Cytokine Profiling of Cyst Fluid and Tumor-Associated Macrophages in Cystic Vestibular Schwannoma

Abstract

Background:

The VS secretome can initiate monocyte recruitment and macrophage polarization to M1 (proinflammatory) and/or M2 (protumorigenic) phenotypes, which in turn secrete additional cytokines that contribute to the tumor microenvironment (TME). Profiling cyst fluid and cerebrospinal fluid (CSF) in cystic VS provides a unique opportunity to understand mechanisms that may contribute to tumor progression and cyst formation.

Hypothesis:

Cystic VS secrete high levels of cytokines into cyst fluid and express abundant M1 and M2 macrophages.

Methods:

Tumor, CSF, and cyst fluid were prospectively collected from ten cystic VS patients. Eighty cytokines were measured in fluid samples using cytokine arrays and compared to normal CSF from normal donors. Immunofluorescence was performed for CD80+ M1 and CD163+ M2 macrophage markers. Demographic, audiometric, and radiographic information were obtained through retrospective chart review.

Results:

Cyst fluid expressed more osteopontin and monocyte chemotactic protein 1 (MCP-1; p<0.0001), when compared to normal CSF. Cyst fluid also expressed more protein (p=0.0020), particularly MCP-1 (p<0.0001), than paired CSF from the same subjects. MCP-1 expression in cyst fluid correlated with CD80+ staining in VS tissue (r=0.8852; p=0.0015) but not CD163+ staining.

Conclusion:

Cyst fluid from cystic VS harbored high levels of osteopontin and MCP-1, which are cytokines important in monocyte recruitment and macrophage polarization. MCP-1 may have a significant role in molding the TME, by polarizing monocytes to CD80+ M1 macrophages in cystic VS. Further investigations into the role of cytokines and macrophages in VS may lead to new avenues for therapeutic intervention.

Introduction

Vestibular schwannomas (VS) are Schwann cell-derived tumors of the cochleovestibular nerve. Patients with VS can present with symptoms of hearing loss, tinnitus and dizziness.¹ Emerging evidence examining the tumor microenvironment (TME) in VS suggests an important role for inflammation in tumor progression.² It is believed that tumors secrete chemoattractant that promote monocyte extravasation, migration into tumor, and differentiation into mature macrophages.³ The cytokines in the tumor secretome can promote macrophage polarization toward a spectrum of macrophage phenotypes. The M1 phenotype exists on one end of the spectrum and is characterized by pro-inflammatory, anti-tumorigenic, and classically activated macrophages. On the opposite end of the spectrum lies the M2 phenotype, which is characterized by immune-suppressive, pro-tumorigenic, alternatively activated macrophages.⁴ The exact role that M1 and M2 macrophages in VS is still under investigation.^5–7 Several investigations have identified associations between cytokines, macrophages, and clinical factors (i.e., hearing loss and tumor progression) in patients with VS.^5,7–11

Cystic VS constitute ~10% of all VS and are notoriously known for their rapid growth and frequent adherence to the brainstem and facial nerve.^12–16 Cystic VS is of particular interest when studying VS tumor biology because analysis of the cyst fluid and adjacent cerebrospinal fluid (CSF) can provide important insight into the VS secretome, the cellular composition of the TME, and their effect on tumor progression and/or hearing loss. Because VS can secrete proteins into the local CSF, determining whether CSF analysis approximates cyst fluid contents will provide future opportunities for establishing CSF biomarkers for prognostic indications. In this study, we analyze the cytokine of CSF and cyst fluid from ten patients with cystic VS and describe macrophage phenotypes within tumor.

Materials and Methods

Tumor, Cyst Fluid and CSF Collection

Patients undergoing microsurgical resection of VS between April 2018 to February 2023 were recruited and enrolled in a prospective study to collect and analyze tumor contents, CSF, and cyst fluid. Tumor tissue, cyst fluid, and CSF from the cerebellopontine angle were collected at the time of surgery from 10 patients with cystic VS. Cyst fluid and CSF were banked at −80 degrees Celsius prior to experimentation. Tumor tissue was fixed in 4% paraformaldehyde overnight and embedded in paraffin prior to tissue section. Normal CSF from healthy donors was obtained for comparison (IPLA-CSFP, Innovative Research). This study was performed through the Institutional Review Board approved protocol (#20150637). Cysts were first identified on magnetic resonance imaging (MRI) by assessing T1-weighted, T1-weighted with contrast, and T2-weighted sequences on axial cuts. CSF was collected immediately after dural opening using a peripheral venous catheter attached to a 10-milliliter syringe. After the CSF was removed and the cerebellum was relaxed, tumor was dissected in an extracapsular plane to distinguish tumoral cysts from CSF trapped between arachnoid membranes. Stereotactic navigation was also utilized to confirm cyst location. Cyst fluid was collected by inserting an 18-gauge needle attached to a 10-milliliter syringe into the cyst and aspirating liquid contents. Of note, cyst fluid was also differentiated from trapped CSF by its thicker, more gelatinous consistency at room temperature.

Human Cytokine Antibody Array

Eighty cytokines and chemokines were assessed in cyst fluid and CSF using human cytokine antibody array kits (AAH-CYT-5-2, RayBiotech), following manufacturer’s protocol. Positive and negative controls were incorporated by the manufacturer into cytokine arrays. In brief, array membranes were blocked in blocking buffer at room temperature for 30 minutes, exposed to cyst fluid or CSF (1:4 dilution) overnight in 4 degrees Celsius. Subsequently, membranes were treated with primary biotin-antibody (1:2 dilution) for 2 hours and horse radish peroxidase-conjugated streptavidin secondary antibody for 2 hours at room temperature. The membrane was developed with enhanced chemiluminescent substrate and imaged with a chemiluminescence imaging system (Bio-Rad Chemidoc XRS or Bio-techne Jess Chemiluminescence Imaging Systems). Pixel intensities were quantified using ImageJ software (National Institute of Health). Pixel intensity measurements were normalized to negative control measurements within samples and then to positive control measurements between samples. Cytokine levels were compared to normal CSF from healthy donors. Cytokine acronyms are presented in Supplementary Table 1.

Immunofluorescence

VS tumor chunks were fixed in 4% paraformaldehyde overnight at 4 degrees Celsius, embedded in paraffin, and sectioned at 5 μm thickness onto glass slides. Samples were deparaffinized and rehydrated by immersing slides in 100% xylene (two washes, 5 minutes each), followed by 100% ethanol (two washes, 3 minutes each) and then 95%, 75%, and 50% ethanol for 1 minute each. Antigen retrieval was performed with ethylenediaminetetraacetic acid buffer and placed in a water bath at 100 degrees Celsius for 30 minutes. Subsequently, sections were blocked with 1% Triton X100, 5% normal donkey serum in phosphate buffered saline (PBS) for 30 minutes. Slides were then incubated with primary antibodies anti-human CD163 mouse monoclonal antibody (M2 macrophage marker) (1:10, MRQ-26, Cell Marker) or anti-human CD80 rabbit polyclonal antibody (M1 macrophage marker) (1:50, 8670, Prosci) in 0.2% Triton X100, 1% normal donkey serum in PBS at 4 degrees Celsius overnight. Slides were then exposed to Alexa594-conjugated anti-mouse and Alexa594-conjugated anti-rabbit IgG secondary antibodies (1:500) (ThermoFisher) for 1 hour at room temperature, respectively. Nuclei were stained with diamidino-2-phenylindole (DAPI; ab104139, Abcam) for 20 minutes at room temperature. Slides were then cover slips with mounting medium (Fluoromount), imaged using fluorescent microscopy (BioTek Lionheart FX Automated Microscope, United States), and analyzed using Gen5 software (BioTek, United States). Expression levels of CD80 and CD163 were calculated by averaging measurements of area fraction (%) for Antoni A and Antoni B areas (each over 60,000 μm² areas).

Clinical and Radiographic Presentation

A retrospective chart review of electronic medical records was performed for the ten VS patients for: (1) age, (2) gender, (3) hearing status (speech recognition threshold and word recognition score), and (4) tumor volume. The American Academy of Otolaryngology Head and Neck Surgery (AAO-HNS) Hearing Classification Scale was used to stratify patients into serviceable hearing (Grade A or B) and non-serviceable hearing loss (Grade C or D). Tumor volumes were measured with MIM software (MIM Software, Inc.) using preoperative axial T1-weighted sequences with contrast.

Statistical Analysis

Expression levels for 80 cytokines were analyzed with analysis of variance with Tukey’s post hoc test for: (1) CSF (relative to pooled normal CSF), (2) cyst fluid (relative to pooled normal CSF), and (3) cyst fluid (relative to CSF from cystic VS patients). Wilcoxon signed rank was used to compare mean cytokine expression levels for all 80 cytokines between cyst fluid and paired CSF from each subject. Pearson’s coefficient was utilized to determine correlations between CD80, CD163, MCP-1, osteopontin, and tumor volume. In addition, linear regression was performed for CD80, CD163, and MCP-1. Bonferroni correction was applied for multiple comparisons, and thus, significance was set at p-value less than 0.0031.

Results

Clinical and Radiographic Characteristics of Patients with Cystic VS

The study included ten patients who underwent microsurgical resection of cystic VS via the retrosigmoid approach (Table 1). The average age at surgery was 60.8 years (range, 52–78 years). The average tumor volume was 17.5 cm³ (range, 5.5–41.3 cm³). Of 9 patients with pre-operative audiograms, 2 had serviceable hearing (AAO-HNS Grade A or B) and 7 had non-serviceable hearing (AAO-HNS Grade C or D). Representative axial T1-weighted with contrast and T2-weighted MRI images are shown in Figure 1. Subject A34 with serviceable hearing had the smallest cystic VS by tumor volume. Subjects A89 and A94 had progressive tumor growth following GammaKnife radiosurgery.

Table 1. Demographic, Audiometric, and Radiographic Information for Cystic Vestibular Schwannoma (VS) Patients.

Patient gender, age at the time of surgery (years), tumor volume (cm³), speech recognition threshold (dB HL), word recognition score (%), and the American Academy of Otolaryngology Head and Neck Surgery (AAO-HNS) hearing class are provided for ten patients with cystic VS. AAO-HNS Hearing Classification Scale stratifies patients into serviceable hearing (Grade A/B) and non-serviceable hearing loss (Grade C/D). F, female. M, male. dB HL, decibels hearing level. NR, no response. NA, not available.

VS Tumor	Sex	Age at Surgery (Years)	Tumor Volume (cm3)	Speech Recognition Threshold (dB HL)	Word Recognition Score (%)	AAO-HNSF Hearing Class
A17	F	58	41.3	NR	0	D
A18	M	65	31.4	NR	0	D
A23	M	52	15.2	110	0	D
A34	M	68	5.5	20	100	A
A76	M	50	23.4	105	0	D
A89	F	63	7.9	NR	0	D
A90	M	55	24.8	NA	NA	NA
A94	M	78	5.9	85	0	D
B01	M	59	8.3	40	80	B
B03	M	60	11.2	90	0	D
Average		60.8	17.5

Figure 1. Magnetic Resonance Imaging (MRI) of Cystic Vestibular Schwannoma (VS).

Representative axial T1-weighted images with contrast and T2-weighted MRI images of cystic VS from ten patients are shown. MRI images demonstrate enhancing masses occupying the internal auditory canal and cerebellopontine angle with cystic components with fluid density.

Cytokine Composition of Cyst Fluid and CSF

Expression levels for 80 cytokines were measured in cyst fluid and CSF of ten patients with cystic VS and compared to normal CSF from healthy donors. CSF expressed higher levels of osteopontin (p<0.0001) but lower levels of angiogenin (p<0.0001) and monocyte chemotactic protein-1 (MCP-1) (p<0.0001), when compared to normal CSF (Figure 2). Cyst fluid expressed higher expression levels of MCP-1 (p<0.0001) and osteopontin (p<0.0001), when compared to normal CSF (Figure 3). Cyst fluid also expressed significantly more angiogenin (p<0.0001), MCP-1 (p<0.0001) (Figure 4) and protein overall (p=0.0020; Figure 5) when compared to paired CSF from the same subjects.

Figure 2. Cytokine Profile of Cerebrospinal Fluid (CSF) from Ten Patients with Cystic Vestibular Schwannoma (VS).

[A-B] CSF expressed significantly less angiogenin (p<0.0001) and monocyte chemoattractant protein-1 (MCP-1) (p<0.0001) and more osteopontin (p<0.0001) than normal CSF from healthy donors. Bar represents mean. Error bars represent standard error mean. *** p<0.0001.

Figure 3. Cytokine Profile of Cyst Fluid from Ten Patients with Cystic Vestibular Schwannoma (VS).

[A-B] Cyst fluid expressed significantly more monocyte chemoattractant protein-1 (MCP-1) (p<0.0001) and osteopontin (p<0.0001) than normal CSF from healthy donors. Bar represents mean. Error bars represent standard error mean. ***p<0.0001.

Figure 4. Difference in Cytokine Expression Between Cyst Fluid and Cerebrospinal Fluid (CSF) in Ten Patients with Cystic Vestibular Schwannoma (VS).

[A-B] Cyst fluid expressed significantly more angiogenin (p<0.0001) and monocyte chemoattractant protein-1 (MCP-1) (p<0.0001) than paired CSF from the same subjects. Box plot represents 25^th-50^th-75^th percentile. Error bars represent minimum and maximum. [C] The mean protein expression level across 80 cytokines was significantly higher in cyst fluid than CSF (p=0.0020). Bar represents mean. Error bars represent standard error mean. ***p<0.0001. **p<0.01.

Figure 5. Immunohistochemistry for CD80 and CD163.

[A-B] Representative fluorescent images (20X lens) demonstrate presence of CD80+ M1 and CD163+ M2 macrophage markers (orange-red color) in varying degrees in 9 cystic VS. Subject A17 did not have sufficient tumor tissue for analysis. DAPI (blue color) represents nuclear staining. [C-D] Linear regression plots show a strong, positive correlation between CD80 expression (measured as percent area) and MCP-1 Expression (pixel intensity) in cyst fluid (p=0.0015). Although moderately strong positive correlations were found between CD163 and MCP-1, the correlation was not significant.

Immunofluorescence

To detect M1 and M2 macrophages in VS tissue specimens, immunofluorescence for CD80 and CD163 macrophage markers was performed, respectively. There was inadequate tumor specimen for A17 for analysis. Representative sections from the other 9 VS demonstrated cytoplasmic staining for CD80 and CD163 at varying degrees (Figure 5A & 5B). Expression levels of CD80 and CD163 were measured as area fraction (%), and Pearson’s correlation was performed between CD80, CD163, tumor volume, and MCP-1 and osteopontin from cyst fluid. A strong, positive relationship (r=0.8852) was found between CD80 and MCP-1 (p=0.0015). Furthermore, 78.36% of the variation in MCP-1 is explained by CD80 (R²=0.7836; p=0.0015). No other significant correlations were identified (p>0.0035).

Discussion

Cytokines can take on several functions in the TME, some of which include (1) induction of angiogenesis, (2) initiation of a pro-inflammatory or anti-inflammatory response, (3) recruitment of monocytes and polarization of macrophages, and (4) promotion of tumor growth.¹⁷ Analyzing the cytokine content of cyst fluid from patients with cystic VS provides unique opportunities to study the VS secretome, the TME, and molecular and cellular mechanisms that may lead to tumor progression and hearing loss. In this study, we focus on elucidating the cytokine profiles of CSF and cyst fluid and their association with tumor-associated macrophages (TAM) to obtain important insight into the TME that may lead to potential therapeutic interventions in the future.

We recruited 10 patients undergoing surgery for cystic VS and analyzed expression of 80 cytokines in both the cyst fluid and CSF. We found that cyst fluid contained more protein overall than CSF (Figure 5C), which may reflect the circulating nature of CSF through the ventricles and constant production and recycling of CSF that can dilute protein levels.¹⁸ Although several proteins were up regulated compared to normal CSF (Figure 3 & 4), osteopontin, a bone sialoprotein linked to tumor progression in solid tumors, was significantly upregulated in both cyst fluid and CSF, warranting further investigation into osteopontin as a novel therapeutic target and CSF biomarker in future studies.¹⁹ MCP-1, also known as chemokine (C-C motif) ligand 2 [CCL-2], was also expressed highly in cyst fluid, especially when compared to its paired CSF, suggesting MCP-1 may have an important role in molding the TME by promoting monocyte recruitment.^2,20–26 Angiogenin, a proangiogenic factor that promotes tumor growth in various cancers, was found in higher levels in cyst fluid than in CSF of cystic VS patients, implying that angiogenin may have a potential role in cystic VS biology and additional investigations are needed.²⁷ Furthermore, we investigated the relationships between tumor volume (including cysts), TAMs, and MCP-1 and osteopontin expression in cyst fluid. This was challenging as it was not feasible or practical to measure the volume of solid tumor without including the multiple cysts that envelope and infiltrate the VS. For this reason, we do not expect to find, nor did we find, a relationship with tumor volume; however, we identified a correlation between MCP-1 in cyst fluid and CD80+ TAMs, further suggesting that MCP-1 may have an important role in macrophage polarization to M1 pro-inflammatory macrophages and cyst formation in VS.

Although several studies have analyzed the expression of proteins in VS, few have analyzed the VS secretome.^2,6,28–30 Dilwali et al. investigated the effect of secreted factors from primary VS cultures by culturing murine cochlear explants in tumor-conditioned media.^11,31 They demonstrated that tumor-conditioned media from VS patients with HL had higher tumor necrosis factor alpha (TNF) and lower fibroblast growth factor (FGF)-2. Furthermore, they also found that secretions from primary VS cells obtained from patients with the most substantial HL caused more auditory hair cell loss and neuronal fiber disorganization in cochlear explants.¹¹ Although most of the cystic VS patients in our investigation had unserviceable hearing loss, we did not find a higher expression of TNF in those patients. There was an increase in the mean expression of TNF in CSF and cyst fluid across the 10 cystic VS, when compared to normal CSF; however, the finding did not reach significance. Our array also did not include FGF2, but both cyst fluid and CSF expressed low levels of FGF-4, FGF-6, FGF-7, and FGF-9 in those with serviceable and unserviceable hearing loss.

The molecular and cellular mechanisms related to formation of cysts in VS are largely unknown. A histologic investigation on 10 solid and 10 cystic VS suggest that microhemorrhages in VS can lead to formation of cysts, as demonstrated by higher levels of hemosiderin deposits, thrombotic vessels, and abnormal vessel proliferation.³² This may reflect a high level of angiogenic factors in the TME. Although we did not find a significant up regulation of secreted VEGF into cyst fluid or CSF, we did show that angiogenin was expressed at higher levels compared to CSF, and thus, the role of angiogenin in cystic VS requires further investigation.

In a retrospective investigation involving 11 cystic and 6 sporadic VS, gene profiling and real time polymerase chain reaction analysis identified down regulation of chromosome 1 open reading frame 130 (C1ORF130), ciliary neurotrophic factor gene (CNTF), and 2 genes encoding alpha chains that form the heterotrimeric type IV collagen molecules (COL4A3 and COL4A4).³³ C1ORF130 encodes myelin glycoprotein, and a reduction in protein expression could potentially lead to cyst formation.^33,34 CNTF is a survival molecule for neurons and oligodendrocytes, and thus, a reduction in CNTF may lead to cell death and cyst formation.^33,35,36 Lastly, COL4A3 and COL4A4 encode important structural components of the basement membrane and thus down regulation may lead to structural instability and cyst formation.^33,37 Although these factors were not evaluated in our cytokine analysis, CNTF is a known chemoattract for blood-derived macrophages.³⁸ Furthermore, osteopontin is an important cytokine that regulates production of type 4 collagen in human retinal capillary cells.³⁹

Another retrospective immunohistochemistry investigation looking at matrix metalloproteinases (MMP) and vascular endothelial growth factor (VEGF) in 24 cystic VS and 38 solid VS demonstrated an upregulation of MMP-14 and VEGF in Antoni B areas of cystic VS.⁴⁰ These findings suggest that MMP-14 may degrade loose collagen in cystic VS, while VEGF may induce extravasation of plasma ultrafiltrate, cystic expansion, and intratumoral hemorrhage.⁴⁰ In a prospective investigation of 9 cystic VS, MMP-2 expression was found ubiquitously in all cyst fluid with gelatin zymography. Furthermore, MMP-2 localized to the tumors cells along the luminal surface of the cyst wall, suggesting additional roles for MMPs in the development of cysts in VS.⁴¹ Our immunofluorescence and cytokine analysis of cyst fluid did not incorporate MMPs; however, osteopontin was upregulated in cyst fluid and CSF and is known to be a substrate of several MMPs, including MMP-2.^42–45 Furthermore, MCP-1 is known to stimulate the expression of multiple MMPs in various cell types, including cancer.^46–49

Macrophages play an important role in molding the TME and a significant role in secreting cytokines. They are a group of monocyte-derived immune cells that are phagocytic in nature and can possess specific phenotypic characteristics.⁵⁰ Macrophages can be polarized by various microenvironmental stimuli to promote tissue homeostasis, defense mechanisms, and wound healing. They have also been found to play a role in inflammation, autoimmune disorders, atherosclerosis and tumorigenesis.⁵¹ In general, activated macrophages are classified as pro-inflammatory M1 or anti-inflammatory M2 macrophages.^52–54 M1 and M2 TAMs can be differentiated by the expression of various membrane proteins.⁵⁵ In this study, CD80 positivity is used to identify M1 TAMs, while CD163 positivity is used to denote M2 TAMs. In this study, both types of macrophages were found to be present in all analyzed tumor specimens.

Several investigators have demonstrated relationships between inflammatory cytokines, TAMs, and tumor growth in VS;^{2,6,9,12,28,56} however, the exact roles that M1 and M2 TAMs play in VS is still under debate. In a study performed by De Vries et al., the expression of pro-tumorigenic M2 TAMs was linked to tumor growth in sporadic VS.⁵ In contrast, Perry and colleagues found strong associations between M1 TAMs and progressive VS growth after subtotal resection, theorizing that programmed death ligand-1 (PD-L1) expression can deactivate anti-tumorigenic M1 TAMs and help VS evade immune response.^7,57 Although we did not find a relationship between the TAMs and secreted cytokines with tumor volume, we did discover a correlation between MCP-1 in cyst fluid and CD80+ TAMs, further suggesting that MCP-1 may have an important role in macrophage polarization to M1 pro-inflammatory macrophages and cyst formation in VS.

Osteopontin upregulates MCP-1 in different cell types, and consistent with our findings, osteopontin and MCP-1 have been linked to monocyte polarization to the classically activated M1 macrophage phenotype.^58–62 Furthermore, osteopontin and MCP-1 have been linked to macrophage-dependent cyst expansion in autosomal dominant polycystic kidney disease, suggesting a mechanistic role for osteopontin, MCP-1, and M1 macrophages in cyst formation in cystic VS.^63–65 In addition, merlin deficiency from NF2 mutations can upregulate pro-inflammatory cytokines and chemokines in schwannoma cell lines through disinhibition of transcription factor, nuclear factor kappa B.⁶⁶ Among these pro-inflammatory cytokines is MCP-1, providing further evidence of a link between schwannoma, MCP-1, and TAMs.

Because of the limited sample size in this pilot study, future investigations with larger sample sizes will be needed to further clarify and understand the relationship between cytokines, macrophage polarization, and clinical parameters such as tumor size, tumor growth rate and hearing loss. Although no equivalent comparison can be made between the cyst fluid of cystic VS and sporadic VS, analysis of CSF between cystic VS and sporadic VS may provide additional information about the VS secretome that contributes to cyst formation. With a better understanding of the VS TME, we can identify important targets for therapeutic intervention to prevent tumor growth and preserve hearing in patients with VS, particularly cystic VS.

Conclusion

Cyst fluid from patients with cystic VS was rich in osteopontin and MCP-1, which are important cytokines that promote monocyte recruitment and macrophage polarization. MCP-1 may have a significant role in molding the TME, by polarizing monocytes to CD80+ M1 TAMs in cystic VS. Furthermore, osteopontin and MCP-1 may be implicated in cyst formation in cystic VS. Further investigations into osteopontin, MCP-1, and TAMs may lead to new avenues for therapeutic intervention in patients with cystic VS.