Female C57BL/6J mice were obtained from the Animal Research Centre (ARC, Murdoch, WA, Australia) and maintained at Curtin University animal facilities under specific-pathogen free conditions. Young mice were aged 3–4 months (equivalent to 18 year-old humans) and elderly mice were 20–24 months (60–70 year-old humans), as defined by the Jackson Laboratory (Yuan et al., 2009). Mice were excluded from the study if they had enlarged organs, a palpable mass or experienced excessive body weight loss (>15% from age 12 months) prior to tumor inoculation. Animals were housed in a standard light/dark cycle and fed a standard chow diet ab libitum. Experiments were performed as per the Curtin University Animal Ethics Committee (AEC) in accordance to the Australian Code of Practice for the use and care of animals for scientific purposes (AEC approval numbers: AEC_2016_04 and AEC_2020_03).

AE17 is a murine malignant mesothelioma cell line generated by inoculation of asbestos fibers and was derived from orthotopic tumor deposits that emerged in elderly C57BL/6J mice and is histologically representative of human mesothelioma (Jackaman et al., 2003). Mesothelioma cell lines, including AE17, have also been reported to express CSF-1 (Fox et al., 2012; Cioce et al., 2014). Cells were maintained in complete medium, containing RPMI 1640 (Invitrogen) supplemented with 10% HyClone™ fetal bovine serum (FBS, Cytiva, Utah, America), 2 mM L-glutamax (Life Technologies, Victoria, Australia), 100 units/ml of penicillin, 100 μg/ml streptomycin (Life Technologies) and 0.05 mM 2-mercaptoethanol (Sigma-Aldrich) at 37°C with 5% CO₂. Cells were collected for tumor inoculation when ≥80% confluent. Mice were inoculated subcutaneously (s.c.) with 5 × 10⁵ cells in 100 μl PBS and body weight, body condition score and tumors monitored daily. Tumor sizes were taken daily using calipers and determined by measurement of tumor width (mm) and length (mm), and calculated as width × length in mm². Each mouse was tracked individually for tumor growth. Mice were monitored until their individual maximum tumor size reached 140 mm², weight loss exceeded >20% or reached experimental endpoint. Based on our previous studies, age impacted tumor growth in that tumors grew faster in elderly compared to young mice (Duong et al., 2018). In order to control for variation between young and elderly tumor growth, samples were collected based on tumor size. Tumors were collected at either early (7–10 days post s.c. injection) or late timepoints (18–23 days post s.c. injection); small tumors measured <30 mm² and large tumors were between 65 and 140 mm² (Jackaman et al., 2003; Jackaman et al., 2011; Jackaman et al., 2016).

To ensure temperature did not impact CSF-1R expression, all reagents were pre-cooled on ice and samples kept on ice throughout staining. Sample collection was also timed so that young and elderly samples spent the same amount of time in processing ex vivo once tissues were out of the animal. Similar numbers of young and elderly samples were also included in each flow cytometry experiment to control for any inter-assay staining variation. Bone marrow, spleen and tumors were collected into collection/staining buffers: ice-cold phosphate buffered solution (PBS) containing 2% fetal bovine serum and 2 mM EDTA (Sigma-Aldrich). Tibiae and femurs were flushed with a 29 g needle to isolate and create single cell suspension of bone marrow cells. Spleens and tumors were disaggregated gently into single cell suspension between two frosted slides in staining buffer. For flow cytometry surface staining, single cell suspensions were blocked with anti-mouse CD16/32 (clone 93, Biolegend) for 15 min on ice, all subsequent steps were performed on ice in the dark unless otherwise stated. A combination of anti-mouse antibodies diluted in staining buffer were incubated for 30 min followed by PBS and incubated for 15 min with Zombie-NIR™ (Biolegend). Intracellular staining was performed after fixing and permeabilization with True-Nuclear™ Transcription Factor Buffer Set (Biolegend) as per manufacturer’s instructions. Surface and intracellular antibodies were diluted in 1x Transcription Factor permeabilization buffer for 1 h in the dark at room temperature. A combination of the following anti-mouse antibodies were used: anti-CD11b BUV395 (BD) or anti-CD11b Alexafluor^® 488 (clone M1/70, Biolegend), Ly6C BV510 (clone HK1.4, Biolegend), anti-Ly6G Brilliant Violet 785™ (clone 1A8, Biolegend), anti-F4/80 PE-Dazzle-594 (clone BM8, Biolegend) or anti-F4/80 Alexafluor^® 647 (clone BM8, Biolegend), anti-CD115 (CSF-1R) Brilliant Violet 421™ (clone AFS98, Biolegend) and anti-CD206-PE-Cy7 (clone C068C2, Biolegend). After staining, cells were washed twice and resuspending in staining buffer for acquisition on BD LSRFortessa™ flow cytometer using FACSDiva (BD Biosciences). Unstained, single stains and fluorescent minus one controls were used for instrument setup, compensation and analysis. Data was analyzed using FlowJo version 10.7 (BD Bioscience).

Whole blood was collected directly into EDTA (Sigma-Aldrich) and plasma stored at −20°C until analysis. Samples were collected from young healthy mice, elderly healthy mice, large tumor-bearing young mice and large tumor-bearing elderly mice. The concentration of CSF-1 in plasma samples was measured using a LEGENDplex™ Mouse Hematopoietic Stem Cell Panel (Biolegend) as per manufacturer’s instructions. Data was collected on a LSRFortessa™ flow cytometer using FACSDiva (BD Biosciences) and analyzed using FlowJo version 10.7 (BD Bioscience).

GraphPad Prism version 9 (California, CA, United States) was used to analyze data. Data presented as means ± SEM. Mann-Whitney U-test was used to determine differences between two populations. A relationship between two variables was determined by Pearson’s correlation coefficient test. p-values of <0.05 were considered statistically significant.

Ex vivo flow cytometry analysis of bone marrow from young and elderly healthy mice (gating strategy shown in Figure 1A) revealed an age-specific expansion of CD11b⁺Ly6G^negF4/80^neg monocytes (Figure 1B), a Ly6C^high subpopulation (Figure 1C) and CD11b⁺F4/80⁺Ly6G^neg macrophages (Figure 1D). We next investigated the spleen as it can serve as a site for extramedullary hematopoiesis and contains a reservoir of monocytes that may be deployed to the tumor (Cortez-Retamozo et al., 2012; Cortez-Retamozo et al., 2013). Similar to bone marrow, healthy aging leads to the expansion of total monocytes (Figure 1E), the Ly6C^high subpopulation (Figure 1F) and CD11b⁺F4/80⁺Ly6G^neg red pulp macrophages (Figure 1G) in the spleen.

As CSF-1R signaling is crucial in monocyte to macrophage differentiation, expansion, and migration (reviewed in Pixley and Stanley, 2004), we next measured expression of this molecule to determine whether altered expression levels may account for the increased proportion of these cells in older tissues. While CSF-1R surface and intracellular expression was unchanged between young and elderly healthy macrophage and monocyte populations in the bone marrow (Figure 2A and Supplementary Figure S1) and splenic monocytes (Supplementary Figure S2), a significant age-specific decrease in surface and intracellular CSF-1R was observed in splenic macrophages (Figures 2B,C, respectively). These data show that CSF-1R expression is reduced with healthy aging in splenic macrophages, highlighting age- and tissue-related differential regulation of CSF-1R expression.

Previously we showed that solid tumors impact the bone marrow macrophage compartment, leading to increased bone marrow macrophage proliferation in elderly-tumor bearing mice compared to young mice (Duong et al., 2018). In this study, we expanded this work to include monocytes with a particular focus on the Ly6C^high subset, which infiltrates tumors and differentiates into TAMs (Movahedi et al., 2010; Van Overmeire et al., 2015). As various stages of tumor progression may impact these cells differently, we investigated the bone marrow at small (early stage) and large (late stage) AE17 mesothelioma tumors. Total bone marrow monocytes increased due to aging rather than tumor growth (Figure 3A). Similarly, the proportion of Ly6C^high monocytes increased with healthy aging in the bone marrow and was not further affected by tumor growth (Figure 3B). Macrophage proportions remained similar between young and elderly mice bearing tumors (Figure 3C). However, when compared to healthy elderly controls, macrophage proportions were reduced in the bone marrow of elderly mice with cancer (Figure 3C), despite displaying increased proliferation (Duong et al., 2018). It is possible this is due to reduced survival or migration of the cells from the bone marrow to support tumor growth. Next, the splenic monocyte and macrophage compartments were examined during AE17 tumor development. The profile of total splenic monocytes was similar to that seen in the bone marrow with an age-related increase in numbers in small and large AE17 tumors (Figure 3D). However, splenic Ly6C^high monocytes increased markedly in elderly mice during tumor progression while their proportions remained steady in young spleens (Figure 3E). This resulted in an almost 3-fold increase in Ly6C^high proportions in the elderly compared to young mice with large tumors. Similarly, splenic macrophage proportions were significantly higher in elderly vs. young AE17-bearing hosts and was further augmented with tumour growth (Figure 3F).

We also examined levels of CSF-1R expression in monocytes and macrophages from the bone marrow and splenic compartments in tumor-bearing mice. Interestingly, intracellular CSF-1R expression in Ly6C^high bone marrow monocytes increased in the elderly compared to young with large tumors (Figure 4A), which may reflect increased internalization of activated CSF-1R (Yeung and Stanley, 2003). Furthermore, intracellular CSF-1R expression increased in macrophages of elderly mice during early tumor growth but matched the levels of younger counterparts when tumors were larger (Figure 4B). Surface expression of CSF-1R in bone marrow macrophages was slightly but not significantly reduced in elderly vs. young mice with large tumors (Figure 4C). In the spleen, intracellular CSF-1R expression in total monocytes remained similar through all cohorts and tumor growth (Supplementary Figure S3). However, splenic macrophages in young tumor-bearing mice demonstrated increased CSF-1R expression that was restricted to the intracellular compartment (Figure 4D) and not seen on the cell surface (Figure 4E) when compared to the healthy setting. Intracellular CSF-1R expression in these cells also increased with tumor growth (Figure 4D). These data suggest CSF-1R internalization in young mice during tumor growth. However, there were no changes to CSF-1R expression during tumor growth in elderly mice (and compared to healthy) suggesting splenic macrophages from tumor-bearing mice could be refractory to CSF-1R signaling (Figure 4D,E). These data suggest that CSF-1R expression may be impacted with aging and tumor progression but varies depending on cell type. Previous studies have reported increased circulating CSF-1 during aging (Suehiro et al., 1999; Larsson et al., 2015; Lira-Junior et al., 2017) which could impact the spleen as a blood filter. Furthermore, the AE17 cell line used in this study expresses CSF-1 (Fox et al., 2012). However, we observed no difference in CSF-1 circulating protein levels during healthy aging (Figure 4F) and circulating CSF-1 levels were decreased in the plasma of elderly large tumor-bearing mice compared to elderly healthy controls (Figure 4F). Combined with the overall changes to CSF-1R expression it is possible there is increased utilization of CSF-1 in elderly tumor-bearing mice.

The data presented thus far indicate an increased supply of monocytes and macrophages during aging, which may be further expanded with tumor growth. Ly6C^high monocytes may be particularly important for the supply of TAMs to tumors. Increased availability of these cells could result in faster tumor growth, as reported in our earlier study (Duong et al., 2018). Furthermore, alterations to CSF-1R expression may impact the supply potential of these cells in terms of numbers and migration. This prompted us to investigate infiltrating monocyte/macrophages and their maturation in the tumor via differential expression of Ly6C and F4/80 (Ramachandran et al., 2012; Crane et al., 2014): with F4/80^intLy6C^high (Ly6C^high) TAMs, representing early monocyte-derived TAMs. The loss of Ly6C expression is associated with maturation and therefore we identified F4/80^intLy6C^low (Ly6C^low) TAMs as an intermediate subset, and F4/80^highLy6C^low/high (F4/80^high) TAMs as a mature population (gating strategy shown in Figure 5A). We found that Ly6C^high TAMs increased significantly in the elderly early during tumor development (Figure 5B), whereas Ly6C^low proportions increased with aging, which was further exacerbated with tumor growth in the elderly (Figure 5C). However, F4/80^high TAMs expanded with tumor growth in both young and elderly and were the largest pool of TAM subsets in the tumor, particularly at later stages of tumor growth (Figure 5D). There was no difference in expression of the pro-tumoral marker CD206 in Ly6C^high or Ly6C^low TAMs during young and elderly tumor growth (Figures 5E,F respectively). In contrast F4/80^high TAMs from elderly tumor-bearing mice displayed increased CD206 expression with increasing tumor growth and compared to young mice (Figure 5G). Together, this may suggest a faster transition of Ly6C^high monocyte-derived TAMs towards Ly6C^low and pro-tumoral F4/80^high TAMs with aging. CSF-1R expression was similar between TAMs in young and elderly, however F4/80^high cells displayed the lowest expression of CSF-1R and intracellular CSF-1R expression was highest in Ly6C^low TAMs (Supplementary Figure S4A,B). This suggests CSF-1R signaling may vary across TAM subsets.