Blood samples were obtained from 40 patients with glioma and 38 healthy controls (all Chinese) at The First Affiliated Hospital of Anhui Medical University (Hefei, China). Glioma tissues and adjacent normal tissues were obtained from 40 patients with glioma (all Chinese) upon excision surgery at the Department of Neurosurgery of The First Affiliated Hospital of Anhui Medical University. The patient characteristics are summarized in Table I. Patients with glioma (average age, 52.7 years; age range, 8–82 years) and healthy controls (average age, 39.7 years; age range, 23–62 years) were recruited and their blood and tumor samples were collected between May 2017 and August 2018. Specifically, adjacent normal tissues were excised from non-functional tissues within 2 cm from the tumor tissues. Blood samples and tumor tissues were collected from the same 40 patients. The staging of glioma was based on the 2016 World Health Organization Classification of Tumors of the Central Nervous System (14). The present study was approved by the Ethics Committee of Anhui Medical University. Informed consent was provided by all participants or their guardians. The histological types of the included patients were limited to glioma (including glioblastoma, mesoglioma, ganglioglioma, astrocytic glioma and spongiocytoma). Patients with incomplete information (age, sex, histological type, stage, recurrence or received chemotherapy) were excluded.

Peripheral venous blood (4 ml) was collected from each participant. Peripheral blood mononuclear cells were isolated using a PBMC isolation kit (cat. no. DKW-KLSH-0100; DAKEWE, Inc.) for use in subsequent flow cytometry analysis. For patients with glioma, blood samples were collected before the surgical procedure.

A flow cytometric gating strategy (CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b⁺) was used to stain macrophage-like cells in the peripheral blood of patients with glioma and healthy controls. CD115 was used to select monocyte-lineage cells, dendritic cells were excluded using CD1c, B cells were excluded using CD19, T and NK cells were excluded using CD2, and granulocytes were excluded using CD15. Macrophage-like cells were selected using the macrophage antibodies CD14, CD16 and CD11b (all Miltenyi Biotec, Inc.). Macrophage-like cells were divided into CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^int and CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^hi subsets for subsequent analyses; CD11b^int (P6, Vioblue 1×10²−1×10⁴) and CD11b^hi (P7, Vioblue 1×10⁴−1×10⁶) were decided based on the fluorescence-activated cell sorting (FACS) strategy shown in Fig. S1. In CD11b⁺ macrophage-like cells, CCR7⁺CD86⁺ was used to mark the M1 type, while CCR7⁻CXCR1⁺, CCR7⁻CD86⁺ and CCR7⁻CCR2⁺ were used to identify the M2a, M2b and M2c macrophage-like cells, respectively (15). The cells were analyzed using the BD FACSAria II flow cytometer and the BD FACSDiva software v8.0.2 (both BD Biosciences). Glioma tissues and adjacent normal tissues were incubated with the following fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, allophycocyanin (APC), PE-Cy7-, APC-Cy7- or Vioblue-conjugated monoclonal antibodies at 4°C for 1 h: Rat anti-human CD115-PE (2 µg/test; cat. no. 347303) was from BioLegend, Inc., whereas mouse anti-human CD1c-FITC (5 µg/test; cat. no. 130-113-863), CD2-FITC (5 µg/test; cat. no. 130-098-685), CD15-FITC (5 µg/test; cat. no. 130-114-010), CD19-FITC (5 µg/test; cat. no. 130-114-171), CD14-APC (2 µg/test; cat. no. 130-110-578), CD16-APC-Cy7 (2 µg/test; cat. no. 130-113-952), CD11b-Vioblue (2 µg/test; cat. no. 130-110-616), CD86-APC (2 µg/test; cat. no. 130-114-095), CCR2-PE (2 µg/test; cat. no. 130-109-654), CCR7-FITC (5 µg/test; cat. no. 130-117-700) and CXCR1-PE-Cy7 (2 µg/test; cat. no. 130-115-950) were from Miltenyi Biotec, Inc.

IHC and H&E staining were performed as previously described (16). Briefly, paraffin-embedded tissue slides obtained from glioma and adjacent tissues were deparaffinized in xylene twice for 5 min each at room temperature. Subsequently, slides were transferred in 100% alcohol twice for 3 min each, and then rehydrated in a descending alcohol series (95, 70 and 50%) for 3 min each. The EDTA antigen retrieval buffer (Beijing Solarbio Science & Technology Co., Ltd.) was heated until boiling and added to the sections; this was repeated two more times every 5–10 min. The slides were left to cool down at room temperature and washed with double-distilled H₂O. Subsequently, the slides were blocked in 5% goat serum (Biological Industries) for 60 min at room temperature, and subsequently incubated with primary monoclonal rabbit anti-human anti-inducible nitric oxide synthase (iNOS) (1:100; cat. no. sc-651) and polyclonal sheep anti-human anti-CD206 (1:50; cat. no. sc-34577; both from Santa Cruz Biotechnology, Inc.) antibodies at 4°C for 12 h. For H&E staining, sections were deparaffinized in xylene, re-hydrated in absolute alcohol and washed briefly with distilled water. Subsequently, they were stained with Harris hematoxylin solution and washed with running tap water, differentiated in 1% acid alcohol and washed with running tap water, blued in 0.2% saturated lithium carbonate solution and washed with running tap water, rinsed in 95% alcohol and counterstained with eosin-phloxine solution, dehydrated in alcohol and cleared in xylene, and finally mounted with mounting medium (xylene-based). ImageJ software (version d 1.47; National Institutes of Health) was used for quantification of IHC.

In areas of iNOS⁺ and CD206⁺ macrophage infiltration, individual macrophage infiltration was measured on a ×200 magnification field using an Olympus BX53 light microscope. An iNOS⁺ and CD206⁺ macrophage cluster was counted as a single infiltrated macrophage. The ratio of M1 and M2 infiltrated macrophages was calculated as the absolute number of iNOS⁺ and CD206⁺ macrophages/total cells per ×200 magnification field (n=8).

GraphPad Prism v7 (GraphPad Software, Inc.) was used for statistical analysis. Variances were first assessed by Bartlett's test. Subsequently, significant differences between two groups were analyzed using a two-tailed unpaired Student's t-test when equal variances were assumed, or a Welch's t-test for unequal variances. P<0.05 was considered to indicate a statistically significant difference.

A total of 40 patients with glioma and 38 healthy controls were recruited in the present study. The clinical and pathological features of the patients are listed in Table I. Macrophage-like cells in the peripheral blood were detected by flow cytometry with CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b⁺ gating (Fig. S1). There were two distinct populations of macrophage-like cells: CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^int and CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^hi cells. The number and percentage of CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^int cells among peripheral blood monocytes from patients with glioma were significantly decreased compared with those from healthy controls (Fig. 1). By contrast, an increased percentage of CD115⁺CD1c⁻CD2⁻CD15⁻CD19⁻CD14⁺CD16⁺CD11b^hi cells was observed in the peripheral blood of patients with glioma compared with those in the blood of healthy controls (Fig. 1).

In CD11b⁺ (including both CD11b^int and CD11b^hi) macrophage-like cells, the percentage of M1 type cells (CCR7⁺CD86⁺) was significantly lower in patients with glioma compared with that in healthy controls (Figs. 2B and 3C). However, there were almost no M2 type cells (CCR7⁻) in the CD11b^int population (Fig. 2A). In the CD11b^hi cell population, the percentage of total M2 type cells was higher in patients with glioma compared with that in healthy controls (Fig. 3A, B and D). Specifically, M2a (CCR7⁻CXCR1⁺) and M2b (CCR7⁻CD86⁺) type cells were upregulated, whereas M2c type cells (CCR7⁻CCR2⁺) were downregulated in the peripheral blood of patients with glioma compared with those in the blood of healthy controls (Fig. 3E-G).

To validate this polarization pattern, glioma and adjacent normal brain tissues were used for detecting the well-established M1 (iNOS) and M2 (CD206) markers (Fig. 4). The typical morphology of glioma and adjacent normal brain tissues is shown in Fig. 4A. Quantification of the IHC results revealed that the percentage of the M2 marker CD206⁺ cells was higher than the percentage of the M1 marker iNOS⁺ cells in glioma tumor tissues (Fig. 4B and C). Additionally, FACS analysis validated that macrophages expressed higher levels of the M2 marker CD206 than the M1 marker iNOS in glioma tissues (Fig. 4D and E). The present results suggested that patients with glioma exhibited a distinct pattern of macrophage polarization in the peripheral blood, namely a lower percentage of M1 type cells and a higher percentage of M2 type cells (especially M2a and M2b type cells) compared with healthy controls.

Subsequently, the patterns of macrophage polarization in patients with different stages of glioma were evaluated. The percentages of macrophage-like cells in the peripheral blood were not significantly different between patients with advanced glioma (stages III and IV) and those with early glioma (stages I and II) (Fig. S2). Since previous studies have suggested that serum YKL-40 is a candidate marker for glioma (17–19), the present study investigated the association between the patterns of macrophage polarization and serum YKL-40 status in patients with glioma. However, the percentages of macrophage-like cells were not significantly different between patients with low YKL-40 expression and those with high YKL-40 expression (Fig. S3).