The authors state that they have no conflicts of interest.
Age-related osteoporosis is characterized by low bone mass, poor bone quality, and impaired osteoblastogenesis. Recently, the Hutchinson-Gilford progeria syndrome (HGPS), a disease of accelerated aging and premature osteoporosis, has been linked to mutations in the gene encoding for the nuclear lamina protein lamin A/C. Here, we tested the hypothesis that inhibition of lamin A/C in osteoblastic lineage cells impairs osteoblastogenesis and accelerates osteoclastogenesis. Lamin A/C was knocked-down with small interfering (si)RNA molecules in human bone marrow stromal cells (BMSCs) differentiating toward osteoblasts. Lamin A/C knockdown led to an inhibition of osteoblast proliferation by 26% and impaired osteoblast differentiation by 48% based on the formation of mineralized matrix. In mature osteoblasts, expression levels of runx2 and osteocalcin mRNA were decreased by lamin A/C knockdown by 44% and 78%, respectively. Furthermore, protein analysis showed that osteoblasts with diminished levels of lamin A/C also secreted less osteocalcin and expressed a lower alkaline phosphatase activity (−50%). Lamin A/C inhibition increased RANKL mRNA and protein levels, whereas osteoprotegerin (OPG) expression was decreased, resulting in an increased RANKL/OPG ratio and an enhanced ability to support osteoclastogenesis, as reflected by a 34% increase of TRACP+ multinucleated cells. Our data indicate that lamin A/C is essential for proper osteoblastogenesis. Moreover, lack of lamin A/C favors an osteoclastogenic milieu and contributes to enhanced osteoclastogenesis.
Senile osteoporosis is a condition of low bone mass and poor microarchitecture, predisposing the elderly to fractures. As people live longer, the prevalence of osteoporosis will increase substantially, leading to a significant medical and economic burden. Senile osteoporosis is characterized by a decreased bone formation rate, relative to bone resorption. At a cellular level, bone loss in the elderly is thought to mainly originate from osteoblast dysfunction, whereas postmenopausal osteoporosis is mainly driven by enhanced osteoclast activity after estrogen deficiency. Using a rodent model of senile osteoporosis, we recently showed that the cancellous bone network is particularly diminished with aging as a result of decreased osteoblast function relative to osteoclast activity. Whereas the pathogenesis of postmenopausal osteoporosis has been defined in greater detail using animal models and human epidemiological data, the molecular and cellular mechanisms of age-related osteoporosis remain poorly defined. Classical mouse models of accelerated aging, such as the senescence accelerated mouse strain P6 (SAMP6), the klotho, or the ataxia telangiectasia mutated gene-deficient mouse, have provided molecular insights into age-related changes of the skeleton. However, homologous genetic defects are not known to exist in humans, and data on studies of monogenetic models of accelerated aging in humans are limited.
Recently, mutations in the lamin A/C gene have been linked to a variety of rare human diseases termed laminopathies, including Hutchinson-Gilford progeria syndrome (HGPS), a prototypic disease of accelerated aging (progeria). HGPS (OMIM 176670) primarily affects mesenchymal tissues, including bone, fat, and skeletal and heart muscle. Lamin A and C are type V intermediate filaments and represent major constituents of the filamentous structures associated with the inner surface of the nuclear membrane. Both proteins are encoded by the LMNA gene and are produced by alternative splicing. Lamin C is directly translated into its mature form. In contrast, lamin A is generated from a prelamin A precursor protein and subsequently undergoes post-translational modifications. Lamin A/C serves several functions, including stabilization of the nuclear membrane, regulation of gene expression, and cell cycle control. Mice carrying mutations in the lamin A/C gene or in genes encoding for lamin A/C–processing proteins show spontaneous fractures, growth retardation, decreased BMD, thin tibial cortices, and thin bone trabeculae. Recently, aged mice have been shown to display lower lamin A/C expression levels in osteoblasts and decreased numbers of lamin A/C–expressing osteoblasts in bone.
Here, we tested the hypothesis that lamin A/C is essential for osteoblastogenesis using primary human osteoblastic lineage cells. We evaluated whether inhibition of lamin A/C in osteoblasts alters osteogenic differentiation, RANKL/OPG expression pattern, and the capacity to modulate osteoclastogenesis. We found that lamin A/C is essential for osteoblastogenesis and that inhibition of lamin A/C favors enhanced osteoclast formation.
MATERIALS AND METHODS
Cell culture and knockdown of lamin A/C
Primary human bone marrow stromal cells (BMSCs) were purchased from Lonza (hMSC; Lonza Group, Basel, Switzerland) and maintained in growth medium in a humidified atmosphere of 95% air and 5% CO2. Cells in passages 3–6 were used. To induce osteogenic differentiation, 6 × 103 cells/cm2 were cultured in basal medium supplemented with ascorbic acid, β-glycerol phosphate, dexamethasone, l-glutamine, penicillin/streptomycin, and growth factors (Osteogenic SingleQuotes; Lonza). Mature osteoblasts were obtained after 21 days of culture. MDA-MB231 cells were purchased from ATCC and were cultured in Mc Coy's 5A medium (Biowest, Frankfurt, Germany) supplemented with 10% FCS and 1% penicillin/streptomycin (both from Invitrogen, Carlsbad, CA, USA). These cells served as a positive control for the mRNA and protein expression of lamin A and lamin C subtypes. Lamin A/C was knocked down in bone marrow–derived osteoblasts twice per week over the entire cell culture period with 50 nM siL (NM_005572; Ambion, Darmstadt, Germany) using siLentFect (Bio-Rad, Munich, Germany) as a transfection reagent. In initial experiments, various siRNA concentrations ranging from 50 to 200 nM and knockdown durations of 24 h to 5 days were used to determine the most efficient siRNA concentration and optimal knockdown duration for lamin A/C. Nontargeting siRNA was used as a negative control with identical concentrations and exposure times (ON-TARGET NonTargeting Pool; Dharmacon RNAi Technologies, Chicago, IL, USA).
Cell proliferation was assessed using a commercially available kit (Biomedica, Vienna, Austria) that measures the amount of formazan derivatives that have been reduced in the mitochondria of viable cells from tetrazolium salts. Briefly, 5 × 103 cells/well were cultured in a 96-well plate in maintenance medium at a total volume of 200 μl/well. Expression of lamin A/C was knocked down for 48 h using 50 nM siL. Subsequently, 20 μl of substrate was added and incubated for 6 h at 37°C. Finally, supernatants were collected, and the absorption was measured in triplicate with a spectrophotometer at 450 nm and a reference wave length of 492 nm.
Two million cells were cultured per well in a 24-well plate in osteogenic medium for 21 days. The expression of lamin A/C was knocked down twice per week. At the end of the culture period, cells were fixed in 70% ethanol for 1 h and stained with 40 mM alizarin red S (pH 4.2; Sigma, Munich, Germany) for 10 min at room temperature. Excess dye was removed by washing the plates with distilled water. The amount of bound calcium was eluted with 100 mM cetylpyridinium chloride (Sigma, Munich, Germany) for 10 min at room temperature. Aliquots were taken and measured in triplicate with a spectrophotometer at 540 nm.
Isolation of osteoclast precursors and co-culture with lamin A/C–deficient osteoblasts
Mononuclear cells were isolated from whole blood of four healthy adult donors after informed consent using a Ficoll gradient centrifugation (Amersham, Buckinghamshire, UK) according to the manufacturer's protocol. Mononuclear cells were seeded into a 75-cm2 flask and were treated with 25 nM M-CSF (R&D Systems, Vienna, Austria) for 2 days to induce macrophage/osteoclast precursor differentiation. For the co-culture assay, 5 × 105 BMSCs were plated per well in a 24-well plate, and after 2 days of culture in osteogenic medium, lamin A/C expression was silenced using 50 nM siL. Eight hours later, the medium was aspirated, and 1 × 106 osteoclast precursor cells were seeded on top of the BMSCs and cultured for another 10 days in the presence of 100 nM RANKL (R&D Systems) and 25 nM macrophage-colony stimulating factor (M-CSF). After fixation in acetone/citrate buffer, cells were stained for TRACP (Sigma, Vienna, Austria). Cells positive for TRACP and showing three or more nuclei were counted as osteoclasts.
Resorption pit assay
Bone slices (Nordic Bioscience, Hervel, Denmark) were rinsed with 70% ethanol and washed twice with sterile HBSS before placing them into a 96-well plate. A total of 8 × 104 BMSCs per well were cultured on the bone slices in osteogenic medium for 2 days. Thereafter, lamin A/C expression was knocked down using 50 nM siL. After 8 h, medium was aspirated, and 1.5 × 105 osteoclast precursor cells were seeded on top of the BMSCs and cultured for another 21 days with 100 nM RANKL and 25 nM M-CSF. Supernatants were collected to measure C-terminal telopeptide cross-linked collagen type I (CTX) released by the resorbing osteoclasts using a Crosslaps ELISA kit from Nordic Bioscience (Hervel, Denmark) according to the manufacturer's protocol. ELISA measurements were performed in triplicate. In addition, bone slices were subjected to electron microscopy.
Scanning electron microscopy
Bone slices were washed twice with PBS and fixed in 2.5% glutaraldehyde (pH 7.4), rinsed three times with PBS, and dehydrated in an ascending ethanol series. The samples were critical point-dried and gold-sputtered. Specimens were examined with a Joel 6310 scanning electron microscope (Tokyo, Japan).
RNA was isolated using the RNeasy kit of Qiagen according to the manufacturer's instructions (Qiagen, Hilden, Germany). Two micrograms of RNA was reverse transcribed and subsequently used for RT-PCR reactions using a standard protocol. The primer sequences and PCR conditions are given in Table 1.
Table Table 1.. RT-PCR Primer Sequences and PCR Conditions
Real-time RT-PCR was performed for lamin A/C (NM_170707), runx2 (NM_004348), RANKL (NM_003701), OPG (NM_002546), and osteocalcin (NM_199173) using assay-on-demand primers and probes following the manufacturer's instructions. The analyses were performed on the ABI Prism Sequence Detection System 7700 (Applied Biosystems, Foster City, CA, USA). All experiments were performed in duplicates and were normalized to the housekeeping control gene (β-actin; NM_001615). PCR conditions were 50°C for 2 min and 94°C for 2 min, followed by 40 cycles with 94°C for 15 s and 60°C for 30 s. The results were calculated applying the ΔΔCT method and are presented in fold increase relative to β-actin expression.
Cells were washed with PBS and scraped in a lysis buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM ethylene glycol-bis-(2-aminoethylether)-N,N,N′,N′-tetra-acetic acid, 10% glycerol, 1% triton X-100, 100 mM NaF, and 10 mM Na4P2O7 × 10 H2O. Lysates were sonicated, centrifuged, and denatured for 5 min at 94°C. Thirty micrograms of proteins was loaded on a 12% SDS-PAGE and transferred onto a 0.45-μm nitrocellulose membrane. After blocking with 5% nonfat dry milk in Tris-buffered saline Tween-20 (TBS-T), membranes were incubated with an anti-human lamin A/C antibody or anti-osteocalcin antibody (both from Santa Cruz, Heidelberg, Germany) for 1 h (1:400) and washed three times with TBS-T, followed by a 1-h incubation with an horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Cell Signaling, Frankfurt, Germany). An HRP-labeled anti-β-actin antibody was used as a loading control (Cell Signaling, Danvers, MA, USA). Membranes were washed with TBS-T, and proteins were visualized with Super Signal (Pierce, Vienna, Austria) enhanced chemiluminescence. The signals were quantified by densitometry.
BMSC-derived osteoblasts were grown on glass slides, and lamin A/C expression was knocked down at a siRNA concentration of 50 nM. After 48 h, cells were washed with PBS and fixed in 4% paraformaldehyde (PFA) in PBS for 20 min. Cells were permeabilized for 10 min in 0.2% Triton X-100/PFA. Subsequently, cells were washed with PBS and blocked with 1% BSA in PBS for 30 min. The glass slides were exposed to an anti-human lamin A/C antibody for 45 min (1:400) and incubated for another 45 min with an Alexa Fluor 488–labeled secondary antibody. After three washing steps, cells were embedded in a small droplet of mounting medium stained with 4′,6-diamidin-2′-phenylindol-dihydrochlorid (DAPI; Vector Laboratories). Slides were examined using a Zeiss Axioplan 2 fluorescence microscope, and photographs were taken and processed with the AxioVision 3.1 program.
Alkaline phosphatase activity assay
Alkaline phosphatase (ALP) activity was measured in cell cultures. Aliquots of each sample were incubated with ALP substrate buffer (100 mM diethanolamine, 150 mM NaCl, 2 mM MgCl2, and 2.5 μg/ml p-nitrophenylphosphate) for 10 min at 37°C. Total protein content was measured using the bicinchoninic acid method.
OPG and RANKL protein measurement
OPG (Immundiagnostik Systems, Frankfurt, Germany) and sRANKL (Biomedica, Vienna, Austria) protein levels were measured in the supernatant using commercially available ELISA kits.
Results are presented as means ± SD. All experiments were repeated at least three times. Statistical evaluations were performed using a one-way ANOVA with posthoc Dunnett's test. p < 0.05 was considered statistically significant.
Expression of lamin A/C in BMSCs and during osteoblastogenesis
To assess physiological levels of lamin A/C during BMSC-derived osteoblast differentiation, we analyzed endogenous gene expression and protein production of lamin A/C in BMSCs at baseline and during different stages of osteoblast differentiation. Whereas mRNA expression of lamin A/C was present in undifferentiated BMSCs and at all stages of BMSC-derived osteoblast differentiation (Fig. 1A), protein levels of lamin A (75 kDa) and lamin C (65 kDa) were hardly detectable in undifferentiated BMSCs. Protein expression of lamin A/C was generally low in osteoblasts compared with the breast cancer cell line MDA-MB231, which was used as a positive control (Fig. 1B). In osteoblasts, lamin protein expression increased by days 7 (lamin A) and 14 (lamin C) of BMSC differentiation. In general, lamin C was more abundantly expressed than lamin A (Fig. 1B).
Establishment of lamin A/C knockdown in BMSC-derived osteoblasts
After defining the pattern of lamin A/C steady-state mRNA levels and protein production during normal osteoblastogenesis in BMSCs, we evaluated the effect of lamin A/C knockdown on this process. To assess the most efficient concentration of lamin A/C–specific siRNA (siL), human primary BMSCs were transfected at day 14 with various concentrations of siL ranging from 50 to 200 nM for 48 h. A concentration of 50 nM siL was sufficient to inhibit lamin A/C mRNA and protein expression and was used for further analysis (Fig. 2A). The endogenous protein expression of lamin A in undifferentiated BMSCs was below the detection limit (Fig. 2A). Nevertheless, the efficiency of lamin A/C knockdown was confirmed by the reduced expression of lamin C. Next, we assessed the duration of lamin A/C silencing in a time course experiment. The greatest inhibitory effect was obtained between 48 and 72 h after lamin A/C knockdown with 50 nM siRNA. After 5 days, normal expression levels of lamin A/C were re-established (Fig. 2B). In addition, lamin A/C knockdown was confirmed by immunofluorescence in BMSC-derived osteoblasts showing the localization of lamin A/C to the nucleus and its incomplete disappearance after lamin A/C knockdown (Fig. 2C).
Proliferation and differentiation of BMSC-derived lamin A/C–deficient osteoblasts
After successfully establishing the knockdown of lamin A/C in primary BMSCs, the proliferative capacity of lamin A/C–deficient cells was studied. Cell proliferation was significantly reduced in undifferentiated BMSCs with decreased levels of lamin A/C compared with nontransfected cells (−26%, p < 0.001) and cells transfected with nontargeting siRNA (−22%, p < 0.001; Fig. 3A).
To assess the role of lamin A/C in BMSC-derived osteoblast differentiation, the mineralization capacity of osteoblasts was determined in mature osteoblasts (day 21) after knocking down lamin A/C twice per week (Fig. 3B). Loss of lamin A/C led to significantly impaired osteoblast differentiation (−48% to nontransfected control, −28% to nontargeting siRNA control, p < 0.05). At the molecular level, reduction of lamin A/C expression led to a significant decrease of runx2 mRNA expression at days 0, 7, and 14 of osteoblast differentiation (Fig. 3C). Differences in the expression of runx2 between control cells and cells transfected with a nontargeting control were negligible. To study effects of lamin A/C on intermediate stages of osteoblast differentiation, ALP activity was analyzed in osteoblasts after lamin A/C knockdown. ALP activity was decreased 2- to 3-fold in osteoblasts with diminished levels of lamin A/C at days 0, 7, and 14 (Fig. 3D). In addition, osteocalcin mRNA steady-state levels were assessed to study the influence of lamin A/C on a late osteoblastic marker. Osteocalcin mRNA levels were detected in mature osteoblasts on days 14 and 21. Loss of lamin A/C resulted in significantly lower osteocalcin steady-state RNA levels (Fig. 3C). This observation was confirmed at protein level in mature osteoblasts (Fig. 3E).
Expression of RANKL and OPG in lamin A/C–deficient osteoblasts
Because the knockdown of lamin A/C impaired osteoblast differentiation and altered the expression of osteoblast markers, we analyzed whether lamin A/C also modulates the expression of the two osteoclastogenesis-regulating proteins, RANKL and OPG. Thus, lamin A/C expression was knocked down in undifferentiated BMSCs at baseline and during osteogenic differentiation of BMSCs, and mRNA and protein expressions of RANKL and OPG were assessed at days 7, 14, and 21. Real-time RT-PCR analysis showed a significant upregulation of RANKL and concurrent downregulation of OPG in lamin A/C–deficient BMSCs at most time points (Fig. 4A). The resulting RANKL/OPG ratio, which reflects the osteoclastogenic potential of the cellular environment, was increased about 2- to 5-fold throughout the course of osteoblast differentiation (Fig. 4A). At the protein level, knockdown of lamin A/C led to a significant increase in RANKL expression in undifferentiated BMSCs (day 0) and immature osteoblasts (day 7). In addition, OPG secretion was decreased in immature (day 7) and mature osteoblasts (day 14) after lamin A/C knockdown (Fig. 4B). The RANKL/OPG ratio at the protein level was significantly increased at days 0, 7, and 14 (Fig. 4B).
Co-culture of lamin A/C–deficient osteoblasts with primary osteoclasts
The functional importance of the increased RANKL/OPG ratio in lamin A/C–deficient osteoblasts was examined in a co-culture system with BMSC-derived immature osteoblasts and osteoclastic lineage cells. In preliminary experiments, BMSCs were co-cultured with monocytes and treated with M-CSF only, to assess the effects of the endogenously produced RANKL and OPG. However, these cultures did not give rise to mature osteoclasts (data not shown). This may be because of the small amount of sRANKL produced by BMSCs in relation to their OPG production (Fig. 4). When exogenous RANKL was added in addition to M-CSF, multinucleated, TRACP+ osteoclasts appeared after 10 days of culture (Fig. 5C). Osteoclast generation, determined by counting TRACP+, multinucleated cells was enhanced in co-cultures with lamin A/C–deficient osteoblasts (+34% to nontransfected control and +41% to nontargeting siRNA control, p < 0.05; Fig. 5A). However, using a resorption pit assay to assess osteoclast resorbing activity, no significant differences were observed in the amount of CTX measured in the supernatant of osteoblasts with or without lamin A/C expression (Fig. 5B). The resorbing activity was also documented by images generated by electron microscopy showing the resorption lacunae (Fig. 5D).
Although the incidence of age-related osteoporosis is rising in industrialized countries, the underlying mechanisms of bone loss with aging remain poorly characterized. In the past, much attention has been paid to the molecular and cellular mechanisms of postmenopausal osteoporosis, whereas comparably few studies addressed the mechanisms of osteoblast impairment in age-related osteoporosis. Studies on senile osteoporosis are limited by a lack of suitable models and long-lasting time spans of animal housing when naturally aged models are used. Because mutations in the LMNA gene that yield reduced levels of mature lamin A in humans are associated with accelerated aging syndromes including those that affect mesenchymal tissues as in HGPS, we hypothesized that lamin A/C might be crucial for BMSC differentiation into osteoblasts and that its loss in osteoblasts may affect osteoclastogenesis through an altered balance between RANKL and OPG.
This is the first in vitro study to show that reduced lamin A/C expression impairs osteoblastic differentiation based on decreased runx2, ALP, and osteocalcin levels. These observations extend a previous study showing that lamin A/C is critical for skeletal muscle satellite cell differentiation, because loss of lamin A/C resulted in decreased levels of MyoD and desmin, proteins crucial for muscle differentiation. Moreover, using a mouse 3T3-L1 adipocyte library in a yeast two-hybrid interaction screen, lamin A/C has been shown to interact with sterol response element binding protein 1, a transcription factor involved in adipogenesis, which has been found to be markedly reduced by lamin A/C mutations found in familial partial lipodystrophy. Thus, lamins are thought to regulate differentiation of mesenchymal cells through the interaction with lineage-specific transcription factors.
Skeletal aging is characterized by decreased numbers of osteoblasts and insufficient osteoblast activity as indicated by observations from bone histomorphometric data that showed a decrease in wall thickness of trabecular bone, as well as a reduction in mineralization surface and mineral apposition rate. A recent study showed higher numbers of lamin A/C–expressing osteoblasts and cardiomyocytes in tissue sections of young versus old mice and reported that the expression of lamin A/C in bone marrow cells declined with aging. Hence, these authors suggested a link between lamin A/C and the normal aging process in that decreasing age–related lamin A/C expression was associated with impaired osteoblast function. In this regard, our study confirms this concept and illustrates that decreased levels of lamin A/C impair osteoblast differentiation.
In our study, loss of lamin A/C also resulted in reduced osteoblast proliferation. This finding is consistent with other reports showing that lamin A/C–deficient human diploid fibroblasts (HDFs) proliferate very slowly in culture and HDFs harboring mutant lamins enter a senescent state prematurely. In contrast, studies using mouse embryonic fibroblasts (MEFs) from Lmna−/− mice showed a rapid growth phenotype resembling those of retinoblastoma protein (Rb)−/− mice, resulting from proteasome degradation of Rb. Also, the knockdown of lamin-associated protein 2α (LAP2α), which is involved in cell cycle regulation together with lamin A/C and Rb, leads to an increase in cell proliferation in HeLa cells. Thus, lamin A/C deficiency seems to differentially regulate cell proliferation in a cell- and species-dependent manner. In osteoblasts, the interactions of the lamin A/C-Rb–LAP2α axis and their functional importance for skeletal diseases remain to be studied in greater detail.
In previous studies, the expression pattern of RANKL and OPG, two crucial regulators of osteoclast differentiation, has been linked to distinct osteoblast differentiation stages, with RANKL being predominantly expressed in immature osteoblasts. Because lamin A/C silencing resulted in osteoblast differentiation arrest, we hypothesized that the expression of RANKL would be increased. The findings presented in our study confirm this hypothesis and suggest that RANKL may account for the increase in osteoclastogenesis. However, our data are not sufficient to definitely conclude this, because in our system, BMSCs alone were not able to drive osteoclastogenesis without the addition of exogenous RANKL. Interestingly, osteoclast activity was not stimulated. Of note, urinary levels of deoxypyridinoline and the number of osteoclasts were also not increased in lamin A/C–deficient mice, which are severely osteoporotic. The reason(s) for this discrepancy are currently unknown and require further study. It may be interesting to determine RANKL and OPG serum levels and skeletal expression in patients with HGPS to assess whether this cytokine network is deregulated in this disease and whether it contributes to osteoporosis in patients with HGPS. A potential limitation of our study was that we did not stably knock down lamin A/C expression during osteoblast differentiation. However, our results showed that the transient lamin A/C knockdown lasted for up to 4 days and because the cells were transfected twice per week with new siRNA, lamin A/C expression was abolished throughout the entire osteogenic differentiation. Finally, we cannot exclude the possibility that factors capable of modulating osteoclastogenesis other than RANKL and OPG may have been affected by lamin A/C knockdown. In addition, the use of BMSCs that were expanded up to passage 6 may have affected their gene expression pattern and osteoclastogenic potential.
In conclusion, our results indicate an important role of lamin A/C for proper osteogenic differentiation and paracrine regulation of osteoclastogenesis. Thus, lamin A/C may be an important regulator of bone remodeling. Whether alterations of lamin A/C contribute to senile osteoporosis in humans requires further study.
The authors thank Ute Niebergall for excellent technical assistance and Doris Moser, PhD, and Guenter Russmueller, MD, for performing the electron microscopy analyses. This work was supported by an Exchange Grant provided by the European Calcified Tissue Society (ECTS) and an ECTS/Amgen fellowship to MR, the Ludwig Boltzmann Institute of Aging Research, Vienna, Austria, and the Austrian Science Fund (P20239-B13) to PP, and grants from the Deutsche Forschungsgemeinschaft (Ho 1875/5-2), DFG Research Center and Cluster of Excellence for Regenerative Therapy Dresden, and Wilhelm Sander-Foundation (2007.005.1) to LCH.