Patient recruitment and ethics statement

The study was conducted according to the principles of the Helsinki Declaration and was approved by the Ethics Committee of Hannover Medical School (Approval number 3377-2016). Participation was entirely voluntary, with patients providing written informed consent before inclusion in the study.

Between January 2022 and February 2023, 43 femoral heads were collected from patients at Hannover Medical School. Inclusion criteria were (1) patients with hip OA undergoing total hip arthroplasty because of severe symptoms, failed conservative treatment, and mobility issues; and (2) elderly patients (> 70 years of age) with displaced femoral neck fractures (Garden III or IV) considered for bipolar hip arthroplasty. Patients with bone or other organ tumors were excluded.

After the inclusion of the femoral heads, they were grouped on the basis of the results of preoperative X-ray examination using the Kellgren-Lawrence (KL) grading system; independent grading was performed by two trained orthopaedic surgeons. Accordingly, the patients and their femoral heads (n = 43) were divided into four groups: KL 1 (n = 6), KL 2 (n = 14), KL 3 (n = 10), and KL 4 (n = 13), i.e., patients with KL scores of 1, 2, 3, and 4 respectively.

Preparation of femoral head cubes

The femoral heads were removed during surgery, fixed in formalin at 4 °C for 24 h, and then stored in 70% ethanol (Th. Geyer GmbH & Co. KG, Renningen, Germany) at 4 °C. Their volume was measured via water displacement. Marked with 1-cm parallel lines, the heads were cut with a surgical oscillating saw (see Fig. S1). Cartilage volume was assessed using the Cavalieri method. Fan-shaped sections were created, preserving a 1-cm outer cartilage circumference. Cubes were cut from 1 cm below the cartilage, totalling 1 cm3 in volume. A systematic uniform random sampling (SURS) selected three cubes from every third counted cube (Boyce et al. 2010), resulting in 129 cubes stored in 70% ethanol.

Embedding femoral head samples

Biopsies were embedded natively in an undecalcified state using Technovit 9100 New® (Heraeus Kulzer GmbH, Wehrheim, Germany). Briefly the samples from the femoral head as described in Sec. “Preparation of the femoral head cubes” underwent dehydration in ethanol and isopropanol series (DVH Chemie-Vertrieb GmbH & Co Hannover KG, Hannover, Germany), xylene incubation (J.T.Baker, Deventer, Netherlands), pre-infiltration, and infiltration before being embedded in Technovit 9100 New® using a polymerization solution (Heraeus Kulzer GmbH, Wehrheim, Germany) (Bernhards et al. 1992). Then the cubes were sliced into 5-µm sections using a microtome.

Histology staining

Hematoxylin and eosin (HE) staining (Carl Roth GmbH, Karlsruhe, Germany) and Safranin O staining were conducted according to Cardiff et al. (2014) and Rosenberg (1971). Immunohistochemistry (IHC) staining was performed according to Magaki et al. (2019). Tissue slides underwent MEA immersion (Merck, Darmstadt, Germany), decalcification, ethanol exposure, Tris-buffered saline rinsing (Sigma-Aldrich Chemie GmbH, Munich, Germany), EDTA incubation (Merck, Darmstadt, Germany), and hydrogen peroxide quenching (Abcam GmbH, Berlin, Germany). IHC staining was performed using the ZytoChem Plus (HRP) Polymer Kit (Zytomed Systems GmbH, Berlin, Germany), involving blocking for 10 min, ready-to-use CD34 antibody (#IR63261-2, Agilent, Santa Clara, USA) application for 1 h, post-block solution (ZytoChem Plus (HRP) Polymer) was added for 20 min, and visualization using ZytoChem Plus (HRP) Polymer (POLHRP-006, Zytomed Systems GmbH, Berlin, Germany) at 20 ℃ for 30 min was performed. TBS was used as the negative control. Although several known antibodies like CD31 or VEGF are well known for blood vessel staining we decided to use the antibody against CD34 as this was able to mark blood vessels in the decalcified tissue. Limitations are potential inaccuracy due to cross staining of endothelial cells and hematopoetic cells.

Counterstaining was done with Mayer’s hemalum solution (Merck, Darmstadt, Germany) and TBS bluing (Sigma-Aldrich Chemie GmbH, Munich, Germany), followed by ethanol and xylene (J.T.Baker, Deventer, Netherlands) treatments for slide preparation. Finally, the slides were mounted with Vitro-Clud embedding glue (R. Langenbrinck GmbH, Emmendingen, Germany).

Evaluation of degenerative cartilage using the Mankin score

This grading system comprises four parts: structure integrity, cellular abnormalities, matrix staining, and tidemark integrity. The scores for all four features are added to obtain a total Mankin score, which can range from 0 (normal cartilage) to 14 (severe OA) (Pauli et al. 2012; van der Sluijs, et al. 1992). In the present study, all 129 slides stained with Safranin O/Fast Green were evaluated using the Mankin score.

Measurement and calculation of volumes of tissues and cellsFemoral head volume

First, the femoral head was freed from the extra tissue and rest of the femoral neck. Total femoral head volume was obtained using Archimedes’ principle using water displacement. The weight change was recorded which corresponded to the femoral head volume (V(FH)) (Scherle 1970).

Cartilage volume

A grid with 200 evenly spaced points (0.5 cm) was used to assess volume density. Points hitting the femoral head and cartilage were counted. The cartilage volume density (P(cartilage)) was calculated and multiplied by the femoral head volume to afford the total cartilage volume (V(cartilage)).

$$}\left( }} \right) = }\left( }} \right) \times \frac}\left( }} \right)}}}\left( }} \right)}}$$

(1)

Subchondral region volume

The volume of the femoral head without cartilage was calculated by subtracting the cartilage volume from the total femoral head volume. The sphere volume formula was used to determine the radius of the femoral head without cartilage (R(FH)). After 1 mm was subtracted from the radius, the inner femoral head volume was recalculated. The subchondral region volume 1 mm below the tidemark (V(SB 1mm)) was obtained by subtracting the inner femoral head volume from the total.

$$}\left( }} \right) = \sqrt[3]}}\left[ }(})}\left( }} \right)} \right]}}}}$$

(2)

$$}\left( }1}} \right) = }(})}(})}\frac}}}}\pi \left( }\left( }} \right)}} \right)^ }$$

(3)

Cartilage thickness

HE-stained femoral head slides were imaged using the VHX microscope (Keyence GmbH, Neu-Isenburg, Germany). The “2 Points” tool in the VHX software was used to select the top and bottom of the cartilage, automatically calculating the thickness. This was performed for three regions (left, middle, right) of each femoral head slice, and the average cartilage thickness for each sample was reported.

Chondrocyte and ECM volumes

The volume densities and total volumes of chondrocytes and extracellular matrix (ECM) were calculated after measuring cartilage thickness. In the VHX software, “Scale” mode was switched to “Mesh” mode with a 100 μm × 100 μm grid. Asterisks at grid intersections, including those on chondrocytes, were counted. The volume densities and total volumes of chondrocytes (Vv and V(chondrocyte, cartilage)) and ECM (Vv and V(ECM, cartilage)) in cartilage were then calculated using standard formulae.

$$\beginc} }_}} \left( }/}} \right) = \frac}\left( }} \right)}}}\left( }} \right)}}} \\ \end$$

(4)

$$}\left( },}} \right) = }\left( }} \right) \times }_}} \left( }}}}}}} \right)$$

(5)

$$\beginc} }_}} \left( }/}} \right) = \frac}\left( }} \right) - }\left( }} \right)}}}\left( }} \right)}}} \\ \end$$

(6)

$$}\left( },}} \right) = }\left( }} \right) \times }_}} \left( }/}} \right)$$

(7)

Subchondral bone and bone marrow volumes

The point counting method was used to calculate volume densities and total volumes of subchondral bone and bone marrow. Subchondral bone volume was measured at 1 mm below the tidemark using ×40 magnification. A 500 μm × 500 μm grid with 160 intersection points was applied to cartilage images, and points intersecting trabecular bone were counted. Volume densities and total volumes for subchondral bone (Vv and V (SB/SR)) and bone marrow (Vv and V (BM/SR)) were calculated using standard formulae.

$$\beginc} }_}} \left( }}}}}}} \right) = \frac}\left( }} \right)}}}\left( }} \right)}}} \\ \end$$

(8)

$$}\left( },}} \right) = }\left( }} \right) \times }_}} }$$

(9)

$$\beginc} }_}} \left( }/}} \right) = \frac}\left( }} \right) - }\left( }} \right)}}}\left( }} \right)}}} \\ \end$$

(10)

$$}\left( },}} \right) = }\left( }} \right) \times }_}} \left( }/}} \right)}$$

(11)

Blood vessel volume

At ×100 magnification, the subchondral region 1 mm below the tidemark was selected using the VHX software’s “Rectangle” tool to calculate its area (A(SR 1mm)). The “Free Line” tool outlined the blood vessels, and the software determined their area (A(vessels)). The volume density of blood vessels (Vv(vessels/SR 1mm)) was calculated by dividing the area of blood vessels by the area of the subchondral region. The total volume of blood vessels (V(vessels, SR 1mm)) was obtained by multiplying the volume density of blood vessels by the volume of the subchondral region.

$$\beginc} }_}} \left( }/}1}} \right) = \frac}\left( }} \right)}}}\left( }1}} \right)}} } \\ \end$$

(12)

$$}\left( },}1}} \right) = }\left( }1}} \right) \times }_}} \left( }/}1}} \right)$$

(13)

Blood vessel surface area

Femoral head images were overlaid with a transparent grid of 34 cycloids aligned vertically to the articular surface. The numbers of intersections (I) with blood vessel boundaries, left endpoints (N), and total cycloid length (LC) were calculated. The surface area density of blood vessels (Sv(vessels/SR 1mm)) was computed, and the total blood vessel surface area in the subchondral region 1 mm below the tidemark was determined by multiplying the surface area density by the subchondral region volume.

$$\beginc} }_}} \left( }/}1}} \right) = \frac}}}}} } \\ \end$$

(14)

$$}\left( },}1}} \right) = }\left( }1}} \right) \times }_}} \left( }}}}}}1}} \right)}$$

(15)

Blood vessel length

To calculate blood vessel length, we assumed the blood vessel to be a long and perfect cylinder. By combining the cylinder volume formula with surface area formula, the formulae for calculating the length density (LV(vessels/SR 1mm)) and total length (L(vessels, SR 1mm)) of the blood vessels were derived as follows:

$$\beginc} }_}} \left( }/}1}} \right) = \left( }}} \right)\left( }_}} \left( }/}1}} \right)^ }}}_}} \left( }/}1}} \right)}}} \right)} \\ \end$$

(16)

$$L\left( },}1}} \right) = }\left( }1}} \right) \times }_}} \left( }/}1}} \right)$$

(17)

Micro CT

Four cartilage-bone biopsies from different KL grades underwent microCT imaging. Previously examined histologically, the goal was to obtain high-resolution scans and overlay the microCT images with their histological counterparts. The biopsies, embedded in polymethyl methacrylate (PMMA) blocks, were positioned in the sample holder for microCT scanning.

The microCT scans were performed utilizing a high-definition micro-CT system (µCT 50, SCANCO Medical AG, Switzerland) with specific imaging parameters. The scan settings included an energy level of 90 kVp, intensity set at 155 µA, voxel size of 7.4 µm, and an integration time of 1000 ms. The subsequent segmentation of the acquired images was accomplished using Scanco’s OpenVMS script, applying standard threshold settings for image segmentation.

Statistical analysis

Statistical analyses and image creation were conducted using GraphPad Prism 9 (v.9.3.1, GraphPad Company, Boston, USA). Data normality was assessed with the Shapiro–Wilk test. Normally distributed data were expressed as mean ± standard deviation and analyzed using one-way analysis of variance (ANOVA), while non-normally distributed data were presented as median and interquartile range (IQR) and analyzed with the Kruskal–Wallis H test. Spearman’s rank correlation and linear regression assessed relationships between cartilage, subchondral bone, and blood vessels. Rank analysis of covariance (ANCOVA) and partial correlation were used to control femoral head volume confounding (Quade 1967). A p value ≤ 0.05 indicated statistical significance.