SEARCH

SEARCH BY CITATION

Keywords:

  • HUVEC;
  • osteoblast;
  • MSC;
  • proliferation;
  • apoptosis

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Angiogenesis is a fundamental process in bone formation, remodeling, and regeneration. Moreover, for the regeneration of bone in tissue engineering applications, it is essential to support neovascularization. This can be achieved by cell-based therapies using primary endothelial cells, which are able to form functional blood vessels upon implantation. In bone composite grafts, coimplanted endothelial cells do not only support neovascularization but also support osteogenic differentiation of osteoblasts and osteoprogenitor cells. In this study, we investigated the effect of endothelial cells on proliferation and cell survival of human primary osteoblasts (hOBs) and human mesenchymal stem cells (MSCs). Human umbilical vein endothelial cells (HUVECs) stimulated hOB and MSC proliferation, whereas proliferation of HUVECs was unaffected by cocultured hOBs or MSCs. The effect of HUVEC cocultivation on hOB and MSC proliferation was more pronounced in direct cocultures than in indirect cocultures, indicating that this effect is at least partially dependent on the formation of heterotypic cell contacts between HUVECs and hOBs or MSCs. Furthermore, HUVEC cocultivation reduced low-serum induced apotosis of hOBs and MSCs by a mechanism involving increased phosphorylation and inactivation of the proapoptotic protein Bad. In summary, our experiments have shown that cocultured HUVECs increase the proliferation and reduce low-serum induced apoptosis of hOBs and MSCs. © 2012 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 30:1682–1689, 2012

Bone formation, regardless if in the setting of embryonic/fetal developement, regenerative processes in vivo after traumatic, degenerative bone loss, and injury or in tissue-engineering applications, is indissolubly linked to the presence of a sufficient blood supply. A sophisticated communication network between osteoblasts and endothelial cells has been found to be indispensable for any three-dimensional bone formation and is delicately balanced by paracrine signaling mechanisms as well as direct cell-contacts.1–5 Although being the subject of intensive research, angiogenesis still is far from being fully controllable or understood. It is commonly accepted, that cocultivation of primary human osteoblasts (hOBs) with human umbilical vein endothelial cells (HUVECs) will set off a series of signaling cascades which—with considerable generalization—support osteogenic differentiation, bone formation, and cell survival.

In particular, it has been observed that hOBs in direct coculture with HUVECs show an upregulation in the expression of alkaline phosphatase (ALP), a marker of early osteogenic differentiation, whereas this effect could not be verified on indirect coculture in a transwell-system or using conditioned media.6–9 We previously demonstrated, that the HUVEC-mediated upregulation of osteoblastic ALP is cell-type specific and is posttranscriptionally regulated via p38 mitogen-activated protein kinase-dependent mRNA-turnover.10 That soluble factors are nevertheless involved in differentiation and maintenance of osteoblast function was already shown by Guenther in 1986,4 who found an endothelium-derived growth factor to stimulate the growth of bone cell populations. In this context, the role of vascular endothelial growth factor (VEGF) on osteoblasts has also been extensively studied and was identified as prolonging the lifespan of hOBs in vitro, as well as upregulating the expression of ALP.11 Wang et al.12 could only partly block the effects of VEGF by anti-VEGF-antibodies, whereas Villars et al.6 did not find VEGF involved in these processes at all. Other factors that are discussed in this context include fibroblast growth factor, Wnt, bone morphogenetic proteins and Notch.13 In summary, it becomes apparent that, despite all efforts, a coherent picture of the processes involved in the interaction of angiogenic and osteogenic cell types does not yet exist.

The aim of this study was to contribute data on the effects of endothelial cells on proliferation and survival of human mesenchymal stem cells (MSCs) and primary osteoblasts.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Cell Culture

Human primary osteoblasts (hOBs) were isolated from cancellous bone of femoral heads from patients undergoing hip surgery. The osteoblasts were obtained corresponding to an established protocol.14 The hOBs were cultured in medium 199 with Earle's salts and L-Glutamine supplemented with 10% fetal calf serum (FCS) and 1% penicillin/streptomycin.

MSC were extracted from the bone marrow of cancellous bone of femoral heads, according to an established protocol.15 Cultivation was realized in MEM Alpha Medium with 10% FCS, 5 ng/ml bFGF, and 50 µg/ml gentamycin.

HUVECs were purchased from PromoCell GmbH (Heidelberg, Germany). The HUVECs were cultured in endothelial cell growth medium (ECGM) supplemented with 10% FCS and supplements. All three different types of cells were cultured in a humidified atmosphere (37°C; 5% CO2) and furthermore only cells from passage 2–5 were used for experiments. The biological material was received with the informed consent of the patients according to hospital's ethics committee guidelines.

Proliferation Experiments in Direct Cocultures

In order to investigate the proliferation of hOBs and HUVECs in direct cocultivation, three groups were formed and compared to each other. In the first group the hOBs were seeded with a cell density of 5.0 × 104 hOBs per six-well cluster plate. The second group consisted of the coculture of each 2.5 × 104 HUVECs and hOBs seeded in a six-well cluster plate. Finally the third group was the HUVEC monoculture with a seeding density of 5.0 × 104 cells.

All three groups were arranged in triplicates and incubated in a humidified atmosphere (37°C, 5% CO2) in ECGM supplemented with 10% FCS and supplements. In order to investigate growth under low-serum medium conditions, cells were cultivated in endothelial cell basal medium (ECBM) supplemented with 1% FCS. The same procedure was performed with MSCs and HUVECs. The medium was refreshed after 4 days and the cell number was detected with a CASY cell counter (Schärfe-System, Reutlingen) after 1, 3, 5, and 7 days of incubation.

Proliferation Experiments in Indirect Cocultures

In this experiment also three groups were designed. Firstly a hOB monoculture, secondly a hOB and HUVEC coculture and thirdly a HUVEC monoculture. 5.0 × 104 cells were seeded into six-well cluster plates either in ECGM with supplements and 10% FCS or in ECBM without supplements and 1% FCS. In order to exclude direct heterotypic cell contacts, the HUVECs in the coculture were seeded into cell culture inserts with 1 µm pore size (BD Labware, Franklin Lakes, NJ). The same procedure was performed with MSCs and HUVECs. The medium was refreshed after 4 days and the cell number was detected with a CASY cell counter (Schärfe-System, Reutlingen) after 1, 3, 5, and 7 days of incubation.

Immunomagnetic Separation

In order to separate HUVECs from hOBs or MSCs after cocultivation, an immunomagnetic separation system (Invitrogen Dynal AS, Invitrogen, Karlsruhe, Germany) was used. In brief, cells were detached from the culture dishes by trypsin/EDTA-treatment. Enzymatic digestion was stopped by addition of 500 µl PBS, 5% FCS. Thereafter, cells were centrifuged (1,000 rpm, 5 min) at room temperature and washed once with PBS, 0.1% BSA. Cells were resuspended in 1 ml of PBS, 0.1% BSA, mixed with 25 µl of magnetic beads (Dynabeads, Invitrogen, Karlsruhe, Germany) coated with an anti-CD31 antibody and incubated on a rotator for 30 min at 4°C. The HUVECs binding to the CD31-coated Dynabeads were separated using a magnetic particle concentrator (Invitrogen Dynal AS, Invitrogen, Karlsruhe, Germany). The unbound cells (hOBs or MSCs) were transfered to a new eppendorf tube, centrifuged (1,000 rpm, 5 min) at room temperature and used for further experiments.

DNA Fragmentation ELISA

Quantitation of fragmented DNA was measured by ELISA (Cell Death Detection ELISA kit; Roche, Mannheim, Germany). To assess DNA fragmentation, hOBs or MSCs grown in monoculture or in coculture with HUVECs (1:1 ratio) were harvested and their DNA was extracted by lysis for 30 min at room temperature with vigorous shaking. The extracts were centrifuged for 10 min at 15,000 rpm and the resulting supernatants were incubated with peroxidase-labeled anti-DNA antibody and biotinylated anti-histone antibody in streptavidin-coated microtiter plates following the manufacturer's instructions. After washing peroxidase substrate ABTS (2,2′-Azino-di[3-ethylbenzthiazolin-sulfonat]) was applied to develop and visualize binding of mono- and oligonucleosomal DNA. Microtiter plates were analyzed at 405 nm using an automated microtiter plate reader.

Quantification of Bad Phosphorylation

A Phospho-Bad Sandwich ELISA Kit (New England Biolabs, Frankfurt, Germany) was used to investigate the phosphorylation status of the proapoptotic protein BAD in hOBs or MSCs grown in monoculture or in coculture with HUVECs (1:1 ratio). Corresponding to manufacturer's instructions, cells were harvested and resuspended with 300 µl ice-cold 1X Cell Lysis Buffer supplemented with 1 mM phenyl-methylsulfonyl fluoride (PMSF) and sonicated on ice. Subsequently the lysate was centrifuged at 12,000 rpm for 10 min at 4°C. The resulting supernatant was transfered to a new eppendorf tube, diluted with sample diluent in a 1:1 ratio and applied to the ELISA plate. The ELISA was incubated overnight at 4°C. After washing, the detection antibody was added and the microplate was incubated at 37°C for 1 h. After washing, the HRP-linked secondary antibody was added for 30 min and after additional washing steps, the TMB substrate was added and incubation was set at 37°C for 10 min. The reaction was stopped with STOP solution and the photometric determination was accomplished by reading the absorbance at 450 nm.

Statistical Analysis

Statistically significant differences between groups were determined by using an unpaired Student's t-test. Statistical significance was defined when p< 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Effect of HUVEC Coculturing on hOB and MSC Proliferation

hOBs were cocultured with HUVECs in normal growth medium or in serum-reduced basal medium either in direct or in indirect coculture for up to 7 days. At defined time points, HUVECs and hOBs were trypsinized, separated by an immuno-magnetic separation process, and counted individually. As shown in Figure 1A, hOBs grown in direct coculture with HUVECs showed a significantly increased proliferation at day 5 and 7 in relation to hOBs grown in monoculture. Interestingly, at the same time, the growth characteristics of HUVECs were almost unaffected by cocultured hOBs, suggesting that the growth supporting effect of HUVEC-hOB coculture is unidirectional and restricted to hOBs. The supportive effect of HUVECs on hOB proliferation was even more pronounced when hOBs and HUVECs were cocultured under low-serum conditions (Fig. 1B). In this case, increased proliferation of hOBs due to HUVEC cocultivation was detectable at days 3, 5, and 7. At day 7, hOB cell numbers were nearly fivefold higher in the cocultivation group in relation to the monoculture group. In contrast, when hOBs and HUVECs were indirectly cocultured, thereby preventing direct heterotypic cell contact, then only a modest effect on hOB proliferation could be detected under low-serum conditions (Fig. 1D) and no stimulative effect was detectable in normal growth medium containing 10% serum (Fig. 1C).

thumbnail image

Figure 1. Proliferation of hOBs and HUVECs in mono- and coculture. hOBs and HUVECs were seeded in monoculture or coculture under direct (A,B) or indirect (C,D) cocultivation conditions. Cells were grown either in normal growth medium (ECGM, 10% FCS plus supplements) (A,C) or in serum-reduced basal medium (ECBM, 1% FCS, without supplements) (B,D) for the indicated time periods. Thereafter, cells were trypsinized and counted directly in case of indirect coculture or—in case of direct coculture—cell populations were firstly separated by immunomagnetic separation and then counted. Values are expressed as increase in cell number in relation to day 1. Shown are mean values + SD from triplicate determinations. Statistically significant differences between the cocultivation groups and the corresponding monocultivation groups at the respective time points are indicated for *p < 0.05, *p < 0.005, and ***p < 0.0005.

Download figure to PowerPoint

Interestingly, MSCs showed very similar growth characteristics as hOBs when cocultured with HUVECs. As shown in Figure 2B, cocultivated HUVECs increased MSC cell numbers under low-serum conditions to a similar extent as observed for hOBs. Again, a weak effect of HUVECs on MSC proliferation was also detected under low-serum conditions in indirect cocultures (Fig. 2D). However, in contrast to the effect of HUVECs on hOB proliferation no effect was detectable on MSC proliferation when cells were cocultured either directly or indirectly in normal growth medium containing 10% serum (Fig. 2A and C). Similar to the results obtained by hOB-HUVEC cocultivation, MSCs exerted only marginal effects on HUVEC proliferation.

thumbnail image

Figure 2. Proliferation of MSCs and HUVECs in mono- and coculture. MSCs and HUVECs were seeded in monoculture or coculture under direct (A,B) or indirect (C,D) cocultivation conditions. Cells were grown either in normal growth medium (ECGM, 10% FCS plus supplements) (A,C) or in serum-reduced basal medium (ECBM, 1% FCS, without supplements) (B,D) for the indicated time periods. Thereafter, cells were trypsinized and counted directly in case of indirect coculture or—in case of direct coculture—cell populations were firstly separated by immunomagnetic separation and then counted. Values are expressed as increase in cell number in relation to day 1. Shown are mean values + SD from triplicate determinations. Statistically significant differences between the cocultivation groups and the corresponding monocultivation groups at the respective time points are indicated for *p < 0.05, *p < 0.005, and ***p < 0.0005.

Download figure to PowerPoint

Effect of HUVEC Cocultivation on Apoptosis of hOBs and MSCs

We next investigated whether cocultivation of hOBs and HUVECs may have an effect on apoptosis induced by serum-starvation. For this purpose, hOBs and HUVECs were either grown in monoculture or in coculture for 2 days in basal medium containing 1% serum (Fig. 3). As evidenced by DNA-fragmentation ELISA, cocultivation of hOBs with HUVECs decreased hOB apoptosis in comparison to hOB monoculture, whereas the apoptosis rate of HUVECs was unaffected by hOB coculturing. In order to investigate whether this effect is mediated by paracrine-acting soluble factors or by a cell-contact dependent mechanism, we cultivated HUVECs and hOBs either in direct or in indirect coculture. As shown in Figure 4A, apoptosis of hOBs was inhibited by cocultivation with HUVECs under both cocultivation conditions to nearly the same extent. The same was true for the cocultivation system consisting of HUVECs and MSCs (Fig. 4B). Also in this case, HUVECs decreased apoptosis of MSCs in direct as well as in indirect coculture in relation to MSCs grown in monoculture.

thumbnail image

Figure 3. Measurement of apoptosis of hOBs and HUVECs in monoculture and direct coculture. hOBs and HUVECs were seeded in monoculture or in direct coculture. Cells were grown in serum-reduced basal medium (ECBM, 1% FCS, without supplements) for 2 days. Thereafter, cells were trypsinized, separated by immunomagnetic separation and nucleosomal fragmentation products were quantitated by DNA fragmentation ELISA. The figure shows the mean ± SD from triplicate determinations. Statistically significant differences between groups are indicated for *p < 0.05.

Download figure to PowerPoint

thumbnail image

Figure 4. Measurement of apoptosis in hOB/HUVEC cocultures (A) and MSC/HUVEC cocultures (B). Cells were either grown in serum-reduced basal medium in monoculture (mono) or in direct or indirect coculture (cocu) for 2 days. Thereafter, cells were trypsinized and lyzed directly for quantification of DNA fragmentation (indirect coculture) or cell populations were firstly separated by immunomagnetic separation and then analyzed for DNA fragmentation (direct coculture). Shown are mean values ± SD from triplicate determinations. Statistically significant differences between groups are indicated for p < 0.05 (*).

Download figure to PowerPoint

Effect of HUVEC Cocultivation on Bad Phosphorylation

It is well known that intracellular apoptotic signaling involves the proapoptotic protein Bad, whose proapoptotic activity is abolished by phosphorylation. We performed Bad phosphorylation ELISAs to investigate whether HUVECs may exert their effect on hOB and MSC apoptosis by modulating the phosphorylation status of Bad (Fig. 5). In direct coculture, HUVECs induced a statistically significant phosphorylation of Bad in hOBs (Fig. 5A) as well as in MSCs (Fig. 5B). However, in indirect coculture, HUVECs were unable to induce Bad phosphorylation in hOBs and MSCs, suggesting that Bad phosphorylation is mediated by a cell-contact dependent mechanism.

thumbnail image

Figure 5. Quantification of Bad phosphorylation in hOBs (A) or MSCs (B) grown in monoculture (mono) or in direct or indirect coculture (cocu) with HUVECs for 2 days in serum-reduced basal medium. Cells were trypsinized and lyzed directly for quantification of Bad phosphorylation (indirect coculture) or cell populations were firstly separated by immunomagnetic separation and then analyzed for Bad phosphorylation (direct coculture) using a phospho-Bad ELISA. Shown are mean values ± SD from triplicate determinations. Statistically significant differences between groups are indicated for p < 0.05 (*).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

It is well accepted that there exists a close relationship between osteogenesis and angiogenesis. During endochondral ossification, hypertrophic chondrocytes secrete the proangiogenic growth factor VEGF, thereby inducing the ingrowth of blood vessels from the metaphysis into the growth plate which leads to chondrocyte resorption and subsequently to bone formation.16, 17 Moreover, angiogenesis is also a prerequisite for fracture repair since abolishment of neovascularization is accompanied with delayed and insufficient bone formation.5 These findings indicate that osteogenesis is highly dependent on sufficient oxygen and nutrient supply delivered by angiogenic and/or vasculogenic processes. In this context, we10 and others6, 7 have also shown that there exists intercellular communication between primary endothelial cells and osteoblasts affecting the osteogenic differentiation of osteoblasts evidenced by increased expression of the early osteogenic differentiation marker alkaline phosphatase (ALP). Interestingly, this effect can only be observed when osteoblasts and endothelial cells are cultivated in direct cell contact which induces the upregulation of osteoblastic ALP expression, indicating that this effect is mediated by the formation of heterotypic cell contacts and not by paracrine-acting soluble factors.

In this study, we investigated the effect of hOB and HUVEC cocultivation in different cocultivation settings (direct and indirect) on cell proliferation and cell survival of the respective cell types. In this context, we have seen that HUVECs, when grown in direct contact with hOBs, increased hOB proliferation in normal growth medium (containg 10% FCS). This effect was even more pronounced when experiments were conducted under low-serum conditions (1% FCS). However, a growth stimulating effect of HUVECs in indirect cocultures could only be detected under low-serum conditions, but not when cells were grown in normal growth medium containg 10% FCS. Our results are in agreement with an article from Villars et al.,6 demonstrating that conditioned medium from HUVECs stimulated the proliferation of human bone marrow stromal cells (HBMSC). Moreover, it was already published that conditioned medium from bovine aortic endothelial cells (BAEC) increased proliferation of bone cells isolated from rat calvaria, evidenced by thymidine incorporation experiments.4 Similarly, Jones et al.18 reported that microvessel-derived endothelial cells increased the proliferation of rat calvarial osteoblasts. However, the authors of the above mentioned articles have used conditioned medium from endothelial cells in their stimulation experiments and have not investigated the effect of direct cell contact between endothelial cells and osteoblasts. We have seen in our experiments that the effect of HUVEC cocultivation on hOB proliferation is much more pronounced when cells were cultivated in direct cell contact, suggesting that heterotypic cell contacts and/or the extracellular matrix produced by endothelial cells provides an additional stimulus for hOB proliferation independent of the effect of soluble factors secreted by HUVECs. Interestingly, although cocultivation of hOBs and HUVECs had a strong effect on hOB proliferation, the growth characteristics of HUVECs were mostly unaffected by hOB cocultivation in our experiments. This result was unexpected since it is known that osteoblasts secrete VEGF which represents a potent mitogen for endothelial cells. Moreover, it was also shown that VEGF gene expression is upregulated in osteoblasts upon cocultivation with HUVECs.19 Obviously, secretion of VEGF by ostoblasts seems to be insufficient to stimulate HUVEC proliferation in hOB/HUVEC cocultures.

Villars et al.6 reported that conditioned medium from HBMSCs increase the proliferation of HUVECs as evidenced by thymidine incorporation assays. This effect was not detectable in our experiments, probably as a result of using different osteoblastic cell types which could exhibit phenotypic or physiological differences or as a result of using different cell proliferation measurement methods with different sensitivities.

We also performed proliferation assays with human bone marrow-derived MSCs which are capable of self-renewal and can differentiate into several types of mesenchymal tissues including cartilage, adipose tissue, and bone.20 In the context of our experiments, MSCs can be considered as osteoblastic progenitor cells and we used this cell type to extend our analysis on the interplay between endothelial cells and preosteoblastic cells. Interestingly, MSCs showed a very similar growth behavior than hOBs when cocultured with HUVECs under low serum conditions. MSC proliferation was enhanced in indirect cocultivation with HUVECs and was even more pronounced under direct cocultivation conditions, suggesting the contribution of soluble factors secreted by HUVECs as well as influences originating from direct heterotypic cell contact between MSCs and HUVECs. Our results are in agreement with data recently published by Bidarra et al.21 who have also detected a stimulative effect of HUVECs on MSC proliferation.

Interestingly, HUVEC cocultivation does not only affect hOB proliferation and survival, but also osteoblastic differentiation. In this context, we have previously shown that HUVECs induce the expression of the osteogenic marker genes ALP, OSF-2, osteonidogen, and BMP-1 in cocultured hOBs.10, 22

In this study, we also investigated the effect of cocultivation on apoptosis induced by serum starvation (1% FCS instead of 10% FCS). These experiments revealed an antiapoptotic effect on hOBs when cocultured with HUVECs. In contrast, apoptosis of HUVECs was not affected by cocultivation with hOBs. Moreover, we demonstrated that cocultivation of HUVECs with hOBs or MSCs reduced apoptosis of hOBs and MSCs in indirect as well as in direct coculture.

It is known that some antiapoptotic signals mediate their effect by phosphorylation of the proapoptotic protein Bad.23 Upon phosphorylation, Bad binds to the 14-3-3 protein thereby preventing the binding of Bad to antiapototic BCL-2 family members.24–26 These events abolish the proapoptotic activity of Bad. Interestingly, we have seen in our experiments that Bad is phosphorylated in hOBs as well as in MSCs upon coculture with HUVECs. However, this effect can only be seen in direct coculture but not in indirect cocultures, although apoptosis inhibition is detectable in both cocultivation settings. This result suggests that there might exist two different intracellular apoptosis pathways, one of which involves proapoptotic Bad, which is inactivated in direct cocultures and the other one is involved in antiapoptotic signaling induced by paracrine-acting factors secreted by HUVECs and is independent of Bad. In this context, it is important to note that apoptosis in general is not only controlled by Bad but also by several other Bad independent pathways.27, 28 It is intriguing to speculate that these pathways are involved in HUVEC mediated antiapoptosis seen in indirect cocultures.

Taking into consideration that HUVECs were able to support growth and survival of HOBs and MSCs, it would be very interesting to investigate the underlying molecular mechanisms. Concerning the growth-inducing effect of HUVECs, it is reasonable to speculate that HUVECs may secrete a wide range of growth and survival factors. Insulin-like growth factor-1 could be very interesting in this context since it is produced by endothelial cells29 and is mitogenic for osteoblasts.30 However, since we have seen in our experiments more pronounced effects under direct cocultivation conditions, it might be intriguing to speculate that gap junction communication between HUVECs and osteoblasts may also play a significant role in the mitogenic and anti-apoptotic effect of cocultivated HUVECs. In fact, it has been demonstrated that gap junction communication is involved in the control of proliferation31 and in the anti-apoptotic response of osteoblasts.32

In summary, our experiments have shown that cocultivated HUVECs increase the proliferation of hOBs and MSCs by paracrine-acting soluble factors but much more pronounced by factors originating from heterotypic cell contacts and/or extracellular matrix modifications. Moreover, under direct cocultivation conditions HUVECs also reduced low-serum induced apoptosis of hOBs and MSCs, mediated by phosphorylation and inactivation of the proapoptotic protein Bad, whereas soluble paracrine-acting factors secreted by HUVECs affect apoptosis of hOBs and MSCs by a mechanism which is independent of Bad.

These results provide the basis for additional studies on the functional significance of heterotypic cell contacts in the control of proliferation and apoptosis of hOBs and MSCs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank Beate vom Hoevel for excellent technical assistance. This work was supported by funding through the Deutsche Forschungsgemeinschaft (FI 790/4-1).

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES