Localization and Quantification of Proliferating Cells During Rat Fracture Repair: Detection of Proliferating Cell Nuclear Antigen by Immunohistochemistry



Bilateral femurs of 12-week-old female Sprague-Dawley rats were fractured, and the fractured femurs were harvested 36 h, 3, 7, 10, and 14 days after the fracture. Localization of cell proliferation in the fracture calluses was investigated using immunohistochemistry with antiproliferating cell nuclear antigen (PCNA) monoclonal antibodies. Thirty-six hours after the fracture, many PCNA-positive cells were observed in the whole callus. The change was not limited to mesenchymal cells at the fracture site where the inflammatory reaction had occurred, but extended in the periosteum along almost the entire femoral diaphysis where intramembranous ossification was initiated. On day 3, periosteal cells or premature osteoblasts in the newly formed trabecular bone during intramembranous ossification still displayed intense staining. On day 7, many premature chondrocytes and proliferating chondrocytes were PCNA positive. Endochondral ossification appeared on days 10 and 14, and the premature osteoblasts and endothelial cells in the endochondral ossification front were stained with anti-PCNA antibodies. Quantification of PCNA-positive cells was carried out using an image analysis computer system, obtaining a PCNA score for each cellular event. The highest score was observed in the periosteum early after the fracture near the fracture site. Immunohistochemistry using anti-PCNA antibodies showed that the distribution of proliferating cells and the degree of cell proliferation varied according to the time lag after the fracture, suggesting the existence of local regulatory factors such as growth factors, and that significant cell proliferation was observed at the beginning of each cellular event.


Fracture repair begins with bleeding at the fracture site. Several cellular events occur during fracture repair. These events include inflammatory reaction, chondrogenesis, intramembranous ossification, and endochondral ossification, resulting in the formation of a fracture callus. In each cellular event, various types of cells proliferate and differentiate. Although several papers have shown the localization of proliferating cells in the fracture callus by3H-thymidine incorporation,1,2 their descriptions principally concerned the early periosteal reaction, and the quantification of cell proliferation was limited to the osteogenic layer of the periosteum where intramembranous ossification occurs.

Proliferating cell nuclear antigen (PCNA), or cyclin, is a 36 kD intranuclear polypeptide synthesized in the late G1 and S phase, and it is localized to the sites of DNA synthesis.3–5 Recently, the monoclonal antibody to PCNA has been employed to demonstrate the proliferative component of paraffin-embedded normal tissues6 and mixed tumors.7–9

Knowledge about cell proliferation and proliferating cell types is essential for the understanding of fracture repair. We detected PCNA in the fracture callus using immunohistochemistry to localize and quantify proliferating cells in each of the cellular events that occur during fracture repair.


Fracture model

Bilateral femurs of Sprague-Dawley rats (female, 12-weeks-old) were fractured as described previously.10 Pentobarbital (65 mg/kg of body weight) was injected intraperitoneally, and the anesthetized rats were prepared for surgery by shaving and cleansing the hind limbs. The knee was exposed through a medial parapatellar incision, and the patella was dislocated laterally, exposing the femoral condyles. A Kirschner wire (1 mm in diameter and 2.8 cm in length) was introduced into the intramedullary canal through the intercondylar portion. The Kirschner wire did not protrude into the knee joint or interfere with the motion of the patella. After closing the knee joint, the mid-diaphysis of the pinned femur was fractured by applying a bending force, as described by Bonnarens and Einhorn.11 Radiographs were taken at the time of sacrifice, and improper fracture calluses were excluded.


Six to eight fractured femurs were harvested at 36 h, 3, 7, 10, and 14 days after the fracture. These were then fixed, demineralized, and embedded in paraffin. Sections were cut through the long axis of each femur in the sagittal plane and stained with Masson's trichrome stain.


Conventional 4-μm histological sections were cut from the paraffin-embedded fractured femurs. The sections were mounted on poly-l-lysine-coated glass slides and then air dried overnight at room temperature. The sections were deparaffinized with xylene and hydrated with serial concentrations of 100, 95, 80, and 70% alcohol. Endogenous peroxidase activity was blocked with 0.5% hydrogen peroxide in methanol for 60 minutes. Following a tris-buffered saline (TBS)/0.1% bovine serum albumin (BSA) wash, hyaluronidase treatment (hyaluronidase 1 mg/ml in sodium acetate buffer, pH 5.5, 0.85% NaCl) was performed for 30 minutes at 37°C. After being washed with TBS/0.1% BSA, the sections were incubated with blocking solutions (fresh 10 ml of TBS/0.5% BSA with 130 μl of normal horse serum) at room temperature for 15 minutes. The sections were then incubated with PCNA monoclonal antibodies, PC10 (Dakopatts, Copenhagen, Denmark) diluted at 1:200. A subsequent reaction was made by using a Vectastain avidin-biotin-peroxidase complex (ABC) kit. The sections were reacted with diaminobenzidine (DAB) solution and then counterstained with methyl green. Purified mouse IgG (Sigma Chemical Co., St. Louis, MO, U.S.A.) was used as a control primary antibody.

Quantification of cell proliferation in the fracture calluses

The intensity and extent of positive staining in PCNA immunohistochemistry were measured using an image analysis computer system (Nexus Qube image analysis processor, Nexus Inc., Tokyo, Japan). This video-based system distinguishes the density of brown DAB reaction product from the methyl green–stained nuclei, and the ratio of the brown-stained area by DAB reaction against the total area of cell nuclei was calculated as the PCNA score.12 The measurements were carried out at least three times at one location for each specimen, and the average measurement obtained was used as the PCNA score. To negate any difference in the immunostaining conditions between each specimen at different dates after the fracture, the PCNA score of the bone marrow at the tibia from the same limb was measured. The tibia was treated with the same procedures, and the immunohistochemistry was carried out on the same glass slides. It was found that there were no significant differences between the PCNA scores of the tibia bone marrow from different specimens (data not shown).


Localization of PCNA-positive cells: Prior to the fracture, there was no staining in the soft tissues surrounding the periosteum. Very few PCNA-positive cells in the periosteum were observed (Fig. 1).

Figure FIG. 1.

Immunohistochemistry of PCNA in an uninjured femur. (A) Magnification ×5. (B) Shows the rectangular region in (A) (×100). Very few PCNA-positive cells in the periosteum were observed (arrow). im, intramedullary canal; ms, muscle; om, periosteum; co, cortical bone.

Inflammatory reaction at the fracture site: Soon after the fracture, inflammatory cells invaded the fracture site, and a large number of cells was observed. Many PCNA-positive mesenchymal cells were detected in the inflammatory reaction (Fig. 2).

Figure FIG. 2.

Cell proliferation during inflammatory reaction. (A) Masson's trichrome staining (×10). (B) Control immunostaining using purified mouse IgG (×100). (C) PCNA immunostaining (×100). Many PCNA-positive cells were observed in the inflammatory reaction region. No staining was observed in the control.

Intramembranous ossification: Thirty-six hours after the fracture, many PCNA-positive cells were detected in the periosteum (Fig. 3). They were observed near the fracture site and even at a distance of approximately 5 mm from the fracture site in both the fibrous layer and the osteogenic layer. On day 3, newly formed osteoid was seen on the cortex. PCNA-positive cells were slightly decreased in the osteogenic layer, while the number of PCNA-positive cells in the fibrous layer was similar to that noted 36 h after the fracture. On day 7, near the fracture site, the periosteal cells in the fibrous layer were clearly stained by PCNA antibodies. On the contrary, there were few positive cells in the newly formed trabecular bone or in the fibrous layer of the periosteum far from the fracture site. On day 10, PCNA-positive cells were very few in the newly formed trabecular bone or in the periosteum even near the fracture site. On day 14, PCNA-positive cells still remained in the fibrous layer of the periosteum close to the fracture site (data not shown).

Figure FIG. 3.

Cell proliferation during intramembranous ossification. (A,D,G,J) Masson's trichrome staining (×5). (B,C,E,F,H,I,K,L) PCNA immunostaining (×50). The corresponding closed squares assigned in (A,D,G,J). (A-C) Thirty-six hours after the fracture. (D-F) Three days after the fracture. (G-I) Seven days after the fracture. (J-L) Ten days after the fracture.

Chondrogenesis: Cartilage tissue appeared on day 7 in the soft callus adjacent to the newly formed trabecular bone (Fig. 4). Small round cells, presumably premature chondrocytes, were stained by PCNA antibodies rather than hypertrophic cells or differentiated chondrocytes. Furthermore, mesenchymal cells around the cartilage tissue, presumably chondroprogenitor cells, were additionally PCNA positive. On day 10, cartilage was predominant in the soft callus, and the majority of chondrocytes were proliferating chondrocytes. Many PCNA-positive cells were observed in the proliferating chondrocytes. On day 14, almost all the cells in the soft callus had an appearance typical of hypertrophic chondrocytes. No hypertrophic chondrocytes were stained by PCNA antibodies.

Figure FIG. 4.

Cell proliferation during chondrogenesis. (A,B,C) Masson's trichrome staining (×20). (D,E,F) PCNA immunostaining (×50). The corresponding open squares assigned in (A,B,C). (A,D) Seven days after fracture. (B,E) Ten days after fracture. (C,F) Fourteen days after fracture.

Endochondral ossification: Endochondral ossification was observed from the 10th day after the fracture (Fig. 5). In the endochondral ossification front between the hard and soft calluses, new bone mineralization formed “mixed spicules” of new trabecular bone with a cartilaginous core. In the mixed spicules, there were many PCNA-positive cells, presumably premature osteoblasts and endothelial cells.

Figure FIG. 5.

Cell proliferation during endochondral ossification. (A) Masson's trichrome staining (×5). (B) PCNA immunostaining of a rectangular area in (A) (×50). (A,B) Fourteen days after fracture.

Quantification of PCNA-positive cells

Many PCNA-positive cells were observed in the early phase of each cellular event, and the proportion of the proliferating cells gradually decreased as revealed by computed imaging analysis (Fig. 6). In the inflammatory reaction site, the PCNA score was about 20% in the early phase, decreasing later. The PCNA score during fracture repair was highest in the periosteal cells near the fracture site, as early as 36 h after the fracture, decreasing to one-third by day 14. In the intramembranous ossification site away from the fracture site, the PCNA score was almost the same as that near the fracture site 36 h after the fracture, although it dropped to a greater extent later. In the chondrogenesis site, the PCNA score could not be measured using our system since cartilage matrix was also stained with methyl green. The degree of cell proliferation at the endochondral ossification front had not changed significantly by day 14 after the fracture, and it was similar to that noted at the intramembranous ossification site near the fracture site on the same day after the fracture.

Figure FIG. 6.

Quantification of cell proliferation in each cellular event. In the early phase of each cellular event, a high PCNA score was detected, with the degree of PCNA score then gradually decreasing. Closed circle, intramembranous ossification near the fracture site; open circle, intramembranous ossification far from the fracture site; open square, inflammatory reaction; open triangle, endochondral ossification.


We detected proliferating cells in the fracture callus by immunohistochemistry for PCNA. Cell proliferation was generally observed especially at the beginning of each cellular event. Tissue regeneration begins with cell proliferation, and identification of the proliferating cells may be equal to the detection of origin cells in each bone or cartilage formation.

Recently, various new techniques and reagents have become available that make the selected measurements of cellular proliferation rather simple. A number of unique proteins are synthesized as cells traverse the cell cycle, some of which can be detected by immunohistochemistry.13,14 One such protein is PCNA, which appears to be an important auxiliary compound for DNA polymerase delta function.4,15,16 More recently, monoclonal antibodies against PCNA17 have become available, which detect the protein even in formalin-fixed, paraffin-embedded tissues.6,7 Many pathologic studies concerning malignant tumors have shown that the ratio of PCNA-positive cells to total cells was clearly correlated with cell proliferation, histologic grade, and poor prognosis.12,18,19 Our study showed that detection of PCNA can be used to localize and quantify proliferating cells in the physiological cell reaction. We attempted to use the PCNA score of bone marrow cells in the tibiae from the same limb, a section of which was placed on the same glass slide as the callus section, to normalize the PCNA scores in the different callus specimens. However, the PCNA scores of bone marrow cells in each sample were very similar, indicating that the reproducability of the immunostaining technique was sufficient to compare the PCNA score in one sample with others directly.

As we have described, PCNA appears in the cells during the G1 to S phase. This means that the actual proportion of proliferating cells is greater than their PCNA score. The actual proportion seemed to be 2- to 3-fold, although the proportion of G1 and S phases against the whole cell cycle depends on the type of cells.20–24 Therefore, 30–40% of the PCNA score soon after the fracture may indicate that almost all the cells proliferated in the callus.

Regarding cell proliferation research during fracture healing, in 1961 Tonna and Cronkite showed mitogenesis in the fracture callus by3H-thymidine incorporation.1 However, their description principally concerned early periosteal reaction, and the quantification of cell proliferation was limited. We succeeded in detecting proliferating cells in the fracture calluses by immunolocalization of PCNA and quantified the cell proliferation for each cellular event such as inflammatory reaction, intramembranous ossification, and endochondral ossification. Comparing an autoradiographic study with our immunohistochemical study in the early cellular events, similar results were obtained. Cell proliferation is reported to peak at 32 h after a fracture using3H-thymidine incorporation. In our study, this was found to be 36 h using immunohistochemistry with anti-PCNA antibodies. In the autoradiographic study, the labeling index at the peak is 10 times that of the control. In the immunohistochemical study, the PCNA score at the peak was also approximately 10 times that of the control. Therefore, PCNA immunostaining seemed to be of great value in detecting proliferating cells as much as3H-thymidine incorporation.

This study showed that the distribution of proliferating cells and the degree of cell proliferation varied during fracture repair (Fig. 7). Current studies have shown that growth factors are involved in the local regulation of fracture repair. Thirty-six hours after a fracture, the distribution of proliferating cells in the callus is wide and homogenous, most likely because a large amount of mitogenic growth factors, including transforming growth factor beta (TGF-β) from platelets during hematoma formation, is secreted and scattered throughout the whole callus.25 On day 3, the proliferating cells remained within the whole callus, although the degree had decreased. The distribution was not homogeneous. In the fibrous layer, cell proliferation was higher than that in the osteogenic layer. The nonhomogenous distribution of proliferating cells may suggest that the cell proliferation was regulated by locally synthesized growth factors. TGF-β has been reported to be synthesized in the periosteal cells during the early phase of intramembranous ossification during fracture repair.25 On day 7, cell proliferation was observed in the limited portion in the callus. In the fibrous layer of the periosteum, PCNA-positive cells were detected, especially near the fracture site. In the cartilage tissue, small immature chondrocytes and proliferating chondrocytes were stained clearly by PCNA antibodies, although large cells adjacent to the newly formed trabecular bone were not stained relatively. The distribution of TGF-β or FGFs, which has been reported previously,25–27 seemed to be similar to the proliferating cell distribution. Fourteen days after the fracture, cell proliferation was observed at only two portions: in the fibrous layer of the periosteum near the fracture site and in the endochondral ossification front. Osteoblasts or endothelial cells seemed to proliferate according to new bone formation and vascular invasion to the adjacent cartilage tissue. These cells are reported to synthesize TGF-β and FGF,25–27 and these growth factors may regulate proliferation. Although the precise mechanisms remain unclear, increasing evidence indicates that local growth factors play an important role in the regulation of the various stages of fracture repair by inducing cell proliferation, differentiation, and matrix synthesis.

Figure FIG. 7.

Distribution of proliferating cells and PCNA score during fracture repair.

Further examinations of local regulatory factors in the callus and analysis of the distribution of proliferating cells in this study may reveal the regulatory mechanism for cell proliferation during fracture repair.


We thank Mrs. Naoko Tanaka for preparation of the histologic specimens and Mr. Yoshiaki Kuriyama for the medical photographs. This work was supported by a Grant-in-Aid for Scientific Research and a JOTF Grant (#0069). We thank Miss K. Miller (Royal English Language Center, Fukuoka, Japan) for correcting the grammar in this paper.