**3. Results**

### *3.1. Detection of Immune Cell Accumulation around Biomaterial*

It was shown that biomaterial triggered the recruitment of macrophages at the site of the bone wound and enhanced wound healing [1,7,32,54]. Rat femur was operated on with biomaterial (with the Masquelet induced membrane technique), and three weeks later the whole bone together with femur and biomaterial was harvested. The non-operated femur from the second leg of this rat was used as a control (called non-operated femur). The Masquelet induced membrane technique consists of two different operative phases. In the first operation, a fixator was placed around the bone, and the missing bone fragments were filled with biomaterial. Three weeks after surgery, a Masquelet induced membrane was formed around the biomaterial. Three weeks later, during the second operation, the biomaterial was removed, but the Masquelet induced membrane remained, and the biomaterial was replaced with a bone graft. To examine the architecture and the regeneration region of the bone, cryosections of femurs were analyzed by Hematoxylin and phloxin (HP) histological coloration (Figure 2). The control femur was used as a reference to observe bone and muscle in their native structure (Figure 2A). In the operated femur, the Masquelet induced membrane, the regenerated region close to the Masquelet induced membrane, and the biomaterial were observed (Figure 2B). Some immune cells were identified by their structure at higher microscope magnification in the regeneration region (Figure 2C), hypothesized as macrophages.

**Figure 2.** Histologic and microscopic analysis of rat femur: ( **A**–**C**) histologic analysis of the femurs; histopathological image of ( **A**) non-operated; and (**B**) operated femurs of rats stained with hematoxylin and phloxin; ( **C**) expanded view: high magnification image of the area within the red rectangle in image (**B**); ( **D**,**E**) identification of macrophages M1 and M2 in the operated femur; anti-CD68 (Alexa488, turquoise fluorescence), labeling the M1 and M2 macrophages; anti-CD206 (Alexa568, red fluorescence) labeling the M2 macrophages and satellite cells; nuclear staining with DAPI (blue fluorescence).

In situ visualization of macrophages is quite problematic. The cluster of differentiation (CD) CD68 protein is one of the most common monocytes/macrophages marker proteins [55], but in other mononuclear phagocyte cells and non-hematopoietic cells (mesenchymal stem cells, fibroblast, endothelial, and tumor cells) weak expression can be

detected [56]. Other markers such as CD206 and CD163 mostly recognize M2 macrophages, but CD206 protein is also expressed in satellite and CD163 protein in dendritic cells [56,57]. In situ identification of human and mouse M2 macrophages can be performed by double immunolabeling with CD206 or CD163 together with CD68 antibodies (Table 1) [47,58].

**Table 1.** Cell phenotypes.


The same region was examined by the traditional immunostaining method (with anti-CD68 and anti-CD206 antibodies) to detect if M1-like or M2-like macrophages were localized in the regeneration region (Table 1). We observed M1-like macrophages expressing CD68 protein only and M2-like macrophages expressing both CD68 and CD206 proteins (Figure 2D). In the absence of a primary antibody (negative control), no expression of these markers was detected (Figure 2E).

To go further in our investigation, immunostaining was performed on both the nonoperated femur and the operated femur (Figure 3) in large sections (2 mm × 1.2 mm). Using bright field microscopy, the non-operated femur showed a representative architecture of the bone (Figure 3A) and an absence of fluorescence signal (Figure 3B). In the operated femur, the three expected regions (the regeneration region, the Masquelet induced membrane, and surrounding muscle) were observed (Figure 3C). Using CD68 and CD206 immunostaining, both M1-like and M2-like macrophages were detected (Figure 3D). However, the M1-like and M2-like macrophage repartition differed among the three zones (Figure 3E–G). A similar amount of M1-like and M2-like macrophages were observed in the regenerating region (Figure 3E) compared to the interface (between the regeneration region and Masquelet induced membrane) region, where mostly M1-like macrophages (Figure 3F) and a predominantly M2-like macrophage population were detected (Figure 3G).

Another widely accepted macrophage marker is the ionized calcium-binding adaptor molecule 1 (Iba1) (Table 1) [59–61]. Labeling with the anti-Iba1 antibody together with the anti-CD206 antibody showed very similar results (Figure S1). The only difference that has been detected is that fewer M1-like macrophages have been detected by Iba1 labeling in the interface region than with CD68 labeling. This result suggests that single CD68 positive cells include other cells than macrophages that have not ye<sup>t</sup> been identified.

In summary, both M1-like and M2-like macrophages were identified in the injured region of the operated femur. After quantification of 1000 cells, 16.55% (CD68/CD206) and 17.5% (Iba1/CD206) were identified as M2-like macrophages compared to 6.17% (CD68/CD206) and 4.1% (Iba1/CD206) of M1-like macrophages (Table 2). To confirm this result, immunostaining with an anti-CD163 antibody was performed because this protein is also expressed by M2-like macrophages [16]. Next, the CD163 and CD206 co-labeling was performed to identify the M2-like macrophages and satellite cells in the same pictures (Figure S2). In this condition, 21.6% of M2-like macrophages were identified (Figure 4A and Table 1), while no expression of the protein was detected in the negative control (Figure 4B).

**Table 2.** Quantification of M1-like and M2-like macrophages in the regeneration region of the operated femur (quantitative analysis, based on random examination of 1000 cells in each of the conditions).


**Figure 3.** Identification of M1-like and M2-like macrophages in the rat femurs: (**A,B**) immunolabeling with anti-CD68 and CD206 antibodies of the non-operated; and (**C**–**G**) operated femurs; bright-field image together with immunolabeling of the non-operated (**A**) and operated (**C**) femurs; (**E**–**G**) expanded view: high magnification image of the area within the red rectangle in image C; anti-CD68 (Alexa488, turquoise fluorescence), labeling the M1 and M2 macrophages; anti-CD206 (Alexa568, red fluorescence) labeling the M2 macrophages and satellite cells; nuclear staining with DAPI (blue fluorescence); thin arrow: M1-like macrophages; thick arrow: M2-like macrophages.

**Figure 4.** Identification of M2-like macrophages in the operated femur: (**A**) immunolabeling of the operated femur with anti-CD163; (**B**) negative control of the immunolabeling; anti-CD163 (Alexa488, turquoise fluorescence), labeling the M2 macrophages; nuclear staining with DAPI (blue fluorescence).

### *3.2. Detection of Macrophages in Other Regions*

Previously we showed the capacity to detect macrophages using immunostaining in a non-decalcified cryo-fixed bone in the regeneration region (region 1). We further examined whether the different regions of the femur far from the injury region contained macrophages with phenotypic differences. In this context, three other regions were investigated: the muscle far from the bone or the wound (region 2), the periosteum (region 3), and the bone (region 4) (Figures 5 and S3). No difference between operated and non-operated femur tissues were observed (Figure 5). In region 2, no macrophages were present; only satellite cells could be detected (Figure 5B,C). In region 3, both M1-like and M2-like macrophages were detected close to osteoclasts, as it is multinucleated cells expressing CD68 (Figure 5D,E). In region 4, several cells expressing CD68 protein were detected. This result seemed to be coherent as macrophages were also localized in the bone marrow. Surprisingly, cells expressing only the CD206 protein were also detected (Figure 5F,G). This result requires further examination. No expression of both CD68 and CD206 markers was detected in the negative control (Figure S3).

### *3.3. Detection of Macrophages Based on Their Autofluorescence Feature*

It is well known that bone tissue has strong autofluorescence [62,63]. This makes it rather difficult to analyze the operated tissue with immunolabeling (using fluorophorelabeled antibody) because there was an abundant autofluorescence in the regenerating region (Figure 6). Autofluorescence may appear from structural proteins, such as collagen and elastin, but other endogenous fluorophores, such as Flavin-type molecules, are also often localized in cells [64,65]. In particular, immune cells, such as macrophages and granulocytes contain a large amount of phagosome/phagolysosome in their cytosol [66]. Sections of the operated femur tissues had relatively intense (excitation with 488 nm (Figure 6A–D,F) and excitation with 568 nm (Figure 6B–D,F)) autofluorescences exclusively in the cells with endosome-like structure (Figure 6E,G). The sections of the operated femur were subjected to different treatments, TrueVIEW Autofluorescence Quenching Kit, Blue Evans, or Black Soudan, to ge<sup>t</sup> rid of endogenous fluorescence (data not shown). Unfortunately, the endogen fluorescence was resistant to all treatments used so far.

**Figure 5.** Identification of M1-like and M2-like macrophages in the rat femurs: (**A**) presentation of the different regions of the femurs; (**B**,**D**,**F**) operated; and (**C**,**E**,**G**) non-operated femurs labeled with anti-CD68 and anti-CD206 antibodies; anti-CD68 (Alexa488, turquoise fluorescence), labeling the M1 and M2 macrophages; anti-CD206 (Alexa568, red fluorescence) labeling the M2 macrophages and satellite cells; nuclear staining with DAPI (blue fluorescence); thin arrow: satellite cells; arrowhead: osteoclast.

**Figure 6.** Observation of the autofluorescence in the operated rat femur: (**A**) detected autofluorescence after excitation with light (488 wavelengths, turquoise fluorescence); (**B**) detected autofluorescence after excitation with light (568 wavelengths, red fluorescence); (**C**) nuclear staining with DAPI (blue fluorescence); (**D**) merged image; (**E**) bright-field image; (**F**) expanded view: high magnification image of the area within the red rectangle in image (**D**); (**G**) expanded view: high magnification image of the area within the red rectangle in image (**E**).

### *3.4. Investigation of Macrophage Fluorescence by Microspectroscopic Analysis*

As the endogen fluorescence was very strong in the bone tissue, further investigation was needed for macrophage characterization. Turquoise fluorescence was detected in macrophages, but the autofluorescence prevented any conclusion as both specific fluorescence (coming from the immune signal) and autofluorescence were detected in a mixed signal (Figure 7A). Microspectroscopical analysis in situ facilitates the differentiation between specific fluorescence and autofluorescence. The ratio between the intensity of specific fluorescence and autofluorescence should be higher than three to conclude that the autofluorescence is not disturbing. This technique was applied to excited signals at 488 nm to separate them from their characteristic emission spectra. The emission spectrum of specific fluorescence reached a maximum at 520 nm, whereas the emission spectrum of autofluorescence peaked at approximately 550 nm (Figure 7B). The intensity of the specific CD68 fluorescence was approximately the same as the intensity of the autofluorescence (Figure 7B). Using this technique, two subtracted images were extracted from Figure 7C: the autofluorescence signal (Figure 7D) and the specific CD68 fluorescence of the macrophages (Figure 7E). In the resulting image, we can distinguish autofluorescence from the CD68 specific signal in cells (Figure 7E).

**Figure 7.** Identification of the in situ autofluorescence by microspectroscopy analysis. In the operated femur, macrophages were labeled with anti-CD68 antibody. Tissues were excited at 488 nm. (**A**) Emission of the autofluorescence and the antigen-specific fluorescence (Alexa Fluor 488) with turquoise fluorescence. (**B**) Microspectroscopycal analysis of Alexa Fluor 488 emission (turquoise line) and autofluorescence (red line). (**C**) Separation of autofluorescence emission (red fluorescence) and antigen-specific fluorescence (Alexa Fluor 488) (turquoise fluorescence). (**D**) Emission of the autofluorescence (red fluorescence). (**E**) Emission of the antigen-specific fluorescence (Alexa Fluor 488) (turquoise fluorescence). Nuclear staining with DAPI (blue fluorescence). Fine arrow: macrophages.

### *3.5. Detection of IL-1β and IL-6 Cytokines in the Operated Femur*

The presence of M1-like and mostly M2-like macrophages was shown earlier in the wounded bone tissue with immunofluorescence techniques. The next step was to determine the M2-like macrophages based on their cytokine expressions. Each subtype of M2-like macrophages (M2a, M2b, M2c, or M2d) secretes a different panel of cytokines. The next step was to identify the types of cytokines produced in the operated femur using RT-qPCR. All three markers showed higher expression in the operated femur (2.2 times more CD68, 1.61 times more CD163, and 2.29 times more CD206 mRNA) (Figure 8). Among the tested genes, two cytokines showed much higher expression, IL-1β (6.89 times more) and IL-6 (4.44 times more), in the operated femur compared to the non-operated femur (Figure 8). The only M2 macrophage that produces these two cytokines is the M2b macrophage (Figure 1). The in situ DIG technique is a commonly used, non-radioactive IHS method because it is very sensitive, but it has the limitation that it only allows analysis of a single gene in a sample.

**Figure 8.** Expression of the different marker genes and cytokines in the operated femur; qRT-PCR analysis of the different gene expressions. This result is the average of three measurements.

### *3.6. Identification of Macrophage Phenotype Using Both I DIG and In Situ HCR Techniques*

In situ HCR was chosen to further characterize the macrophages. The advantage of this technique is that it is more sensitive than FISH, and unlike in situ DIG, it allows the identification of several genes at once. Our next question was whether we could detect M2blike macrophages in situ using this technique in the operated rat bone. Since cytokines are usually secreted [48,67], immunostaining is not suitable to distinguish between different subtypes. Indeed, it is quite challenging to identify which cells produce the secreted protein. For this reason, an in situ hybridization technique was used to reveal the subtypes of M2-like macrophages as it allows the detection of the mRNA coding for those cytokines.

For the first time, in situ hybridization with a digoxigenin-labeled cRNA probe (in situ DIG) was used because the labeling remains stable, and it is an advantage to examine the labeling architecture. This technique is the most widely used non-radioactive ISH method because it is very sensitive, unfortunately, only one mRNA per section can be detected [68]. The expression of β-actin, IL-1β, IL-6, CD68, and CD163 genes was identified (Figure 9). The β-actin was used as a positive control (Figure 9A). In the absence of a specific probe, no expression was observed (Figure 9B). The mRNA expression of CD68, CD163, IL-1β, and IL-6 genes (Figure 9C–F) was detected around the biomaterial. Both CD68 (Figure 9C) and IL-1β mRNA expression (Figure 9D) was very intense. The CD163 (Figure 9E) and IL-6 (Figure 9F) mRNA expressions were much weaker, but a higher magnification showed a clear labeling (Figure 9G,H). In addition, β-actin, CD68, CD163, IL-1β, and IL-6 signals were quantified in the operated femur. Upregulation of these genes was detected in the Masquelet induced membrane region.

The same labeling was performed on the non-operated femur sections. In the nonoperated section, abundant expression of the β-actin mRNA was detected (Figure S4A), and in the negative control (Figure S4B), expression was not observed. The CD68 (Figure S4C) and IL-1β mRNA (Figure S4D) expressions were intense in the bone marrow, but no expression in the other part of the femur was detected. No expression was found with a small magnification of the CD163 (Figure S4E) and IL-6 (Figure S4F) mRNA. The CD163 expression was detectable only with higher magnification in the periosteum (Figure S4G). IL-6 mRNA was present in the bone marrow (Figure S4H).

**Figure 9.** In situ hybridization in the operated rat femur: (**A**) expression of β-actin mRNA (positive control); (**B**) negative control; (**C**) CD68 mRNA; (**D**) IL-1β mRNA; (**E**) CD163 mRNA; (**F**) IL-6 mRNA; (**G**) expanded view: high magnification image of the area within the red rectangle in image (**E**); (**H**) expanded view: high magnification image of the area within the red rectangle in image (**F**); (**I**) the stained area of the picture was quantitatively analyzed using ImageJ. The Masquelet induced membrane region was compared with the control region (**C**) (n = 3). \*\* *p* < 0.01, \* *p* < 0.5, ns = not significant, compared to control region.

Because co-labeling was not possible with the in situ DIG technique, in situ hybridization coupled with hybridization-chain-reaction detection (in situ HCR) was performed on cryosections of the bone. Indeed, three different mRNA expressions were co-detected, taking into account the rapid loss of signal of fluorophores within two weeks after labeling. Using the in situ DIG method, a high expression of IL-1β and IL-6 was detected in the regenerating bone. To examine whether M2-like macrophages were expressing these two cytokines, in situ HCR was performed with CD163, IL-1β, and IL-6 probes altogether (Figure 10A). Macrophages co-expressing the three signals were identified in the regenerated region of the operated femur, while no expression was detected in the negative control (Figure 10B). By random examination of the fluorescence of 100 cells labeled with a CD163 probe, the percentage of the M2b-like macrophages was approximately 68% of the M2-like macrophages.

**Figure 10.** Identification of the M2b-like macrophages in the operated femur by in situ hybridization combined with hybridization-chain-reaction detection (in situ HCR): (**A**) in situ hybridization combined with hybridization-chain-reaction detection (in situ HCR); (**B**) negative control: Probe-IL-6 (Alexa488, turquoise fluorescence), Probe-IL-1β (Alexa546, red fluorescence), Probe-CD163 (Alexa647, yellow fluorescence), nuclear staining with DAPI (blue fluorescence).
