Introduction

The extracellular matrix (ECM) is an intricate three-dimensional non-cell network of proteoglycans (PGs) and fibrous proteins that surrounds and supports cells within tissues [1]. This complex scaffolding not only provides structural integrity but also plays a pivotal role in regulating various cellular processes, including proliferation, migration, differentiation, and apoptosis. Its composition and organization are dynamically modulated in response to developmental cues, physiological changes, and pathological conditions. In the context of cancer, such as gliomas, the ECM takes center stage as a crucial microenvironmental component that profoundly influences disease progression, metastasis, and therapeutic responses.

Gliomas, intrinsic brain tumors that originate from neuroglial progenitor cells, are characterized by their aggressive infiltrative behavior and propensity for recurrence [2]. Glioma cells exploit the ECM’s dynamic properties to their advantage. Through an array of molecular mechanisms, gliomas actively remodel the ECM, transforming it into a permissive environment that facilitates tumor growth, invasion, and evasion of therapeutic interventions. This remodeling involves the synthesis, deposition, degradation, and modification of ECM components, all of which collectively contribute to the establishment of a niche that supports glioma recurrence and progression.

Despite recent advances in surgery combined with radiotherapy and chemotherapy, patient survival rates have increased only slightly [3, 4]. One of the primary reasons for this poor prognosis is the diffuse infiltration of small clusters of tumor cells into the brain tissue, which prevents complete surgical tumor resection [5]. There is usually no clear boundary between the tumor and the surrounding brain parenchyma, which complicates complete surgical resection. Consequently, within months after surgery, recurrent neoplasms are triggered in the proximity of the resection zone, mostly within 2 cm of the original region [6, 7]. Individual or small clusters of tumor cells can migrate centimeters from the gross tumor into the surrounding healthy brain tissue [8].These roaming glioma cells actively remodel the ECM around them by degrading current ECM and replacing with Glioma-Modified ECM (GM-ECM). A total RNA-seq analysis showed an enhancement of cell-ECM interaction gene expression in recurrent versus parental cell populations [9]. The GM-ECM helps roaming gliomas survive adjuvant treatment after surgery and furtherly promotes residual roaming glioma to migrate, recurrent and progress.

Residual roaming glioma cells create and remodel ECM to be more abundant, denser and stiffer

The GM-ECM in gliomas significantly differs in composition and architecture from that in normal tissue. Considering its physical properties, the GM-ECM is more abundant, denser, and stiffer. Glioma cells have the ability to extensively remodel their microenvironment both through deposition of new matrix components and degradation of existing ones [10]. One key class of proteins responsible for regulating the turnover of collagens are the matrix metalloproteinases (MMPs). MMPs possess a broad range of overlapping substrate specificities and family members are able to degrade diverse substrates including collagens, fibronectin, laminin, proteoglycans, cell surface proteins, and pro-forms of growth factors such as TGF-β [11]. In the context of collagen turnover, 14 MMPs have been reported to degrade various collagens [12] (Fig. 1)(Table 1).

Fig. 1
figure 1

Residual diffuse glioma cells remodel a more abundant, denser, and stiffer glioma-modified extracellular matrix (GM-ECM). Created with BioRender.com

Full size image
Table 1 Difference of ECM component between normal brain and peri-glioma environment
Full size table

GM-ECM volume in the brain increases from 20 to 48% in high-grade gliomas, mainly due to Hyaluronan (HA). HA is a non-sulfated glycosaminoglycan present in the extracellular matrix, composed of N-acetylglucosamine/glucuronic acid disaccharide repeats of variable length. Hyaluronan is one of the most highly anionic substances in the body, attracting Na+ and other cations which bring with them an osmotic influx of water. Thus, the extracellular space in the brain expands to a volume approximately 10 times greater than that of the matrix substance itself, ensuring a water-rich environment.

Hyaluronan is increased approximately four-fold in primary brain tumors, reaching levels comparable to those present during the central nervous system (CNS) development [13]. The encoded gene for HA synthase (HAS), especially HAS2, was aberrantly up-regulated in tumor, relative to HAS genes in normal tissue, consistent with the finding that HAS overexpression results in increased tumorigenic features [14].

ECM glycoproteins such as fibronectin, tenascin, laminin, vitronectin and collagen of different types have been shown to be produced by glioma cells [13, 29, 30]. These overproduced glycoproteins increase GM-ECM density. Tenascins are a family with 4 known members in vertebrates: tenascin-W, tenascin-C (TNC), tenascin-X and tenascin-R (TNR). TNR is produced by normal healthy neurons and released into the ECM [31]. It is expressed in the ECM of non-tumoral brain parenchyma but exhibits a loss of expression in glioma, which may reflect deregulation of the normal ECM [15]. TNC is a highly expressed tenascin during normal fetal development and usually undetectable in adult brains. However, persistent levels of TNC have been characterized in adult neural stem cell niches such as the radial glia and astrocyte stem cell compartment of the subventricular zone, or associated with chronic neuropathological conditions, like glioma. Glioma cells are the main source of TNC production, and overexpression of TNC involves Notch pathway, which is a common upregulated pathway in glioma [16, 17]. TNCs are very large multimeric glycoproteins that consist of six polypeptide monomers and are combined into the hexamer at their N-termini. Nearly all kinds of solid tumors express high levels of TNC, but the highest concentrations have been found in gliomas. In fact, TNC was originally discovered as a major glioma cell secretion product. Its enrichment in human glioma ECM is correlated with malignancy.

Normal mature tissue cells rarely express TNC but in gliomas, TNC is expressed by malignant tumor cells [21]. TNC acts by binding to chondroitin sulfate proteoglycans(CSPGs) such as brevican or neurocan, which are also special in brain development [22]. By knocking down TNC in primary GBM cells and injecting into nude mice then, the patient derived model (PDX) with glioma shows not only increased survival, but also softer tumor area and lower mechanosignaling, which demonstrates that TNC acts as a driver molecular in increasing ECM stiffness.

Immunohistochemistry in patient glioma samples revealed that the elevated ECM stiffness in the patients with poorer prognosis was accompanied by a substantial increase in HA expression as well as increased levels of TNC. Additionally, ECM stiffness did not correlate with levels or distribution of type I collagen, vasculature or cellularity [18]. GBM that had TNC immunopositivity survived for a significantly shorter period than those in which TNC expression was absent, TNC in gliomas can be identified as a predictor of poor prognosis and disease progression [19, 20].

There is no consensus of fibronectin’s (FN) function in glioma, but its presence is well recognized in glioma tumors and their surrounding GM-ECM. A clear positive IHC staining of FN is found in all malignant gliomas, a much lower intensity was observed in glioma with lower-grade malignancy, and no staining was observed in healthy brain tissues [23].Glioma cells can promote fibronectin fibrillogenesis through CD93, a transmembrane receptor which is often overexpressed in tumor vessels in many cancers including glioma [24].

The GM-ECM is filled with overproduced glycoproteins and other components that drive the stiffening phenomena through various physical and chemical approaches. Normal brain is a flexible, soft organ with ECM stiffness of 0.2 to 1.2 kPa, which increases up to 45 kPa around the GM-ECM [32].Another clinical study, using Intraoperative Shear Wave Elastography for in vivo measurement of brain tumor stiffness shows that normal brain tissue has been characterized by a reproducible mean stiffness of 7.3 ± 2.1 kPa, that lower-grade glioma stiffness is different from high-grade glioma stiffness (p = 0.01) and that normal brain stiffness is very different from low-grade gliomas stiffness (p < 0.01) [33]. The increase of stiffness in glioma is also supported by the atomic force microscopy (AFM), which measured the stiffness of freshly removed human brain tumor tissue and indicated a significant stiffness difference (p ≤ 0.001) between glioblastoma and healthy tissue distributions, with stiffness levels of glioblastoma tissue almost threefold higher [34].

It’s difficult to obtain fresh human brain tissue samples for stiffness measurements due to logistical and regulatory challenges. However, we can still study the stiffness around glioma in several ways. There are some up/down-regulated pathways or gene alterations from glioma that influence the ECM to make it stiffer, one of which is PIEZO1. PIEZO1 is overexpressed in aggressive human gliomas and its expression inversely correlates with patient survival. PIEZO1 localizes at focal adhesions to activate integrin-FAK signaling, regulate the ECM and reinforce tissue stiffening. In turn, a stiffer mechanical microenvironment elevates PIEZO1 expression to promote glioma aggression [35]. The PIEZO1 promoter is generally hyper-methylated in the IDH mutant gliomas, correlating with decreased PIEZO1 mRNA expression in these tumors. This finding suggests that generally more aggressive IDH wild-type gliomas are epigenetically more poised to overexpress PIEZO1.

Another influencing pathway is Cancer-Associated Fibroblasts (CAFs) and its downstream. CAFs promote ECM stiffness in response to signals from yes-associated protein 1 (YAP1), which is one of the key effectors in the Hippo pathway acting in cooperation with other oncogenic factors like COX-2 for promotion of immunosuppression and drug resistance in cancer cells [36].

There is s potential feedback mechanism in gliomas when regulating the stiffness in the GM-ECM and then sensing it, possibly through the connection between TNC expression and NOTCH pathway. TNC can bind to integrin α2β1 on the glioma cell and upregulates JAG1 expression which interacts with its receptor NOTCH [22]. Glioma cells can sense the stiffness of the ECM through TNC by working as a ligand for integrins α2/7/8/9β1 and αvβ1/3/6 [37].

GM-ECM protects residual glioma cells from adjuvant therapy by building a safe harbor

Alterations that increase ECM abundance, density, and stiffness can negatively affect response to therapy in multiple ways. One of the most obvious effect is that an excessive accumulation of dense and rigid ECM, which histologically often encapsulates clusters of tumor cells, can act as a barrier, shielding the cells from therapeutic agents [38, 39] (Fig. 2).

Fig. 2
figure 2

The modified extracellular matrix of glioma protects glioma cells through multiple ways. Created with BioRender.com

Full size image

When culturing GBM cells in a 3D hydrogel system with a gradient of stiffness, the cells in stiffer regions of the gradient hydrogel show less proliferation and spreading. These stiffer regions lead to the formation of more densely packed cellular spheroids. Because these spheroids have a compact structure, they can act as a barrier to drug diffusion. This physical barrier may prevent chemotherapeutic agents, such as Temozolomide (TMZ), from effectively penetrating the spheroid and killing the cancer cells inside, leading to the drug resistance [40].

High concentration of microproteins in the ECM such as chondroitin sulfate proteoglycans (CSPGs) can cause accumulating interstitial pressure and severe tortuosity in the extracellular space within the tumor, further limiting the penetration and distribution of therapeutic agents, especially those with larger molecular sizes [41].

Hyaluronan (HA), upregulated and enriched in the GM-ECM, can bind large amounts of water in the ECM which leads to an increase in interstitial fluid pressure (PIF). Some studies indicate that transcapillary transport and diffusion within the tumor might be hindered by high PIF resulting from high HA contend and/or vessel leakage. In two of these studies [42, 43], improved vascular perfusion and reduced vessel collapse were observed after hyaluronidase treatment. This might indicate that the high PIF in hyaluronan-rich tumors restricts drug transport mainly by compressing the supplying vessels and rather than interfering with interstitial drug diffusion. By depleting HA, tumor harbors increased macromolecular permeability by a trigger of fenestrations and interendothelial junctional gaps in tumor endothelia.

The GM-ECM not only acts as a physical barrier obstructing various pharmaceutical agents to protect residual glioma cells from destruction, but it also acts as an active biocomponent to cloak these cells, shielding them from immune cell detection and assault. Neither T cells nor dendritic cells are able to penetrate dense fibrils in GM-ECM. T-lymphocyte phenotype is dependent upon collagen fiber density: loose matrices support cytotoxic T-lymphocytes and dense matrices or those incorporating specific glycoproteins leads to immune-inhibitory phenotypes. T-lymphocyte movement is driven by chemokine gradients in loose ECM, but in dense ECMs, T-lymphocytes doesn’t demonstrate chemokine-directed movement. The clinical relevance of this shielding function of the stromal ECM that keeps immune cells at distance from the tumor cells was most strikingly demonstrated by previous research using a urothelial cancer patient cohort that shows non-response to PD-L1 checkpoint inhibition correlated with CTL’s low activity in the stromal ECM [44].

Beside the physical shielding, changes and regulatory mechanisms at the molecular and pathway levels are also involved in the protection of residual glioma cells. High ECM density causes tumor hypoxia, potentially increasing collagen deposition in the ECM, leading to higher levels of ECM density and stiffness [45]. An excessive accumulation of dense and stiff ECM, which histologically often encapsulates clusters of tumor cells, can act as a barrier. This effect is directly linked to a reduced overall supply, as this barrier also impairs diffusion of oxygen, nutrients, and metabolites. Increased hypoxia and metabolic stress lead to activation of antiapoptotic and drug resistance pathways. Presence of hypoxia reprograms tumor cells through multiple proteins, such as hypoxia-inducible factor (HIF)-1α, which can protect glioma cells in multiple ways [46,47,48].

In addition to passive defense, the GM-ECM can also take the initiative to resist recognition and attack by immune cells. TNC secreted by tumor cells in GM-ECM does not only promoting stiffness, but also protects glioma cells from T cell attack. By deposited in the vicinity of T cells in the GM-ECM and interacting with α5β1 and αvβ6 integrins on T lymphocytes associated with reduced mTOR signaling, TNC inhibits T cell proliferation and protects the glioma cell from the immune cell [49]. TNC can also provide drug resistance, experiment shows TNC-knockdown GBM neurospheres found to be more sensitive to temozolomide (TMZ) treatment [50].

Integrins belong to a large family of heterodimeric transmembrane adhesion receptors, named after their roles in “integrating” cell function with the surrounding stroma [51]. The expression of integrins is upregulated in various cancers, including GBM [52]. Cells can sense tissue stiffness through signals from FAK that again cooperates with integrins. FAK signals increase pro-survival pathways like AKT and MAPK. This has been shown to confer resistance to glioma treatments such as, rapamycin, an mTOR inhibitor [53].

FN is a high-molecular-weight glycoprotein in the ECM that is known for being over-expressed in several cancers [54]. Research looking into FN and the glioma stem-like cells (GSCs) indicates that FN can suppressed p53-mediated apoptosis and upregulated P-glycoprotein expression, which gives GSCs chemoresistance to alkylating agents such like carmustine [55].

Matrix stiffness can also promote glioma cell stemness by activating BCL9L/Wnt/β-catenin signaling. Higher matrix stiffness can enhance the stemness of glioma cells, resulting in sustained tumor growth by activating the BCL9L/Wnt/β-catenin signaling pathway [56]. CSCs promote tumorigenesis and tumor development, It has been well documented that Wnt/β-catenin signaling is a critical determinant of CSC pluripotency and self-renewal. This nest preserves tumorigenesis energy for further epigenetically recurrence.

GM-ECM facilitates tumor recurrence and progression mainly by enhancing stemness and promoting migration

After initially modified by residual glioma cells and then consequently preserves glioma cells during adjuvant therapy, GM-ECM now will promote tumor progression and recurrence at a macroscopic level, mainly by supporting cell stemness and migration. GM-ECM can promote glioma recurrence in a variety of ways, such as providing structural physical support or participating in biological regulation of multiple pathways, but its role is often concentrated in two perspectives: promoting stemness or migration (Table 2) (Fig. 3).

Table 2 How GM-ECM component functions in promoting glioma progression
Full size table
Fig. 3
figure 3

ECM involving athways promoting glioma recurrence through stemness or aggression. Created with BioRender.com

Full size image

TNC is usually present in neural stem cell niches in the embryonic developmental stage and shortly after birth. However, it has also been found enriched in the recurrent glioma ECM, proving that the GM-ECM provides a suitable ecological niche for glioma stem cells to survive and proliferate [22]. Collagen and fibronectin enriched in the GM-ECM can provide structural support and activating critical signaling pathways such as Integrin αvβ3 Signaling, PI3K/AKT/SOX2 and CDC42/F-actin/YAP-1/Nupr1/Nestin signaling pathways, resulting in regulating cell proliferation and contributing to glioma stemness [57].

Recent studies have illuminated the critical role of thrombospondin in ECM in glioma progression. One study demonstrated that the formation of microtubes (MTs) was significantly enhanced in the resection cavity following neurosurgery in experimental models [58]. This MT formation may partly elucidate why recurrent tumors frequently arise near the resection site. In following investigation, the stimulation of GBM stem-like cell cultures with transforming growth factor-beta 1 (TGF-β1) promotes MT formation has been observed. Notably, thrombospondin-1 (TSP-1) emerged as a crucial mediator in this process, acting downstream of TGF-β1 [59]. Additionally, another study has shown that extracellular vesicles can mediate tumor progression through the thrombospondin-1 pathway, proves the pivotal role of thrombospondin-1 facilitating glioma progression [60].

Type I collagen/fibronectin (FN) in the GM-ECM can strengthen the tumorigenic potential and proliferative characteristics of glioma cells by promoting GSCs adherence through a concentration-dependent manner. Moreover, FN can drive GSCs differentiation and furtherly promote differentiated cell growth by the elevation of Ki-67 [55]. Comparing GM-FCM and normal ECM, FN seems to have a higher concentration in GM-ECM, and can promote migration of glioma cells [61].

By initially shaping GM-ECM into a stiffer environment, stiffer GM-ECM can enhance cell invasiveness and promote relapse through a feedback regulation. The stiffness of GM-ECM gradually increases in accordance with aggressiveness, from a hundred pascals of gliosis to several thousand in lower grade glioma, to tens of thousands of pascals in high grade glioma [18].

The stiffness of the peri-glioma GM-ECM creates a microenvironment that fosters aggressive behaviors, including enhanced proliferation, migration, and invasion. Previous research shows that GBM cells show higher proliferation and migration rates when cultured on stiff two-dimensional substrates [8]. On a stiffer ECM, these complexes become larger and more stable, promoting a stronger connection between the cell and the ECM. This enables glioma cells to migrate more efficiently and invade surrounding healthy brain tissue. Human GBM tissue biopsies indicated a significant correlation between the proportion of highly stiff areas within a GBM tissue (E > 1,400 Pa) and a poor patient prognosis score. By contrast, the tissues that contained a high proportion of soft ECM regions (E < 200 Pa) had the best patient prognosis score [18].

Increasing ECM stiffness increases the percentage of cells in S phase [62], representing a high level of cellular DNA synthesis and replication, in other word, increased proliferation. Increasing matrix stiffness from 0.08 kPa to 119 kPa produced a five-fold increase in phosphorylated EGFR (pEGFR) and nearly two-fold increases in phosphorylated Akt (pAkt) and total PI3K, increasing microenvironmental stiffness broadly activates EGFR signaling in GBM tumor cells to regulate proliferation [62]. Another study also indicates that increasing ECM rigidity can induce a suite of phenotypic changes in human glioma cells that includes increased cell spreading, faster motility, and enhanced proliferation [63].

Stiffness modulates expression of glioma EGFR pathway and its downstream effector Akt, to promote cell cycle progression and proliferation [62]. Research shows integrin signaling is a vital pathway in regulating stiffness of ECM [64]. Increasing matrix stiffness led to delayed U87 cell proliferation inside hydrogels, but cells form denser spheroids with extended cell protrusions. Cells cultured in stiff hydrogels also showed upregulation of HA synthase 1 and matrix metalloproteinase-1 (MMP-1) [70]. Recurrent cells grown on 0.5kPa showed higher Young’s moduli suggesting the ability of these cells to make the surrounding ECM stiffer, which further promotes glioma recurrence [9].

IDH is one of the most important molecular in glioma classification. It is widely known that IDH mutant gliomas usually have better prognosis compared to IDH wildtype ones. One of the perspectives from the GM-ECM that can explain this result is that mutant IDH1 restricts glioma aggression by reducing HIF1α-dependent TNC expression to decrease ECM stiffness and mechanosignaling. But still, recurrent IDH-mutant glioma still have a stiffer TNC-enriched ECM, which possibly due to ECM stiffness can bypass protective activity of IDH mutational status through a HIF1α and TNC mediated via a tension-dependent positive feedback loop [18].

Other than making tissue stiffer to protect glioma from treatment, integrin also helps aggression. Extensive data demonstrate that the expression of integrins, together with corresponding ECM components, facilitates the infiltration of tumor cells through normal brain tissue [71]. The integrin arsenal expressed on the cell surface is important for the cellular phenotype. Integrins constitute a large family of dimeric transmembrane receptors, which are composed of α and β subunits that initiate intracellular signaling cascades by binding to adapter proteins [65]. α3β1 has been shown to be consistently over-expressed and to be a key regulator of glioma cell migration [66]. αvβ3 (observed at the periphery of high-grade gliomas) and αvβ5 integrins (expressed more at the center of the tumor) are involved in glioma malignancy [67]. Expression of αvβ3 coincides with expression of the metalloproteinase MMP-2 in tumor cells at the invasion front.

Besides constructing a defensive shield, HA also contributes to invasiveness. HA acts as a ligand for CD44 and might thereby play a role in EMT, resulting in increased invasiveness and metastasis [68, 69]. Targeting these receptors, such as CD44 and RHAMM, has been shown to inhibit tumor invasion and migration.

Laminins is also a major glioblastoma cells production in constructing GM-ECM, other with collagen type IV and FN [23]. Laminins can be produced by GFAP positive cells during glioma cell invasion in humans [25, 26]. Laminin gamma 1 gene (LAMC1) may play an important role in glioma invasion [27], but in the meantime Laminin alpha 5 also can significantly lower the invasion of mobile U251MG cells [28].

Conclusion

In summary, this review illustrates how glioma cells engineer and remodel the ECM, and how this modified matrix assists glioma cells in resisting therapy, preserving stemness, fostering growth, increasing invasion and leads to recurrence. Despite the challenges posed by the complex glioma-ECM dynamics, as we refine our understanding of the ECM’s influence on glioma behavior, targeting key ECM elements in offers a viable strategy to postpone or even prevent glioma recurrence.