Systemic and local immunosuppression in glioblastoma and its prognostic significance

The effectiveness of tumor therapy, especially immunotherapy and oncolytic virotherapy, critically depends on the activity of the host immune cells. However, various local and systemic mechanisms of immunosuppression operate in cancer patients. Tumor-associated immunosuppression involves deregulation of many components of immunity, including a decrease in the number of T lymphocytes (lymphopenia), an increase in the levels or ratios of circulating and tumor-infiltrating immunosuppressive subsets [e.g., macrophages, microglia, myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Tregs)], as well as defective functions of subsets of antigen-presenting, helper and effector immune cell due to altered expression of various soluble and membrane proteins (receptors, costimulatory molecules, and cytokines). In this review, we specifically focus on data from patients with glioblastoma/glioma before standard chemoradiotherapy. We discuss glioblastoma-related immunosuppression at baseline and the prognostic significance of different subsets of circulating and tumor-infiltrating immune cells (lymphocytes, CD4+ and CD8+ T cells, Tregs, natural killer (NK) cells, neutrophils, macrophages, MDSCs, and dendritic cells), including neutrophil-to-lymphocyte ratio (NLR), focus on the immune landscape and prognostic significance of isocitrate dehydrogenase (IDH)-mutant gliomas, proneural, classical and mesenchymal molecular subtypes, and highlight the features of immune surveillance in the brain. All attempts to identify a reliable prognostic immune marker in glioblastoma tissue have led to contradictory results, which can be explained, among other things, by the unprecedented level of spatial heterogeneity of the immune infiltrate and the significant phenotypic diversity and (dys)functional states of immune subpopulations. High NLR is one of the most repeatedly confirmed independent prognostic factors for shorter overall survival in patients with glioblastoma and carcinoma, and its combination with other markers of the immune response or systemic inflammation significantly improves the accuracy of prediction; however, more prospective studies are needed to confirm the prognostic/predictive power of NLR. We call for the inclusion of dynamic assessment of NLR and other blood inflammatory markers (e.g., absolute/total lymphocyte count, platelet-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, systemic immune-inflammation index, and systemic immune response index) in all neuro-oncology studies for rigorous evaluation and comparison of their individual and combinatorial prognostic/predictive significance and relative superiority.

The effectiveness of tumor therapy, especially immunotherapy and oncolytic virotherapy, critically depends on the activity of the host immune cells.However, various local and systemic mechanisms of immunosuppression operate in cancer patients.Tumor-associated immunosuppression involves deregulation of many components of immunity, including a decrease in the number of T lymphocytes (lymphopenia), an increase in the levels or ratios of circulating and tumorinfiltrating immunosuppressive subsets [e.g., macrophages, microglia, myeloidderived suppressor cells (MDSCs), and regulatory T cells (Tregs)], as well as defective functions of subsets of antigen-presenting, helper and effector immune cell due to altered expression of various soluble and membrane proteins (receptors, costimulatory molecules, and cytokines).In this review, we specifically focus on data from patients with glioblastoma/glioma before standard chemoradiotherapy.We discuss glioblastoma-related immunosuppression at baseline and the prognostic significance of different subsets of circulating and tumor-infiltrating immune cells (lymphocytes, CD4+ and CD8+ T cells, Tregs, natural killer (NK) cells, neutrophils, macrophages, MDSCs, and dendritic cells), including neutrophil-to-lymphocyte ratio (NLR), focus on the immune landscape and prognostic significance of isocitrate dehydrogenase (IDH)-mutant gliomas, proneural, classical and mesenchymal molecular subtypes, and highlight the features of immune surveillance in the brain.All attempts to identify a reliable prognostic immune marker in glioblastoma tissue have led to contradictory results, which can be explained, among other things, by the unprecedented level of spatial heterogeneity of the immune infiltrate and the significant phenotypic diversity and (dys)functional states of immune subpopulations.High NLR is one of the most repeatedly confirmed independent prognostic factors for shorter overall survival in patients with glioblastoma and carcinoma, and its combination with other markers of the immune response or systemic inflammation significantly improves the accuracy of prediction; however, more prospective studies are needed to confirm the prognostic/predictive power of NLR.We call for the

Introduction
In the United States, glioblastoma (diagnosed during 2013-2017) accounts for 48.6% (all ages combined) of all malignant brain and other central nervous system tumors, and five-year survival for patients diagnosed with glioblastoma in 2009-2015 is 7% (3-27%, varied by age) (1).Although the incidence of glioblastoma among primary brain tumors in different countries varies greatly (from ≈8.5% to 69%) (2,3), five-year relative survival estimates are comparable and are <7% overall (4).The standard of care for the treatment of glioblastoma (maximal safe resection followed by radiotherapy with temozolomide chemotherapy) has not changed much since 2005 (Stupp protocol) (5,6), and glioblastoma remains incurable (1) despite significant advances in our knowledge of its genetics and molecular biology over the past two decades.
The blood brain barrier is thought to be a key factor limiting the effectiveness of chemotherapy, including targeted agents, in the treatment of glioblastoma/glioma (7)(8)(9).Even temozolomide, with features such as 100% oral bioavailability, rapid absorption, excellent biodistribution, and ability to cross the blood-brain barrier because of its small size and lipophilic properties (10), reaches levels in tumor tissue that are only 20% of systemic drug levels (11).In 2016, of the ongoing 98 phase I/II and II glioma clinical trials, 63 studies (64.29%) were reported to include at least one drug able to pass the blood brain barrier (12).Unfortunately, according to systematic reviews and meta-analyses, almost all clinical trials involving targeted drugs and personalized chemotherapy in adult patients with glioblastoma have been unsuccessful (13)(14)(15)(16)(17).However, recent phase II trials (NCT02684058 and NCT04775485) found that dabrafenib plus trametinib could be an effective therapy as first-line treatment for pediatric patients with low-grade glioma with BRAF V600 mutations (18), and type II RAF inhibitor tovorafenib could be an effective therapy for BRAF-altered, relapsed/refractory pediatric low-grade glioma (19).Moreover, patients with IDH1 wild-type high-grade gliomas harboring BRAF or NF1 mutations and receiving trametinib monotherapy or in combination with dabrafenib had longer progression-free and overall survival than patients who did not receive genotype-matched targeted therapy (20).
Great hopes are currently placed on combination immunotherapy, including oncolytic virotherapy (16,(21)(22)(23)(24)(25).The immune system plays a primary role in the control of tumor development and the effectiveness of anticancer therapy (26-28).Historically, the brain has been considered an immune-privileged organ based on the lack of traditional lymphatic vessels in the brain and experiments with transplantation of foreign tissue into brain tissue and lack of rejection, as well as experiments with peripherally injected dyes that stain peripheral organs, but not the brain, due to the blood-brain barrier, restricting the access of macromolecules and cells into the brain parenchyma (29).Increasing evidence demonstrates that meningeal lymphatic vessels draining to the cervical lymph nodes play an important role in immune surveillance of the brain (30-32) and are essential for mounting an efficient immune response to brain tumors (33,34).Various types of immune cells, including T cells and dendritic cells, were observed within meningeal lymphatics in both normal and pathological conditions (30-32).In mice with intracranial glioma or metastatic melanoma, dorsal meningeal lymphatic vessels were found to undergo extensive remodeling, and their specific pharmacochemical ablation impaired intratumoral fluid drainage, dendritic cell trafficking, and the efficacy of immunotherapy (33).Despite the partial disruption of the blood-brain barrier in glioblastoma (7), which promotes the infiltration of immune cells (25), nevertheless, human glioblastoma exhibits a predominantly "cold" ("immune-desert"/"immuneexcluded") phenotype, characterized by the absence or exclusion of T cells in the tumor microenvironment (33,35) and T cell dysfunction, including tolerance and exhaustion (36,37).
A tumor subdues the immune system, exerting both a local complex inhibitory effect on the tumor tissue microenvironment and systemic immunosuppression through the secretion of many soluble factors (38,39).Profound immunosuppression and lymphopenia pose a challenge to current treatment strategies, including chemotherapy (38, 40, 41) and especially immunotherapy, the effectiveness of which may critically depend on the state of the patient's immune system (41,42).This review focuses on understanding the state of the immune system and the prognostic significance of different immune cell subtypes in patients with glioblastoma before standard therapy.In our accompanying review in Frontiers in Immunology (43), we comprehensively discuss the prognostic significance of standard therapy-related (iatrogenic) systemic immunosuppression and its implications for immunotherapy and oncolytic virotherapy.We provide compelling clinical data indicating that standard therapy affects various immune cell subsets, promoting tumor-related immune deficiency in patients with glioblastoma.Low posttreatment total lymphocyte count (TLC) is a prognostic factor for shorter survival in glioblastoma, and radiation-induced lymphopenia is a prognostic factor for mortality in virtually all solid cancers.Chemotherapy and corticosteroids may exacerbate radiation-induced lymphopenia.Dexamethasone use is a prognostic factor for shorter survival in glioblastoma.In addition, there is growing evidence that immunosuppression associated with standard therapy may be a barrier to immunotherapy, and lymphopenia is significantly associated with response and survival outcomes in patients with advanced cancer receiving immune checkpoint inhibitor therapy.Finally, we discuss how detailed blood and/or tumor immunophenotyping may be valuable for immunotherapy/oncolytic virotherapy research in terms of identifying new or validating the proposed immunological-based prognostic/predictive variables, and suggest what changes/ interventions to the standard therapy paradigm should be considered to maintain lymphocytes counts.These reviews should help inform more rational clinical trial design and treatment decisions to potentially improve the effectiveness of immunotherapy/oncolytic virotherapy.
One of the contributors to systemic immunosuppression is CD4+CD25+FOXP3+ regulatory T cells (Tregs), which are involved in immune tolerance of tumors and compromise cytotoxic T cell function (67).In the majority of studies, elevated Treg fractions were documented among peripheral blood CD4+ T cells of patients with glioblastoma (60,(68)(69)(70), even in cases with severe CD4+ T cell lymphopenia (<200 cells/mL) and regardless of steroid use (70).Patients with elevated Treg fractions but not normal Treg fractions showed significant proliferative dysfunction of CD4+ T cells, reduced quantities of Th1-type cytokines, and increased quantities of Th2-type cytokines (70).
Another contributor to systemic immunosuppression is myeloid-derived suppressor cells (MDSCs), a heterogeneous population of early myeloid progenitors and precursor cells that can suppress immune responses mediated by CD4+ and CD8+ T cells (71).The number of circulating MDSCs is higher in patients with glioblastoma than in healthy donors or patients with low-grade gliomas (72-75).
In contrast to changes in the counts of T cell subsets, MDSCs, and neutrophils (discussed below), natural killer (NK) cells (45,46,(76)(77)(78) and natural killer T (NKT) cells (79,80) were within the normal range in patients with glioblastoma/glioma before standard therapy in the majority of studies, although their cytotoxic activity may have been impaired (44,81,82).
3 A high neutrophil-to-lymphocyte ratio is an independent prognostic factor for shorter survival

Factors influencing the ratio of neutrophils and lymphocytes
Neutrophils account for 50-70% of circulating leukocytes in humans (83).The neutrophil-to-lymphocyte ratio (NLR), derived from the absolute neutrophil and lymphocyte counts of a full blood count, is an easily accessible and measurable marker (84).Changes in the balance between neutrophils and lymphocytes reflect an increase in systemic inflammation and a decrease in anti-tumor adaptive immunity (84).Baseline NLR increases with glioma progression (grade I-IV glioma), with the highest NLR values observed in patients with grade IV gliomas, followed by grade III and grade I-II gliomas (85-92).In addition, in a retrospective study of adult patients with not otherwise specified subtype of glioblastoma (n=89), a weak positive association was found between tumor size and preoperative NLR values (Spearman r=0.3212, p=0.0493) (93).In another retrospective study of patients with glioma (n=64), higher pretreatment NLR was significantly associated with larger tumor diameter (p=0.02)(94).Similarly, a positive correlation between NLR and tumor size in patients with papillary thyroid carcinoma has been repeatedly documented (95).In carcinomas in general, NLR is higher in patients with more advanced or aggressive disease, as evidenced by increased tumor stage, nodal stage, and number of metastatic lesions (84).This correlation might be due to the fact that tumor cells secrete granulocyte colony-stimulating factor (G-CSF) and/or granulocytemonocyte colony-stimulating factor (GM-CSF), which are not only direct growth factors for tumor cells but may also contribute to increased NLR in patients, shifting bone marrow hematopoiesis from the lymphocyte lineage toward the granulocyte lineage (96-98).
Any type of damage to brain tissue, including surgery-and therapy-related damage, tends to enhance G-CSF/GM-CSF synthesis (96).In addition, patients with glioblastoma treated with steroids have higher neutrophil counts and NLR (48,54,55,78,99,100); however, no significant influence of steroid use on neutrophil counts (101) or only a weak correlation between dexamethasone dose and NLR (53, 57) were also reported.It is advisable to measure NLR prior to surgery or other treatments that may increase the neutrophil count.It should also be noted that NLR may change not only under the direct influence of tumor progression, surgery, or steroid treatment, but also of local or systemic infection; inflammatory diseases; thyroid, renal, or hepatic dysfunction; diabetes mellitus; heart diseases; hypertension; obesity; psychologic stress; and other complications in cancer patients (102,103).
Regional anesthesia for patients with glioma has been proposed as a strategy to reduce postoperative systemic and local inflammatory responses (104).In a retrospective study of patients with glioblastoma (n=119) (104), local anesthesia of the nerves of the scalp during craniotomy, referred to as a "scalp block" (105), was shown to reduce postoperative NLR (104).This reduction was associated with longer median progression-free survival (16.7 versus 6.5 months for patients without a scalp block) (104).However, in another retrospective large cohort study (n=808), the use of a scalp block in glioma resection was not associated with improved progression-free and overall survival (106).Moreover, the use of different anesthetics, including isoflurane, desflurane, and propofol, during glioblastoma surgery is not associated with overall survival (107, 108).

A high neutrophil-to-lymphocyte ratio is a prognostic factor in solid tumors
In retrospective studies, lower neutrophil counts before radiochemotherapy were associated with better overall survival (n=164, n=369, and n=2002) (55, 57, 99) independent of steroid use (55).Higher neutrophil counts at relapse were also prognostic for worse overall survival, but only in patients who did not receive bevacizumab (109).However, there are studies that report no prognostic role for neutrophil counts in patients with glioblastoma (52,56,58,59).
Higher NLR was a significant prognostic factor for shorter progression-free survival in some glioblastoma/glioma retrospective studies (110-113) but not in others (51,54,(114)(115)(116) on univariate and/or multivariate analyses.In large meta-analyses, higher NLR in patients with carcinomas was a significant prognostic factor of cancer-specific, progression-free and/or disease-free survival (117,118).
In the majority of glioblastoma/glioma studies (Table 1), high NLR was established as an independent factor for worse outcomes without including MGMT promoter methylation status, steroid use, IDH mutation status, or other prognostic variables in multivariate analyses.Nevertheless, baseline NLR was still prognostically independent of MGMT promoter methylation status (58,129), IDH mutation status (51,92,121), and steroid use (122, 124) in patients with glioblastoma, and patients with increased NLR and requiring steroids had the poorest outcomes (73).Moreover, in a recent prospective glioma study (n=73, 37% with grade III astrocytoma and 63% with grade IV glioma), patients were divided into four groups based on the median baseline NLR and NLR decrease during chemoradiotherapy (53).Patients with baseline NLR <3.5 with NLR decrease during treatment (n=14), baseline NLR <3.5 without NLR decrease during treatment (n=23), baseline NLR ≥3.5 with NLR decrease during treatment (n=24), and baseline NLR ≥3.5 without NLR decrease during treatment (n=12) had median overall survival of 36.5, 19.2, 14.7 and 7.1 months, respectively (53).On univariate analysis, patients with baseline NLR <3.5 and NLR decrease during treatment had lower mortality risk than those with baseline NLR ≥3.5 (HR 0.512; 95% CI 0.291-0.904)or no decrease during Systematic reviews and meta-analyses have provided evidence that high baseline neutrophil count or high NLR were independently associated with adverse overall survival in various types of solid tumors (84, 117,118,[145][146][147][148][149][150][151], while normalization of post-treatment NLR was associated with improved survival (84).Moreover, systematic reviews and/or meta-analyses examining the correlation between NLR and outcomes in patients treated with immune checkpoint inhibitors have reported that higher NLR is a prognostic factor of worse disease control rate, objective response rate, progression-free survival, and/or overall survival in patients with head and neck squamous cell carcinoma (152, 153), gastric carcinoma (154,155), melanoma (156), metastatic renal cell carcinoma (157,158), non-small cell lung carcinoma (159), hepatocellular carcinoma (160) and cancer patients in general (161-163).High post-treatment NLR has also been associated with poor survival outcomes in cancer patients • Total: 208 There is no standardized cutoff for the prognostic/predictive NLR value.An NLR value ≥4 was associated with poorer overall survival in the majority of the studies of patients with glioblastoma/ gliomas with statistically meaningful sample sizes (Table 1).Of note, in a meta-analysis of 75 eligible studies covering more than 20 cancer types, the median cutoff for high NLR with the strongest prognostic effect was 4.0 (range 1.9-7.2) (117), whereas in a prospective study in an unselected general population (individuals aged 45 years, n=8711), the reference NLR value (mean and 95% intervals) was 1.76 (0.83-3.92) (172).
Neutrophils are plastic populations of immune cells with different functions (173).In rodent models, the immunosuppressive neutrophil populations may promote tumor progression by potentiating tumor invasion, angiogenesis, and metastasis (173).In patients with cancer, the proportion of immunosuppressive neutrophils is dramatically increased (173-175).Neutrophils can suppress the activation and proliferation of cytotoxic T cells (173-176).The predominance of immunosuppressive neutrophils over lymphocytes, which demonstrate the disproportion of the CD4 +/CD8+/Treg ratio, provides a clue as to why NLR is a prognostic marker of worse survival across many solid tumor types.

A composition and prognostic significance of the immune infiltrate of glioblastoma
It is worth noting that adult glioblastoma/gliomas in general differ significantly from pediatric gliomas.An overview of the composition of the tumor immune infiltrate across different types of pediatric glioma is given elsewhere (177).

Neutrophils
Glioblastoma tissue is abundantly infiltrated by neutrophils (178).Fossati et al. found a strong correlation between glioma tumor grade, the extent of neutrophil infiltration, and the preoperative circulating neutrophil counts (101).Over 70% of all glioma samples analyzed (n=105) showed significant neutrophil infiltration (40-50% of low grade gliomas and 87% of glioblastomas) (101).In another study, neutrophil infiltration was observed in 86% of II-IV glioma samples (n=232), and the level of neutrophil infiltration was significantly correlated with glioma grade (179).Increased neutrophil infiltration was associated with shorter overall survival in patients with glioblastoma (n=152) (58).Tumor-infiltrating neutrophils are an independent prognostic factor for overall survival across different tumor types (146,180).
Ionized calcium-binding adaptor molecule-1 (IBA-1) is a panmarker for all microglia and macrophages.High IBA-1 intensity was correlated with longer survival (193).However, in another study, the number of IBA+ cells was positively correlated with the overall tumor size and edema but not with overall survival (72).CD204+ (scavenger receptor) (193) or CD163+ (scavenger receptor) (194) M2-like microglia/macrophage density was correlated with worse survival, whereas lower expression of CD163 and higher expression of CCL3 (C-C Motif Chemokine Ligand 3), an M1 marker, was correlated with better survival (192).In contrast, Zeiner et al. found that high levels of CD68+ (a panmacrophage marker), CD206+ (mannose receptor C type 1), and CD163+ tumor-infiltrating macrophage subpopulations in the vital tumor core of patients with IDH1R132H-non-mutant glioblastoma (n=241) were associated with improved survival (187).Finally, Karimi et al. revealed that increased levels of MPO +CD163−P2Y12−CD68+ macrophages were associated with prolonged survival of patients with glioblastoma (184).

NK and NKT cells
NK cells and invariant NKT cells are scarcely present in glioblastoma/glioma tissue (60,183,184), and the role of these immune cell subtypes has not been clearly established in patients with glioblastoma/glioma.Nevertheless, there is increasing evidence of NK or NKT cell-based immunotherapy efficacy in rodent glioma models (197,198).In addition, a local administration of activated haploidentical NK cells (199) or irradiated CAR-NK cells (NK-92/ 5.28.z) targeting HER2 (200) in patients with recurrent glioblastoma was feasible and safe.

Dendritic cells
There are plasmacytoid dendritic cells, type 1 and type 2 classical dendritic cells, monocyte-derived dendritic cells, and a new dendritic cell subset, DC3 (201).The heterogeneity and functionality of the dendritic cell compartment in patients with glioblastoma (subsets, counts, and functionality) are poorly characterized and reviewed elsewhere (202).

T lymphocytes
The density of tumor-infiltrating CD4+, CD8+, and Tregs increases with glioma grade (203)(204)(205)(206), with higher levels of infiltrated CD4+ cells than CD8+ cells (203,207).However, Innocenti et al. reported no difference between CD4+ and CD8+ cell counts in glioblastoma tissue samples (n=59) (208).In general, tumor-infiltrating lymphocytes are differentially distributed in glioblastoma samples, from absent to abundant (183, 209-211).T cell infiltrates are mainly located in the perivascular areas and zones of tumor invasion into the surrounding brain parenchyma and are only infrequently found within the tumor tissue and in the perinecrotic areas (211-215).The percentages of immunological synapses established by tumor-infiltrating lymphocytes with tumor cells are very low (213), and the tumor-infiltrating lymphocytes have a suppressed and functionally impaired state/phenotype (described as tolerance and exhaustion) (36,37,60,69,216,217).These observations indicate that although the density of infiltrated lymphocytes varies considerably between patients, tumorinfiltrating lymphocytes cannot readily migrate into the immunosuppressive tumor microenvironment, are mainly arrested in the perivascular or peritumoral space, and are functionally compromised.

CD4+ and CD8+ T lymphocytes are highly heterogeneous immune cell populations
The inconsistent correlations between tumor-infiltrating lymphocytes and survival (Table 2) may stem from substantial functional and phenotypic complexity and plasticity of CD4+ and CD8+ T cell subsets discovered by high-dimensional single-cell mass cytometry (cytometry by time-of-flight, CyTOF) (244).Detailed qualitative and quantitative assessments of tumor immune infiltrates using mass cytometry to measure the frequencies and ratios of immune cell subsets and their functional and activation status may be required for accurate prognostication (184,(245)(246)(247). Distinct subsets of T cells are anti-inflammatory, pro-inflammatory, or both, and a dual role in cancer immunity has been ascribed to CD3+CD4+ cell subsets (246).For example, in follicular B-cell lymphomas, 12 subsets of intratumoral CD4+ T cells were identified, and specific subpopulations were correlated with poor or improved patient survival (246).In patients with clear cell renal cell carcinoma, 22 subsets were identified among infiltrating T cell lymphocytes, and a distinct immune composition correlated with survival (247).In addition, different prognoses may be obtained depending on the degree of infiltration of immunosuppressive FOXP3(high) Tregs or non-suppressive FOXP3(low) cells (248).For example, colorectal carcinomas with abundantly infiltrated FOXP3(low) T cells demonstrated a much better prognosis than tumors with predominant FOXP3(high) Treg infiltration (248).
It is also evident that not only the density of tumor-infiltrating lymphocytes but also the ratio between immunosuppressive immune cell subsets and cytotoxic lymphocytes have an effect on prognostic significance.In some studies, increased CD8+/CD4+, CD4+/CD8+, CD3+/Treg, or CD8+/Treg ratios, rather than the absolute counts of individual populations, were correlated with survival outcomes in patients with glioblastoma/glioma (203,229,232).Furthermore, the immunosuppressive cellular immune landscape in the glioblastoma/glioma microenvironment goes beyond well-defined immune cell subtypes such as Tregs, neutrophils, MDSCs, and tumor-associated macrophages (215, 249).Li et al. revealed that immunosuppressive CD3+CD4 +FOXP3− type 1 regulatory T cells occurred at high frequencies within glioblastoma tissue (249).In addition, Waziri et al. found that a great portion of CD3+ T cells within glioblastoma tissue was represented by CD3+CD4+CD56+ T cells (215).However, these T cells did not represent classical invariant NKT cells as they were neither stained with antibodies against an invariant TCR Vb24 nor with CD1d tetramer loaded with a-Gal-Cer (215).These CD3+CD4+CD56+ T cells might be immunosuppressive, since immunosuppressive CD3+CD4+CD56+CD25+FOXP3+ T cells were identified at a high frequency in hepatocellular carcinoma, and higher infiltration of these cells was inversely correlated with survival (250).
• Total: 199 • Total: 342 • Total: 68 • Total: 108 • Total: 43 A lack of standardization and recommendations for assessing tumor-infiltrating lymphocytes (technical and methodological differences associated with staining and analysis, including high inter-observer variability) in glioma tissue, as well as non-uniform patient cohorts, are potential factors contributing to inconsistent correlations between tumor infiltrating lymphocytes and overall survival (Table 2).It is known that pre-analytical variables such as tissue collection, fixation (time, buffer composition, and temperature), processing (dehydration reagents, temperature, and paraffin embedding), storage, staining (manual versus automated, quality, and quantity of antibodies), and other conditions may affect the accurate assessment of CD3+, CD4+, and CD8+ lymphocytes within a tissue (251,252).Moreover, since the immune infiltrate and individual subpopulations of immune cells are heterogeneously distributed in glioblastoma tissue (e.g., central areas versus marginal areas; perivascular areas versus perinecrotic areas), histochemical sections of different tumor regions will give different insights into immune cell density and diversity.New high-throughput multiplex immunohistochemistry and immunofluorescence technologies are expanding the ability to obtain additional information on cellular composition and spatial arrangement with greater reproducibility using standardized quantitative protocols (253).Therefore, indepth spatial immunophenotyping of glioblastoma tissue, together with consensus guidelines for the assessment of glioma-infiltrating lymphocytes, is required to more accurately establish the prognostic power of infiltrating lymphocytes.In general, this also applies to subsets of myeloid cells (254, 255).

The immune landscape and prognostic significance of mesenchymal, proneural, and classical molecular subtypes
In 2006, using unsupervised expression profiling on a cohort of 76 grade III and IV astrocytoma samples to classify tumors into prognostic groups, Philips et al. defined subtypes based on a 35gene signature termed proneural, mesenchymal, and proliferative (276).The proneural subtype, containing the majority of grade III astrocytoma samples, was distinguished by markedly better prognosis compared with other subtypes (276).In 2010, Verhaak et al. described a gene expression-based molecular classification (840-gene set classifier) of glioblastoma samples into the proneural, neural, classical, and mesenchymal subtypes (77% subtype concordance) (277).Somatic mutations and/or DNA copy number aberrations/overexpression of EGFR, NF1, and PDGFRA/ IDH1 predominantly defined the classical, mesenchymal, and proneural subtypes, respectively.MGMT promoter methylation status was not associated with subtypes (277).In meta-analyses, IDH mutations are associated with better progression-free and overall survival in patients with glioma (278-280).Surprisingly, overall survival of patients with the proneural subtype was not significantly different from other gene expression subtypes defined  (290).Currently, the combination of these parameters allows for improved prediction of outcome (291).
In 2017, applying unsupervised transcriptome analysis after filtering glioblastoma overexpressed genes to 369 IDH-wild type glioblastoma samples, Wang et al. confirmed three subtypes previously designated mesenchymal, proneural, and classical (≥93% subtype concordance; defined 50-gene signatures for each subtype with overlap with Verhaak et al.'s 840-gene set ranging from 42% to 54%), while the neural phenotype was suggested to be non-tumor specific (292).In addition, strong associations between subtypes and genomic abnormalities in previously reported subtype-defining genes were also confirmed (292).Again, patient survival did not differ significantly between the three subtypes (292).In 2019, using a 500-gene set classifier (only 108 genes matched Verhaak et al.'s 840-gene set) on six different datasets (three TCGA-cohorts and three Asian-cohorts), Teo et al. also confirmed the classical, mesenchymal, and proneural subtypes, with similar survival outcomes between subtypes (293).Finally, in 2023, White et al. also found no significant difference in overall survival between subtypes in the GLIOTRAIN (n=123), TCGA (n=164), and CGGA (n=693) cohorts after stratifying the mesenchymal, proneural, and classical subtypes according to Wang et al.' report (294).Interestingly, the authors introduced novel glioblastoma tumor microenvironment subtypes for IDH wild type glioblastoma (TME High , TME Med , and TME Low ), characterized by significantly different expression of genes specific to all immune and endothelial cell markers.However, stratification into these new subtypes showed no association with overall survival in the GLIOTRAIN, TCGA, CGGA, and DUKE cohorts (294).
Using The Cancer Genome Atlas (TCGA) database, Doucette et al. analyzed the mRNA expression levels of immune system genes (cytokines, cell markers, and signaling pathways) between four glioblastoma subtypes and found that the mesenchymal subtype was the most proinflammatory due to the preferential enrichment of both pro-inflammatory and immunosuppressive genes compared with other subtypes (295).Rutledge et al., using histochemical analysis, found that tumor-infiltrating lymphocytes were enriched in the mesenchymal subtype and strongly associated with mutations in NF1 and RB1, while lymphocytes were depleted in the classical subtype, EGFRamplified, and homozygous PTEN-deleted tumors.However, no association with survival was found (209).Wang et al. documented the increased presence of macrophages/microglia and neuroglial cells in mesenchymal subtype, and NF1 deactivation was associated with macrophage/microglia recruitment (292).The activated natural killer cell gene signature was significantly reduced in the mesenchymal subtype, the resting memory CD4+ T cell gene signature was significantly reduced in the proneural subtype, and the activated dendritic cell gene signature was significantly greater in the classical subtype (292).In an immunohistochemical study of the immune infiltrate of the four glioblastoma subtypes, Martinez-Lage et al. found that the mesenchymal and proneural subtypes were the most and least immunogenic, respectively (194).The percentage of CD3+, CD4+, and CD5+ lymphocytes differed significantly between the mesenchymal and proneural or classical subtypes but not other subtypes.The percentage of CD8+ did not differ between the four subtypes.The percentage of CD163+ and CD68+ macrophages/ microglia in the mesenchymal subtype was also significantly different compared with the classical or proneural subtypes, with no difference between them (194).In another immunohistological study of three glioblastoma subtypes defined according to the classification of Wang et al., the numbers of IBA1+ microglia/ macrophage cells and CD3+ and FOXP3+ lymphocytes were significantly higher in the mesenchymal subtype compared with other subtypes, with no significant difference between the proneural and classical subtypes (296).The number of CD8+ differed significantly only between the mesenchymal and proneural subtypes (296).On the contrary, Han et al. reported that the number of CD4+, CD8+, and FoxP3+ lymphocytes did not vary greatly between the subtypes defined according to the classification of Verhaak et al. (203).Finally, using flow cytometry, Gabrusiewicz et al. found that the frequency of MDSCs, microglia, and macrophages in the proneural and neural subtypes was not significantly different; the classical subtype had a markedly higher percentage of MDSCs than macrophages, whereas the mesenchymal subtype was predominantly infiltrated with microglia (72).
Taken together, glioma/glioblastoma transcriptome subtypes in different studies were defined using different numbers of only partially overlapping gene sets, and in each case the subtypes (except for the classification of Philips et al.) did not differ prognostically unless at least proneural G-CIMP status was not taken into account, despite the fact that the proneural subtype is characterized by the presence of prognostic IDH mutations much more often than other subtypes, on the one hand, and the existence of the greatest immunological difference between the proneural/ IDH mutant and mesenchymal subtypes, on the other hand.These observations might explain why stratification of glioblastoma patients based on the transcriptomic subtypes has not translated into clinical practice.Considering the enormous inter-and intratumoral cellular heterogeneity and plasticity, as well as the evolution of the cancer genome and phenotype, clinically meaningful subtyping of glioblastoma based on the transcriptome alone is challenging.However, there is some retrospective clinical evidence that mesenchymal or TME High glioblastoma might respond better to immunotherapy (vaccine, checkpoint inhibitor or oncolytic virus) (294, 297).

Future directions
There is growing interest in evaluating neoadjuvant immunotherapy in neuro-oncology (298)(299)(300).Neoadjuvant immunotherapy is advantageous over adjuvant immunotherapy, since it is applied before lymphotoxic standard radio-/chemotherapy (43).However, in patients with glioblastoma/glioma before standard therapy, tumor-related immunosuppression involves deregulation of many components of immunity, including changes in expression of different soluble and membrane proteins, reduced T lymphocyte counts (lymphopenia), increased NLR, increased levels or ratios of circulating and tumor-infiltrating immunosuppressive immune subsets (e.g., macrophages, MDSCs, and Tregs), and defective functions of antigen-presenting, helper, and effector immune cell subsets due to altered expression of receptors, costimulatory molecules, and cytokines (Figure 1).In this condition, the effectiveness of neoadjuvant immunotherapy for glioblastoma/ glioma might be suboptimal.
The spatial complexity and phenotypic diversity of the immune infiltrate (multicellular spatial organization) in patients with glioblastoma/gliomas at presentation, as well as the prognostic and/or predictive significance of immune cell subsets/signatures remain poorly characterized and largely unknown (255,301,302).As discussed above, the accurate prognostic assessment of immune infiltrate by traditional immunochistochemistry is influenced by too many analytical variables and factors.Although multidimensional single-cell approaches such as multiplexed ion beam imaging, imaging mass cytometry, and mass cytometry has become powerful tools for characterizing different immune cell subsets in immunology and identifying specific immune signatures associated with survival and response to immunotherapy in cancer patients (184, 245-247, 303-306) and may also help better decipher immune deregulation in patients with glioblastoma/gliomas and identify putative predictive or prognostic markers of therapeutic response and improved survival, but at present they should still be considered discovery tools in preclinical research and clinical trials rather than for routine clinical practice (303)(304)(305)(307)(308)(309)(310)(311)(312).It is also worth considering the factor of the availability and quality of tissue for assessing spatially distributed immune infiltrate in glioblastoma.Moreover, in various reports, approximately 15-20% of cases present as unresectable glioblastoma (diagnostic "biopsy-only") (313).However, the actual number of such patients may be higher.In the National Cancer Database, the percentage of  What is needed is a cost-effective, easily accessible and repeatedly validated prognostic and predictive factor, available to all patients, that can be dynamically assessed (at baseline, during and/or after treatment) and rapidly integrated into all ongoing and planned clinical studies of various forms of therapy in all clinical centers around the world.In general, immune-related biomarkers derived from blood rather than tumor tissue may be more suitable in terms of these requirements.NLR has prognostic value in patients with solid tumors across cancer types, stages of disease, and treatment strategies, including immunotherapy.However, although NLR has often been correlated with the effectiveness of immune checkpoint inhibitors in carcinomas, NLR has only rarely been taken into account when evaluating the effectiveness of vaccines, oncolytic viruses, and other immunotherapies in cancer patients.For patients with glioblastoma, this may also be due to the fact that the vast majority of clinical trials of all forms of therapy in general are non-randomized and/or uncontrolled (316,317).It is also important to note that evaluation of the prognostic/predictive significance of NLR has been largely limited to retrospective studies, with very few of the studies based on prospectively collected samples [e.g (318-320)], including glioblastoma/glioma (53,112).Since the number of prospective studies evaluating NLR in oncology is limited, NLR has not yet entered routine clinical practice as a stratification/predictive factor.
The prognostic/predictive power of NLR may be further improved by combining NLR assessment with other biomarkers.According to recent meta-analyses and systematic reviews, absolute/total lymphocyte count (ALC/TLC, especially posttreatment) (321), platelet-to-lymphocyte ratio (PLR) (133,322), lymphocyte-to-monocyte ratio (LMR) (323), systemic immuneinflammation index (SII), calculated by platelet count×neutrophil count/lymphocyte count (129,324), and systemic immune response index (SIRI), calculated by neutrophil count×monocyte/ lymphocyte count (325-327) are the emerging prognostic factors in glioblastoma/glioma.In a retrospective study, Yang et al. developed and compared the SII-NLR, SII-PLR, and NLR-PLR, and SII-NLR-PLR scoring systems and found that the combination of these inflammatory markers demonstrated greater predictive accuracy for overall survival at one and two years than any single indicator in patients with glioblastoma (n=208), with the best scoring system being SII-NLR (129).The authors constructed a nomogram including age, Karnofsky Performance Status (KPS), extent of resection, MGMT promoter methylation status, chemoradiotherapy, and SII-NLR score to predict 2-year survival in patients with glioblastoma (the c-index of the nomogram was 0.848 (95% CI 0.836-0.861)and 0.843 (95% CI 0.830-0.855)excluding MGMT promoter methylation status) (129).
Finding methods to reduce NLR is another important area of research in clinical oncology.In patients with recurrent glioblastoma (n=18), treatment with recombinant interleukin-7 restored and maintained total lymphocyte counts without serious toxicity and irrespective of steroid and temozolomide use (328).Interleukin-7 is currently considered the most potent therapeutic candidate for the treatment of lymphopenia in cancer and noncancer patients (41).As we have discussed, NLR may be affected by many physiological and pathological confounding factors, including psychological/emotional stress in cancer patients.It is known that cortisol and epinephrine may increase neutrophil and decrease lymphocyte counts (329-331).In a randomized trial of lung cancer patients (n=80), psychological intervention was shown to significantly reduce NLR compared with usual care (332).

Conclusion
Glioma progression and molecular characteristics (e.g., IDH mutations or mesenchymal gene signature) have distinct effects on major immune cell subsets, and conversely, different proportions of immune cell subsets and their polarization or activation states may have different effects on tumor progression, response to therapy, and survival.All attempts to identify a reliable prognostic immune marker in glioblastoma tissue have led to markedly contradictory results, which can be explained, among other things, by the unprecedented level of spatial heterogeneity of the immune infiltrate and the significant phenotypic and functional diversity of immune subpopulations.High NLR has been repeatedly established as an independent prognostic factor for shorter overall survival in patients with glioblastoma and carcinomas, and its combination with other markers of the immune response significantly improves the accuracy of prediction; however, more prospective studies are needed to confirm the prognostic/predictive power of NLR.We call for incorporating the dynamic assessment of NLR and other emerging blood inflammatory markers (e.g., platelet-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, systemic immune-inflammation index, and systemic immune response index) into all neuro-oncology trials to carefully evaluate and compare their individual and combinatorial prognostic/ predictive significance and relative superiority.

FIGURE 1
FIGURE 1 Glioblastoma/glioma-related systemic and local immunosuppression.Before standard therapy, patients may have reduced counts of circulating T lymphocytes (lymphopenia); however, in the majority of studies, the increased fractions of regulatory T cells (Tregs) are documented.The numbers of circulating myeloid-derived suppressor cells (MDSCs) and neutrophils are higher in patients with glioblastoma than in healthy donors or patients with low-grade gliomas.The counts of circulating natural killer (NK) and NKT cells are within the normal range in the majority of studies.The density of tumor-infiltrating CD4+, CD8+, and Tregs increases with glioma grade; however, lymphocytes are differentially distributed in tissue samples, from absent to abundant.Glioblastoma tissue is abundantly infiltrated by neutrophils and macrophages, while NK cells and NKT cells are scarcely present.Serum levels of Th1-type cytokines (IL-2, IL-6, IL-12, TNF-a, and IFN-g) are reduced, while serum levels of Th2-type cytokines (IL-4 and IL-10) are increased.Serum, cerebral spinal fluid, or tumor cyst fluid from patients can suppress the proliferation and/or function of lymphocytes and other immune cells from healthy donors.Consequently, patients have defective functions of antigen-presenting, helper, and effector immune cell subsets due to altered expression of different soluble and membrane proteins (receptors, costimulatory molecules, and cytokines).The tumor-infiltrating lymphocytes have suppressed and functionally impaired state/phenotype.The Figure was generated using Servier Medical Art, provided by Servier, licensed under a Creative Commons Attribution 3.0 unported license.

TABLE 1
High NLR is an independent prognostic factor for overall survival in patients with glioblastoma/gliomas.No data on steroid use, MGMT promoter methylation status, and IDH mutation status (patients diagnosed between 2004 and 2009) • Neutrophil and lymphocyte counts in isolation were not prognostic • Multivariate analysis adjusted for age, gender, ECOG performance status, extent of resection, tumor location, full Stupp protocol, and second line therapy No data on MGMT promoter methylation status and steroid use • 31 patients had mutant IDH (patients diagnosed between 2009 and 2014) 865.9 vs. 9.7 months (p=0.02);TR1.86,95% CI 1.18-2.93,p=0.01 • Multivariate: TR 1.65, 95% CI 1.15-2.35,p<0.01 • Retrospective • Relatively small sample size • Blood sampling time: post-therapy prior to second surgery • 67.3% of patients used steroids prior to second surgery • No data on MGMT promoter methylation status and IDH mutation status (patients diagnosed between 2004 and 2011) • No data on confounding variables in multivariate analysis Han et al. (2015) (58) • Total: 152 • Male: 95 (62.5%); female: 57 (37.5%) • Mean age: 50.•Retrospective • No data on IDH mutation status (patients diagnosed between 2010 and 2014) and steroid use • NLR ≥4 was associated with increased tumor neutrophil infiltration/ decreased CD3+ infiltration • Neutrophil and lymphocyte counts in isolation were not correlated with survival • Multivariate analysis adjusted for age, KPS, extent of resection, and MGMT promoter methylation status • Retrospective • No data on the extent of resection, steroid use, MGMT promoter methylation, and IDH1 mutation status (patients diagnosed between 2009 and 2012) • Heterogeneity in patient population and treatment • No data on confounding variables in multivariate analysis • Retrospective • Relatively small sample size • No data on steroid use, MGMT promoter methylation status, IDH mutation status, and post-surgery therapy (patients diagnosed • Retrospective • No data on steroid use, MGMT promoter methylation status, and IDH mutation status (patients diagnosed between 2005 and 2013) • ≈50% of patients had comorbidities with potential impact on NLR • No correlation of absolute neutrophil and lymphocyte counts with overall survival • Multivariate analysis adjusted for KPS, tumor location, first-line and second-line therapy, and presence of comorbidities (Continued) • A disproportionate number of patients with NLR ≤4 • Multivariate analysis adjusted for age, sex, KPS, pathology, extent of resection, standard treatment, and IDH mutation • Retrospective • Relatively small sample size • No data on MGMT promoter methylation status • 58% of patients had TERT promoter mutation and 42% had IDH mutation • mdNLR assessed by an immunomethylomic approach was associated with survival independent of chemotherapy and steroid use • Multivariate analysis adjusted for age, grade, and mutation status • Retrospective • 37 (19.3%)patients had methylated MGMT promoter status; 127 (66.1%) with unknown status • No data on steroid use • 38 (19.8%) patients had mutant IDH1; 124 (64.6%) with unknown status (patients diagnosed between 2006 and 2018) • Multivariate analysis adjusted for age and adjuvant therapy • Retrospective • No data on MGMT promoter methylation status and steroid use • 103 (26.3%) were triple-positive (1p/19q codeletion, IDH and TERT mutations), 19 (4.8%) had both IDH and TERT mutations, 100 (25.5%) had

TABLE 2
Correlation between the density of tumor-infiltrating lymphocytes and overall survival of patients with glioblastoma/gliomas.

TABLE 2 Continued
Stepanenko et al. 10.3389/fimmu.2024.1326753Frontiers in Immunology frontiersin.orgpatients with unresectable glioblastoma among patients diagnosed from 2004 to 2013 with known survival and extent of resection is 28.5% (314).Kole et al. documented that among 1325 patients with biopsy-only glioblastoma who received radiochemotherapy, the median overall survival was 9.2 months (313), and Harlay et al. reported that among 139 patients who underwent biopsy only, 54 (39%) and 68 (49%) were amenable to standard radiochemotherapy or chemotherapy only, with the median overall survival of 14 months (95% CI, 9.65-18.71)and 8 months (95% CI, 4.62-7.67),respectively (315).This relatively large group of patients is particularly in need of new treatment strategies and should not be ignored.