ R-Ab levels of patients with APA and women with pre-eclampsia were similar (3.43±1.20 vs. 3.66±1.79) ⁵¹ . Another indirect ELISA using the peptide AFHYESQ located in the extracellular loop 2 was performed in the studies of Kem et al. and Li et al. ^{39 40} . Kem et al. described elevated AT ₁ R-Ab levels in 4 of 13 patients with PA (31%) while the latter study focused on the subtypes of PA and measured elevated AT ₁ R-Ab levels in 42% of BAH, 23% of APA and 7% of NT which contradicts the initial findings of the same group ⁴⁰ .

Using a commercially available ELISA kit with full-length AT ₁ R as antigen, Sabbadin et al. found higher AT ₁ R-Ab levels in patients with PA compared with healthy controls which is in accordance with previous findings ³⁵ . However, the authors could not distinguish patients with APA (n=15) and BAH (n=29) ³⁵ . In contrast, a larger study comprising a cohort of 80 patients with PA (40 APA, 40 BAH), 40 with PH, 23 with pre-eclampsia and 25 NT observed equal levels of AT ₁ R-Abs in all groups except for patients with pre-eclampsia using two different commercial available ELISA kits ( Table 1 ) ³⁸ . Using the same CellTrend ELISA kit, Piazza et al. described higher AT ₁ R-Ab levels in patients with APA (n=27) compared to healthy controls (n=7) ³⁷ . Overall, it is clear that these studies using ELISA-based assays have yielded highly contrasting results.

AT ₁ R-Ab levels pre- and post-adrenalectomy were investigated in 14 patients with APA ³⁷ who were biochemically cured following surgery according to the PASO criteria ⁵² . The authors found no significant decrease in AT ₁ R-Ab levels at one month after adrenalectomy indicating that the resected adrenal was not the source of antigens stimulating the immune response ³⁷ .

When summarizing the results of AT ₁ R-Ab quantification by ELISA it can be stated that there is a high variability in the AT ₁ R-Ab levels of patients with PA with studies reporting contrasting AT ₁ R-Ab levels for the different subtypes of PA ^{39 51} . Of note is that elevated AT ₁ R-Ab levels were also described in healthy individuals without initiating subsequent AT ₁ R activation ⁴⁰ . Furthermore, Kem et al. found for instance more frequently elevated AT ₁ R-bioactivity in patients with PA by a cell-based assay using AT ₁ R-transfected CHO cells than autoantibodies were quantified by ELISA which is in agreement with a second study from the same group ^{39 40} . The authors hypothesized that the use of a linear peptide for the secondary extracellular loop instead of full-length AT ₁ R could miss other potential binding sites for the autoantibodies in ELISA ³⁹ . This is supported by the findings of multiple binding sites for IgG on the AT ₁ R ⁵³ and the recommendations to avoid using immobilized peptides in ELISA for GPCR-autoantibody detection ⁵⁴ . However, using the full-length AT ₁ R does not guarantee the functional activity of the captured autoantibodies which can be demonstrated in cell based-functional assays using whole serum and/or purified IgG ^{38 39 40} .

The rat cremaster arteriole assay or AT ₁ R-transfected CHO cells both demonstrated elevated receptor activation when exposed to serum of patients with PA compared to controls, which was largely normalized by the AT ₁ R-blockers losartan or candesartan ³⁹ . Similarly, candesartan was able to reduce AT ₁ R-Ab-stimulated aldosterone production in vitro in HAC15 cells ³⁹ . AT ₁ R-activity is more frequently observed in patients diagnosed with BAH (75%) compared to APA patients (46%), while serum of NT induced no AT ₁ R-bioactivity ⁴⁰ . However, one research group failed to measure AT ₁ R-bioactivity in patients with APA and controls using this assay which might be due to the lack of activating function of the autoantibodies ³⁷ . AT ₁ R-transfected U2OS cells also showed no group differences when incubated with serum of patients with APA, BAH or pre-eclampsia or NT in the absence of losartan ³⁸ . Notably, the administration of losartan only barely reduced AT ₁ R-activation in the BAH group in comparison to APA, pre-eclampsia and NT indicating a losartan-independent activation pathway of the receptor ³⁸ . Purified IgGs also caused an increase in AT ₁ R-bioactivity in the BAH compared to APA group, independent of losartan ³⁸ . This can be explained by different binding sites for IgG and losartan at the AT ₁ R while angiotensin II and losartan share a common binding site on the AT ₁ R ⁵³ .

Autoantibody levels were not correlated with age, gender, BMI, blood pressure, baseline aldosterone and aldosterone-to-renin ratio (ARR) ^{35 51} . Mineralocorticoid receptor antagonist (MRA) treatment of patients with PA was also not associated with AT ₁ R-Ab levels ³⁵ . Agonistic AT ₁ R-Abs displayed vasoconstrictive effects and correlate with the mean arterial pressure ^{39 40} . The previously mentioned elevated agonistic activity of AT ₁ R-Abs in patients diagnosed with BAH and the higher responsiveness to angiotensin II in those patients ⁵⁵ suggest an allosteric function of the autoantibodies which promotes the binding angiotensin II to its receptor. This is supported by the reduction of aldosterone levels after captopril-challenge in AT ₁ R-Ab-positive compared with AT ₁ R-Ab-negative patients with hypertension or PA ⁵¹ . In addition, treatment of HAC15 cells with angiotensin II and affinity-purified IgGs increased aldosterone production compared with angiotensin II treatment alone ³⁹ .

As previously indicated, upon activation, the AT ₁ R can initiate two distinct independent signaling pathways. More prominent is the G-protein signaling pathway in which PIP ₂ is cleaved to DAG and IP ₃ resulting in increased intracellular Ca ²⁺ and aldosterone production ¹⁵ . Secondly, biased signaling mediated by β-arrestin1 is independent of G-proteins and may occur in parallel to the G-protein signaling pathway ¹⁵ . Aldosterone production appears to be stimulated via extracellular signal-regulated kinase (ERK)-mediated activation of steroidogenic acute regulatory protein (StAR) to promote transfer of cholesterol to mitochondria ¹⁵ . When AT ₁ Rs are activated by angiotensin II, both pathways appear to be initiated ¹⁵ . This has been demonstrated in vivo in rats with overexpression of β-arrestin1, which showed suppressed aldosterone production only through the administration of candesartan or valsartan and not by losartan or irbesartan despite belonging to the same pharmacological class ^{56 57} .

In conclusion, some evidence supports a pathological role for autoantibodies against the AT ₁ R in different diseases. It has not been established if AT ₁ R-Abs play a causative role or are a consequence of the pathology. Further studies are warranted to address the functional relevance of AT ₁ R-Abs in PA and the significance of AT ₁ R biased signaling.

Funding Statement

Funding: M. Reincke is supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 694913). T. A. Williams and M. Reincke are supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) Projektnummer: 314061271-TRR 205.