Introduction

Osteonecrosis of the femoral head (ONFH) is one of the most common orthopedic diseases in clinical practice [1, 2]. It is characterized by an interruption of blood supply to the femoral head due to venous stasis or arterial insufficiency, leading to ischemia of the internal tissues of the femoral head [3]. This, in turn, can result in subchondral bone fracture, necrosis, and collapse of the weight-bearing portion of the femoral head, causing severe functional impairment, pain, and even disability of the hip joint [4]. Non-traumatic avascular necrosis of the femoral head (NONFH) is the main cause of disability in young people, with the misuse of hormones and alcohol greatly increasing its incidence [5, 6]. Currently, femoral head collapse is considered to be the final stage of NONFH [7]. The severe pain and functional impairment caused by femoral head collapse greatly affect patients' quality of life, making the prevention and delay of femoral head collapse an effective approach to preventing and treating NONFH [8, 9].

Current research indicates that the construction of animal models for ONFH can be broadly categorized into several types of modeling methods, including physical methods [10], alcohol-induced methods [11], hormone-induced methods [12], and surgical trauma-induced methods [13]. Nevertheless, these models suffer from the drawback of inducing widespread femoral head necrosis, and to date, no animal model has been established to fully recapitulate the pathological and physiological processes of human ONFH. This represents a significant impediment to the advancement of femoral head necrosis research [14]. Current research suggests that the initiation of collapse in end-stage ONFH may be attributed to microfractures in trabeculae during bone remodeling, which lead to gradual changes in internal mechanical properties and consequent concentration of stress on the trabeculae. This results in the acceleration of bone absorption and establishes a vicious cycle [15]. Based on hip joint movement characteristics, the femoral head's stress column can be classified into the medial, superior, and lateral columns. The lateral column of the femoral head is believed to be a crucial predictive factor for the collapse of ONFH [16].

Currently, the rabbit ONFH model lacks specific investigations into local compression of the femoral head. A novel animal model for ONFH is needed based on a new biomechanical understanding. Therefore, this study aimed to establish a new animal model of ONFH that can simulate the local compression of the human femoral head, providing a fundamental basis for ONFH-related research.

Material and methods

Femoral head compression devices design, fabrication, and characterization

Design and fabrication

Utilizing a 3D printing process (ZRapid Tech. China) and resin as the principal material to manufacture a pressurization device. A compression device was developed with dimensions of a long axis of 6 mm and a short axis of 2 mm, informed by the rabbit acetabulum’s thickness and the gap within it. The thickness of the compression device was found to be positively correlated with the pressure it exerted on the rabbit femoral head. To explore this relationship, four compression devices of varying thicknesses (4.5 mm, 4.0 mm, 3.5 mm, and 3.0 mm) were designed and fabricated (Fig. 1A–D). To secure the compression nail, a rotational clamp device was devised using a rectangular opening to form a cross-rotation (Fig. 1E).

Fig. 1
figure 1

A 4.5 mm intraacetabular femoral compression nails, B 4.0 mm intraacetabular femoral compression nails, C 3.5 mm intraacetabular femoral compression nails, D 3.0 mm intraacetabular femoral compression nails. E Implantation fixation system for femoral head compression device

Full size image

Assessment of the physical properties

The material tensile strength and fracture elongation were measured according to GB/T 1040.1-2006 using a universal material testing machine (Tinius Olsen, U.S.A.). The material bending strength and bending modulus were measured according to GB/T 9341-2008, and the material notch impact strength was measured according to GB/T 1843-2008. The material hardness was determined according to GB/T 2411-2008. The material thermal expansion coefficient was determined using a plastic polymer material expansion coefficient tester (GCPZY-100, China) according to GB/T 1036-89. GB/T standards, also known as Guo Biao/Tui Jian Biao, are the National Standards of the People's Republic of China. These standards, formulated and published by the Standardization Administration of China, encompass technical specifications, test methods, and product standards across various fields.

Isolation, cultivation, and identification of bone marrow mesenchymal stem cells

Bone marrow mesenchymal stem cells (BMSCs) were isolated from the bone marrow of both femurs and tibias of 5-week-old male rats. The isolated BMSCs were then cultured in a complete DMEM medium containing 10% fetal bovine serum (FBS, Gibco, U.S.A.) and 1% (v/v) penicillin/streptomycin and incubated at 37 °C and 5% CO2. After 24 h, the medium was replaced and non-adherent cells were discarded. The medium was changed every 3 days until the cell fusion rate reached 80%. Passage-3 BMSCs were used for all experiments. The identification of cell surface markers (CD29, CD90) of BMSCs was performed using flow cytometry (FACS Calibur, U.S.A.). The BMSCs were cultured in an osteogenic differentiation medium for 14 days and stained for alkaline phosphatase activity at 7 days and with Alizarin Red at 14 days to assess their osteogenic potential.

Cytotoxicity of compression nails

Cell toxicity was evaluated using the CCK-8 assay (TargetMol, U.S.A.). The sterilized femoral compression nails were immersed in α-minimum essential medium (α-MEM, Gibco, U.S.A.) complete medium for 72 h, and then the BMSCs cell suspension was evenly planted in a 96-well plate with a volume of approximately 100 μL per well. The plate was placed in a culture incubator for 24 h (37 °C, 5% CO2). One group was randomly selected and added with 100 μL/each well of α-MEM complete medium, while the other group was added with 100 μL/each well of plant material soaking liquid. The CCK-8 cell toxicity test was performed on 1, 3, 5 and 7 days after inoculation, and the data were recorded and the curve was plotted using an enzyme label instrument (450 nm, 37 °C, FLUO star Omega, Germany). Simultaneously, cell counts were performed alongside the CCK-8 assay, and the data were used to construct additional growth curves. Moreover, the total protein content of the cells was quantitatively determined using the bicinchoninic acid (BCA, CWBIO, China) assay and plotted on a curve.

Experimental animals and design

The animal ethics of this experiment were approved by the Ethics Committee of Xi'an Jiaotong University School of Medicine (Ethics Approval No.: 2021–1413). Thirty-six female New Zealand white rabbits, weighing 4.0–5.0 kg and aged 12 months, were purchased from the Animal Experiment Center of Xi'an Jiaotong University and housed under standard conditions. The rabbits were randomly divided into six groups (A1–A4 and B1–B2 n = 6). Except for the B2 group which is the control group, all rabbits received a single ear vein injection of lipopolysaccharide (LPS) at a dose of 10 μg/kg at 0 h, followed by an intramuscular injection of methylprednisolone (MPS) at a dose of 20 mg/kg at 24 h, 48 h, and 72 h. After 3 weeks, 3D-printed compression devices with different curvatures (3.0 mm, 3.5 mm, 4.0 mm, and 4.5 mm) were implanted in groups A1-A4, group A1 was implanted with a 4.5 mm compression device, Group A2 with a 4.0 mm compression device, Group A3 with a 3.5 mm compression device, and Group A4 with a 3.0 mm compression device (Fig. 2A–D). Confirm the placement of the compression device intraoperatively using an X-ray (Fig. 2E), with a local injection of 0.2% ropivacaine for analgesia. Group B animals underwent only an incision without implantation of the compressive device and received an intramuscular injection of penicillin at a dose of 400,000 U per day for 3 days after surgery. At 8 weeks post-surgery, the modeling effect was evaluated using micro-computed tomography (micro-CT, Siemens Healthineers, Germany). After euthanasia, the femoral head specimens were subjected to micro-CT scanning and reconstruction following decalcification, and pathological sections were prepared.

Fig. 2
figure 2

Operational procedures for creating the femoral head necrosis model. A Selection of the iliac crest and greater trochanter as the sites for an arched incision; B Sequential dissection of muscle and other soft tissues to expose the acetabulum; C Create a window approximately 3 mm away from the acetabular rim; D Implantation of the femoral head compression device. E Intraoperative X-ray imaging confirms the accurate placement of the bone window

Full size image

Gross observation of the specimens

After 8 weeks, femoral head specimens were harvested to observe the contour, color, shape, and smoothness of the femoral head in each group.

Micro-computed tomography (micro-CT) imaging system was used for evaluation

Femoral heads were subjected to fixation in paraformaldehyde for 48 h before imaging using a micro-CT system, which was configured with scanning parameters including a voltage of 70 V, a current of 100 mA, and a pixel size of 10 μm. Subsequently, the acquired images were analyzed and processed using VG Studio MAX software. Three-dimensional reconstruction techniques available in VG Studio MAX were utilized to quantitatively assess various parameters within the subchondral bone region of the femoral head. These parameters included the ratio of bone volume (BV) to total volume (TV), trabecular number, and trabecular thickness.

Histological analysis

Cultured specimens were subjected to fixation in 4% paraformaldehyde for 24 h at 4 °C, followed by decalcification in a 20% EDTA solution. Subsequently, the specimens were dehydrated and embedded in paraffin, and 10-μm-thick sections were prepared. The sections were then stained with hematoxylin and eosin (HE) for observation under an optical microscope, and the number of empty lacunae in the subchondral bone region of the femoral head was quantified.

Statistical analysis

All experimental data analysis was performed using independent-sample t-tests, with statistical analysis conducted using SPSS 24.0 software. A significance level of P < 0.05 was considered statistically significant.

Results

Characterization of the physical properties

The femoral head compression device has an opaque white appearance and was constructed using a 3D printing technique with a layer thickness of 0.1 mm and a density of 1.13 g/cm3. Physical characterization of the device material was performed, and the results are summarized in Table 1.

Table 1 The physical properties of the femoral compression screw
Full size table

Identification of bone marrow mesenchymal stem cells

To identify the isolated cells, the characterization of the cells using flow cytometry revealed positive expression rates of 99.9% and 99.7% for CD29 and CD90, respectively, with negative expression rates of 1.6% and 1.7% (Fig. 3).

Fig. 3
figure 3

Flow cytometry analysis was performed using surface markers CD29 and CD90 to characterize the cells

Full size image

Upon culturing the aforementioned isolated cells in the osteogenic induction medium (OIM) for a duration of 7 days, cells were qualitatively observed to exhibit greater ALP staining compared to their non-induced counterparts. The OIM solution was prepared by supplementing the previously prepared complete DMEM basal medium with 1% glutamine, 0.2% ascorbic acid, 1% β-glycerophosphate, and 0.01% dexamethasone. Cytoplasmic granules and precipitates were observed in abundance and appeared deeply stained (Fig. 4A, B). These results indicate the potential for osteogenic differentiation of the isolated BMSCs.

Fig. 4
figure 4

ALP staining status after 7 days of osteogenic induction and the safranin-O staining status after 14 days of osteogenic induction. A Untreated BMSCs exhibited lower alkaline phosphatase activity as demonstrated by ALP staining. B) BMSCs subjected to osteogenic induction displayed higher alkaline phosphatase activity, as indicated by ALP staining. C Safranin-O staining of untreated BMSCs revealed less calcium deposition. D osteogenically induced BMSCs displayed substantial calcium deposition

Full size image

The ARS staining solution was utilized for the detection of osteoblastic differentiation, as it can bind with calcium deposits on the surface of osteoblasts that have undergone successful osteogenic induction. Following 14 days of osteogenic induction in vitro, the isolated and cultured cells were subjected to ARS staining. The results of ARS staining revealed significantly larger mineralized nodules in osteogenically induced BMSCs, as compared to non-induced BMSCs (Fig. 4C, D). Taken together, these findings suggest that the isolated BMSCs possess the capability to differentiate into osteoblasts upon osteogenic induction.

Assessment of cytotoxicity of the femoral head compression device

Cellular toxicity analysis of the immersion solution of the device was performed using the CCK-8 assay, cell counting method, and BCA assay. The results demonstrated that the cell viability of cells cultured in the immersion solution of the pressure device remained unaffected on days 1, 3, 5, and 7 of cell culture (Fig. 5A). Additionally, no significant statistical difference in cell count was observed between the two groups (Fig. 5B). Moreover, the quantitative analysis of total cellular protein showed no significant statistical difference between the two groups (Fig. 5C). Therefore, these findings collectively indicate that the materials of the device are non-cytotoxic.

Fig. 5
figure 5

Device toxicity assessment. A The results of the CCK-8 assay revealed no significant statistical difference in cell viability between the two groups at days 1, 3, 5, and 7 of cell culture. B The cell counting results demonstrated no significant statistical difference in cell numbers between the two groups at days 1, 3, 5, and 7 of cell culture. C The BCA assay results indicated no significant statistical difference in total protein content between the two groups at days 1, 3, 5, and 7 of cell culture

Full size image

Macroscopic examination

After 8 weeks, the femoral head specimens were obtained and morphological changes were observed in each group. The femoral head of the control group B2 had a normal appearance with a smooth and intact cartilage surface, no collapse, and a white color (Fig. 6F). The femoral head of group B1 had a normal contour, relatively smooth cartilage without defects, and a darker color compared with the control group (Fig. 6E). The femoral head of group A4 was darker in color (Fig. 6D) than that of the control group, with small areas of cartilage peeling off (yellow arrow). In group A3, a focal cartilage defect (blue arrow) was observed in the compression screw site of the femoral head (Fig. 6C). Although the contour of the femoral head in group A2 was normal (Fig. 6B), large areas of cartilage defects and slight collapse of the femoral head (black arrow) were observed in the compression screw site. The femoral head contour was completely lost in group A1 (Fig. 6A), and severe femoral head collapse and cartilage damage were observed (red arrow).

Fig. 6
figure 6

External morphology of the femoral heads in each group. A The femoral head in the A1 group: complete collapse with accompanying bone defect (red arrow); B the femoral head in the A2 group: significant cartilage loss and minor femoral head collapse (black arrow); C the femoral head in the A3 group: localized cartilage defect (blue arrow); D the femoral head in the A4 group: minor cartilage delamination (yellow arrow); E the femoral head in the B1 group; F the femoral head in the B2 group

Full size image

Results of micro-CT analysis

After 8 weeks of modeling, the rabbits in each group were euthanized, and the femoral heads on the operated side were harvested and subjected to micro-CT scanning. The micro-CT results (Fig. 7A) revealed a notable reduction in trabecular bone volume and density in the subchondral bone of all modeled groups compared to the blank control group B2. Furthermore, there was a marked decrease in the thickness and density of the trabecular bone, as well as an increase in spacing. Specifically, the femoral head collapse was observed in A2, and severe femoral head collapse with the bone defect was observed in A1, while no significant collapse was observed in the femoral heads of B1, B2, A3, and A4 groups. Subsequent analysis of micro-CT data using software (Fig. 7B) confirmed that all modeled groups exhibited a significant decrease in the bone volume to total volume ratio, a reduction in the number of trabecular bones, and an increase in the trabecular spacing when compared to the control group B2.

Fig. 7
figure 7

The micro-CT images depict the modeling effects of the femoral head compression device implanted in hip sockets with varying radii of curvature. A Coronal micro-CT scan results in the image, B micro-CT 3D reconstruction image

Full size image

Histological analysis

After euthanizing New Zealand white rabbits at 8 weeks post-surgery, the femoral heads on the operated side were decalcified, sectioned, and subsequently subjected to HE staining. The histological evaluation of the HE-stained sections (Fig. 8) revealed a significant loss of subchondral bone in all experimental groups as compared to the control group, as evidenced by the presence of numerous osteolytic lesions (black arrows). Notably, the area of femoral head collapse in the A1 group showed marked infiltration of inflammatory cells (blue arrow). These histological findings corroborate the micro-CT results, revealing a correlation between the degree of femoral head necrosis in groups A1–A4 and the curvature radius of the loading device. Specifically, we observed that a larger curvature radius was associated with a higher degree of femoral head necrosis, further supporting the existing correlation between these factors.

Fig. 8
figure 8

The findings obtained from HE staining: osteolytic lesions (black arrow); infiltration of inflammatory cells (blue arrow)

Full size image

Discussion

Over the past century, the pathogenesis of ONFH has evolved from the initial speculation that it was associated with vascular injury to later theories involving fat embolism and other mechanisms. Currently, research has advanced to the genetic level, and investigations have never ceased [17]. Research tools for ONFH have evolved from gross specimens to include various animal models such as mice [18], rabbits [13], pigs [19], and ostriches [20], among others, which have contributed to a diversity of research methods. However, a review of the current research reveals that there are still some shortcomings in the study of ONFH. For instance, as mentioned in Introduction, there are significant differences between the hip joint-femoral head structure of quadrupeds and that of humans, and their load characteristics differ considerably. Therefore, it is difficult to simulate the real state of femoral head load-bearing from a biomechanical perspective [21]. Moreover, bipeds are mostly non-mammals, which have significant differences in hormone sensitivity compared to mammals, making it difficult to simulate the natural course of ONFH induced by hormones [12]. Animal models are the foundation of basic medical research, and therefore, based on the needs of the research, it is desirable to improve the existing models through some technical means, to open up new directions for the study of ONFH.

We selected rabbits as the animal model for a novel ONFH model primarily due to their unique anatomical characteristics. Specifically, the rabbit hip socket is positioned near the surface on both sides of the spine, allowing for the posterior entry and exposure of the lateral aspect of the hip socket without the need for a muscle incision. This technique reduces surgical trauma and subsequent errors during modeling. Additionally, the larger femoral head size of rabbits, coupled with stronger femoral ligaments and joint capsules, facilitates consistent positive pressure interaction with the compression device. Notably, this force remains unaffected by positional changes [22], enabling an accurate simulation of femoral head stress in the upright human posture and enhancing the femoral head necrosis model.

In the extant literature, a plethora of techniques have been utilized to induce ONFH, such as vascular occlusion [13, 23], steroid administration [24,25,26], cryogenic insult [27], dislocation of the hip joint [28], exposure to high pressure [29], hypersensitivity reactions [30], endotoxin injection [31], and various combinations thereof. However, the majority of animal models fail to develop advanced-stage femoral head collapse and instead perish as a result of invasive systemic injuries leading to sepsis or death [32]. Therefore, the judicious production of localized subchondral necrosis within the femoral head has emerged as a pivotal strategy in the establishment of a robust ONFH model. To this end, the fabrication of internal bone devices represents a promising avenue of investigation in this field. Reed et al. developed a cryoprobe that was guided by radiography to induce necrosis in the subchondral bone [33]. Goetz further validated the effectiveness of this technique by conducting a finite element analysis on the necrotic area [34, 35]. Another study by Fan used a more advanced cryo-heating probe to create an ONFH model [36]. However, due to the size and cost of this equipment, the use of this technique is limited in establishing ONFH models.

In the present study, we developed a pressurized device with dimensions comparable to the hip joint gap to apply a compressive load to the femur. Additionally, a rotating clamping device was designed to apply the load to the upper lateral aspect of the femoral head. The pressurized device was fabricated using 3D printing technology with resin as the material. The experimental findings indicate that the resin material employed for modeling does not exhibit any cellular toxicity. Femoral head specimens modeled with steroids (A1-A4, B1) exhibit a considerable number of empty osteocyte lacunae on HE-stained sections compared to the control group (B2). The application of the pressurized device (A1-A4) results in localized cartilage defects and subchondral bone fractures in the pressurized device region, leading to femoral head collapse. Additionally, the extent of femoral head damage varies with the curvature radius of the pressurized device, which can be modified to simulate different stages of femoral head necrosis. For instance, in the early stages of femoral head osteonecrosis, human pathological observations have revealed subtle morphological alterations accompanied by localized cartilage detachment. This corresponds to the femoral head damage state observed in Group A4 of our study. During the intermediate phase of femoral head osteonecrosis, which represents the period when a majority of patients seek medical intervention, human pathological investigations have demonstrated mild to moderate collapse of the femoral head while preserving the joint space. This parallels the femoral head damage state observed in Groups A3 and A4 of our study. In the advanced stage of femoral head osteonecrosis, human pathological examinations have documented significant femoral head collapse, mirroring the femoral head damage state observed in Group A1 of our study.

Nonetheless, this experiment is subject to certain limitations. Firstly, the study solely focused on a single observation point at 8 weeks post-surgery, which precluded the continuous monitoring of the progression of femoral head necrosis. Secondly, the procedure for inserting the pressurized device necessitated adequate preoperative training.

Conclusion

This study has developed a novel method for modeling femoral head necrosis that incorporates stress factors using 3D printing technology and principles of biomechanics. This approach enhances the fidelity of the model to the clinical scenario and improves its utility as an experimental platform for osteonecrosis of the femoral head. The technique holds promise for advancing our understanding of the pathogenesis and potential therapeutic interventions for this debilitating condition.