β-Cell regeneration declines with aging, but the molecular mechanisms controlling β-cell replication in humans are not well understood. We compared the expression of selected cell cycle proteins in prenatal and adult tissue and examined the association of these proteins with β-cell replication. Pancreatic tissue from a total of 20 human fetuses and adults was stained for Ki67, cyclin D3, p16 and p27, and insulin. The β-cellular expression of these cell cycle proteins was determined. The frequency of β-cell replication was lower in adult compared with prenatal β-cells (<0.5 vs. 3.4 ± 0.5%, respectively; P < 0.0001). p16 was sporadically expressed in prenatal β-cells (8.0 ± 1.1%) but highly enriched in adult β-cells (63.1 ± 5.2%, P < 0.0001). Likewise, the expression of p27 was much lower in prenatal β-cells (1.7 ± 0.4 vs. 44.1 ± 5.4%, respectively, P < 0.0001), and cyclin D3 expression increased from 24.2 ± 4.1 to 47.25 ± 5.0%, respectively (P < 0.001), with aging. The expression of all three proteins was significantly correlated with each other (P < 0.01 and r > 0.75, respectively). The strong expression of cyclin D3 in adult human β-cells and its correlation to p27 and p16 suggest a positive role in human β-cell cycle regulation. p16 and p27 appear to restrict β-cell replication with aging. The age dependency of cell cycle regulation in human β-cells might explain the reduced β-cell regeneration in adult humans.
- cyclin D3
pancreatic β-cell mass is reduced in diabetes, and the extent of β-cells within the pancreas appears to largely determine the functional insulin-secretory capacity, fasting, and postchallenge glycemia (6, 30, 33). Therefore, β-cell regeneration is an area under active investigation for the treatment of diabetes (29, 31).
A number of glucose-lowering drugs have been suggested to enhance β-cell proliferation and increase β-cell mass (for review, see Ref. 29). This postulate has been primarily based on studies carried out in rodents. However, most rodent studies have used animals at very young ages (i.e., between 1 and 4 mo), and two recent reports have provided compelling evidence that the capacity for β-cell regeneration declines significantly with age. The latter has been associated with an increased expression of the cell cycle inhibitor p16 in mice (26, 48). Whether β-cell proliferation can be therapeutically enhanced in adults with diabetes is therefore questionable.
The molecular machinery driving the replication of eukaryotic cells in a concerted fashion consists of numerous protein kinases, kinase activators, and kinase inhibitors (46). Over the past decade, elegant studies involving knockout (23, 28) or transgenic (51) animals as well as overexpression of cell cycle proteins in isolated human islets (10, 16) have led to a better understanding of cell cycle regulation in pancreatic β-cells. Although the knowledge about the cell cycle regulation in human β-cells in vivo is still limited, previous studies have provided compelling evidence for the functional significance of some cell cycle regulators for β-cell replication.
Regarding the INK family of kinase inhibitors, an inverse association with β-cellular replication has been proven for p16. For example, mice with a p16 knockout exhibit increased levels of β-cell replication (26). In contrast, an experimental reduction in expression of the p16 inhibitors Ezh2 and Bmi-1 caused a decrease in β-cell mass and/or mild diabetes (8, 12). In support of a role for p16 in human β-cell turnover, the p16-gene locus has been associated with type 2 diabetes (e.g., Refs. 5, 50). Furthermore, an age dependency of nucleic β-cellular p16 expression was observed in six postnatal human pancreatic tissue samples (8).
Among the KIP/CIP proteins, p27 appears to be the most important regulator of rodent β-cell regulation. Thus deletion of the p27 locus led to an increase in β-cell mass and replication (20, 49). In contrast, p27 overexpression caused a diabetic phenotype (41, 49), and p27 accumulation was observed in the β-cells of diabetic animals (49). Evidence for the importance of p27 in the human islet comes from studies showing that germline mutations in the p27 locus are associated with the multiple endocrine neoplasia syndrome, typically comprising pancreatic cell neoplasias (1, 21, 40).
In terms of positive regulators of β-cell proliferation in rodents, a central role has been ascribed to cyclins D1 and D2 (19, 28), whereas the importance of cyclin D3 has been challenged (28). However, recent studies employing adenoviral overexpression of cyclin D3 have suggested a strong positive effect on β-cell replication in human islets (17). Furthermore, these studies identified cyclin D3 as the most abundant D-type cyclin in human β-cells, in contrast to findings in rodents.
Aside from the suggested role of p16, p27, and cyclin D3 in β-cell proliferation, a molecular interrelationship of the two kinase inhibitors with cyclin D molecules has been proposed (e.g., Refs. 14, 38). Therefore, we sought to gain more insights into the roles of these molecules in β-cell proliferation by comparing their protein expression in tissue specimens from adult and prenatal human pancreas. This approach was chosen because prenatal β-cells represent the only physiological system in which β-cell replication is constantly upregulated. On that basis, it was possible to compare the expression rates of these cell cycle proteins between two conditions exhibiting high (prenatal state) or very low (adult state) rates of β-cell replication (24, 34).
We addressed the following questions: 1) What is the quantitative difference in β-cell replication between prenatal and adult human pancreatic tissue? 2) Are there differences in the expression of p16, p27, and cyclin D3 between prenatal and adult human β-cells? 3) Is there an association between the expression of these cell cycle markers with each other or with β-cell replication?
Pancreatic tissue was obtained from 20 human embryos and fetuses between the 12th and 34th weeks postconception. The sex could be determined in 16 cases and was male in 10 cases and female in 6 cases. Specimens were provided by the Institute of Pathology and the Division of Clinical and Functional Anatomy, Medical University of Innsbruck. They were obtained from miscarriages and legal abortions after parental consent and in compliance with the local governmental and institutional guidelines. Cases were included only if careful histological examination by a trained pathologist excluded pancreas dysplasia. In addition, pancreatic tissue samples from 20 adult humans (9 females and 11 males, ages 27–84 yr, mean age 61 ± 16 yr) who had undergone pancreatic surgery between the years 2004 and 2007 were examined. The reasons for undergoing pancreatic surgery included tumors of the papilla of Vater (n = 9) and the choledochus (n = 6), perforating cholecystitis (n = 1), physical pancreatic injury (n = 1), duodenal carcinoma (n = 1), carcinoma of the uncinatus (n = 1), and benign pancreatic adenoma (n = 1). In the latter cases, only pancreatic tissue distant from the pathologies was examined. Examination of these tissues was approved by the Ethics Committee of the Ruhr-University Bochum, Bochum, Germany (registration no. 2392).
Fluorescence and Immunohistochemical Tissue Staining
Immunohistochemical staining for Ki67, insulin, and DNA.
After being heated at 37°C overnight, pancreatic tissue sections were deparaffinized in two 10-min xylene incubations and hydrated through graded alcohol series (100, 95, and 70% alcohol, and water). To block endogenous peroxidase activity, 70% alcohol contained 3% hydrogen peroxide.
Antigen retrieval was performed in DakoCytomation Target Retrieval Solution (pH 9; Dako, Glostrup, Denmark) by being heated for 20 min. After being cooled and rinsed in distilled water, sections were blocked for unspecific protein binding with Tris-buffered saline containing 3% BSA and 0.2% Triton X-100 for 10 min and for endogenous biotin with the use of the DakoCytomation Biotin Blocking System. The tissues were incubated in mouse primary antibody against human Ki67 (M7240, 1:50; Dako) for 1 h, and Ki67 was stained brown with the use of the Dako-Envision-Kit according to the manufacturer's instructions. Slides were incubated in a guinea pig antiserum against pig insulin (A0564, 1:400; Dako) for 30 min. Insulin was then stained red using the Dako REAL, Alkaline Phosphatase/Red detection system. For a blue nuclear counterstaining, specimens were incubated in hematoxylin (Dako) for 5 min and then incubated in tap water for 5 min. Finally, samples went through ascending alcohol concentrations (70, 96, and 100%) and were mounted with Entellan (Merck, Darmstadt, Germany). With omission of the primary antibodies, no staining was observed (Supplemental Fig. 1). (Supplemental data for this article is available online at the American Journal of Physiology-Endocrinology and Metabolism website.) Tonsil tissue strongly expressing Ki67 at the protein level (e.g., Refs. 25, 47) was used as a positive control (Supplemental Fig. 2).
Immunohistochemical staining for cyclin D3, insulin, and DNA.
A monoclonal mouse antibody (ab-28283; Abcam) that has previously been shown not to recognize the other D-type cyclins (2, 3) was applied to the sections at room temperature (1:600, 1 h). Deparaffinizing, epitope retrieval, blocking, target visualization, and insulin detection were performed as described above. Omission of the primary antibodies yielded no staining (Supplemental Fig. 1). Malign melanoma and breast cancer cells expressing cyclin D3 at the protein level (3, 18) were used as positive controls (Supplemental Fig. 3).
Immunohistochemical staining for p27, insulin, and DNA.
Sections were incubated in DakoCytomation Target Retrieval Solution (pH 6) and cooled to room temperature. After being blocked as described above, a monoclonal mouse antibody (sc-56454; Santa Cruz Biotechnology, Santa Cruz, CA) was applied in a dilution of 1:400 for 1 h at room temperature. Target visualization and insulin detection were performed as described above. Omission of the primary antibodies yielded no staining (Supplemental Fig. 1). Skin, tonsil, and colon carcinoma cells expressing p27 at the protein level (37, 45, 47) were used as positive controls (Supplemental Fig. 4). Furthermore, the specificity of the antibody was checked by using a second p27 antibody (DB017; Delta Biolabs). The latter yielded an analogous staining pattern in adult and embryonic tissue that could be abolished by incubation of the antibody in the respective blocking peptide before its application (Supplemental Fig. 5).
Fluorescence staining for p16, insulin, and DNA.
After antigen retrieval by heating at pH 9 and protein blocking as described above, sections were incubated with a rabbit primary antibody against human p16 (LS-B647/11712, 1:500; Lifespan) for 60 min at room temperature and with a guinea pig insulin antibody (A0564, 1:400; Dako) for 30 min. Subsequently, p16 was detected by using Cy3-labeled secondary goat antibodies (111-165-003, 1:100; Jackson ImmunoResearch, Suffolk, UK). Insulin was detected by means of FITC-labeled donkey anti-guinea pig secondary antibodies (706-095-148, 1:100; Jackson ImmunoResearch). Finally, the nuclei were stained blue by incubation in a 0.33 μg/ml DAPI (Invitrogen, Paisley, UK) solution in PBS for 5 min. After 5-min incubations in three changes of distilled water, the sections were mounted with Dako Fluorescence Mounting Medium and stored at 4–8°C in the dark. Omission of one or both primary antibodies yielded no staining (Supplemental Fig. 6). Skin and tonsil cells expressing p16 at the protein level (15, 22) were used as positive controls (Supplemental Fig. 7). Furthermore, the specificity of the antibody was checked by using a second p16 antibody (DB018; Delta Biolabs). The latter yielded an analogous staining pattern in adult and embryonic tissue that could be abolished by incubation of the antibody in the respective blocking peptide before its application (Supplemental Fig. 8).
Four-color fluorescence staining for p16, Ki67, insulin, and DNA.
After antigen retrieval and blocking, sections were incubated with mouse primary antibody against human Ki67 (M7240, 1:100; Dako) overnight at 4°C. As a secondary antibody, Cy3-labeled goat anti-mouse antibody was used (1:100). Subsequently, sections were incubated with a rabbit primary antibody against human p16 (LS-B647/11712, 1:500, 60 min; Lifespan). A donkey secondary antibody labeled with the dye DyLight649 (711-495-152, 1:100; Jackson ImmunoResearch) was applied for 30 min. The infrared emission was visualized through a cyan color channel. Insulin/DNA staining was performed as described above. Omission of one or both primary antibodies yielded no staining signal.
Four-color fluorescence staining for cyclin D3, p27, insulin, and DNA.
After antigen retrieval and blocking, sections were incubated with mouse primary antibody against human cyclin D3 (ab-28283, 1:2,000; Abcam) overnight at 4°C. As a secondary antibody, Cy3-labeled donkey anti-mouse antibody was used (1:50, 30 min; Jackson ImmunoResearch). Subsequently, sections were incubated with a rabbit primary antibody against human p27 (DB017, 1:500; Delta Biolabs) overnight at 4°C. A donkey secondary antibody labeled with the infrared dye DyLight649 (711-495-152, 1:100; Jackson ImmunoResearch) was applied for 30 min. The insulin/DNA staining was performed as described above. Omission of one or both primary antibodies yielded no staining.
Tissue Analysis and Imaging
All sections were examined with an AxioImager Z1 microscope (Carl Zeiss MicroImaging, Oberkochen, Germany). Immunohistochemically stained samples were imaged using the AxioCam MRc. For fluorescence microscopy, we used an X-Cite 120 XL FL illumination system with a metal halogenide vapor pressure lamp and a set of four filters including FS 49 for blue, FS 44 for green, FS 43 He for red fluorochromes, and FS 50 for fluorochromes emitting light in the near infrared. A monochrome AxioCam MRm and the ApoTome mode of the microscope were used for imaging fluorescence stainings.
To perform morphometric analyses, we determined the relative β-cell area per pancreatic section by scanning the total tissue area at ×100 magnification and quantifying the respective insulin-positive area by means of a color threshold (35). For all applications, Axiovision microscope software (Carl Zeiss MicroImaging) running on a Pentium 4 computer (Intel, Santa Cruz, CA) was used.
Quantification of Ki67, p16, p27 and Cyclin D3 Expression Frequency in β-Cells
To measure the expression frequencies of Ki67, p16, p27, and cyclin D3 in β-cells, 10 random fields were stained for insulin and the respective proteins were imaged at ×40 objective magnification. Because of restrictions in tissue material, only one tissue slide was investigated for each case.
We calculated Ki67, p16, p27, and cyclin D3 index as the percentage of cells with positive nuclear staining in the total number of β-cells counted. A mean of 543 ± 33 β-cells was investigated for the parameters in prenatal and adult tissue samples. Note that a quantification of β-cell replication in the prenatal human pancreas has previously been reported in a study specifically addressing this point (34).
Results are means ± SE. Statistical comparisons were carried out using Students' t-test. Correlation analysis was performed using the program Prism4 (GraphPad Software, La Jolla, CA).
During prenatal pancreas development, the fractional insulin-positive area increased age dependently from 1.5% at gestational week 12 to ∼6.6% around the time of birth (r = 0.54, P = 0.013; Fig. 1, A–D). In adult pancreatic tissue, the fractional β-cell area was significantly smaller than in the prenatal tissue samples (0.86 ± 0.28%, n = 18 vs. 3.02 ± 0.41%, n = 20, respectively; P = 0.0001; Fig. 1, A–C, E). However, because the exocrine pancreas appears to grow at a faster rate than the endocrine islets after birth (32), this difference in fractional β-cell area does not imply an actual reduction of β-cell mass in the adult cases.
Whereas the expression of Ki67 in β-cells was detectable at high levels during all stages of pancreatic development (Fig. 2A), only very few cases of replication were observed in adult β-cells (Fig. 2B). The mean frequency of β-cell replication in adult β-cells was significantly lower than in prenatal β-cells (<0.5%, n = 18 vs. 3.4 ± 0.5%, n = 20, respectively; P < 0.0001; Fig. 2C).
Expression of p16 in Pancreatic β-Cells
There was sporadic p16 expression in islet β-cells throughout the prenatal growth period, with a mean fraction of 8.0 ± 1.1% (n = 16) of positive β-cells (Fig. 3A). In adults, p16 was highly enriched in β-cells (Fig. 3B) but rarely expressed in exocrine acinar cells. The expression frequency in adult β-cells (63.1 ± 5.2%, n = 17) significantly exceeded that in prenatal β-cells (P < 0.0001; Fig. 3C). There was no detectable correlation between the β-cellular p16 expression and the respective age in either the prenatal or adult tissue samples (details not shown). Of note, p16 was never detected in actively dividing β-cells, as determined by p16 and Ki67 coexpression (Fig. 4).
Expression of p27 in Pancreatic β-Cells
Whereas abundant p27 expression could be observed in the vicinity of the β-cells in the prenatal pancreas, only 1.7 ± 0.4% (n = 18) of the β-cells themselves stained positively for p27 (Fig. 5A). In contrast, in the adult tissue, p27 was clearly enriched in the islet core containing the majority of β-cells (Fig. 5B), and the relative frequency of β-cells expressing p27 (44.1 ± 5.4%, n = 16) was significantly increased compared with that in the prenatal tissue samples (P < 0.0001: Fig. 5C). There was no detectable age dependency of p27 expression in β-cells from prenatal or adult cases (details not shown). p27 was also abundantly expressed in the exocrine pancreas, albeit without any obvious differences between the prenatal and postnatal tissue samples.
Expression of cyclin D3 in Pancreatic β-Cells
Cyclin D3 expression was relatively frequent in the prenatal pancreatic tissue, with a mean expression frequency of 24.2 ± 4.1% (n = 20) in β-cells (Fig. 6A). This expression frequency further increased to 47.3 ± 5.0% (n = 16; Fig. 6B) of all β-cells in the adult tissue (P = 0.0009; Fig. 6C). In contrast, cyclin D3 expression was only sporadically observed in the exocrine tissue of adult cases (Fig. 6B). There was no detectable age dependency of cyclin D3 expression in β-cells from prenatal or adult cases (details not shown).
Associations Between the Cell Cycle Regulators
Correlation with the Ki67 expression frequency.
There was no general association between the replication marker Ki67 and any of the cell cycle proteins examined (p16, p27, and/or cyclin D3). This was also the case when the prenatal tissues, in which the frequency of Ki67 expression was relatively high, were examined separately (details not shown).
Association of the different cell cycle proteins with each other.
In the prenatal tissue, there was no significant association between expression of the two cell cycle inhibitors p16 and p27 in β-cells and the respective expression of the cell cycle promoter cyclin D3 (details not shown). In contrast, expression of these three cell cycle regulators in β-cells was significantly correlated with each other in the adult tissue specimens. Thus the respective Pearson's correlation coefficient amounted to 0.75 (P = 0.0034) for the interrelationship of p16 with p27 (Fig. 7A) and 0.76 (P = 0.0025) for the correlation of p16 with cyclin D3 (Fig. 7B). Likewise, there was a strong association between p27 and cyclin D3 (r = 0.78, P = 0.0003; Fig. 7C). Simultaneous staining for cyclin D3 and p27 revealed frequent coexpression of p27 and cyclin D3 (Fig. 8).
The control of cell cycle progression in eukaryotic cells is highly complex and regulated by a plethora of proteins. However, the relative expression rates of these factors differ remarkably between cell types, thereby potentially explaining the unequal propensity for regeneration in different tissues. To identify potential factors involved in the control of β-cell replication in humans, we compared the expression of selected cell cycle proteins in pancreatic tissue specimens from human fetuses, a model of physiologically high β-cell replication, and from adult humans, where β-cell replication occurs at very low rates. By these means, it was possible to relate the expression of certain cell cycle proteins to the changes in the activity of β-cell replication.
Consistent with previous reports (29, 32, 34), the frequency of β-cell replication was ∼150-fold higher in prenatal β-cells compared with that in adult islets. This reduced β-cell proliferation was associated with an increased nuclear abundance of the cell cycle regulators p16, p27, and cyclin D3. The expression of all three cell cycle proteins was significantly correlated with each other.
Although previous studies have already suggested that cyclin D1 and cyclin D2 promote β-cell proliferation (28), the role of cyclin D3 in this regard has not yet been examined in detail. However, a possible involvement of cyclin D3 in embryonic development has been suggested based on studies in cyclin D1−/− cyclin D2−/− mice, whose islets showed a normal neonatal development (9, 28). The present expression studies reveal for the first time a very sporadic expression of the cyclin/cdk complex inhibitors p16 and p27 in prenatal β-cells, whereas cyclin D3 was detectable in about 24% of these cells. It is therefore possible that cyclin D3 promotes β-cell proliferation in prenatal pancreas tissue in the absence of p16 and p27. Thus, together with the findings from previous studies, our results support a role for cyclin D3 in embryonic pancreatic β-cell proliferation.
In addition to the moderate prenatal cyclin D3 expression, the present study revealed a very high labeling index for cyclin D3 in adult human β-cells. This finding is consistent with previous studies in adult human islet homogenates (16) and an early immunohistochemical expression survey reporting the strong β-cellular accumulation of cyclin D3 in adult tissue specimens (13).
As a recent study showed, overexpression of cyclin D3 along with cdk4 or cdk6 provides an exuberant effect on β-cell replication in isolated human islets (17), but the role of cyclin D3 in human β-cell proliferation under native conditions remains obscure. Thus the negative association of high cyclin D3 expression frequencies with a low replicative capacity in adult human tissue appears surprising. However, cyclin D3 accumulation has also been observed in neuronal cells (13), suggesting that cyclin D3 expression can also increase in terminally differentiated cells.
To explain the low frequency of β-cell replication despite the upregulation of cyclin D3 in adult human β-cells, we examined the expression of the G1-phase cell cycle inhibitors p16 and p27. The strong correlation observed between the protein expression of cyclin D3 and p27, a member of the KIP1 family of kinase inhibitors, in β-cells is in line with previous reports from neoplasms of various tissues (e.g., Refs. 14, 44) and supports a molecular interrelationship between these two proteins.
We also observed a significant correlation between the expression frequencies of p16 and cyclin D3, suggesting that p16 might also antagonize cyclin D3 activity, thus restricting proliferation in adult human β-cells. In fact, members of the INK4 family of kinase inhibitors are known to inhibit cdks by preventing the binding of their corresponding cyclins (39), and p16 has been described to form more stable cdk complexes than other INK proteins (38). Furthermore, in contrast to p27-cdk-cyclin complexes, p16 binding appears to always inactivate cdks (43).
Another interesting finding from the present study is the strong increase in p16 and p27 expression in adult compared with prenatal β-cells. A similar age dependency has also been observed for p16 mRNA and protein levels in rodent studies (27, 48). Furthermore, a role of p16 as an important negative, age-dependent regulator of β-cell proliferation is supported by the finding that a β-cell-specific p16 knockout increases β-cell proliferation in old, but not in young, mice (26). In humans, information about p16 expression in β-cells has been sparse (16, 36). However, consistent with the present findings, one previous study has reported the absence of p16 expression in the islets of two newborn children (36). The present study extends these findings by demonstrating very low p16 expression rates in β-cells from a large number of pancreases at different stages of prenatal development.
A nuclear expression (11) and accumulation of p27 in adult human β-cells (7) has been reported in previous studies, and p27 has been detected in human islets by immunoblotting (16). Furthermore, expression studies in mice revealed an increased p27 abundance in adult vs. 4 wk-old animals (49, 52). This age dependency of p27 expression is not unexpected, because p27 appears to be a typical feature of differentiation whose levels are high in resting cells (20, 39, 46) and decrease in G1/S due to degradation (39). Furthermore, a number of experiments have suggested a role of p27 in initiating or sustaining mitotic quiescence (42). In line with this, p27 overexpression has led to the development of overt diabetes in mice (49), and p27 knockout animals exhibited an ∼2-fold increase in β-cell mass (20). Together with these experiments, the presently observed correlation between cyclin D3 and p27 supports a potential role of p27 as an important restricting factor in β-cell replication and extends this concept to the cell cycle regulation in human β-cells.
Although the present experiments suggest important roles of p16, p27, and cyclin D3 in the regulation of human β-cell proliferation, it is important to emphasize that in reality, the control of cell cycle entry and progression is orchestrated by numerous different factors and that the decision of each cell to enter the cell cycle likely depends on the fine balance of these multiple factors. In the present studies, we have focused on cyclin D3, p16, and p27. However, this certainly does not exclude the possibility that other cell cycle regulators contribute to a large extent, as well.
Furthermore, although the overall expression rates of all three cell cycle proteins examined in the present study were quite high, thereby yielding a high number of positively stained cells, the overall number of β-cells quantified was still relatively low. This was done by necessity, because the amount of pancreatic tissue available, especially in the early prenatal cases, was limited. It is possible, however, that a more detailed study involving larger amounts of tissue as well as more homogeneous groups of samples would have yielded slightly different results.
Although β-cell replication certainly contributes to the expansion and maintenance of β-cell mass, a number of previous studies have suggested that new β-cells may also arise from the transdifferentiation of stem cells residing in the ductal epithelium or elsewhere (4). It is therefore important to emphasize that the present data characterizing the expression of cell cycle proteins in human β-cells only apply for the mechanism of replication, whereas they do not allow for any conclusions regarding the regulation of islet neogenesis from non-β-cell sources.
In conclusion, the strong expression of cyclin D3 in adult human β-cells alongside its close correlation to p27 and p16 suggests that cyclin D3 might act as an important cell cycle regulator in human β-cells and that the activity of this factor might be restricted by the accumulation of p16 and p27 with aging. This age dependency of cell cycle regulation in human β-cells should be borne in mind when findings in young animal models are translated to the situation in adult humans.
This work was supported by Deutsche Forschungsgemeinschaft Grant Me 2096/5-1 (to J. J. Meier).
There is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
We are grateful to Nathalie Fiaschi-Taesch and Andrew F. Stewart for excellent discussions and helpful suggestions. The excellent technical assistance of Martha Rozynkowski, Mechthild Schweinsberg, Birgit Baller, and Sabine Richter is gratefully acknowledged.
- Copyright © 2011 the American Physiological Society