**3. Results**

### *3.1. Animals and Blood Parameters at the End of IMT Stage 1*

All rats tolerated surgical procedures well and gained weight steadily from day 4 after stage 1 surgery onwards. Two animals (one PMMA and one metakaolin) were excluded from the analysis due to infection-related fixator failure. Given the inflammatory potential of the aluminosilicate present in metakaolin, we determined blood cell counts for the animals to assess systemic inflammation at the time of death. White blood cell counts and red blood cell parameters were similar between the PMMA and metakaolin groups (Figure 1A). Serum P1NP and TRAP-C concentrations and ratios were similar in the two groups, suggesting that bone remodeling activity four weeks after the creation of the bone defect was similar in the PMMA and metakaolin groups (Figure 1B).

**Figure 1.** Panel ( **A**) shows the hematological parameters of animals four weeks after spacer implantation. Blood was collected into EDTA-containing tubes when the animals were killed, and blood parameters were determined with an optical hematology analyzer (MS-9, Melet Schloesing) with rat-specific analysis software. Panel (**B**) shows the serum levels of bone turnover markers, as determined by ELISA, four weeks post-spacer implantation. Concentrations of a bone formation marker P1NP (**top**) and a bone resorption marker TRAP-C (**middle**) were determined, and turnover for bone remodeling was evaluated by calculating the P1NP/TRAP-C ratio (**bottom**).

### *3.2. Comparison of Biological Properties between Metakaolin- and PMMA-Induced Membranes* 3.2.1. Membrane Architecture and Cellularity

We previously showed that bioactive IMs are organized as bilayered structures and have a rich cellular network. Figure 2 illustrates typical sections of PMMA-induced (Figure 2A) and metakaolin-induced (Figure 2B) membranes, with an inner layer in contact with the biomaterial, including fibroblasts, lymphocytes and macrophages. A thick outer layer principally consists of fibroblasts with a dense vascular network in contact with the muscle. The quantification of DAPI-stained nuclei showed cell density to be slightly higher in metakaolin-IMs than in PMMA-IMs, although this difference was not statistically significant (4553 ± 51 nuclei/mm<sup>2</sup> in the PMMA group versus 5882 ± 695 nuclei/mm<sup>2</sup> in the metakaolin group, *p* = 0.15).

**Figure 2.** Representative hematoxylin-eosin-saffron-stained sections of (**A**) PMMA-IMs and (**B**) metakaolin-IMs showing their histological organization. \* Indicates the site of the spacer before its removal. The right panel (**C**) illustrates the semi-automatic counting process of DAPI-stained nuclei and the comparison of cell density (the mean number of nuclei per mm<sup>2</sup> ± SEM) in PMMA-IMs and metakaolin-IMs.

### 3.2.2. Gene Expression within Membranes

We compared the expression of key inflammation-related genes involved in wound healing between PMMA-IMs and metakaolin-IMs. Real-time RT-PCR analysis (Figure 3) showed that the relative levels of insulin-like growth factor-1 (*IGF-1*), vascular endothelial growth factor (*VEGF*), interleukin-6 (*IL-6*) and interleukin-1-beta (*IL-1β*) expression was similar in PMMA-IMs and metakaolin-IMs. However, in metakaolin-IMs, transforming growth factor-β (*TGF-β*) mRNA levels were significantly upregulated (fold change = 2.74, *p* = 0.016), potentially enhancing bone healing and regeneration.

### 3.2.3. Secretion of Proteins by the IM and BMP-2 Expression within Membranes

IMs form a biological chamber containing secreted angiogenic and osteogenic factors around the bone defect. We, therefore, performed a descriptive mass spectrometric analysis to compare the secretome profiles of PMMA-IMs and metakaolin-IMs. We detected a total of 688 proteins in both groups (Figure 4A), 683 (99.3%) of which were not differentially secreted between PMMA-IMs and metakaolin-IMs (i.e., the frequency of secretion of these proteins was similar in the two groups). In contrast, the secretion frequency differed between the

two groups for five proteins (0.72%): four were more frequently secreted by metakaolin-IMs, and one was more frequently secreted by PMMA-IMs. The four proteins more frequently secreted by metakaolin-IMs were identified as cysteine- and glycine-rich protein 3, the GON7 subunit of the KEOPS complex, carboxylic ester hydrolase and synaptogyrin. These proteins are involved in various metabolic pathways, including myogenesis and apoptosis. The protein most frequently secreted by PMMA-IMs was the neurotrophin tyrosine kinase receptor 1 TrkA L0 variant, which is involved in the MAPK pathway.

**Figure 3.** Relative levels of *Igf-1*, *Vegf*, *Il-6*, *Il-1β* and *Tgf-β* mRNA in four-week-old PMMA-IMs or metakaolin-IMs. Data are expressed as the mean ± SEM, \* *p* < 0.05.

**Figure 4.** (**A**) Secretome profiles of PMMA-IMs and metakaolin-IMs. In total, 683 proteins with similar frequencies of secretion in the PMMA and metakaolin groups were identified by mass spectrometry. By contrast, five proteins were differentially secreted: four proteins were more frequently secreted by the metakaolin-IMs, and the other was more frequently secreted by PMMA-IMs. (**B**) Representative histological slide of in situ BMP-2 immunostaining in PMMA-IMs (left panel) and metakaolin-IMs (right panel). The diagram shows the percentage of the area of the collected membranes positive for BMP-2; \* *p* < 0.05.

We also investigated the expression of the pro-osteogenic mediator BMP-2 within the membranes by immunohistochemistry (Figure 4B). BMP-2-expressing cells were uniformly distributed throughout the membranes, but BMP-2 staining was more intense in metakaolin-IMs than in PMMA-IMs. Furthermore, the percentage of the membrane area positive for BMP-2 was 1.9 times higher in metakaolin-IMs than in PMMA-IMs (25.25% ± 4.83% versus 48.41% ± 7.11%, *p* = 0.0.21).

### *3.3. Macrophage Distribution in IMs*

We characterized the macrophage populations in IMs by immunofluorescence analysis to detect both CD68 and CD206, with CD68 used as a phenotypic marker of the M1-like subtype and CD68+/CD206+ double labeling as a marker of the M2-like subtype (Figure 5A). CD68-/CD206+ cells were defined as satellite cells. Semi-automatic quantification revealed that the total macrophage population was significantly larger in metakaolin-IMs than in PMMA-IMs (25.77% ± 5.48% versus 48.11% ± 5.77%, *p* = 0.02; Figure 5B). This larger total macrophage population reflected a significant expansion of the M1-like population (20.01% ± 3.81% versus 36.30% ± 4.45%, *p* = 0.02) and a smaller, non-significant expansion of the M2-like subtype (5.75% ± 2.40% versus 11.81% ± 1.72%, *p* = 0.07).

**Figure 5.** Identification and quantification of M1-like and M2-like macrophages in PMMA-IMs and metakaolin-IMs: (**A**, **top panel**) Representative immunolabeling with anti-CD68 (red) and anti-CD206 (green) antibodies and DAPI (blue) nuclear staining. (**A**, **bottom panel**) Illustration of semi-automatic macrophage quantification. Green objects represent satellite cells, and red and yellow objects correspond to M1-like and M2-like macrophages, respectively. Histograms show (**B**) the % total (M1-like + M2-like) macrophages in IMs, (**C**) the % M1-like macrophages and (**D**) M2-like macrophages. \* *p* < 0.05.

### *3.4. Bone Healing after IMT Stage 2 Surgery*

We compared the bone-healing properties of PMMA-IMs and metakaolin-IMs, by performing a quantitative analysis of callus volume within the osteotomy region 10 weeks after bone graft implantation in the IM cavities. We observed a small, non-significant difference in new bone volume within the defect, with a slightly greater volume in the

 

metakaolin group (16.65% ± 3.59% versus 26.37% ± 6.5%, *p* = 0.22, Figure 6A) than in the PMMAgroup.

**Figure 6.** Bone formation was assessed by microCT. (**A**) The quantitative and comparative analysis showed no difference in bone volume between the metakaolin and PMMA groups. (**B**) Representative three-dimensional reconstructions of the region of interest in the two groups.
