Type I collagen is the principal constituent of extracellular bone matrix and a crucial determinant for mechanical properties of bone tissue [1, 2]. Posttranslational collagen modifications result in the formation of a mature functional matrix, which is essential for subsequent matrix mineralization [3–6]. Lysyl oxidase (protein-lysine 6-oxidase, LOX) is a copper-dependent enzyme that initiates cross-linking of collagen and elastin by catalyzing oxidative deamination of ε-amino groups of lysine and hydroxylysine residues [3, 7]. Several LOX isoforms, lysyl oxidase-like proteins 1–4 (LOXL1–4) have been identified [8–13], and amine oxidase activities of some of them have been demonstrated [14, 15]. In bone tissue, pyridinolines and deoxypyridinolines are the primary cross-links of mature type I collagen which provide mechanical integrity, rigidity, and strength [16, 17]; and diminished LOX enzyme activity results in an increased risk of bone deformities and fractures [18, 19].

Cells secrete 50-kDa pro-LOX, which is then processed by procollagen C-proteinases in the extracellular region, resulting in the 32-kDa mature LOX enzyme and its 18-kDa LOX propeptide [20]. LOX has been detected in the intracellular space [21–23], and it has been suggested to regulate gene transcription and to play a dominant role in stabilizing a normal cell phenotype [24–28].

In osteoblastic cells (MC3T3-E1), we have shown that LOX seems to be specifically regulated in the course of osteoblast cell differentiation and that this regulation is required for normal collagen deposition [29]. A stage-dependent distribution of LOX in differentiating osteoblasts with colocalization with the nuclear region as well as with the tubular network was found, which could indicate a role of LOX in the regulation of osteoblast development [21].

A lack of Lox gene expression leads to perinatal lethality in mice [30, 31]. Lox knockout mice (Lox −/−) develop to term but die soon thereafter just before or at birth [30, 31]. Due to defective collagen and elastin cross-linkage, they suffer from severe cardiovascular and pulmonary defects [30–32]. In the present study, we examined the effects of Lox gene deficiency on the skeletal phenotype and osteoblast development.

Materials and Methods

Histological Analysis

For alcian blue and alizarin red staining, embryonic day (E) 18.5 Lox −/− mice as well as wild-type (wt) mice (C57BL/6) were removed from the uterus, fixed, and stored in 70% ethanol. Prior to staining, skin and eyes were removed and calcified tissues were stained with alcian blue (0.3% alcian blue 8GX [EMD Chemicals, Gibbstown, NJ], 70% ethanol) and alizarin red solution (0.1% alizarin red S [Wako Chemicals, Richmond, VA], 95% ethanol, 1 volume glacial acetic acid, 17 volumes 70% ethanol) for 3 days. Samples were placed in 1% KOH for 24–72 h and then stored in 70% glycerol.

Moreover, three-dimensional histological reconstruction of two heads of each genotype was performed. Samples were fixed, and embedded paraffin sections were cut as 10-μm-thick serial sections using a rotary microtome (model 2065; Reichert-Jung, Heidelberg, Germany) in the frontal plane. Hematoxylin–eosin sections were evaluated, every eighth section was photographed, and scanned images were aligned. Calcified tissues using the contours of well-characterized landmarks (i.e., skull base structures) were reconstructed by computer software (Analysis Software; SIS, Münster, Germany) [33].

Measurement of Collagen Fibril Diameter

Tissues were fixed in Karnovsky solution (1% glutaraldehyde, 1% tannin in 0.2 M phosphate buffer, pH 7.4) and then postfixed with 1% osmium tetraoxide in 0.1 M phosphate buffer. The samples were rinsed, dehydrated, and embedded in epon/araldite 502 resin (Ted Pella, Redding, CA). Sections of 30–50 nm were stained with uranyl acetate and lead citrate, and the images were observed on a CM-12 transmission electron microscope (Philips Electron Optics, Eindhoven, The Netherlands). Images were recorded at ×35,000 on SO-163 electron image film (Eastman Kodak, Rochester, NY). Five hundred fibril diameters were measured in randomly chosen areas using Analysis Software.

Primary Calvarial Osteoblast Cultures

Lox −/− and wt calvariae were digested (0.2% collagenase), minced, and cultured in growth medium containing of α-MEM supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and nonessential amino acids in six-well plates as has been previously established [34, 35]. In the present study, cells from each calvaria were cultured separately because the genotype of each embryo was determined by Southern blotting at a later time point. First and second cell passages were used for the experiments.

BrdU Incorporation

Primary osteoblasts were plated on 96-well plates and cultivated for up to 48 h at 37°C in growth medium. DNA synthesis was assessed in monolayer cultures by colorimetric immunoassay (Roche, Basel, Switzerland) at 405 nm. The assay is based on measuring 5-bromo-2′-deoxyuridine (BrdU) incorporation following 2-h labeling into newly synthesized DNA of replicating cells, by enzyme-linked immunosorbent assay (ELISA).

Cell Apoptosis

Rapid binding of annexin V to phosphatidylserine was used for the early identification of cells undergoing apoptosis, as described previously [36]. Cells, plated on 24-well plates were serum-starved for 24 h and then grown in the presence and absence of 1 μM staurosporine for 6, 16, and 24 h at 37°C. Then, cells were incubated with FITC-labeled annexin V (1 μg/ml) and propidium iodine (PI, 2 μg/ml) for 15 min at 15–25°C and analyzed by fluorescence microscopy. In addition, a photometric ELISA was applied for detection of cytoplasmatic histone-associated DNA fragments (mono- and oligonucleosomes) in apoptotic osteoblastic cells (Cell Death Detection ELISA, Roche) [37, 38]. Osteoblasts were cultivated in the presence and absence of 1 μM staurosporine for 6, 16, and 24 h at 37°C. Then, cells were washed, lysed for 30 min, and centrifuged at 15,000 rpm for 10 min. The supernatant was transferred into a streptavidin-precoated microtiter plate and incubated with the immunoreagent (anti-histone biotin, anti-DNA peroxidase) for 2 h. After washing, substrate solution was added and absorbance was determined at 405 nm.

Also, lactate dehydrogenase (LDH) release, as a marker of cell necrosis, was determined in the supernatants by a colorimetric Cytotoxicity Dectection Kit (Roche).

Mineral Nodule Formation

Osteoblast differentiation was induced by growth medium supplementation with 5 mM ß-glycerophosphate and 50 μg/ml ascorbic acid for up to 21 days. Mineralized cultures were fixed and stained with 2% alizarin red (pH 4.0) for 20 min. For quantification, incorporated alizarin red stain was eluted with 10% cetylpyridinium chloride, and optical density was measured at 562 nm using a spectrophotometer.

Quantitative Real-Time PCR

Total cellular RNA from calvariae of E18.5 murine embryos was isolated (RNeasy Mini Kit; Qiagen, Hilden, Germany) according to the manufacturer’s protocol, and real-time PCR was performed using sequence-specific primers and the ABI Prism 7000 sequence detection system (Applied Biosystems, Lincoln, CA). Specific primers for LOX (Mm00495386_ml), LOXL1 (Mm01145734_ml), LOXL2 (Mm00804740_ml), LOXL3 (Mm00442953_ml), LOXL4 (Mm00446385_ml), bone differentiation markers bone sialoprotein (BSP, Mm00492555_ml), COL1A1 (Mm 00801666_g1), Runx2/Cbfa1 (Mm00501578_ml), and ß2-microglobulin (Mm00437762_m1) were Taqman probes and purchased from Applied Biosystems. ß2-Microglobulin was used as internal control. The mRNA expression relative to ß2-microglobulin was determined, and the fold changes were calculated using the wt values as calibrator by means of the 2−∆∆CT method [39].

Statistical Methods

Statistical analyses were performed using a statistical software package (SPSS for Windows, version 14.0; SPSS, Inc., Chicago, IL). Comparisons between data of Lox −/− and wt were made using an unpaired Student’s t-test assuming equal variances, and P < 0.05 was considered statistically significant.

Results

Skeletal Phenotype

Alcian blue and alizarin red staining of the Lox −/− compared to the wt mice demonstrated normal morphology of calcified structures (Fig. 1). Three-dimensional reconstructions (Fig. 2) indicated normal development of nasal, alveolar, as well as teeth-related structures in Lox −/− mice. However, handling of Lox −/− samples compared to wt revealed fragile tissues and high sensitivity to KOH maceration.

Fig. 1
figure 1

Lateral views of E18.5 calcified tissues of Lox −/− and wt mice stained with alizarin red and alcian blue. Experiments were performed three times. Representative images of one experiment are shown

Full size image
Fig. 2
figure 2

Three-dimensional reconstructions of Lox −/− and wt alveolar as well as of teeth-related structures. Reconstructions were performed two times for each genotype. Representative images (a, b coronal, c frontal, and d dorsal views) of one experiment are shown. Bone tissue is presented in brown, while Meckel’s cartilage and teeth are given in blue and white, respectively

Full size image

Decreased Collagen Fibril Diameter

Data in Fig. 3 demonstrate decreased collagen fibril diameters in Lox −/− compared to wt bone tissues. Quantitative analyses showed that the mean fibril diameter significantly (P < 0.001) decreased from 34.06 ± 4.18 nm in wt to 31.01 ± 3.78 nm in Lox −/−.

Fig. 3
figure 3

Electron microscopic analysis of collagen fibril diameters of craniofacial bone tissues. For each genotype, 500 fibril diameters were measured in randomly chosen areas. Representative images are shown for Lox −/− and wt. Bar = 30 nm, magnification ×35,000

Full size image

LOX mRNA Expression

To establish that LOX gene expression is diminished in Lox −/− calvarial bone tissue and to analyze mRNA expression of LOX isoforms (LOXL1–4), quantitative real-time PCR was performed. Figure 4 shows a significantly (P < 0.05) diminished mRNA expression of LOX and its isoforms in Lox −/− compared to wt.

Fig. 4
figure 4

Real-time RT-PCR analysis of LOX and its isoforms, LOXL1–4, from Lox −/− and wt mice. Data are presented as mean ± SD obtained from three measurements of pooled RNA samples. *Significant differences between Lox −/− and wt mice (P < 0.05, unpaired Student’s t-test assuming equal variances)

Full size image

BrdU Incorporation

Primary osteoblasts from Lox −/− and wt calvariae were cultivated for up to 48 h. Figure 5 demonstrates a statistically significant (P < 0.05) decrease of BrdU incorporation in Lox −/− osteoblasts compared to wt.

Fig. 5
figure 5

BrdU incorporation in primary Lox −/− and wt osteoblastic cells after 24 and 48 h measured by ELISA at 405 nm. Data are presented as means ± SD of three experiments (n = 6 cultures/genotype). *Significant differences between Lox −/− and wt cells (P < 0.05, unpaired Student’s t-test assuming equal variances)

Full size image

Apoptosis

Evaluation of cell apoptosis in Lox −/− compared to wt is given in Fig. 6. At 6, 16, and 24 h, no differences in annexin-V binding on outward-facing phosphatidylserine or PI binding or in the detection of cytoplasmatic histone-associated DNA fragments were found. Also, in the presence of 1 μM staurosporine, as a potent inducer of cell apoptosis, no differences in annexin-V binding and DNA fragmentation were noted between Lox −/− and wt osteoblasts (data not shown). To exclude nonspecific annexin-V binding and DNA fragmentation, we measured LDH release in the supernatants, as a marker of cell necrosis. No differences were found between Lox −/− and wt osteoblasts at any time point studied.

Fig. 6
figure 6

Annexin V-positive (a) and propidium iodine (PI)-positive (b) cell count in primary Lox −/− and wt osteoblastic cells after 6, 16, and 24 h measured by fluorescence microscopy. Measurement of DNA fragmentation (c) in primary Lox −/− and wt osteoblastic cells after 6, 16, and 24 h by ELISA at 405 nm. Spectrophotometric measurement of LDH release (d) in primary Lox −/− and wt osteoblastic cells after 6, 16, and 24 h at 490 nm. Data are presented as means ± SD of three experiments (n = 3 cultures/genotype)

Full size image

Analysis of Mineral Deposition

We next examined whether decreased cell proliferation in Lox −/− cells affects later stages of osteoblast differentiation such as mineral nodule deposition. Significantly (P < 0.05) less alizarin red stain was eluted from primary Lox −/− osteoblast cultures compared to wt, indicating less mineralized nodule formation in Lox −/− cells at day 14 as well as at day 21 (Table 1).

Table 1 Means and standard deviations of optical density (OD) of three experiments (n = 3 cultures/genotype)
Full size table

Analysis of Osteoblast Differentiation Markers

Osteoblasts undergo different phases of differentiation [40]. In accordance with decreased mineralized nodule formation observed in Lox −/− osteoblasts, Fig. 7 shows a statistically significant (P < 0.05) decrease of COL1A1 as well as of BSP and Runx2/Cbfa1mRNA expression compared to wt.

Fig. 7
figure 7

Real-time RT-PCR analysis of bone markers COL1A1, bone sialoprotein (BSP), and Runx2/Cbfa1from Lox −/− and wt mice. Data are presented as mean ± SD obtained from three measurements of pooled RNA samples. *Significant differences between Lox −/− and wt cells (p < 0.05, unpaired Student’s t-test assuming equal variances)

Full size image

Discussion

Extracellular posttranslational modifications of fibrillar type I–III collagens by LOX are crucial for collagen cross-linkage and for the accumulation of a functional collagen matrix in bone tissue [3, 5, 7]. In the present study, besides normal morphology of calcified structures, significant changes in collagen fibril formation as well as altered osteoblast differentiation were found in Lox −/− compared to wt mice.

Significantly diminished truncated and nonfunctional LOX mRNA expression was demonstrated in Lox −/− samples. In other collagen-producing cell types, such as mouse fibroblasts and vascular smooth muscle cells isolated from Lox −/− mice, lysyl oxidation itself has been reduced approximately by 80% [32] and the overall amount of immature collagen cross-links has been reported to be reduced by 40% in Lox −/− embryos [31], thus indicating that LOX’s role in cross-linking is essential and may be important in osteoblasts. Also, amine oxidase activities have been demonstrated for other LOX isoforms [14, 15], indicating that multiple LOX isoforms may be involved in cross-linkage. The present data show that LOX and LOXL1 are clearly the prominent forms at the RNA level in bone tissue. On the contrary, LOXL2 and LOXL3 showed only minor expression, which may suggest that they are less relevant in bone metabolism. In MC3T3-E1 cells, it was shown that LOX, LOXL1, and LOXL4 were regulated in the course of osteoblast differentiation, whereas constituent expression of LOXL3 and no expression of LOXL2 were found [41]. It has been suggested that LOX isoforms may partly compensate for the lack of LOX [30], which may have contributed to the fact that normal bone morphology was found in Lox −/− bone tissues by histological means. However, in the present study, in the absence of Lox, other genes including LOX isoforms, LOXL1–4, were significantly downregulated as well, which may suggest regulatory interactions between different isoforms in bone tissue. Whether the effects on osteoblast development seen in the present study are primarily due to Lox deficiency or to its isoforms, or a combination of the two, is still unknown. Moreover, the observed effects of Lox depletion on osteoblasts could also potentially depend on both changes in LOX enzyme activity as well as altered functions of LOX propeptide, as recently shown in other cell types [24, 27, 28].

The formation of covalent cross-links by enzymatic LOX activity is required for the formation of mature and insoluble collagen [3, 5, 7]. In the present study, significantly decreased collagen fiber diameters were measured in Lox −/− tissues, indicating impaired collagen matrix accumulation. A functional collagenous bone matrix may be critical for cell–cell and cell–matrix interactions affecting cell differentiation. Previous studies suggest that LOX plays an important role in osteoblastic differentiation [21, 29, 42, 43]. The present study demonstrates for the first time that Lox gene deficiency affects the osteoblastic phenotype. DNA synthesis was significantly decreased in Lox −/− osteoblasts compared to wt cells. From our previous data, it is known that LOX can be localized in the intracellular space, and LOX has been suggested to regulate gene transcription [21, 24, 25, 29]. LOX binds to intracellular cell components, in particular to the microtubular network [21], which is considered a prerequisite for development of a proper mitotic spindle apparatus and normal cell cycle progression [44]. Further studies will examine involved transcriptional control of the observed effects of Lox depletion on cell division. In addition, the balance of osteoblast proliferation and apoptosis determines the size of the osteoblast population [45]. In the present study, decreased cell proliferation was accompanied by normal regulation of apoptosis in Lox −/− osteoblasts, suggesting an overall decrease of the size of the osteoblast population.

Osteoblasts pass through phenotypic changes with distinct patterns of gene expression as they differentiate [40]. The present data indicate that the Lox gene influences not only early but also later stages of osteoblast differentiation as decreased mineral nodule formation in Lox −/− cultures was noted. In addition, to gain insight into the molecular mechanisms associated with this inhibition, we determined the expression levels of osteoblast transcription factors and phenotypic markers of various stages of osteoblast differentiation. Our data show that early markers of osteoblast differentiation, such as type I collagen as well as the marker of fully differentiated osteoblasts, BSP, were significantly downregulated in bone tissue lacking the Lox gene. Additionally, Runx2/cbfa1, a master transcription factor for the osteoblast lineage, which positively controls the expression of type I collagen and BSP [46], was significantly decreased. Altogether, inhibition of differentiation could be a consequence of inhibited cell proliferation. Besides, Lox depletion seems to directly affect osteoblast differentiation, as shown by downregulation of the expression of osteoblastic differentiation markers, and thus cannot completely be explained by diminished cell density.

Further studies are warranted to examine the interaction between enzymatic collagen cross-linkage and cellular effects as well as its effects on the formation of functional bone tissue. Based on the present data, we suspect that LOX promotes bone development through mechanisms in addition to its well established role in collagen cross-linking, and further studies are in progress to address novel mechanisms of action in normal developing osteoblast cultures.