Transforming growth factor-β1
Allergic asthma is characterized by airway hyperreactivity (AHR), eosinophilic airway inflammation and elevated serum IgE levels. T-helper 2 (Th2) cells play a critical role in the pathogenesis of asthma, but the immunological mechanisms that inhibit Th2 cell function in vivo are not well understood. Conflicting results regarding the protective role of Th1 cytokines and TGF-β in asthma have been reported. To further investigate the role of TGF-β1 in asthma, we examined mice heterozygous for deletion of the TGF-β1 gene (TGF-β1+/– mice) in a murine asthma model. While TGF-β1+/– mice seem phenotypically normal, they express only about 30% of wild type TGF-β1 protein levels as shown before. The reduced expression of TGF-β1 is accompanied by a strikingly increased eosinophilic inflammation and mucus secretion in response to ovalbumin (OVA) sensitization. Moreover, TGF-β1+/– mice develop significantly enhanced Th2-cytokine levels, decreased IFN-γ production and increased levels of OVA-specific IgE in serum. In contrast, AHR in response to methacholine is not altered significantly. Our data demonstrate that reduced expression of TGF-β1 exacerbates pathology in an experimental asthma model and support the view that the elevated levels of TGF-β1 in asthmatic airways might be, at least in part, a result of anti-inflammatory compensation by this cytokine.
Allergic asthma is an inflammatory disease dependent on Th2-type immune responses and characterized by airway hyperreactivity (AHR) to specific and nonspecific stimuli. So far therapeutic attempts focused on the reconstitution of the balance between Th1 and Th2 cells, since IFN-γ-secreting Th1 cells have been proposed to protect against asthma and allergic disease by dampening the activity of Th2 effector cells. Evidence for this hypothesis was mainly derived from epidemiological studies, demonstrating a negative correlation between the asthma incidence and a cleaner environment with fewer Th1-promoting childhood infections in westernized countries 1. Moreover, recent animal experiments using bacterial components, plasmid DNA, CpG oligonucleotides and other approaches 2–4 demonstrated that induction of Th1 cells is associated with the inhibition of Th2 cell development and effectively down-modulates AHR and airway inflammation.
On the other hand, both Th1 and Th2 cells can be found in the lungs of asthmatic patients 5, and the administration of IL-12 has failed to show significant effects in the early and late phase of the asthmatic response 6. Moreover, adoptive transfer of antigen-specific Th1 cells into SCID mice and immunized BALB/c mice failed to inhibit Th2 cell-induced AHR and inflammation, even when Th1 cells were given in great excess 7. Unexpectedly, in this model, Th1 cells were even harmful and increased airway inflammation. These conflicting data underline the need for a better understanding of the regulatory mechanisms and cytokines involved in the development of AHR and airway inflammation.
One key regulator in the maintenance of immunological homeostasis is TGF-β, a pleiotropic cytokine with significant anti-inflammatory and immunosuppressive properties. TGF-β inhibits the production of proinflammatory cytokines from macrophages, B cells, and T cells and is a potent inhibitor of T cell-mediated immune responses, both in vitro8–10 and in vivo11–13. The essential role for TGF-β1 in the maintenance of immune homeostasis is demonstrated in TGF-β1–/– mice, which develop a severe, multifocal inflammatory response in all of the pups that are born alive 14, 15. In contrast, mice heterozygous for deletion of TGF-β1 gene (TGF-β1+/– mice) seem phenotypically normal and show no significant difference in the lifespan compared to wild type (WT) animals 16.
In a previous study with T cells genetically engineered to produce TGF-β, we demonstrated that, in contrast to Th1 cells, these TGF-β-producing cells were able to abolish AHR and airway inflammation induced by OVA-specific Th2 cells 17. In humans, the levels of TGF-β1 in the airways of asthmatics are reported to be higher than in normal subjects 18. So far, it is not clear whether this resembles the profibrotic effects of TGF-β in the airways of asthmatics or a protective anti-inflammatory effect of this cytokine in allergic asthma.
To further investigate the role of TGF-β1 in asthma, we examined mice heterozygous for deletion of TGF-β1 gene (TGF-β1+/– mice) in a murine asthma model. While TGF-β1+/– mice seem phenotypically normal 14, Tang et al. 19 have demonstrated that TGF-β1+/– mice express only about 30% of TGF-β1 protein levels compared to WT mice. This finding was confirmed in our study. In these animals, the inflammatory response to OVA sensitization is clearly increased and mucus secretion is enhanced as compared to WT controls. Moreover, the Th2 cytokines IL-4, IL-5 and IL-13 are strikingly elevated, which is accompanied by an increase in IgE production in TGF-β1+/– mice. In contrast, IFN-γ was significantly decreased. Although AHR in response to methacholine is elevated at high methacholine concentrations, these differences do not reach the level of significance, demonstrating that 30% of TGF-β1 expression as presented in TGF-β1+/– mice is sufficient to control AHR in response to OVA sensitization, while it is not sufficient to control airway inflammation, mucus secretion and IgE production. Bearing in mind the profibrotic potential of TGF-β1, our data support, at least in part, a protective role of TGF-β1 in asthma.
Decreased TGF-β1 protein levels in mice heterozygous for deletion of the TGF-β1 gene (TGF-β1+/– mice)
Although TGF-β1+/– mice seem phenotypically normal, TGF-β1 protein levels in TGF-β1+/– mice were significantly reduced and ranged from 26±15% of normal lung, to 29±18% of normal serum, to 29±6% of normal spleen (Fig. 1). These data confirm the data of Tang et al. 19, who showed a comparable or even stronger reduction of TGF-β1 protein levels in mice heterozygous for deletion of the TGF-β1 gene compared to WT animals (10-30%). OVA sensitization did not significantly change TGF-β1 protein levels in serum, lung and spleen of TGF-β1+/– or WT mice.
Stronger inflammatory response and mucus production in OVA-immunized TGF-β1+/– mice as compared to WT mice
After having demonstrated that loss of one TGF-β1 allele causes significantly reduced TGF-β1 protein levels, we tested the effect of reduced TGF-β1 expression in OVA-immunized animals in a murine asthma model. First, TGF-β1 WT and TGF-β1+/– mice were sensitized twice, i.p. to OVA protein adsorbed to Alum, and then challenged intranasally (i.n.) with OVA in PBS six times over a course of 24 days (see Sect. 4). Two days after the last i.n. challenge with OVA, mice were killed, and bronchoalveolar lavage (BAL) was performed. The inflammatory response to OVA sensitization was strikingly increased in the TGF-β1+/– animals with lower TGF-β1 levels as compared to WT controls (Fig. 2). Cell numbers in the BAL fluid of OVA-sensitized TGF-β1+/– animals exceeded those of WT animals almost threefold (35±6.1×105vs. 13±1.4×105, p<0.01) and the percentage of eosinophils in the BAL was increased almost fourfold in TGF-β1+/– animals as compared to controls (26.3±4.5×105vs. 6.8±1.2×105, p<0.01).
In addition to examination of the BAL fluid, we compared lung histology of OVA-sensitized TGF-β1+/– animals and OVA-sensitized WT controls. For this, mice were killed 2 days after the last i.n. OVA challenge, lungs were removed and lung histology was examined after fixation and staining of one part of the lung with hematoxylin and eosin (H&E), and another part with periodic acid-Schiff's reagent (PAS) for better mucus demonstration. Lung sections from WT mice that were immunized with OVA showed strong airway inflammation with dense peribronchiolar and perivascular infiltrates, consisting of lymphocytes, eosinophils and some neutrophils (Fig. 3). The inflammatory response to OVA sensitization was stronger in TGF-β1+/– animals with lower TGF-β1 protein levels as compared to WT controls. Interestingly, lungs of naive TGF-β1+/– control mice revealed a slight cellular infiltrate that was not seen in WT controls. PAS staining demonstrated strong mucus production in OVA-immunized animals as quantified by the mucus index described in Sect. 4 (Fig. 4). Mucus production was increased in the TGF-β1+/– animals as compared to WT controls. Our data demonstrate that reduced expression of TGF-β1 leads to increased airway inflammation, airway eosinophilia and mucus production after OVA sensitization.
No significant differences in airway hyperreactivity in TGF-β1+/– mice as compared to WT mice
The effect of TGF-β1 expression on AHR in response to methacholine was tested in a whole-body-plethysmograph for mice 1 day after the last i.n. challenge with OVA in the 24-day immunization protocol. The response to methacholine was similar in OVA-sensitized TGF-β1+/– animals as in WT animals, although there was a slight increase in AHR at 80 mg/ml methacholine (Fig. 5). This difference did not reach the level of significance.
Reduced TGF-β1 expression leads to strikingly increased Th2 cytokine levels and decreased IFN-γ levels after OVA sensitization
Allergic asthma is a Th2 cytokine-dominated disease, and the pathology in asthma is closely related to the Th2 cytokines IL-4, IL-5 and IL-13 that all play a crucial role for the development of AHR and inflammation in this disease. To investigate the effect of reduced TGF-β1 protein levels on cytokine production, cells from OVA-sensitized TGF-β1+/– and WT animals were isolated from spleen and restimulated in vitro with OVA. Levels of IL-4, IL-5, IL-10, IL-13 and IFN-γ in culture supernatants were determined by ELISA. IL-4, IL-5 and IL-13 were strikingly increased in TGF-β1+/– mice compared to WT controls, while IFN-γ levels were significantly decreased. These data demonstrate a Th2 shift in TGF-β1+/– mice and show that in our system Th1 and Th2 responses are not equally suppressed by TGF-β1. IL-10 concentrations are increased in WT mice compared to TGF-β1+/– mice, although these differences were not significant (Fig. 6).
Effects of reduced TGF-β1 expression on IgE, IgG1 and IgG2A levels after OVA sensitization
One hallmark of allergic asthma is the production of allergen-specific IgE. To investigate the effect of reduced TGF-β1 expression on OVA-specific IgE, IgG1 and IgG2a production, blood was taken 2 days after the last i.n. challenge with OVA and OVA-specific IgE levels were measured by ELISA. We found that IgE levels in OVA-sensitized TGF-β1+/– mice were significantly higher than those in OVA-sensitized TGF-β1 WT controls (p⩽0.05). Furthermore, we observed increased IgG1 levels and significantly reduced IgG2a levels in OVA-sensitized TGF-β1+/– mice, also indicating a stronger Th2 response in TGF-β1+/– mice compared to WT controls (Fig. 7). These results confirm the selective increase of Th2 cytokines in TGF-β1+/– mice.
TGF-β1 is a member of a large family of evolutionarily conserved proteins with highly pleiotropic properties 20. A loss-of-function mutation in TGF-β1 mice results in a severe, multifocal inflammatory response in all of the pups that are born alive 14, 15, which demonstrates the essential role for TGF-β1 in maintenance of immune homeostasis. While TGF-β1+/– mice seem phenotypically normal, we and others 19 have demonstrated that TGF-β1+/– mice express only about 30% of WT TGF-β1 protein levels. In our study, lower expression of TGF-β1 in OVA-sensitized TGF-β1+/– animals led to strikingly increased levels of the Th2 cytokines IL-4, IL-5 and IL-13, which all play a crucial role in the pathogenesis of allergic asthma. The increased Th2 cytokine levels were accompanied by an increased inflammatory response in the lung, enhanced mucus production and significantly elevated serum IgE and reduced IgG2a levels. At the same time, IFN-γ, which is thought to counterbalance Th2 cytokine effects, was significantly decreased. These data demonstrate a Th2 shift in TGF-β1+/– mice and show that, in our system, Th1 and Th2 responses are not equally suppressed by TGF-β1.
IL-10 concentrations are decreased in TGF-β1+/– mice compared to WT animals. The production and action of IL-10 and TGF-β, which both are able to down-regulate immune responses, likely involve a positive feedback loop in which TGF-β enhances IL-10 production and vice versa. For example, treatment of colitis by administration of a plasmid encoding TGF-β provides a beneficial effect via stimulation of IL-10 production 21. This effect may be due to the ability of Smad3, a transcription factor required for signaling downstream of the TGF-βR, to interact with GATA3 and form a complex contributing to the transcriptional activation of the IL-10 promotor 22. Intestinal epithelial cells transgenic for IL-10 display an increase in TGF-β production 23, supporting previous in vitro data demonstrating that exogenous IL-10 stimulates the production of TGF-β 24. In contrast, lymphocytes from mice deficient for IL-10 produce less TGF-β than cells from WT mice 23. Similarly, reduced expression of TGF-β in TGF-β1+/– mice might lead to the reduced production of IL-10 in TGF-β1+/– mice observed in this study.
Our data support the hypothesis that the elevated levels of TGF-β1 in asthmatic airways 18 might be, at least in part, a result of anti-inflammatory compensation.
In addition to the data from this study, our previous study clearly supports a protective role of TGF-β1 in allergic asthma. Here we have demonstrated that transfer of OVA-specific T cells engineered in vitro to express latent TGF-β abolished AHR and airway inflammation induced by OVA-specific Th2 cells in severe combined immunodeficient (SCID) and BALB/c mice 17. In this model, the inhibitory effect on AHR was abolished when mice were treated with a neutralizing monoclonal antibody to TGF-β.
Similarly, a recent study identified the T cells as a central effector cell of TGF-β1-mediated regulation of AHR 25. Upon antigen-specific challenge, mice with impaired TGF-β1 signaling in T cells demonstrated increased AHR and lung inflammation dominated by eosinophils compared with WT mice. In contrast to these data, we did not see significant changes in AHR in TGF-β1+/– mice with reduced expression of TGF-β1. This finding demonstrates that 30% of TGF-β1 expression, as presented in TGF-β1+/– mice, is sufficient to control AHR in response to OVA sensitization. In contrast, it does not appear to be sufficient for controlling airway inflammation, mucus secretion, and cytokine and IgE production.
Since TGF-β1 has been shown to be produced by a variety of cells within the lungs of asthmatic patients 26–28, the higher levels of TGF-β1 in the airways of asthmatics as compared to normal subjects 18 are very likely produced by different kinds of cells, not only the T cells. TGF-β1 is produced by T and B lymphocytes, macrophages, dendritic cells (DC) and platelets 29–31. In the lung of mice, TGF-β1 mRNA is found in bronchiolar epithelium, Clara cells, mesenchymal cells, vascular endothelium, and alveolar cells, including macrophages 32. Additionally, a very high expression of TGF-β1 messenger RNA transcripts is found in the smooth muscle cells of the large vessels 33. In humans, bronchial epithelial cells contain a large amount of TGF-β1 protein 34. The TGF-β isoforms and their receptors have also been demonstrated in macrophages 32, 34–36, mesenchymal cells 32, vascular and airway smooth muscle cells 34–36 and bronchial glands 37. While some studies also detected TGF-β1 in normal epithelium 32, 35, 37, others 36 reported epithelial presence of TGF-β1 only in association with fibrosis.
The presence of TGF-β in the normal lung suggests its participation in the normal regulation of physiological processes to maintain lung homeostasis. These functions may include local immunomodulation, regulation of cell proliferation and differentiation, as well as the control of normal tissue repair. On this background, the investigation of TGF-β1+/– mice with reduced TGF-β1 expression in the lung caused by different cell types resembles the natural situation more than an investigation of transgenic mice with impaired TGF-β signaling.
In our study, lower TGF-β1 levels were accompanied by increased mucus secretion. Low levels of mucus are produced in the normal lung and mucus metaplasia with mucus hypersecretion is a characteristic feature of asthma. While the role that TGF-β1 plays in these responses has not yet been evaluated, a variety of inflammatory mediators, including the Th2 cytokines IL-4, IL-5, IL-13 and IL-9, are believed to contribute to the pathogenesis of the observed mucus abnormalities in this disorder 38–41. Therefore, it is likely that the strikingly increased Th2 cytokine levels in TGF-β1+/– mice are responsible for the increased mucus production, so that TGF-β1 indirectly controls mucus production via the control of other cytokines.
Another key feature of allergic asthma is the elevated level of allergen-specific IgE. In this study, we have shown that reduced TGF-β1 levels lead to increased antigen-specific IgE levels in antigen-sensitized mice. The mechanism for this phenomenon is not clear, but increased levels of IL-4 and IL-13 in TGF-β1+/– mice might be responsible, since IL-4 and IL-13 stimulate B cells to produce IgE 42. Another possible mechanism is the regulation via Id2. Recently, evidence has been provided for the inhibitory and selective role of Id2 in IgE class switch recombination in response to TGF-β1. Id2 might act as a molecular safeguard to suppress the IgE class switch recombination to prevent serious complications such as allergic hypersensitivity during the normal course of immune responses 43.
To our knowledge, so far no data have been published on the immune response of TGF-β1+/– mice in asthma or other allergic diseases. Moreover, only very little is known about the immune response of TGF-β1+/– mice. One study on cardiac transplant arteriosclerosis showed that TGF-β1+/– transplant recipients had an increased expression of the transcription factor STAT 4, IFN-γ and IL-2 and unaltered or reduced expression of the transcription factor STAT 6, IL-4, and IL-10 44. In contrast to the sparse data on the immune response of TGF-β1+/– mice, several tumor studies have been done with TGF-β1+/– mice, which have demonstrated a significant role of TGF-β1 as tumor suppressor protein 19, and in the carcinogenesis of lung 45–47 or other carcinomas 48, 49. Although TGF-β1+/– mice have also been used in different studies with varying disease models 50–52, the immune response of TGF-β1+/– mice has not been in the focus of these studies.
On the basis of the data from our animal study with TGF-β1+/– mice, the elevated levels of TGF-β1 in asthmatic airways might be, at least in part, a result of protective, anti-inflammatory compensation in asthmatic airway inflammation. In addition, as TGF-β has been implicated in tissue repair 53, the high levels of TGF-β in asthmatic airways might be involved in repair processes of asthmatic airways that are damaged by inflammation. TGF-β1 in the airways of asthmatics could thus function as a healing molecule, promoting the process of tissue repair and diminishing airway inflammation. Under ideal circumstances the actions of TGF-β might lead to the restoration of normal tissue architecture. On the other hand, TGF-β1 is clearly a profibrotic cytokine 54. Therefore, under circumstances such as persistent activity of TGF-β1 induced by chronic inflammation, which might be caused by repeated stimulation with allergens, TGF-β1 expression might lead to the detrimental effect of TGF-β1 known as tissue fibrosis or airway remodeling, which results in chronic airway obstruction.
Although TGF-β1 might have damaging effects, probably depending on concentration and duration of tissue contact, our data and the data from previous studies clearly support the protective potential of TGF-β1 in allergic asthma.
Materials and methods
B6.129S2-TGF-β1tm1Doe were obtained from the Jackson Laboratory (Bar Harbor, ME) and bred and maintained in our animal facility at the BioCenter in Halle/Saale (Germany). These mice have been backcrossed to C57BL/6J for at least 22 generations. C57BL/6J were used as WT mice. For all experiments sex-matched 8–10-week-old animals were used; experiments involved groups of four to five mice and were performed according to institutional and state guidelines. The local Committee on Animal Welfare (Bezirksregierung Halle) approved animal protocols used in this study.
Screening of mice by PCR
Neonates from heterozygote mice were screened for the TGF-β1 WT and KO allele by genomic PCR. Briefly, DNA was extracted from tail biopsies of 4-week-old mice. The DNA was examined by PCR analysis using two primer pairs (A: 5′-GAGAAGAACTGCTGTGTGCG-3′ and 5′-GTGTCCAGGCTCCAAATATAGG-3′; and B: 5′-AGACAATCGGCTGCTCTGAT-3′ and 5′-GTGTCCAGGCTCCAAATATAGG-3′), the first pairing to exon 6 sequences of TGF-β1, the second amplifying neomycin-specific sequences in the mutant mice. PCR synthesis was run for 35 cycles (94°C for 1 min, 60°C for 30 s and 72°C for 2 min) followed by final extension at 72°C for 10 min. PCR products were visualized on agarose gels.
Measurement of TGF-β1 protein levels in mice heterozygous for TGF-β1 gene and WT animals
Freshly isolated tissues or serum were homogenized and extracted with acid-ethanol solution as described 55. ELISA for TGF-β was performed with DuoSet® (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. The detection limit was 8 pg/ ml.
Immunization of mice with OVA
Mice were sensitized to OVA protein for the induction of AHR and airway inflammation using the following protocol: OVA (20 μg) adsorbed to 2 mg aluminium potassium sulfate (Alum) was administered i.p. on days 1 and 14, followed by 20 μg OVA in 40 μl normal saline given i.n. on days 14, 16, 17 and 20, 21, 24. Control mice received i.p. injections of alum alone and normal saline i.n. One day after the last i.n. challenge (day 25), AHR was measured in conscious mice after inhalation of increasing concentrations of methacholine (see below). On day 26, mice were killed; blood was taken, the left lung was tied off for histology and BAL was performed in the right lung.
Measurement of airway responsiveness
Airway responsiveness was assessed by methacholine-induced airflow obstruction from conscious mice placed in a whole body plethysmograph (model PLT UNR MS, Emka Technologies, Paris). Pulmonary airflow obstruction was measured by Penh using the following formula: Penh = [(Te/RT-1) × (PEF/PIF)], where Penh = enhanced pause (dimensionless), Te = expiratory time, RT = relaxation time, PEF = peak expiratory flow (ml/ s), and PIF = peak inspiratory flow (ml/s) 56. Enhanced pause (Penh), minute volume, tidal volume, and breathing frequency were obtained from chamber pressure, measured with a transducer connected to amplifier module (model AC264), and analyzed by system XA software (version 1.565). Mice were exposed for 2 min to aerosolized 0.9% NaCl produced by a sonicator (model LS 290–990N), followed by increasing amounts of methacholine. Data are expressed as the percentage of baseline Penh values after 0.9% NaCl exposure.
Collection of BAL fluid and lung histology
Animals were killed by CO2 asphyxiation. Left lung was tied off for histology. The trachea was cannulated and the right lung was lavaged three times with 400 μl PBS with 1% BSA. Cells in the lavage fluid were counted using a hemocytometer, and BAL cell differentials were determined on slide preparations stained with May-Grünwald-Giemsa in a blinded fashion (Merck, Darmstadt, Germany). At least 200 cells were differentiated by light microscopy based on conventional morphological criteria. Left lung was fixed in 10% formalin; one part was stained with H&E (Merck) and the other part with Alcian blue-PAS (Sigma, Taufkirchen, Germany). The degree of mucus production in the lungs was assessed by a mucus score, which is a product of PAS-positive bronchi and a factor correlating to the number of positive goblet cells (<30% PAS-positive goblet cells factor 1; 30–60% factor 2; >60% factor 3).
To detect or exclude fibrosis, lung tissue of WT and TGF-β1+/– mice was stained with methyl violet (Merck) in annilin alcohol and treated with Lugol's solution (iodine-potassium iodide solution, Merck). Fibrosis was not detectable in any of the investigated lungs.
Splenocyte culture and assay for cytokines
One day after the airway function test, spleens were removed and passed though a nylon mesh. Single-cell suspensions of splenocytes (5×106 cells/ ml) were stimulated in vitro by incubation with 200 µg/ml OVA in culture medium (RPMI medium supplement with 10% FCS, 100 U/ml penicillin, 100 U/ml streptomycin). Supernatants were collected in triplicate after 36 h and ELISA for IL-4, IL-5, IL-10, IL-13 and IFN-γ were done using DuoSet® (R&D Systems) according to manufacturer's instructions.
Serum levels of OVA-specific IgE, IgG1 and IgG2A were measured with a sandwich ELISA according to a standard protocol. Briefly, 96-well microtiter plates (Nunc, Roskilde, Denmark) were coated overnight with 10 µg/ml OVA diluted in bicarbonate buffer (pH 9.6). After washing and blocking plates, samples were incubated for 2 h. Subsequently, plates were washed and horseradish peroxidase-conjugated goat anti-mouse IgE antibody (Bethyl Laboratories, Inc., Montgomery, TX), or IgG1 or IgG2A antibody (Southern Biotechnology Associates, Inc. Birmingham, AL) was added. Tetramethylbenzidine was used as substrate, and the absorbance was determined at 450 nm using a Bio-Rad microplate reader (Bio-Rad, Munich, Germany). The titer was calculated by logarithmical regression defined as the reciprocal dilution of the sera, where the extinction was twofold the background extinction. The background extinction was defined by the use of sera from non-immunized mice, diluted 1:100.
All statistics were done using an unpaired Student's two-tailed t-test. Error bars represent SEM.
We thank the Deutsche Forschungsgemeinschaft (DFG, HA 2799/2–1) and the Bundesministerium für Bildung und Forschung (BMBF, 01 ZZ 0109) and the Bayrisches Staatsministerium für Wissenchaft, Forschung und Kunst (KKF01-03/Bu and Kap. 1528-Bu) for supporting this project. We thank Jana Bergmann and Marita Reiprich for excellent technical assistance and Christoph Werner, Günther Richter and Martin Staege for reading and discussing the manuscript.