Page 77 - Read Online

Page 693 Sharma et al. Cancer Drug Resist 2023;6:688-708 https://dx.doi.org/10.20517/cdr.2023.82

mainly in bone marrow . This may partially explain the T cell lymphopenia in glioblastoma patients.
[81]
However, treatment (radiation and TMZ) associated T cell lymphopenia was also very common [83,84] .

Glioblastoma tumors can produce IL-6 and drive myeloid immunosuppression by inducing PD-L1
expression on MDSCs . Glioblastoma can also utilize the natural immune tolerance mechanisms to recruit
[85]
regulatory T cells (Tregs) through the expression of indoleamine 2,3-dioxygenase (IDO) , as well as the
[86]
tumor-associated macrophages (TAMs) expression of TIM4 . Besides soluble factors, extracellular vesicles
[87]
containing various signaling molecules, including growth factors, non-coding RNAs, cytokines, and other
functional proteins, have been found to play an important role in the regulation of glioblastoma TME .
[88]
Those mechanisms involve an extensive network of DCs, TAMs, MDSCs, and T lymphocytes with complex
and dynamic crosstalk [Figure 2].

Heterogeneity in tumor microenvironment
Tumor heterogeneity has been well-known in glioblastoma biology at multiple levels , including genetics/
[89]
epigenetics (molecular subtypes), molecular signaling (tumor driver mutations), cellular components
(clonal and subclonal tumor cells vs. tumor microenvironment), and temporal (primary vs. secondary).
scRNAseq analysis of infiltrating neoplastic cells in human glioblastoma revealed vast genomic and
transcriptomic heterogeneity . Another work in brain endothelial cells derived from human glioblastoma
[90]
using a similar approach (scRNAseq) showed five distinct endothelial cell phenotypes representing different
states of EC activation and BBB impairment and association with different anatomical locations within and
around the tumor .
[91]

With the advancement of multi-omics platforms, tumor heterogeneity at both inter- and intra-tumoral
levels has been much better depicted in glioblastoma [92-94] . The inter-tumoral heterogeneity can be readily
appreciated by the molecular subtyping of human glioblastoma tumors by their transcriptional profile and
phenotypical response to therapy [2,95,96] . Consistent with the four molecular subtypes of glioblastoma, a more
recent scRNAseq analysis showed that glioblastoma cells can differentiate into four principal states,
including astrocyte-like, oligodendrocyte progenitor-like, neural progenitor cell-like, and mesenchymal-like
state . These four cellular states are influenced by the tumor microenvironment and oncogenic drivers
[97]
with certain plasticity .
[97]
The intra-tumoral heterogeneity in glioblastoma is characterized by the presence of clonal and subclonal
differentiated tumor cells, glioma stem cells (GSCs), and various components of the tumor
microenvironment (stromal, endothelial, and infiltrating immune cells). A recent study by Schaettler et al.
using scRNAseq revealed the differences between primary and secondary glioblastoma in their genomic
abnormality and neoantigen formation, as well as the spatially differential T cell clones within the
glioblastoma . The authors used TCR β-chain CDR3 sequences as unique barcodes of individual T cell
[98]
clones, as TCR β-chain CDR3 is highly diverse with a significant role in antigen recognition . Their results
[99]
demonstrated a topological clonal diversity of T cells in glioblastoma . Besides microglia, another
[98]
representative cell population that further complicates glioblastoma heterogeneity is a large variety of
myeloid cells in the TME . They mainly comprise TAMs, MDSCs, DCs, neutrophils, and undifferentiated
[100]
monocytes [78,101] . Another study using scRNAseq and multiplexing tissue-imaging techniques demonstrated
a spatially differential tumor microenvironment characterized by inflammatory signaling and hypoxia in
glioblastoma . The authors revealed that CD73, a critical regulator of local purinergic signaling with an
[102]
essential role in inflammatory response , was mainly expressed in glioblastoma cells with a positive
[103]
correlation between levels of CD73 and HIF1α expression in the hypoxic tumor regions, where the CD73+
glioma cells co-localize with CD39+ microglia to form a spatially compartmentalized microenvironment to