Abstract
We have developed methodology that enables alkaline phosphatase (ALP) to be histochemically stained reproducibly in decalcified paraffin-embedded bone and cartilage of rodents. Proximal tibiae and fourth lumbar vertebrae were fixed in periodate-lysine-paraformaldehyde (PLP) fixative, decalcified in an EDTA-G solution, and embedded in paraffin. In the articular cartilage of the proximal tibia, ALP activity was localized to the hypertrophic chondrocytes and cartilage matrix of the deep zone and the maturing chondrocytes of the intermediate zone. The cells and matrix in the superficial zone did not exhibit any enzyme activity. In tibial and vertebral growth plates, a progressive increase in ALP expression was seen in chondrocytes and cartilage matrix, with activity being weakest in the proliferative zone, higher in the maturing zone, and highest in the hypertrophic zone. In bone tissue, ALP activity was detected widely in pre-osteoblasts, osteoblasts, lining cells on the surface of trabeculae, some newly embedded osteocytes, endosteal cells, and subperiosteal cells. In areas of new bone formation, ALP activity was detected in osteoid. In the bone marrow, about 20% of bone marrow cells expressed ALP activity. In adult rats, the thickness of the growth plates was less and ALP activity was enhanced in maturing and hypertrophic chondrocytes, cartilage matrix in the hypertrophic zone, and primary spongiosa. This is the first time that ALP activity has been successfully visualized histochemically in decalcified, paraffin-embedded mineralized tissues. This technique should prove to be a very convenient adjunct for studying the behavior of osteoblasts during osteogenesis.
Alkaline Phosphatase (ALP) is a membrane-bound metalloenzyme which catalyzes the hydrolysis of phosphomonoesters at an alkaline pH. At least four isozymes of ALP have been identified in humans: nonspecific liver/bone/kidney, intestinal, placental, and germ-cell ALP. The ALP produced by osteoblasts (OBs) and neutrophilic granulocytes is of the nonspecific liver/bone/kidney type ALP. The importance of ALP in bone formation and mineralization was first recognized by Robison (1923), and bone ALP has become the clinically most relevant enzyme in the diagnosis of bone disease (reviewed by Hoof and Bore 1994). The activity and localization of ALP are a valuable index for tissue development and differentiation. ALP has been thought to be a marker for the post-mitotic neutrophilic granulocyte (Pedersen 1982) and has been used for diagnosing neoplastic changes, myelosis, and leukocytosis (Wilson et al. 1981; Beutler 1995).
The localization of ALP in calcified tissues presents problems that do not exist with other histochemical methods. Its activity is quite sensitive to fixation and is denatured by moderately high temperatures (Hasselgren et al. 1978; Farley et al. 1993). A further complication is that decalcification procedures remove the magnesium and zinc ions that are necessary for ALP activity (Yoshiki et al. 1972). Because the localization of ALP is central to studies on osteogenesis, fracture repair, and bone metabolism, many attempts have been made by histologists and bone biologists to demonstrate ALP activity in calcified tissues, and some of the problems encountered during ALP localization in calcified tissues have been solved. The diffusion of reaction products and potential false localization due to the abundance of phosphate groups in tissue components can be avoided by the use of azo-dye methods. Enzyme activities in decalcified tissue samples can also be restored by preincubating with magnesium chloride (Yoshiki et al. 1972). ALP has been demonstrated immunohistochemically in decalcified, paraffin-embedded bone and cartilage sections (Tulli et al. 1992; Hoshi et al. 1997). However, the appropriate antibodies are not available for all species and this method is considerably more time-consuming and expensive than purely histochemical methods. ALP has been demonstrated histochemically in decalcified frozen bone and cartilage sections (Yoshiki et al. 1972; de Bernard et al. 1986; Hoshi et al. 1997) and in methyl-methacrylate-embedded undecalcified bone sections (Erben 1997). Unfortunately, frozen sections do not maintain tissue integrity as well as paraffin-embedded sections and plastic sections are not available to many laboratories for purely practical reasons.
A method has been described by which ALP can be localized in ethanol-fixed undecalcified bone. However, it was remarked that the material had to be processed immediately after fixation and that ALP activity was lost after decalcification (Roach 1999). The demonstration of ALP activity in decalcified, paraffin-embedded mineralized tissues would be highly desirable because paraffin is a cheap and popular medium for embedding tissues and large numbers of tissue blocks can be processed in a comparatively short time and with the minimum of supervision. In addition, in decalcified tissues, sectioning and staining present fewer difficulties than other media or frozen tissues. However, to our knowledge this has not yet been achieved. Here we report that by simply combining the use of periodate-lysine-paraformaldehyde (PLP) as the fixation agent (Bourque et al. 1993) and the replacement of magnesium ions as described by Yoshiki et al. (1972) ALP can be sensitively and reproducibly localized in decalcified, paraffin-embedded mineralized tissues.
Materials and Methods
Animals and Harvest of Tissues
Wistar rats (200–300 g) or embryonal mice were sacrificed by cervical dislocation according to home office guidelines. The left proximal tibiae and the fourth lumbar vertebral bodies were removed, dissected free of soft tissue, and fixed with PLP fixative (2% paraformaldehyde containing 0.075 M lysine and 0.01 M sodium periodate solution, pH 7.4, stored at 5C) for 24 hr at 4C as previously described (McLean and Nakane 1974).
Decalcification of Tissues
After fixation, the specimens were washed for 12 hr at 4C in each of the following series of solutions: 0.01 M PBS containing 5% glycerol, 0.01 M PBS containing 10% glycerol, and 0.01 M PBS containing 15% glycerol. The specimens were then decalcified in EDTA-G solution (14.5 g EDTA, 1.25 g NaOH, and 15 ml glycerol were dissolved in distilled water and the pH was adjusted to pH 7.3. The solution was then made up to 100 ml and stored at 4C) for 10–14 days at 4C as previously described (Mori et al. 1988). The EDTA-G solution was replaced every 5 days and the progression of decalcification was checked every 24 hr by micro-X-ray. Using this protocol, rat tibia would normally be fully decalcified in 10–14 days.
Washing of Decalcified Tissues
To remove EDTA and glycerol from the decalcified tissues, they were washed at 5C for 12 hr in successive washes of 15% sucrose and 15% glycerol in PBS, 20% sucrose and 10% glycerol in PBS, 20% sucrose and 5% glycerol in PBS, 20% sucrose in PBS, 10% sucrose in PBS, 5% sucrose in PBS, and 100% PBS as previously described (Bourque et al. 1993).
Dehydration and Paraffin Embedding
Tissues were dehydrated in a graded series of alcohols and embedded in low-melting-point paraffin using a Shandon Citadel 2000 automatic tissue processor (Shandon Scientific; Runcorn, UK).
Tissue Sectioning
The tissue blocks were trimmed down to the tissue surface. The block was then placed in a microtome and 5-μm sections were cut at frontal for tibiae and at sagittal for vertebrae. Tissue sections were floated in a water bath at 48C and collected on poly-L-lysine-coated glass slides. Sections were stained with hematoxylin and eosin (HE) and histochemically for ALP.
Histochemical Demonstration of ALP Activity in Paraffin Sections
ALP histochemistry was performed following a modified version of a previously described method (Yoshiki et al. 1972). Briefly, tissue sections were deparaffinized, hydrated through a xylene and graded alcohol series, and preincubated overnight in 1% magnesium chloride in 100 mm Tris-maleate buffer (pH 9.2) and then incubated for 2 hr at room temperature in ALP substrate solution (freshly prepared 100 mM Tris-maleate buffer, pH 9.2, containing 0.2 mg/ml naphthol AS-MX phosphate and 0.4 mg/ml Fast Red TR). After washing with distilled water, the sections were counterstained with Vector methyl green nuclear counterstain (Vector Laboratories; Peterborough, UK) and mounted with Kaiser's glycerol jelly.
Preparation of Rat Bone Marrow Cells (BMCs) and Cytospin Preparations
Tibiae and femurs of 200-g male Wistar rats were removed under aseptic conditions and the BMCs were flushed out with standard medium. The cells were dispersed by repeated pipetting and a single-cell suspension was achieved by forcefully expelling the cells through a 21-gauge syringe needle. One-ml cell suspensions were centrifuged at 3000 rpm for 5 min. The supernatant was aspirated off and the pellet was fixed with 1 ml 10% neutral phosphate-buffered formalin or PLP fixative for 20 min and resuspended. Cell suspensions of 150 μl were cytospun onto slides at 500 rpm for 5 min. After air-drying, the slides were stored at —20C until staining.
Cytochemical Demonstration of ALP Activity in BMCs
ALP cytochemistry is much easier than ALP histochemistry because the cells do not undergo a long stringent process of decalcification and paraffin embedding. The cells were cytospun onto slides and incubated for 15 min at RT in the ALP substrate solution as described above. After washing with distilled water, the cells were counterstained with Vector methyl green nuclear counterstain and mounted with Kaiser's glycerol jelly. After staining, the numbers of total cells with nuclei, ALP strong positive cells, and ALP weak positive cells within five fields of view (×400) were recorded. The percentages of ALP strong and weak positive cells were calculated and the results are expressed as means ± SD.
Results
Effect of Fixative on ALP Activity in BMC Cytospin Preparations
In a preliminary experiment using cytospin preparations of whole BMCs, the effects of fixation with either 10% neutral phosphate-buffered formalin or PLP fixative were compared. It was found that ALP activity in BMCs was well preserved after fixation with PLP fixative (Figure 3G but was totally absent after fixation with 10% neutral phosphate-buffered formalin (results not shown). In the cytospin preparations the BMCs could be divided into strong and weakly positive cells according to the intensity of ALP staining (Figure 3H. The percentage of ALP strong and weak positive BMCs was 4.4 ± 0.2% and 16.5 ± 0.4% of total BMCs, respectively. It was also found that after fixation with PLP fixative, the cytospin preparations of BMC could be stored at —20C for up to 2 years and still retain their ALP activity.
Histochemical Demonstration and Localization of ALP Activity in Decalcified, Paraffin-embedded Bone and Cartilage Sections
Following on from the above experiment using cytospin preparations of whole BMCs, tibiae were fixed with either 10% neutral phosphate-buffered formalin or PLP fixative, decalcified, paraffin-embedded, and then sections were stained using an ALP protocol in which Mg2+ ions, lost during the decalcification process were replaced by preincubation with MgCl2. As before, those fixed with 10% neutral phosphate-buffered formalin were totally devoid of ALP activity in both the growth plate and the primary spongiosa (Figure 1). In contrast, those fixed with PLP successfully retained their ALP activity (see below).
In the articular cartilage of proximal tibiae from 200-g rats, ALP activity was localized to hypertrophic chondrocytes and cartilage matrix of the deep zone and maturing chondrocytes in the intermediate zone. Resting chondrocytes and matrix of superficial zone and the matrix of intermediate zone did not exhibit any enzyme activity (Figure 2A. In older 285-g rats, ALP activity was stronger in hypertrophic chondrocytes and cartilage matrix of the deep zone. However, the distribution of ALP activity in articular cartilage was similar to that of the 200-g rats (Figure 2B.
Effect of fixation with 10% neutral phosphate-buffered formalin on ALP expression in the growth plate (
Localization of ALP in articular cartilage and growth plates. In the articular cartilage of the proximal tibia, ALP activity was localized to hypertrophic chondrocytes and cartilage matrix of deep zone and maturing chondrocytes in intermediate zone in 200-g rats (
In 200-g rat tibial and vertebral growth plates, a progressive increase in ALP expression was seen in chondrocytes and cartilage matrix, with activity being weakest in the proliferative zone, higher in the maturing zone, and highest in the hypertrophic zone. No ALP activity was detected in resting chondrocytes and the matrix of the reserve zone (Figures 2C-2G). In 285-g rats, the growth plates were narrower and ALP activity was stronger in maturing and hypertrophic chondrocytes, the cartilage matrix of the hypertrophic zone, and the primary spongiosa, compared to that of 200-g rats.
In bone, ALP activity was detected widely in pre-osteoblasts, osteoblasts, lining cells on the surface of trabeculae, some newly embedded osteocytes, endosteal cells, and subperiosteal cells (Figures 3A-3F). In areas of new bone formation, ALP activity was detected in the osteoid (Figure 3D. However, no ALP activity could be dectected in deeply embedded osteocytes and calcified bone matrix. In general, ALP activity seen in bones from older 285-g rats was stronger compared to that of 200-g rats.
Expression of ALP in Z/AP and Z/AP;Creactin Mice
To further study the utility of this method, ALP expression was investigated in tibiae and vertebrae from Z/AP and Z/AP;Creactin mouse embryos. Z/AP transgenic mice express ubiquitously the lacZ reporter gene under the control of the CMV enhancer and the chicken β-actin promoter (Lobe et al. 1999). When these mice are crossed with Creactin mice, Cre-mediated excision takes place and the lacZ gene is removed, allowing the expression of a second reported gene, human placental alkaline phosphatase. As shown in Figures 4A and 4C, in the absence of Creactin the sections are entirely negative for ALP expression but do stain for lacZ. In contrast, in the presence of Creactin there is intense staining of alkaline phosphatase activity but no staining of lacZ (Figures 4B and 4D).
Reproducibility of the Histochemical Technique for ALP
To examine the reproducibility of the histochemical technique for ALP developed in this study, specimens were harvested and processed at different time points and stained simultaneously, or the sections from the same rats were stained repeatedly three times. ALP activity was demonstrated very well in all specimens harvested at different time points. When we stained repeatedly for three times in tibial and vertebral sections from 45 rats, ALP activity in the same section was demonstrated at the same levels.
Discussion
So far as we are aware, this is first time that ALP activity has been demonstrated histochemically in PLP-fixed, EDTA-G decalcified, paraffin-embedded mineralized tissues. This histochemical technique for ALP was found to be highly reproducible and can be performed on large numbers of blocks in conventional histology laboratories. This is a significant advantage over frozen or plastic-embedded samples because paraffin-embedded bone can be used to successfully localize many other enzymes, antigens, RNA, and DNA and this development means that these can now be co-localized with ALP, a central enzyme in bone metabolism. This technique would therefore be ideal for studying osteogenesis in vivo.
Fixation was found to be a critical step in the successful demonstration of ALP activity in decalcified paraffin-embedded tissues. Our results show that the use of neutral phosphate-buffered formalin destroys all ALP activity, whereas PLP fixation preserves ALP activity very well, particularly after long-term fixation of larger tissue blocks. The ALP activity was still well preserved in BMCs stored at —20C for up to 2 years after fixation with PLP. Although ALP activity is preserved using other fixatives and can subsequently be demonstrated in plastic-embedded, decalcified frozen and undecalcified paraffin-embedded tissues (Yoshiki et al. 1972; de Bernard et al. 1986; Erben 1997; Hoshi et al. 1997; Roach 1999), PLP fixation gives not only an optimal cell surface antigenic preservation (McLean and Nakane 1974; Pollard et al. 1987; Bourque et al. 1993) but also an excellent preservation of ALP activity and tissue architecture.
The replacement of magnesium chloride leached out during the decalcification process was also found to be crucial for the demonstration of ALP activity in decalcified mineralized tissues. To achieve this, the sections were preincubated in 1% magnesium chloride in 100 mM Tris-maleate buffer before staining. Omission of this step results in total lack of staining for ALP. This is consistent with the work of Yoshiki et al. (1972), who showed that ALP activity in decalcified paraffin-embedded tissues could be reactivated by replacing magnesium ions.
By employing the histochemical technique for ALP established in this study, the localization of ALP on bone and cartilage was investigated in rats of different ages. The distribution of ALP activity as assessed in the present study in decalcified paraffin-embedded bone and cartilage matches that as assessed in decalcified frozen bone and cartilage (Yoshiki et al. 1972; de Bernard et al. 1986; Hoshi et al. 1997) and is also similar to the distribution of ALP protein as assessed immunohistochemically (Tulli et al. 1992; Hoshi et al. 1997). However, slight differences between histochemical and immunohistochemical localization can be seen. For example, resting chondrocytes apparently expressed ALP protein when stained immunohistochemically (de Bernard et al. 1986; Hoshi et al. 1997), whereas this was not the case when stained histochemically. Both results may be correct and the differences due to the presence of inactive ALP. The availability of a histochemical method will, however, improve our ability to interpret such data.
Localization of ALP activity in bone tissue. ALP was localized to osteoblasts and pre-osteoblasts (
Expression of ALP in Z/AP and Z/AP;Creactin mice. Tibiae and vertebrae from Z/AP and Z/AP;Creactin mouse embryos were stained for ALP as described in Materials and Methods. Tibiae and vertebrae taken from Z/AP mice were entirely negative for ALP expression but did stain for lacZ (
Tissue architecture was well preserved when whole proximal ends of tibiae and vertebral bodies were processed, which allowed the observation of several areas of cartilage and bone simultaneously. ALP activity was evident not only in the cartilage of growth plates, osteoblasts, endosteal cells, and bone marrow cells, as described previously, but also in periosteal cells and articular cartilage. Many subperiosteal cells expressed ALP, indicating that these cells may be osteoblast precursors. ALP was localized in the same cell type in articular cartilage as that in growth plate. However, less ALP was detected in the matrix of deep articular cartilage than in the matrix of the growth plate hyper-trophic zone. We also demonstrated that ALP activity was detectable in the osteoid of areas of new bone formation but not in calcified bone matrix.
ALP activity can also be localized to large numbers of bone marrow cells. Because it is now believed that osteoblast precursors reside in the stromal compartment of the bone marrow (Owen 1985; Friedenstein 1990), it has been previously suggested that ALP-positive BMCs may represent osteoprogenitor cells (Weinreb et al. 1990; Tulli et al. 1992). However, the frequency of osteoprogenitors in the bone marrow is much lower than that of ALP-positive BMCs. Our results show that in cytospin preparations, about 20% of BMCs stain for ALP activity, whereas osteoprogenitor cells are known to be present at a frequency of only 0.1% or less. In addition, it has also been shown that osteoprogenitors derived from bone marrow and then grown in culture are not necessarily ALP-positive (Walsh et al. 2000). Conversely, ALP has also been shown to be expressed by other cells in the bone marrow, including post-mitotic neutrophilic granulocytes (Pedersen 1982; Garattini and Gianni 1996). As in bone, the ALP expressed by neutrophilic granulocytes is a product of the nonspecific liver/bone/kidney type ALP gene and is expressed widely in differentiated granulocytes, including myelocytes, metamyelocytes, and segmented neutrophils (Stewart 1974). It is therefore not possible to identify osteoprogenitors in bone marrow using ALP alone.
Footnotes
Acknowledgements
Supported by Schering AG, Berlin, and by Research into Ageing.